DEPARTMENT OF HEALTH AND HUMAN SERVICES
Food and Drug Administration


21 CFR Parts 1 and 101

[Docket Nos. 90N - 0135, 91N - 0162, 78P - 0091, 87P - 0194/CP,
AND 90P - 0052]
RIN 0905 - AD08

Food Labeling: Mandatory Status of Nutrition Labeling and
Nutrient Content Revision, Format for Nutrition Label

AGENCY: Food and Drug Administration, HHS.

ACTION: Final rule.

                                                                 

                                                                 

                                                                 

                                           


SUMMARY: The Food and Drug Administration (FDA) is amending its
regulations to: (1) Require nutrition labeling on most foods that
are regulated by FDA, (2) revise the list of required nutrients
and food components and the conditions for declaring them in
nutrition labeling, (3) specify a new format for declaring
nutrition information, (4) allow specified products to be exempt
from nutrition labeling, and (5) prescribe a simplified form of
nutrition labeling and the circumstances in which such simplified
nutrition labeling may be used. This final rule also responds to
citizen petitions on the declaration of dietary fiber in
nutrition labeling and on methodologies for determining protein
content.

DATES: Effective February 14, 1994. The Director of the Office of
the Federal Register approves the incorporation by reference in
accordance with 5 U.S.C. 552(a) and 1 CFR part 51 of certain
publications in 21 CFR 101.9(c)(1)(i)(A), (c)(1)(i)(B),
(c)(1)(i)(C), (c)(1)(i)(E), (c)(6), (c)(7)(ii), (c)(7)(ii)(B),
and (g)(2), effective (February 14, 1994).

FOR FURTHER INFORMATION CONTACT: Virginia L. Wilkening, Center
for Food Safety and Applied Nutrition (HFF - 200), Food and Drug
Administration, 200 C Street SW., Washington, DC 20204, 202 - 205
- 4561.

SUPPLEMENTARY INFORMATION:

I. Background

    In the Federal Register of July 19, 1990 (55 FR 29847), FDA
published a proposed rule entitled ``Food Labeling; Mandatory
Status of Nutrition Labeling and Nutrient Content Revision''
(hereinafter identified as ``the mandatory nutrition labeling
proposal'') to amend its food labeling regulations to require
nutrition labeling on most food products that are meaningful
sources of nutrients. FDA also proposed to revise the list of
nutrients and food components that must be included in nutrition
labeling by adding calories from fat, saturated fatty acids,
cholesterol, and dietary fiber to that list. It proposed to make
the listing of thiamin, riboflavin, and niacin optional rather
than mandatory. In addition, FDA addressed the conditions under
which other nutrients could be, or are required to be, included
in nutrition labeling and proposed to allow manufacturers to
voluntarily include a nutrition profile of selected food
components in nutrition labeling.
    During the comment period for these proposed regulations,
Congress passed, and the President signed into law, the Nutrition
Labeling and Education Act of 1990 (the 1990 amendments) (Pub. L.
101 - 535, November 8, 1990). The 1990 amendments amended the
Federal Food, Drug, and Cosmetic Act (the act) by adding section
403(q) (21 U.S.C. 343(q)) which specifies, in part that: (1) With
certain exceptions, a food is to be considered misbranded unless
its label or labeling bears nutrition labeling, (2)certain
nutrients and food components are to be included in nutrition
labeling, although the Secretary of Health and Human Services
(the Secretary) can add or delete nutrients by regulation if he
finds such action necessary to assist consumers in maintain if
necessary, regulations, (4) a simplified nutrition label is to be
used when the food contains insignificant amounts of most
nutrients, and (5) FDA is to develop regulations governing
labeling of foods to which section 411 of the act (21 U.S.C. 350)
applies.
    In response to these requirements of the 1990 amendments, FDA
published in the Federal Register of November 27, 1991 (56 FR
60366; as amended (57 FR 8178, March 6, 1992)) a proposal
(hereinafter identified as the supplementary proposal) to modify
its July 19, 1990, proposal by: (1) Adding sugars and complex
carbohydrates to the list of required nutrients in nutrition
labeling, (2) prescribing a simplified form of nutrition labeling
and the circumstances in which such simplified nutrition labeling
must be used, (3) allowing specified products to be exempt from
nutrition labeling, and (4) establishing regulations for the
nutrition labeling of vitamin and mineral supplements. The agency
also responded to a citizen petition regarding methodologies for
determining protein quality. Interested persons were given until
February 25, 1992, to comment.
    Subsequently, FDA published in the Federal Register of July
20, 1992 (57 FR 32058; amended at 57 FR 37190, August 18, 1992),
a proposal (hereinafter identified as the format proposal) to
adopt a new format, specifically the PERCENT DV (Daily Value)
with DRV (Daily Reference Value) format, for use in presenting
nutrition information on the food label. Interested persons were
given until August 19, 1992, to comment. In addition, on July 23,
1992, a notice was published in the Federal Register (57 FR
32750) of a public meeting to be held on the format proposal in
Bethesda, MD, on August 17, 1992.
    On October 6, 1992, Congress passed the Dietary Supplement
Act of 1992 (H.R. 6181) (hereinafter referred to as the ``DS
Act'') that, in section 202(a)(1), establishes a 1 year
moratorium on the implementation of the 1990 amendments with
respect to dietary supplements of vitamins, minerals, herbs, or
other similar nutritional substances. Section 202(a)(2) requires
the Secretary, and by delegation FDA, to issue new proposed
regulations that are applicable to dietary supplements no later
than June 15, 1993, and final regulations by December 31, 1993.
In addition, section 203 instructs FDA to not promulgate
regulations that require the use of, or are based upon,
recommended daily allowances of vitamins or minerals before
November 8, 1993 (other than regulations establishing the United
States Recommended Daily Allowance (U.S. RDA) specified in 21 CFR
101.9(c)(7)(iv) as in effect on October 6, 1992).
    FDA received approximately 1,500 responses to its July 19,
1990, mandatory nutrition labeling proposal, approximately 3,000
responses to the November 27, 1991, supplementary proposal, and
approximately 1,000 responses to the July 20, 1992, format
proposal, each of which contained one or more comments. Responses
were received from consumers, health professionals, health
promotion organizations, trade and retail associations, State and
local governments, foreign governments, professional societies,
consumer advocacy organizations, industry, and universities. The
comments generally supported the proposals. Several comments
addressed issues covered by other proposals that are a part of
this overall food labeling initiative, and they will be addressed
in those final documents, while other comments were outside the
scope of these proposals and will not be discussed here. Many
comments dealt with issues pertaining to meat and poultry
products whose labeling is regulated by the U.S. Department of
Agriculture (USDA). Of those comments, comments pertaining to the
content or format of the nutrition label are included in the
following discussions. However, comments pertaining to issues
covered exclusively by USDA, such as specific exemptions applicab
    A number of comments to both the July 19, 1990, and November
27, 1991, proposals suggested modifications in, or were opposed
to, various provisions of the proposals. A summary of the
suggested changes, the opposing comments, and the agency's
responses follow.

II. Mandatory Nutrition Labeling--Legal Authority

    1. Most comments agreed that the 1990 amendments clearly
established FDA's authority to mandate nutrition labeling on most
foods. One comment, however, argued that a requirement that
labels say or not say certain things curtails freedom of the
press.
    The agency disagrees. FDA's authority to regulate the content
of food labels has been broadly upheld against First Amendment
challenges. This issue is discussed at length in both the final
rule on nutrient content claims entitled ``Food Labeling:Nutrient
Content Claims, General Principles, Petitions, Definition of
Terms'' and the final rule on health claims entitled ``Food
Labeling; General Requirements for Health Claims for Food,'' both
of which are published elsewhere in this issue of the Federal
Register. Those discussions are incorporated in this document by
reference. As those discussions make clear, there is no merit to
this comment. Therefore, FDA is taking no action on the basis of
this comment.
    2. One comment objected to FDA being given authority to
mandate nutrition labeling on most foods on the basis that
current nutrition labeling rules were legally questionable.
    The question of FDA's authority to require nutrition labeling
was a fundamental issue that led Congress to pass the 1990
amendments. As discussed in the legislative history, Congress
concluded that legislation was needed to strengthen FDA's
authority to require nutrition labeling on foods and to avoid the
possibility of protracted litigation over the comprehensive
nutrition labeling regulations that the agency adopts (Ref. 16).
Therefore, there can be no question about FDA's authority to
require nutrition labeling on most food products.

III. Content of the Nutrition Label

A. General Issues

1. Voluntary Declaration of Additional Information
    3. A number of comments objected to the voluntary declaration
of nutrients beyond those required in nutrition labeling.
Numerous comments stated that the declaration of additional
information on the food label would be confusing, or that it
might mislead the consumer into believing that a product with
additional nutrients listed is more nutritious or has greater
public health significance than is the case. Some comments
objected on the basis that the additional information would
clutter the labeland diminish the consumer's focus on mandatory
nutrients. A few comments expressed concern that voluntary
declaration of additional nutrients on the label will require
smaller print on the food label to accommodate the inclusion of
all the mandatory and voluntary information, and that the smaller
type size would compromise the usefulness of the label
information to the elderly or visually impaired.
    A number of comments supported the voluntary listing of
additional nutrients, pointing out that the 1990 amendments
require that the regulations permit the label or labeling of food
to include nutrition information, which is in addition to the
information required by section 403(q) of the act and which is of
the type described in subparagraph (1) or (2) of that section. A
few comments supported the view that voluntary listing of
additional nutrients may provide valuable information to an
individual or aid the consumer in making an informed choice in
food selection. Other comments supported voluntary listing of
additional nutrients stating that some nutrients may satisfy
nutrient needs of some individuals or pose a health risk to
others. One comm
    Numerous comments that basically supported listing of
additional information also supported limiting the information
allowed. Some comments supported allowing voluntary information,
but they suggested that FDA standardize the manner in which it is
included on the label to the extent of requiring that it be
separate from the nutrition label or in different type size. Five
comments requested voluntary listing of specific nutrients
including: Potassium; vitamins E, K, and B6; copper; manganese;
iodine; maltodextrin; and l-glutamate, l-cysteine, and
l-tryptophan. Two comments supported the listing of additional
nutrients but recommended restricting the allowed nutrients to
those for which Recommended Dietary Allowances havebeen set by
the National Academy of Sciences (Ref. 23) or for which Reference
Daily Intakes (RDI's) have been determined by FDA. One of these
comments further suggested restricting the allowed nutrients to
exclude nutrients that do not have RecommendedDietary Allowances
but only have Estimated Safe and Adequate Daily Dietary Intakes
(ESADDI's), which are also set by the National Academy of
Sciences (Ref. 23). One comment suggested that additional
information on the food label be restricted to information
permitted by the Council of the European Communities (EC).
    FDA, in its mandatory nutrition labeling proposal of July 19,
1990, proposed to allow the voluntary declaration of several
nutrients (e.g., potassium and soluble fiber) and any naturally
occurring vitamins and minerals for which RDI's had been proposed
in 101.9(c)(10)(iv) (21 CFR 101.9(c)(10)(iv)), which was
redesignated as 101.9(c)(11)(iv) in the November 27, 1991,
proposal. Additionally, section 2(b)(1)(C) of the 1990 amendments
states that regulations shall permit the label or labelingof food
to include nutrition information that is in addition to the
information required by section 403(q) of the act and that is of
the type described in subparagraph (1) or (2) of that section.
Section 403(q)(2) of the act refers to information that will
assist consumers in maintaining healthy dietary practices.
    FDA believes that it is required by statute to allow
additional information on the food label insofar as it assists
consumers in maintaining healthy dietary practices. However, the
agency raised questions in the supplementary proposal about how
thepresence of these additional nutrients on the label would be
interpreted by consumers, and whether the listing of some
voluntary nutrients would actually be misleading (56 FR 60366 at
60372). The comments confirmed that unlimited additional
information on the nutrition label would have the potential of
being confusing or misleading.
    FDA requested comments on whether it is necessary to include
limits on the voluntary information that may be provided on the
nutrition label. The comments that FDA received on this issue
have lead the agency to conclude that it has a responsibilityto
limit the number of nutrients permitted to be voluntarily listed
on the food label. Such a limitation reflects the statement in
the House report (Ref. 16, p. 18) that the regulations that FDA
adopts should assure that the information that is included
voluntarily does not interfere with the consumer's understanding
of the information that is required to be included on the
nutrition label. The agency finds that limits are necessary so
that the emphasis is on the required information, and that the
additional information does not clutter the food label or mislead
or confuse the consumer.
    Therefore, to limit the information that may be provided on
the nutrition label, FDA is amending the proposed regulations to
delete ``calories from unsaturated fat,'' ``calories from
carbohydrates,'' ``calories from protein,'' and quantitative
declaraion on the nutrition label of ``calories from saturated
fat,'' ``polyunsaturated fat,'' ``monounsaturated fat,''
``soluble fiber,'' ``insoluble fiber,'' ``sugar alcohol,''
``other carbohydrate,'' ``potassium'' and those vitamins and
minerals for which RDI's have been established.
    With respect to other nutrients suggested by individual
comments for consideration for voluntary listing on the food
label, the agency has not been persuaded that there are large
numbers of consumers who desire a voluntary listing of the food
components in question (e.g., maltodextrin or single amino acids.
Therefore, FDA will not allow voluntary listing of these other
substances or food components on the nutrition label. To
implement this section, FDA has added a sentence to 101.9(c)
that states that no nutrients or food components other than those
set forth in that section as either mandatory or voluntary may be
included in the nutrition label. The inclusion of any other
nutrient or food component would violate section 403(q) of the
act and misbrand the food.
    Also, while FDA supports efforts toward international
harmonization of food labeling where possible, the 1990
amendments direct FDA to permit that a broad spectrum of
nutrients be on the food label unless the agency finds that the
information is notnecessary to assist consumers in maintaining
healthy dietary practices, a finding that FDA has generally not
made. As a result, the spectrum of required and permitted
nutrients exceeds those permitted by the EC.
    FDA is not requiring that additional nutrients declared
voluntarily be put in separate boxes or a different type size
because it believes these actions would confuse consumers and
would complicate and clutter the label needlessly. In some
instances additional nutrients, whose declaration is usually
voluntary, will be required to be declared. For example, in the
case of fortified foods, enriched pasta must declare amounts of
thiamin, riboflavin, and niacin, and margarine must declare
vitamin D when it is added. In other cases, if certain claims are
made, additional nutrients will be required to be declared. For
example, when nutrient content claims are made about cholesterol,
declaration of poly- and monounsaturated fats are required (see
101.9(c)(2) (i) and (ii). Placing these nutrients in the
principal box for nutrition labeling when required, and in a
separate box (or different type size) when voluntarily added to
the nutrition label would easily confuse consumers. Also,
separating subcomponents that can voluntarily be declared, such
as soluble and insoluble fiber, from the primary component,
dietary fiber, for which declaration is mandatory, would unduly
complicate the label.
    However, in response to comments concerned that information
on additional nutrients would clutter the label and to comments
on the format proposal, FDA is providing in new 101.9(d)(8) for
a linear array of vitamins and minerals. This form of
presentation, which is discussed in more detail in section V. of
this document, is similar to that recommended in the IOM report
(Ref. 1) which places more emphasis on the macronutrients.
2. Order of Nutrients
    4. Several comments from industry, health promotional
organizations, and academia supported the order of nutrients
proposed by FDA in 101.9(c) (56 FR 60366 at 60386 through
60390). One industry comment stated that the proposed sequence
fairly prioritized the Dietary Guidelines for Americans (Ref. 4)
and placed the proper emphasis on those dietary factors that
affect the health of consumers. This comment, along with one from
a health promotion organization, also endorsed the separation of
vitamins and minerals from other nutrients seen in proposed
formats (57 FR 32058), stating that this feature represented a
logical break in the list of nutrition information and would both
improve label readability and facilitate consumers' search for
specificients from those most important to consumers to those
least important to consumers but questioned whether protein
should be included.
    On the other hand, several comments argued that the proposed
order of nutrients has features that would mislead consumers. One
comment characterized the proposed order as ``an unwarranted
effort to overemphasize some nutrients, such as fat, at the
expense of the other important label components'' and suggested
that the decision on whether to emphasize one nutrient over
another should be left to nutrition education programs that
consider the total diet over a long period of time. Other
industry comments criticized the proposed order of nutrients,
stating that it would be consistent with the ``good food/bad
food'' concept and would convey a negative impression to
consumers. One industry comment supporting the current order of
nutrients argued that protein should not be listed near the end,
stating that beneficial nutrients should be listed at the
beginning of the nutrient list. The comment suggested that from
an educational standpoint, it is more positive to educate on the
good points of nutrition labeling than to focus on negative
aspects.
    A number of comments advocated that the current order of
nutrients be maintained, or that any modified order resemble the
current order as closely as possible. Several comments supporting
the current order of nutrients stated that consumers are already
accustomed to the current order, and that changing the order
would lead to unnecessary confusion and diminish consumers'
understanding of the nutrition label.
    A few comments suggested alternative nutrient orders. A
comment from a professional organization stated that those
nutrients whose overconsumption is related to increased risk of
disease should be placed at the top of the list of required
nutrients. One industry comment recommended that nutrients be
regrouped to first list those nutrients whose Daily Value is
dependent on calorie intake (i.e., total fat, saturated fat,
carbohydrate, dietary fiber, and protein), followed by those
whose Daily Value remains the same for varying calorie intakes
(i.e., cholesterol and sodium). Another comment requested that
sodium be listed with the vitamins and minerals rather than among
the organic macronutrients.
    A comment from a manufacturer addressed the issue of where to
place the voluntary nutrients on the label. The comment suggested
that voluntary nutrients should be sequenced in a logical manner
with respect to the nutrients whose declaration is mandatory. The
following examples were cited: Unsaturated fat should follow
saturated fat (both should be indented), potassium should follow
sodium, soluble and insoluble fiber should follow dietary fiber,
and vitamins and minerals should follow those thatare mandatory.
    The agency is not persuaded by arguments stating that listing
nutrients in order of public health importance will cast foods as
either ``good foods'' or ``bad foods.'' Listing nutrients in this
manner will instead facilitate selection of an overalldiet that
is consistent with dietary guidelines based on what nutrients are
present in a particular food and in what amounts. No data were
presented to show that use of this nutrient order on the
nutrition label is likely to be confusing to consumers.
    The agency also does not agree with the request that sodium
be placed with vitamins and minerals rather than with the organic
macronutrients. Sodium is an electrolyte that is distinct from
both organic nutrients and vitamins and minerals.
However,excessive intake is associated with a potential increase
in the risk of chronic diseases, as are excessive intakes of the
other mandatory organic nutrients (i.e., macronutrients such as
fat) in the nutrition label. Vitamins and minerals generally are
as
    FDA agrees that the placement of voluntary nutrients should
be sequenced in a logical manner with respect to the mandatory
nutrients. FDA has provided in new 101.9(c) that voluntary
nutrients that are subcomponents are to be declared immediately
beneath the primary components, and that potassium (the second
electrolyte) is to be declared adjacent to sodium.
    The agency believes that a revised order according to the
public health significance of a nutrient will adequately convey
nutrient information with no appreciable increase in consumer
effort. This action is based on the order provided in section
403(q)(1) of the act (see Ref. 16, p. 13) and the comment
recommending that nutrients whose overconsumption is related to
increased risk of disease should be placed at the top of the list
of required nutrients.
    Accordingly, new 101.9(c) is modified to require mandatory
and voluntary nutrients to be arranged in the following order:
Calories, calories from fat, calories from saturated fat, total
fat, saturated fat, polyunsaturated fat, monounsaturated fat,
cholesterol, sodium, potassium, total carbohydrate, dietary
fiber, soluble fiber, insoluble fiber, sugars, sugar alcohol,
other carbohydrate, protein, vitamin A, vitamin C, calcium, iron,
and other vitamins or minerals in the order listed in
proposed101.9(c)(11)(iv), redesignated as 101.9(c)(8)(iv). This
order deviates from that provided in section 403(q)(1) of the act
only by reversing dietary fiber (and its subcomponents) and
sugars. The reason for this reversal is discussed in comment 58
of this document.
    Consequently, the paragraphs in 101.9(c) are renumbered as
discussed below for each nutrient. Redesignations also occur as a
result of moving paragraphs (c)(1) and (c)(2) pertaining to
serving size and servings per container, respectively, to new
paragraph (d).
    The agency believes that this amended order of nutrients,
which lists them in order of public health significance, will
benefit consumers. The agency's decision is a reasonable
outgrowth of its commitment to present nutrition information in
the context of a total daily diet, and it reflects the agency's
commitment to link nutrient information with the dietary guidance
considered important to public health (Ref. 4).

B. Calories

1. Total Calories
    5. The majority of comments supported the proposal for
mandatory declaration of calories with voluntary use of metric
terminology (i.e., declaration of the number of kilojoules in
addition to calories in proposed 101.9(c)(3), redesignated
as101.9(c)(1), and voluntary use of the term ``energy''
parenthetically as a synonym for calories, as provided
in101.9(c)(11)(v), redesignated as 101.9(c)(8)(v)).
    Other comments expressed a preference for metric labeling.
The comments argued that American consumers should become
accustomed to the metric system of measurement and recommended
the exclusive use of metrics to ensure compatibility with
European markets. The comments suggested that the avoirdupois
system of measurement used in the United States is outmoded and
impedes international commerce and the exchange of scientific
information. Several comments suggested that ``energy'' should be
used in lieu of calories and requested that the conversion factor
for calories to kilojoules be stated on each label.
    Still other comments, taking the opposite position, suggested
that metric units be disallowed to avoid consumer confusion and
for the sake of simplicity.
    Although FDA agrees that efforts should be made to
familiarize consumers with metric units, the agency disagrees
with the comments that urged the exclusive, mandatory conversion
to metrics at this time. The technical amendments (August 3,
1992) to the Fair Packaging and Labeling Act (15 U.S.C. 1451 et
seq., Pub. L. 102 - 329) require the use of the most appropriate
units of both the customary inch/pound system of measure and the
metric system on food labels for measuring quantity. These
amendmentsories. Upon implementation, this act should further an
awareness of metric measurement among American consumers and
permit a greater concordance in units of measurement with the
international market and scientific community. Until that time,
the agency is not persuaded that the mandatory use of metric
terminology, or the declaration of factors to convert calories to
kilojoules, is justified. Accordingly, the agency is not making
the requested changes.
2. Calories From Fat
    6. Many comments were received from consumers, state and
local governments, universities, professional associations,
consumer groups, manufacturers, and health associations on the
issue of calories from fat. The majority agreed with the proposal
that the declaration of calories from fat should be mandatory on
the nutrition label.
    Several other comments suggested that calories from fat be
voluntarily listed or disallowed because this information might
be confusing or misleading to consumers and might establish
artificial ``good food'' and ``bad food'' categories. These
comments stated that consumers may tend to exclude foods with a
significant amount of calories from fat, possibly creating
nutritional deficiencies. Further, these comments stated that it
is important that consumers view fat as part of a day's diet
rather than in the context of individual foods. A few comments
suggested that declaring the calories from fat is unnecessary
because calories from fat can be easily calculated by multiplying
the number of grams (g) of fat by nine, the number of calories
per g of fat. A few comments suggested that there be further
study of the effectiveness of the declaration of calories from
fat in nutrition labeling as a nutrition education tool. One
comment suggested that low fat foods, such as fruits and
vegetables that contain less than 2 g of fat, be exempted from
the requirement to list calories from fat.
    FDA is not persuaded by the arguments that the declaration of
calories from fat should be voluntary or disallowed. The
declaration of calories from fat is required by section
403(q)(1)(C)(ii) of the act. While section 403(q)(2)(B) of the
act allows the Secretary to delete nutrient information that is
not necessary to assist consumers in maintaining healthy dietary
practices, no data were presented that would support making such
a finding with respect to the declaration of calories from fat.
    It is well established that diets that are high in fat pose
significant health risks. Dietary fat contributes more than twice
the calories per g than does protein or carbohydrate.
Overconsumption of fat is associated with higher rates of obesity
(Refs. 2 and 3), and there is evidence from epidemiological and
animal studies that high fat intakes are associated with some
types of cancer (Refs. 2 and 3). The most common and consistent
dietary recommendation for the general population is for
caloriesfrom total fat to be reduced to less than or equal to 30
percent of total calories (Refs. 3 and 4).
    Currently, the consumption of total fat in the general
population is approximately 37 percent of total calories, an
amount well above the recommended level (Ref. 2). Further,
consumption of total fat in the United States is significantly
higher than that consumed in countries with much lower rates of
coronary heart disease, such as Japan, China, and the
Mediterranean countries (Ref. 2).
    Based upon this body of evidence, FDA believes that reducing
total fat intake is an important public health priority. The
agency is not persuaded that the declaration of calories from fat
will automatically lead to consumers viewing foods in
strict``good food,'' ``bad food'' categories, or that consumers
cannot make appropriate decisions regarding the consumption of
foods that may have a significant number of calories from fat in
their diets. No evidence was presented demonstrating a
relationship
    Although calories from fat can, in fact, be readily
calculated (FDA is requiring that information on the number of
calories per g of fat, carbohydrate, and protein, be included as
part of the nutrition label (see 101.9(d)(10)), the declaration
of calories from fat will be beneficial in assisting consumers to
moderate their fat intake by providing an additional method,
other than g of fat, for monitoring their fat intake.
    However, the agency concurs that fat should be viewed as a
part of the complete daily diet. Foods that may have a
significant number of calories from fat may readily be included
in the daily diet when the overall fat intake for the day is
moderate. The agency intends to build this concept into its
consumer education program, discussed later in this document.
Further, FDA welcomes further study on the health implications of
overconsumption of calories from fat and the effectiveness of
this method of depicting fat content.
    7. Many comments advocated the mandatory declaration of the
percent of calories from fat. Other comments suggested that
calories from total fat should be replaced by the percent of
calories from fat. The comments stated that the process of
determining the percent of fat is time consuming and unfamiliar
to many consumers. Further, the comments argued that it is
unlikely that substantial numbers of consumers would or could
keep running totals of their fat intake in order to calculate the
percent of daily fat consumed. The comments argued that the best
way to determine whether a food is high or low in fat is to have
fat content declared by percent of total calories.
    A few comments suggested that percent of calories from fat
for individual foods is incomplete information, while the percent
of calories from fat for a complete meal or the daily diet is
useful information. These comments suggested that the percentof
calories from fat be voluntary and limited to meal-type products,
such as frozen dinners, and disallowed for other foods.
    The agency is not persuaded by the comments that the
declaration of percent calories from fat is warranted. As
discussed in the July 19, 1990, mandatory nutrition labeling
proposal (55 FR 29487 at 29493 and 29494), information on the
percent of calories from fat is only valuable in the context of a
total daily diet. Recommendations from various health
organizations to limit dietary fat intake to 30 percent or less
of calories pertain not to individual foods but to the entire
day's intake.
    In addition, the percent of calories from fat in low calorie
foods can be quite misleading. For instance, in radishes, over 25
percent of calories are from fat. Despite this relatively high
percentage, radishes contain very low amounts of fat and readily
fit within a daily diet that meets current dietary
recommendations.
    The agency agrees that calculating the percent of calories
from fat consumed in a day may be difficult for many consumers.
The agency notes that the PERCENT DV format (see section V.G.2)
facilitates monitoring of dietary fat because the Daily Value for
fat is set at 30 percent of calories from fat. Consumers need
only add the percent DV for total fat with a target of no more
than 100 percent or a target percentage adjusted for their
individual caloric intake. Alternatively, consumers can determine
the maximum number of g of fat recommended per day at their
calorie level and track the number of g of fat. There are several
publications listing recommended daily maximum amounts of fat
according to caloric intake or that have simple arithmetical
methods for deriving this information (Refs. 26 through 29). In a
similar fashion, the DRV for fat, which is established in the
companion document entitled ``Food Labeling; Reference Daily
Intakes and Daily Reference Values'' (hereinafter identified as
therganizations to develop and publish similar approaches.
    8. One comment objected to that section of 101.9(c)(3)(i) in
the July 19, 1990, mandatory nutrition labeling proposal that
allowed ``calories from fat'' to be omitted and replaced with the
statement ``Not a significant source of calories from total fat''
if the product contains less than 1 g of fat per serving. The
comment objected to similar provisions for saturated fat,
cholesterol, and dietary fiber on the basis that it made the
regulations complex and confusing. These provisions were carried
forward in the November 27, 1991, supplementary proposal with the
1-g criterion being changed to \1/2\ g, and similar provisions
being added for complex carbohydrates and sugars.
    These provisions were included in the July 19, 1990,
mandatory nutrition labeling proposal (55 FR 29487 at 29502),
which was published before the adoption of the requirement in
section 403(q)(5)(C) of the act for a simplified format, to
minimize thespace required for nutrition labeling. This provision
is similar to that allowed in current 101.9(c)(7)(i) for
vitamins and minerals that are present in amounts less than2
percent of the U.S. Recommended Daily Allowance (U.S. RDA). FDA
did not delete these provisions in the November 27, 1991,
supplementary proposal because the agency believed they might be
helpful in minimizing space requirements on foods that do not
qualify for the simplified format under proposed 101.9(f).
    FDA finds that the added flexibility that these provisions
provide outweighs any added complexity they may create. USDA's
final nutrition labeling regulations, which are being published
simultaneously with these final rules, include similar
provisions. Under FDA's regulations, with the exception of the
core nutrients (i.e., calories, total fat, total carbohydrate,
protein, and sodium), nutrients that are present in insignificant
amounts may be omitted from the list of nutrients and grouped in
a summary statement (e.g., ``Not a significant amount of calories
from fat, saturated fat, and cholesterol''). Therefore, the
agency is retaining in 101.9(c)(3)(i), redesignated as
101.9(c)(1)(ii), for calories from total fat; in
101.9(c)(4)(i), redesignated as 101.9(c)(2)(i), for saturated
fat; in 101.9(c)(5), redesignated as 101.9(c)(3), for
cholesterol; in101.9(c)(6)(ii)(A), redesignated as
101.9(c)(6)(ii), for sugars; in 101.9(c)(7), redesignated as
101.9(c)(6)(i), for dietary fiber; and in 101.9(c)(11)(iii),
redesignated as101.9(c)(8)(iii), for vitamins and minerals a
provision that allows the nutrients to be omitted and replaced
with a statement ``Not a significant source of
&llmdash;&llmdash;&llmdash;&llmdash;&llmdash;'' when present in
insignificant amounts. An example of this shortened format is
given in appendix A of this document.
3. Calories From Saturated Fat
    9. Several comments agreed with the proposal that declaration
of calories from saturated fat should be voluntary. A few
comments suggested that this information should be mandatory and
referred to national dietary guidelines recommending that
saturated fat be limited to less than 10 percent of total daily
calories. A few comments requested that declaration of percent of
calories from saturated fat be made mandatory.
    Several comments believed that the declaration of calories
from saturated fat should be disallowed. The comments argued that
this information is redundant, confusing, and misleading.
    FDA acknowledges that research has established the role of
saturated fats in the etiology of atherosclerotic vascular
disease and recognizes that there are national consensus
recommendations regarding the levels of intake for saturated fat.
However,section 403(q)(2)(A) of the act permits the Secretary to
require the inclusion of information on additional nutrients in
nutrition labeling if he determines that such information ``will
assist consumers in maintaining healthy dietary practices.'' The
agend that the mandatory declaration of calories from saturated
fat or the percentage of calories from saturated fat meet this
criterion.
    First, this information may be obtained by simple calculation
if needed (i.e., calories from saturated fat can be calculated by
multiplying the g of saturated fat by nine, the number of
calories per g of fat; the percentage of calories from saturated
fat can then be determined by dividing the number of calories
from saturated fat by the total calories). Secondly, concerns
have been expressed in comments that consumers will be faced with
so much information that they will avoid using any of it.
Tominimize the possibility of this happening, FDA believes that
it is preferable to have consumers concentrate on the number of
calories from total fat. By controlling dietary intake of
calories from fat, intake of calories from saturated fat will
also be controlled.
    However, in recognition of dietary recommendations that
Americans should consume less than 10 percent of calories from
saturated fat (Refs. 3, 4, and 30) FDA is continuing to allow
voluntary declaration of calories from saturated fat
in101.9(c)(3)(ii)(A), redesignated as 101.9(c)(1)(iii).
4. Calories From Unsaturated Fat
    10. Several comments agreed with proposed101.9(c)(3)(ii)(B)
that the declaration of calories from unsaturated fat should be
voluntary. A few comments suggested that the declaration of
calories from unsaturated fat should be mandatory. These comments
stated that caloric information on unsaturated fat would be
helpful in monitoring unsaturated fat intake to maintain
consumption of unsaturated fat at not more than the 20 percent of
total calories.
    Several other comments suggested that this information be
disallowed because it will not be useful to the consumer in
evaluating a total day's food intake, and because the information
is potentially confusing.
    A few comments requested that declaration of calories from
monounsaturated and polyunsaturated fats be mandatory. One
comment requested that declaration of the percent of calories
from monounsaturated and polyunsaturated fats be mandatory. These
comments stated that caloric or percentage information on
monounsaturated and polyunsaturated fats would be helpful in
limiting consumption of each of these two classifications of
fatty acids to not more than 10 percent of total calories each.
    The agency has decided not to permit declarations regarding
calories from unsaturated fats because there is considerable
uncertainty and controversy about the term ``unsaturated fat''
and its definition, specifically whether the ``trans'' isomers of
monounsaturated fat should be included in this category of fats.
These isomers have been implicated in the development of coronary
heart disease and cancer (Ref. 31) and are discussed in the
subject health claims rules published elsewhere in thisissue of
the Federal Register.
    Further, the agency is not persuaded that it should allow the
voluntary declaration of calories from monounsaturated and
polyunsaturated fats. Definitions of monounsaturated and
polyunsaturated fats include cis isomers only. Trans isomers are
excluded. The declaration of calories from monounsaturated and
polyunsaturated fats would therefore underrepresent the total
caloric value of these fats because of the exclusion of the trans
isomers. Such an underrepresentation would be misleading to
consumers. Therefore, the agency is not allowing the declaration
of calories from polyunsaturated and monounsaturated fats in the
nutrition label.
    11. One comment suggested that 101.9(c)(3)(ii)(A) and
(c)(3)(ii)(B) be modified to clarify that when the declaration of
calories from saturated fat is declared adjacent to the
declaration of g of saturated or unsaturated fat, that it be in a
colu
    The agency has reconsidered the proposed format in the
supplementary proposal that would have allowed a separate column
for listing calories. As discussed in section V. of this
document, FDA is incorporating additional columns into the
nutrition label to declare the percent of daily value and the
daily value list. For this reason, the agency believes a column
headed ``calories'' would add to label clutter and, therefore,
has not made the suggested change. FDA has
modified101.9(c)(1)(ii) to delete the option that calories from
total fat be declared in a column headed ``calories.''
5. Calories From Carbohydrate
    12. Several comments requested that the declaration of
calories from carbohydrates be made mandatory so that consumers
can monitor and adjust their intake of calories from carbohydrate
to approach the recommended 50 to 60 percent of total calories. A
few comments requested that the declaration of percent of
calories from carbohydrate be made mandatory. Several comments
agreed with the proposal to allow the voluntary listing of
calories from carbohydrate. Several other comments requested that
FDA not permit the declaration of calories from carbohydrate
because this information is potentially confusing to consumers.
These comments suggested that this information would not be
helpful in evaluating a total day's food intake. A few comments
argued that too much information is burdensome to the consumer,
and that if it results in the manufacturer using smaller type
size, it could make the information more difficult for the
elderly to read. Further, the comments suggested that there was a
danger of ``information overload'' and ``label clutter.''
    Based on the comments to the July 19, 1990 and November 27,
1991, proposals, the agency has reconsidered its proposal to
permit the voluntary declaration of calories from carbohydrate
and has decided not to permit this declaration. As discussed in
the mandatory nutrition labeling proposal of July 19, 1990 (55 FR
at 29493), FDA's intent is to require the listing of only those
nutrients that present public health concerns and for which
quantitative intake recommendations have been made. FDA proposed
to permit the voluntary declaration of calories from carbohydrate
because of general recommendations that suggested that intake of
carbohydrate should be increased to 50 to 60 percent of total
calories but recognized that carbohydrate is not of pressing
public health significance.
    Based on its evaluation of the comments, FDA has become
concerned that it will overemphasize the public health
significance of carbohydrate if it allows the declaration of
calories from this nutrient. Additionally, the legislative
history of section (2)(b)(1)(C) of the 1990 amendments (Ref. 16)
makes clear that while FDA must allow the declaration of
additional nutrients in nutrition labeling, it must ensure that
such information does not interfere with the consumer's
understanding of the information required by the act. Thus, FDA
considers it important to ensure the comprehensibility of the
nutrition label.
    The 1990 IOM report ``Nutrition Labeling: Issues and
Directions for the 1990s'' (Ref. 1) emphasizes the importance of
considering information quantity and complexity when determining
the components of the food label (Ref. 1). The report suggests
that too much information compromises the ability of many
consumers to understand the label.
    The agency is persuaded that because the amount of calories
from carbohydrate is not of pressing public health significance,
it should not provide for inclusion of this information in
nutrition labeling. Accordingly, FDA has deleted proposed
101.9(c)(3)(ii)(C) from the final regulation.
    Consumers interested in determining the calories from
carbohydrate for the vast majority of individual foods may simply
multiply the number of g of carbohydrate by four, the number of
calories per g of carbohydrate. Consumers attempting to compare
thntake can use the Percent Daily Value format in the same way
described for monitoring fat intake. Because the Daily Value for
carbohydrate is set at 60 percent of calories, consumers need
only add the percent DV for total carbohydrate with a target of
100 percent or a target of a percentage adjusted for their
individual caloric intake. Alternatively, consumers can sum the g
of carbohydrate consumed for the day, multiply the total by four,
divide the result by the total calories consumed in that day, and
multiply by 100 to obtain percent.
    13. Although FDA chose not to propose the declaration of
calories from sugars and complex carbohydrates, a few comments
addressed this topic. Some of these comments stated that the
declaration of calories from sugars and complex carbohydrates
should be voluntary, and that this information, especially for
sugars, was of interest to consumers. Other comments felt that
the declaration of calories from sugars and complex carbohydrates
should be mandatory. Both sets of comments felt that this
information is potentially valuable to diabetics and parents of
young children who are concerned about dental caries and
excessive sugar intake. A few additional comments argued that the
declaration of calories from sugars and complex carbohydrates is
unnecessary and should not be permitted.
    Interest in having calories from sugars and complex
carbohydrates declared in the nutrition label was slight, and no
data were presented to support the requests for such information.
Further, dietary guidelines have not recommended specific
quantitative amounts for caloric intake from sugars or complex
carbohydrates. Therefore, the final rules do not permit the
inclusion of such information in the nutrition label. FDA advises
that the calculation of calories from sugars, which was of the
most interest to the comments, can be easily calculated by
multiplying the number of g of sugars present by four, the number
of calories per g of sugars.
6. Calories From Protein
    14. A few comments requested that the declaration of calories
from protein as well as the percent of calories from protein be
made mandatory to permit consumers to evaluate the quality of the
food. Other comments agreed with the proposal for the voluntary
declaration of calories from protein. On the other hand,
additional comments suggested that this information would be
confusing and misleading. These comments pointed out that
concerns about protein intake are of limited public health
significance in the United States and suggested that the
declaration of calories would not be helpful in evaluating a
total day's diet. The comments urged, therefore, that this
declaration should not be permitted. One comment suggested that
consumers would betempted to overconsume protein if calories from
protein were listed.
    Upon consideration of the comments, FDA has reassessed its
position. The agency agrees that the declaration of calories from
protein and the percent of calories from protein are of limited
usefulness to the consumer because the diets of the majority of
Americans exceed the Recommended Dietary Allowances (Ref. 23) for
protein. This lack of usefulness appears to outweigh any of the
potential benefits of allowing the declaration of calories from
protein. For this reason, and in an effort to reduce unnecessary
information that might interfere with the consumer's
understanding of required information, FDA is amending the final
regulations by deleting proposed101.9(c)(3)(ii)(D) which allowed
for the voluntary declaration of calories from protein.
    Consumers interested in determining the calories from protein
for an individual food may simply multiply the number of g of
protein by four, the number of calories per g of protein.
Consumers interested in determining the percent of calories from
pro
7. Increments for Calories
    15. The agency received only a few comments concerning the
proposed change in 101.9(c)(3) to delete the use of 2-calorie
increments for expressing caloric content up to and including 20
calories per serving. Most of the comments agreed with
theproposal which would express caloric content to the nearest
5-calorie increment up to and including 50 calories and to the
nearest 10-calorie increment above 50 calories. While one comment
disagreed with the proposed change to delete the 2-calorie
increments on the basis that it would permit less accurate
information for very low calorie foods, another comment
considered 2-calorie differences as inconsequential to the
consumer. Another suggestion was made to round all calorie levels
to the nearest 5-calorie increment.
    FDA is not persuaded by the comments that there is sufficient
reason to maintain the use of 2-calorie increments for foods
containing 20 or fewer calories or to use only 5-calorie
increments. FDA acknowledges the concern expressed about very low
calorie products. However, only a relatively small number of
products will be affected by the change. In fact, the agency
traditionally has been tolerant of slight differences in the
declared and actual amounts of calories. Current 101.9(e)(6),
redesignated as 101.9(g)(6), states that ``Reasonable
deficiencies of calories * * * under labeled amounts are
acceptable within current good manufacturing practice.'' Thus,
FDA is adopting this aspect of 101.9(c)(3), redesignated
as101.9(c)(1), as proposed.

C. Total Fat, Fatty Acids, and Cholesterol

1. Total Fat
    In the mandatory nutrition labeling proposal, FDA proposed to
require the declaration of fat, saturated fat, and cholesterol.
In addition, FDA proposed definitions for saturated fat,
unsaturated fat, polyunsaturated fat, and monounsaturated fat.
The agency did not define ``fat'' (i.e., total fat) for nutrition
labeling purposes. For compliance purposes, FDA has used as its
definition the sum of compounds with lipid characteristics that
are extracted by the Association of Official Analytical Chemists
International (hereinafter referred to as AOAC) methods or by
other reliable and appropriate analytical procedures (current
101.9(e)(2)).
    16. The agency received a number of comments concerning the
agency's standards for assessing total fat. A few comments from
food manufacturers and trade associations agreed with the
customary method of estimating dietary fat. Comments from other
food manufacturers, trade associations, college and university
nutrition professionals, consumer advocate groups, other
Government agencies, and foreign governments, disagreed with the
agency's position regarding the determination of fat content.
Some of these comments expressed uncertainty about what current
declarations of fat represent. It became evident that some
persons considered that the agency had implied a definition of
total fat as the sum of all triglycerides by stating in
current101.25(c)(2)(ii) that the amount of fatty acids was to be
calculated as triglycerides. This statement led some comments to
assume that mono- and diglycerides did not need to be included in
the declaration of total fat.
    Several comments suggested that the definition of fat should
include all dietary lipids, especially mono-, di-, and
triglycerides, phospholipids, and free fatty acids. These
comments pointed out that advances in food technology have led to
the development of fats and oils that reduce the triglyceride
content found in foods by replacing triglycerides with mono- and
diglycerides and phospholipids. These new forms of fats provide
calories and should be included in total fat values declared in
nutrition labeling.
    One comment suggested changing the definition of fat to
``substances possessing the physiological properties of fat.''
This comment stated that this definition would encompass all
types of dietary fats. Another comment suggested that the
definition be the ``sum of fatty acids from a total lipid
extraction.'' These comments pointed out that dietary lipids not
only contribute to the total dietary caloric intake but have
other physiologic functions attributable to fats. These functions
include transporting of lipids and fat soluble vitamins in the
body and structural functions in cell membranes, as well as
serving as essential fatty acids and as precursors of certain
hormones and eicosanoids.
    Several comments suggested that FDA's position on total fat
is not consistent with the definition found in Codex Alimentarius
or with that used by the Canadian government and the EC.
According to the comments, the international definition of fat
isnot restricted to triglyceride-releasable fatty acids but
includes total free fatty acids and other lipids, including
phospholipids.
    A couple of comments suggested that the definition of fat
should exclude some types of lipids (i.e., phospholipids, plant
sterols, and novel lipids) because these lipids constitute only a
small portion of total fat consumed, and, according to the
comments, these types of lipids have not been reported as having
a causal role in disease.
    The agency is concerned about the obvious confusion caused by
the lack of a precise definition for total fat. Because of the
importance given to dietary recommendations to reduce the intake
of total dietary fat, it is critical that all parties (i.e.,
Government agencies, food manufacturers, health professionals,
nutrition scientists, and consumers) clearly understand what the
values declared on the nutrition label represent.
    Concerns that the total fat value not be underrepresented
have persuaded the agency that it is not adequate to continue
only using a reference to AOAC methods or ``other reliable,
appropriate analytical procedures.'' Such an approach allows for
the use of many methods that measure different analytes. For
example, according to AOAC procedures, ``fat'' content can be
determined by ether or chloroform-methanol extraction. In the
case of an ether extraction, results yield a value for neutral
lipidswhich are primarily triglycerides (a complex lipid composed
of glycerol and three fatty acids) and some mono- and
diglycerides. In contrast, the chloroform-methanol extraction
method extracts all classes of lipids. The two methods, which are
both acceptable according to current regulations, may result in
different values for total fat being obtained for the same
product and different values being declared on the nutrition
label.
    The agency believes that the use of the implied definition of
total fat as the sum of triglyceride fatty acids from saturated,
polyunsaturated, and monounsaturated fatty acids, would in some
cases underestimate both total g of fat and the caloric intake
from fat. The agency agrees that all forms of fatty acids that
contribute to energy intake of foods should be included in the
calculation of total fat, particularly in view of dietary
recommendations that target total fat intakes at 30 percent or
less of calories.
    For these same reasons, the agency disagrees with the
suggestion that some lipids (e.g., mono- and diglycerides and
phospholipids) be excluded from the definition of fat.
    Therefore, the agency has decided to define total fat as
total lipid fatty acids, that is, the sum of fatty acids from
mono-, di-, and triglycerides, free fatty acids, phospholipid
fatty acids, and sterol fatty acids. This definition includes all
sources of fatty acids that provide energy, preventing
underestimates of energy from total fat. It also acknowledges
that certain lipid components, such as cholesterol and other
sterols, do not contribute metabolizable calories and constitute
only a verysmall amount of the total weight of lipids.
    This definition represents all fatty acids obtainable from a
total lipid extraction. The definition does not require that a
single extraction method be used. The extraction method will
depend upon the type of lipid being sought in the food and
thetype of food (i.e., the food matrix). Analytical procedures
are discussed further in section IV. of this document.
    The agency finds that this definition is more consistent
with, although not identical to, international definitions for
dietary fat. The Codex Alimentarius considers any source of
dietary energy to be a nutrient (that would include
nontriglyceride sources of fatty acids), and declaration of
dietary fat would reasonably include all sources of fatty acids.
The EC definition of fat is total lipids, including mono- and
diglycerides and phospholipids. The difference between the EC
definition and the agency's definition is that the agency
excludes the sterol fraction, not a large difference in
quantitative terms. Furthermore, the agency's definition reflects
dietary goals for consumption of only 30 percent of calories from
fat, because the sterols are not absorbed and therefore do not
contribute calories.
    However, the agency also recognizes that the definition of
total fat as total lipid fatty acids does not account for the
weight of glycerol to which the fatty acid chains are linked in
the formation of mono-, di-, and triglycerides. Unless the
glycerol is included in the weight of the total fat, it will be
reported as carbohydrate. In this case, foods in which the fat is
mostly triglyceride, e.g., corn oil and lard, will appear to have
95 percent total fat and 5 percent carbohydrate, while other
products such as muscle meats which have never been reported to
contain carbohydrate may now contain measurable amounts. These
values would conflict with common perceptions of food composition
because nutrient data bases and food composition tables routinely
include the weight of glycerol in the declaration of total fat.
    Therefore, the agency has decided to require that the
declaration of total fat be expressed as the amount of
triglyceride that would provide the analytically measured amount
of total lipid fatty acids in the food. This position is
supported by a recent report in The Referee, a publication of the
AOAC International (Ref. 32). Likewise, because food composition
data bases do not include glycerol in the declaration of fatty
acids (i.e., values represent free fatty acid), the agency is not
requiring that the amount of saturated fatty acids or other
classes of fatty acids be expressed as triglycerides (see comment
30 of this document).
    While the inclusive term ``total lipid fatty acids expressed
as triglycerides'' would be the more accurate term to use in the
nutrition label, the agency will continue to require use of the
term ``total fat'' to be consistent with the terminology used in
dietary recommendations and to avoid consumer confusion.
    17. Several comments from manufacturers, trade associations,
a consumer advocacy group, and a research firm addressed the
issue raised in the preamble to the supplementary proposal (56 FR
60366 at 60371) of the increased use of fats containing verylong
(longer than 18 carbons) chain fatty acids in the food supply.
These compounds provide the potential for marketing novel
compounds in which fatty acids are linked to carbon structures in
a manner that reduces their digestibility. As a result, these
compounds have the technical effect of fat with less calories
than traditional fats.
    Comments requested that fat be defined to exclude various
types of very long chain fatty acids because of their poor
absorbability and reduced digestibility. A recent article was
cited as evidence of the poor absorption of the very long chain
fattyacids (Ref. 33). One comment stated that the definition of
fat should exclude extractable compounds that do not have the
physiological effects of fatty acid compounds. Two comments
suggested the omission of these fatty acids from fat and calorie
declaraat have reduced digestibility and therefore fewer calories
than the fats they replace. The declared amount of fat would then
be the total analytically determined fat times the fat
digestibility coefficient.
    FDA acknowledges the effect that the use of certain very long
(longer than 18 carbons) chain saturated fatty acids with reduced
digestibility have on the fat and calorie content of foods. In an
effort to encourage innovation in the creation of products that
provide lower fat and lower calorie contents to enable the
consuming public to have a healthier diet and thus to meet one of
the primary objectives of the Surgeon General's report on
Nutrition and Health (Ref. 2), the agency is willing to consider
the digestibility of novel fat compounds. However, the agency has
concluded that because of the diversity of possible products, it
is not appropriate to modify the definition of total fat in
101.9(c)(2) to allow for alternate values because of reduced
digestibility of very long chain fatty acids. Rather, the agency
will address the digestibility of new ingredients containing
these fatty acids on a case-by-case basis.
    Because the digestibility of a substance is one of the
identifying characteristics of the substance, the agency requests
that manufacturers who wish to declare adjusted values of total
fat based on reduced digestibility include information on the
digestibility of the compound, analytical assay procedures for
the compound, and data on interference with required methods of
analysis in food additive petitions (21 CFR part 171) or in
petitions for affirmation that the use of the substance is GRAS
(21 CFR 170.35). The agency will include the specific
digestibility coefficients that can be used in determining the
quantitative declaration of fats and the caloric contribution
from fats as part of the statement of identity for the substance
in the listing regulation in part 172 for food additives and in
part 184 for substances whose use is affirmed as GRAS. However,
FDA recognizes that mechanisms other than food additive or GRAS
petitions may be appropriate to bring issues involving the
digestibility of a substance to the attention of the agency.
Interested persons may wish to use the mechanism in 101.9(g) to
request to use specific digestibility coefficients.
    18. Several comments said that there is a need for adequate
analytical methods for assaying novel forms of fat in new low-fat
foods. They noted the difficulty of isolating new ingredients by
the traditional or AOAC determinations for fat. As one comment
stated, the current acid hydrolysis analysis may not be
appropriate for these type of substances.
    FDA agrees with the concern about analytical methodology and
is aware that different methods might be needed for each product
or type of product. Because of this concern, and as noted in the
preceding comment, the agency finds that it is necessary that
manufacturers delineate, in the documentation submitted to FDA in
support of the lower fat content declarations, the methodology
needed to assay for the novel fat compound. Use of the method by
the manufacturer and the agency in lieu of conventional methods
found in the AOAC or other recognized sources should alleviate
labeling compliance concerns.
    19. One comment urged the agency to allow manufacturers to
use calculations from product formulas to arrive at the calorie
or fat content of products containing these alternate
ingredients.
    As discussed in previous comments, the supporting
documentation submitted to the agency to substantiate different
caloric levels for novel fats should contain adequate information
regarding the digestibility coefficient, analytical methodology,
and other factors to ascertain an accurate label value for the
fat and calorie declarations. FDA will use the information and
analytical methodology for each such fat to determine whether the
values for fat and calories stated on the label are correct.
Manuf product formulas, to determine fat and calorie values if
they have a reasonable basis on which to believe that the values
so obtained will be consistent with values determined
analytically. However, they do so at the risk that FDA will
disagree.
2. Saturated Fat
    a. Definition. FDA proposed (55 FR 29487 at 29495) to make
the amount of fat, saturated fatty acids, and cholesterol
mandatory elements of nutrition labeling. At the same time, FDA
proposed in 101.9(c)(4)(i) to continue to define saturated fatty
acids as the sum of lauric (C12:0), myristic (C14:0), palmitic
(C16:0), and stearic (C18:0) acids, the major saturated fatty
acids in the U.S. food supply. FDA requested comments on the
questions of what fatty acids should be considered as saturated
fatty acids, and on what basis these decisions should be made (55
FR 29487 at 29495).
    20. Many comments, including comments from food manufacturers
and distributors, trade associations, professional organizations,
nutritionists, and state health departments agreed with FDA's
proposal to include only the four saturated fatty acids with 12
to 18 carbons (lauric, myristic, palmitic, and stearic acids) in
the definition of saturated fat for labeling purposes. The
reasons given included: (1) Lauric, myristic, palmitic, and
stearic acids comprise the vast majority of saturated fatty acids
in the American diet; (2) C12:0 - C16:0 fatty acids raise total
and low density lipoprotein (LDL)-cholesterol; (3) although some
clinical and metabolic evidence suggests that C18:0 fatty acid
(stearic acid) does not have the same blood total and
LDL-cholesterol-raising effect as C12:0 - C16:0 saturated fatty
acids, the effect of C18:0 on blood total and LDL-cholesterol is
not conclusive enough to warrant deletion of C18:0 from the
definition; (4) stearic acid may be associated with other risk
factors of cardiovascular disease such as thrombosis and platelet
aggregation; and (5) a separate analysis for stearic acid would
be costly.
    Many other comments, including comments from other food
manufacturers, trade associations, and health professionals,
suggested that FDA include only lauric acid (C12:0), myristic
acid (C14:0), and palmitic acid (C16:0) in the definition of
saturatedfat for labeling purposes. These comments stated that
there is no evidence that stearic acid has a cholesterol-raising
effect, and that the postulated role of stearic acid in
thrombosis is open to dispute. With regard to the latter point,
the comments cited the conclusion of a workshop on Dietary Fatty
Acids and Thrombosis that there is no direct evidence of a
prothrombotic effect of long chain fatty acids (e.g., C18) in
humans (Ref. 34).
    One comment from a trade association suggested that FDA not
include palmitic acid in the definition of saturated fat because
palmitic acid does not raise blood total and LDL-cholesterol. The
comment cited recent research articles (Refs. 35 and 36) as the
evidence.
    A comment from a major food manufacturer suggested that FDA
not include saturated fatty acids with less than 10 carbons in
the saturated fat category. Other comments suggested FDA exclude
lauric acid or myristic acid from the saturated fat categoryfor
labeling purposes. The reasons given were: (1) That the medium
chain fatty acids (C6:0 - C10:0), lauric acid (C12:0), and
myristic acid (C14:0) are readily absorbed and oxidized and may
not raise blood total and LDL-cholesterol, and (2) that medium
chain fatty acids are minor sources of saturated fat in the
American diet. In contrast, a consumer and a state agency
suggested that FDA include saturated fatty acids with carbon
numbers less than 12 in the saturated fat category because they
may elevate blood cholesterol.
    Some comments from a major food manufacturer and a state
public health department stated that saturated fatty acids with
carbon chains longer than 18 (i.e., C20 - C24) should not be
categorized as saturated fatty acids. The reasons given included:
(a small part of saturated fat content in the U.S. diet, and (2)
these fatty acids are poorly absorbed, have no or little
physiological effects, and therefore do not contribute to heart
disease.
    Several comments, from a food manufacturer, a consumer, and
foreign governments, suggested that FDA use a chemical definition
of saturated fatty acids for labeling purposes. Two other
comments from the meat industry suggested that FDA use a chemical
definition if it is not possible to set a definition on the basis
of the relationship of saturated fatty acids to the risk of
cardiovascular disease. The reasons included in these comments
were:
    (1) Underrepresentation of the content of saturated fat.
Several comments stated that the proposed definition, limiting
saturated fat to only the four saturated fatty acids with 12 to
18 carbons, would result in underrepresentation of the saturated
fatty acid content of foods, particularly of those foods that
contain significant amounts of saturated fatty acids with less
than 12 carbons or more than 18 carbons. They further stated that
this underestimation of saturated fat contradicts the current
dietary recommendation that Americans consume less than 10
percent of calories as saturated fats. The examples that they
presented of foods in which the definition of saturated fat as
C12 - C18 would underrepresent saturated fat were milk,
underrepresented by 8 percent; coconut oil, by 14 percent; and
palm kernel oil, by 7.2 percent.
    (2) Oversimplification. A consumer stated that FDA's proposal
is an oversimplification and suggested that all saturated fatty
acids be included.
    (3) International harmonization. The Canadian Government and
the EC stated that FDA's proposal to restrict the definition of
saturated fat to only lauric, myristic, palmitic, and stearic
acids is at variance with the Codex Alimentarius Commission,EC,
and Canadian definitions. They stated that the proposed
deviations from international definitions present serious
problems for food companies in the EC and would be confusing for
consumers. They suggested that FDA include all saturated fatty
acidswithout double bonds in the saturated fat definition. A
major food manufacturer also stated that it already had
encountered minor problems with different definitions in labels
for Canada and the United Kingdom and suggested that FDA consider
international harmonization in its decision on the definition of
saturated fat.
    FDA is persuaded by the comments that there is substantial
controversy regarding the inclusion or exclusion of specific
fatty acids in a definition of saturated fat that is based on
effects on blood total and LDL-cholesterol.
    Therefore, the agency has reconsidered its position of
linking the definition of saturated fatty acids to effects of
particular fatty acids on blood total and LDL-cholesterol levels
and has determined that a chemical definition is appropriate for
the following reasons.
    First, a chemical definition avoids much of the controversy
regarding the blood cholesterol effects of palmitic acid, very
long (longer than 18 carbons) chain fatty acids, and short and
medium chain fatty acids, because the definition is not subjectto
changes in knowledge about the physiologic effects of particular
fatty acids. The chemical definition also avoids uncertainties
about physiologic effects other than those regarding blood total
and LDL-cholesterol, such as effects on thrombosis, and other
possible health effects of very long chain, medium chain, and
short chain saturated fatty acids.
    Secondly, FDA agrees that the amount of saturated fat in some
foods could be underrepresented when the definition of saturated
fat is confined to the sum of four fatty acids. This
underreporting of saturated fat may be increased as new foods
containi
    Thirdly, the agency also notes that the chemical definition
is in line with EC and Canadian definitions of saturated fat and,
hence, will promote international harmonization.
    Finally, the agency notes that the chemical definition of
saturated fat is more consistent with dietary recommendations to
reduce fat consumption to 30 percent of calories and saturated
fat consumption to less than 10 percent of calories. Food
composition data tables have generally been used in epidemiologic
investigations that relate diet to risk of chronic diseases, and
these tables group all the chemically defined saturated fatty
acids together as a class. Thus, the term ``saturated fat'' used
in these dietary recommendations pertains to the chemical
classification of fatty acids, not FDA's current, more restricted
definition.
    Accordingly, FDA has amended 101.9(c)(4)(i), redesignated as
101.9(c)(2)(i), to define saturated fat as the sum of all fatty
acids containing no double bonds.
    b. Voluntary declaration of specific saturated fatty acids.
21. Some comments specifically requested that the agency provide
for labeling that distinguishes those fatty acids associated with
increased blood total and LDL-cholesterol levels from those not
associated with increased cholesterol. One approach that was
identified in the comments and in a published commentary (Ref.
13) would allow a declaration of ``cholesterol-raising fatty
acids,'' so that a manufacturer could show that a particular food
contained little or none of these fatty acids.
    The agency recognizes that there is substantial uncertainty
as to which saturated fatty acids are cholesterolemic and which
are not. Conclusions of authoritative documents and review papers
are not consistent on this issue (Refs. 2, 3, 37 through 41). The
effects of most individual saturated fatty acids on blood total
and LDL-cholesterol are not fully understood. The agency finds
that the only saturated fatty acid that has been consistently
reported as cholesterol-raising is myristic acid. The effects of
palmitic acid and lauric acid are not as clearly associated with
increased blood cholesterol, although the prominence of palmitic
acid in the diet makes any contribution of this fatty acid
important in the control of blood cholesterol. On the other hand,
it has consistently been reported that stearic acid, when
substituted for other saturated fatty acids in the diet, has a
neutral effect on blood total and LDL-cholesterol concentration
(Refs. 37 through 40). As a result, the agency is concerned that
there is not an adequate basis for deciding which fatty acids
should be included in the term ``cholesterol-raising fatty
acids.''
    In addition, the agency is concerned that the term
``cholesterol-raising fatty acids'' will be confusing to
consumers. Since consumers are unfamiliar with the term
``cholesterol-raising fatty acids,'' there is a possibility that
they would misinterpret it and would avoid foods with such a
declaration on the nutrition label, even if the intent of the
labeling was to indicate the absence of these fatty acids. Also,
given that the only fatty acid declaration the agency is
requiring is saturated fat (defined as the sum of all saturated
fatty acids), any added declaration of ``cholesterol-raising
fatty acids'' would be on a voluntary basis. Under these
circumstances, manufacturers could be expected to only include
this declaration when the level of such fatty acids is low to
emphasize the absence of such components from the product.
    A variation of this term that avoids the negative connotation
and applies positively to the composition of the product is the
term ``noncholesterol-raising fatty acids.'' However, this term
suffers from the other problems with respect to
cholesterol-raising fatty acids (i.e., the scientific uncertainty
concerning what fatty acids to include and the likelihood of
increased consumer confusion).
    FDA is also concerned that either of the terms
``cholesterol-raising fatty acids'' or ``noncholesterol-raising
saturated fatty acids'' could be seen as a health claim. Section
403(r)(1) of the act states that information that is required or
permitted under 403(q) of the act to be included in the nutrition
label is not a nutrient content or health claim. Because of the
relationship between fatty acids and increased blood cholesterol
and, thereby, heart disease, however, the agency is concerned
that the use of either of the subject terms goes beyond the
factual reporting of nutrients that is characteristic of the
nutrition label.
    For the reasons enumerated above, FDA has concluded that it
is not appropriate to distinguish among fatty acids by the terms
``cholesterol-raising fatty acids'' or ``noncholesterol-raising
saturated fatty acids.''
    Another approach to distinguishing among fatty acids is to
declare specific saturated fatty acids without any reference to
effects on blood total and LDL-cholesterol. This approach is
consistent with the agency's intention of providing factual
information on the nutrition label. Because some comments
strongly opposed the inclusion of stearic acid in the declaration
of saturated fat because of the consumer's association of
saturated fat with increased blood cholesterol levels, it is
reasonable to indicate the extent of the saturated fat content of
the food that is stearic acid and, thus, not associated with
increased blood cholesterol. As noted above, a consensus that
seems to be emerging is that stearic acid, when substituted for
other saturated fatty acids, does not raise or lower blood total
and LDL-cholesterol level. Consumer education programs could
advise consumers that when a large portion of the saturated fat
in a product consists of stearic acid, the fat content of the
food is not likely to increase blood total and LDL-cholesterol
levels.
    The agency, however, has some reservations about allowing for
the voluntary labeling of stearic acid in that: (1) Other
saturated fatty acids that may raise blood total and
LDL-cholesterol are not addressed; (2) only one risk factor of
cardiovascular disease, blood cholesterol level, is addressed;
(3) it may complicate and overcrowd the label; and (4) it would
require a consumer information program to have any meaning to
consumers.
    In addition, recognizing particular saturated fatty acid
effects on blood cholesterol may require that the agency redefine
the saturated fat threshold criterion for cholesterol claims in
101.62(d) (cholesterol claims are not allowed on foods
containing more than 2 g saturated fat, defined as the sum of all
fatty acids containing no double bonds, per serving), as
described in a companion document on nutrient content claims
published in this issue of the Federal Register. Because of the
agency's reservations about the meaningfulness of labeling of
individual fatty acids and the need to reconsider criteria for
cholesterol claims if such action was to be taken, the agency
concludes that more information, including public comment, is
necessary before taking further action on this approach. The
agency intends to further address this issue at a later date, and
would welcome submission of information and views on this
question.
3. Polyunsaturated and Monounsaturated Fat
    a. Use of the term ``unsaturated fat''. FDA proposed in both
the July 19, 1990, and November 27, 1991, documents in
101.9(c)(4)(ii) to permit the voluntary declaration of the
quantitative amount of unsaturated fat in nutrition labeling. The
agency proposed to make the declaration of unsaturated fat
mandatory if claims were made about fatty acid or cholesterol
content or if the manufacturer voluntarily declared the number of
calories from unsaturated fat. Alternatively, the agency proposed
to allow separate declarations of polyunsaturated and
monounsaturated fats.
    22. The agency received comments that either agreed or
disagreed with the proposed definition and voluntary use of the
term ``unsaturated fat''. Comments that supported the use of the
inclusive term did so because neither monounsaturated nor
polyunsand because both types of unsaturated fats decrease the
risk of coronary heart disease relative to saturated fat.
    Comments objecting to the term ``unsaturated fat'' argued
that the term is not useful, that it offers no additional
information that could not be obtained by subtracting the
saturated fat content from total fat, and that it obscures the
presence of essential fatty acids. Other comments were concerned
that the term was misleading in that it suggests that all
unsaturated fats are synonymous by including both cis and trans
isomers and both poly- and mono-unsaturated fats together. These
comments argued that in light of the current uncertainty and
controversy surrounding the physiological effects of trans fatty
acids (which are a particular type of unsaturated fatty acid
having some physical properties of saturated fatty acids), use of
the term ``unsaturated fat'' would not only be misleading to the
consumer but possibly could have an adverse effect on the health
of some individuals. This opposition to the inclusion of trans
isomers in the definition of ``unsaturated fat'' was the most
frequently repeated concern.
    Comments suggesting that it would be misleading to group
poly- and mono-unsaturates together argued that this action would
imply that the two types of fatty acids are the same and have
similar effects on blood total and LDL-cholesterol when, in fact,
they do not. It has been reported that monounsaturates do not
effect blood total and LDL-cholesterol levels and do not reduce
high density lipoprotein-cholesterol when substituted for
saturated fats. On the other hand, polyunsaturates have often
been reported to reduce blood total and LDL-cholesterol levels
and to decrease blood pressure.
    A few comments suggested that if the term ``unsaturated fat''
is permitted, the declaration of the cis forms of polyunsaturates
and monounsaturates should either be permitted or required at the
same time. Comments also argued that there is noscientific
consensus supporting the use of the inclusive term, and that it
was not a term used in international trade.
    FDA is persuaded by these comments that the use of the term
``unsaturated fat'' is potentially confusing to consumers, does
not provide useful information, and could result in consumer
deception. Accordingly, the agency is revising the regulation by
not providing for the voluntary declaration of unsaturated fat in
nutrition labeling. As a result, the proposed listings of
polyunsaturated fat and monounsaturated fat in101.9(c)(4)(ii)(A)
and (c)(4)(ii)(B), respectively, are redesignated as
101.9(c)(2)(ii) and (c)(2)(iii). In addition, each paragraph has
been modified to incorporate provisions that had been included in
proposed 101.9(c)(4)(ii). The revised listings provide that the
disclosure of the level of polyunsaturated fat and
monounsaturated fat is voluntary unless claims are made on the
label about fatty acid or cholesterol content, and that, if
either polyunsaturated or monounsaturated fat is declared, the
other must also be declared.
    b. Trans Fatty Acids. In its July 19, 1990 proposal on
mandatory nutrition labeling (55 FR 29487 at 29496), the agency
tentatively concluded that there is no basis for declaring trans
isomers of fatty acids on the nutrition label. This conclusion
was based on a consensus report that noted that current evidence
does not support a blood cholesterol-raising effect for trans
isomers when they are substituted for saturated fatty acids in
the diet. The agency requested comments on this issue. Later that
year, new research and commentary was published (Refs. 12 and 13)
which led FDA to request in the supplementary proposal (56 FR
60366 at 60371) comments on the significance of the new findings
and a reevaluation of any comments submitt
    23. Several comments, from a major food manufacturer, health
professionals, a professional health organization, a state
agency, a trade association, and a consumer, suggested that FDA
include trans fatty acids in the saturated fat category because
research suggests that trans fatty acids raise blood total and
LDL-cholesterol. On the other hand, several comments were against
the inclusion of trans fatty acids in the saturated fat category
because the evidence of a cholesterol-raising effect of trans
fatty acids is not conclusive. Several comments suggested that
trans fatty acids should be declared separately because it may
increase blood cholesterol.
    FDA disagrees that there is sufficient evidence that
indicates that trans fatty acids raise blood total and
LDL-cholesterol. In 1985, a report of the Federation of American
Societies for Experimental Biology on ``Health Aspects of Dietary
trans Fatty Acids''(Ref. 42) concluded that human studies
indicate that trans isomers are little, if any, more
cholesterolemic than cis isomers. In animals (rabbits, swine, and
monkeys), trans fatty acids are cholesterolemic but
notatherogenic. Since the publication of the Federation of
American Societies for Experimental Biology report, a scientific
review article, (Ref. 40) concluded that reports are inconsistent
regarding the effects of trans unsaturates on blood total and
LDL-cholesterol levels in humans.
    Recently, two studies in The Netherlands (Refs. 12 and 43)
have shown that a high intake of trans fatty acids may elevate
blood total and LDL-cholesterol concentration. Concerns have been
raised about the applicability of these studies to U.S. diets
because of certain methodologic limitations, because the level of
trans fatty acid tested was 2 - 3 times higher than the current
average consumption of the U.S. population, and because the
methods for generating trans fatty acids might have been
different from those used in the United States.
    In contrast, another study (Ref. 44) seems to indicate that
trans fatty acids do not raise blood total and LDL-cholesterol in
mildly hypercholesterolemic, normotensive men, although diet
differences other than trans fatty acids may have been
responsible for the effects.
    Finally, the agency is aware of preliminary results from a
very recent unpublished study designed to address the criticisms
of the studies from The Netherlands (Ref. 45) that suggests that
trans fatty acids raise LDL-cholesterol.
    In the absence of the fully analyzed data from this study,
the agency considers it premature to require the labeling of
trans fatty acids because of their effects on total or
LDL-cholesterol. However, even if there was a need for labeling
of trans fatty acids, the agency does not agree that trans fatty
acids should be included in the category of saturated fats. The
agency has argued against inclusion or exclusion of particular
saturated fatty acids in the definition of saturatedfat solely on
the basis of their effect on blood total and LDL-cholesterol. In
addition, the agency recognizes that inclusion of trans fatty
acids in the definition of saturated fat is not consistent with
the EC, Codex, or Canadian definitions ofsaturated fat.
    Because of the current uncertainties, the agency does not
agree that a separate declaration of trans fatty acids is
appropriate at this time. Because new data are rapidly emerging
(Ref. 45) that imply that trans fatty acids raise
LDL-cholesterol, however, the agency recognizes that it may be
necessary to readdress the labeling of trans fatty acids in the
near future.
    24. One comment suggested that not all foods voluntarily
declaring levels of monounsaturates and polyunsaturates need to
be analyzed to differentiate cis and trans fatty acids because
only those containing hydrogenated fats would contain     The
agency agrees with this comment in the case of vegetable oils and
other plant lipids. However, naturally occurring trans fatty
acids are found in some animal lipids (e.g., dairy products). If
there is adequate and reliable reason to believe that a nutrient
is not present in a food, there is no need to analyze for that
nutrient. However, a manufacturer is responsible for ensuring
that its labeling is truthful and not misleading.
    c. Definition of polyunsaturated fatty acids. FDA proposed in
101.9(c)(4)(ii)(A) and (c)(4)(ii)(B) in July 19, 1990 and
November 27, 1991 to define polyunsaturated and monounsaturated
fats as cis, cis-methylene interrupted polyunsaturated fatty
acids and cis-monounsaturated fatty acids, respectively. These
definitions exclude trans isomers. The definition of
polyunsaturated fat is consistent with
current101.25(c)(2)(ii)(a). FDA has not previously defined
monounsaturated fat for labeling purposes.
    25. Comments from food manufacturers, trade associations,
health promotion organizations, consumer groups, and
international agencies supported the agency's definition for
polyunsaturated fat. However, some comments urged FDA to change
the definition to reflect the various physiological roles of
specific types of polyunsaturated fats, in particular to allow
for the identification of omega-3 (n - 3) and omega-6 (n - 6)
fatty acids, indicating that these are essential in the diet and
that the ratioof consumption of these fatty acids can be
important.
    FDA is not persuaded that there is a need to require further
breakdown of polyunsaturated fats in the nutrition label. As
discussed above, FDA is concerned that additional information on
the nutrition panel may confuse consumers and interfere with
their understanding of other required information.
    However, the agency agrees that there are valid reasons to
consider the voluntary labeling of omega-3 and omega-6 fatty
acids. These chemical distinctions are important nutritionally,
because the omega-3 fatty acids (with the first double bond at
the third carbon from the methyl end of the fatty acid) and
omega-6 fatty acids (with the first double bond at the 6th carbon
from the methyl end) are not interchangeable during metabolism in
the body; rather each must be supplied by diet. Each subcategory
has members that are considered essential nutrients (]-linolenic
acid and linoleic acid, for the omega-3 and omega-6 classes,
respectively) (Ref. 23). Dietary omega-3 and omega-6 fatty acids
are precursors for biologically active compounds, e.g.,
prostaglandins, eicosanoids, and the nutritional balance of
omega-3 and omega-6 fatty acids modulates the production of many
of these biologically important substances. Furthermore, omega-3
and omega-6 fatty acids are important components of cell
membranes.
    Although the National Research Council has not yet
established a Recommended Dietary Allowance for omega-3 and
omega-6 fatty acids, they recognized in 1989 that ``[T]he
possibility of establishing Recommended Dietary Allowance's for
these fatty acids should be considered in the near future.''
(Ref. 23)
    FDA agrees that information on the amount of omega-3 and
omega-6 fatty acids may be useful to allow interested consumers
to select foods that provide these fatty acids. It is not
difficult to consume a diet rich in omega-6 fatty acids because
vegetable oils are rich in these fatty acids. However, vegetable
oils vary widely in their content of omega-3 fatty acids, and
labeling may be useful to identify those foods that contain
substantial amounts of omega-3 fatty acids to encourage a
balanced intake of these two classes of fatty acids.
    However, FDA is not fully persuaded about the usefulness of
additional label information on omega-3 and omega-6 fatty acids,
and whether there are many consumers who desire this information.
As discussed above, the agency is concerned that additional the
voluntary declaration of these subcomponents of polyunsaturated
fats at this time. The agency intends to address this issue at a
later date and would welcome submission of information and views
on this issue.
    26. The majority of comments supported the voluntary
declaration of polyunsaturated fats. However, a few comments
suggested that their declaration be mandatory rather than
voluntary. One of these comments was concerned with possible
safety issues associated with increased consumption of
polyunsaturated fats. The comment alleged that polyunsaturated
fats convey a potential source of free-radical peroxidation
products, and that consumers should be informed of the amounts of
polyunsaturated fat present in a food. Other comments merely
stated that mandatory declaration of polyunsaturated fat would
provide consumers with valuable and more complete nutritional
information.
    FDA is not persuaded that there is a need to require the
inclusion of polyunsaturated fats on the nutrition label. These
fatty acids do not meet the criteria for mandatory declaration
set forth in the mandatory nutrition labeling proposal (55 FR
29487 at 29493) that the nutrient or food component be of
particular public health significance, and that quantitative
intake recommendations for the nutrient be given in major
scientific consensus reports. The comments largely support this
view, and the agency therefore rejects the suggestion that the
declaration of polyunsaturated fats be mandatory.
    The agency disagrees with the contention that commonly
consumed amounts of polyunsaturated fats would pose any safety
concerns. While the potential exists for formation of oxidative
products as a result of the increased number of double
bonds--present in polyunsaturated fats, any risk would only occur
at very exaggerated levels of consumption.
    d. Monounsaturated fats.
    27. Comments both agreed and disagreed with the proposed
definition of ``monounsaturated fat.'' Those opposed generally
requested that the definition not exclude trans fatty acids on
the basis that they have not been proven to have an adverse
effect on health or disease in humans, or that cis and trans
isomers have similar metabolic and physiologic properties. One
comment asked the agency to include trans fatty acids in the
definition of monounsaturated fats until an expertpanel can
determine if trans or unusual cis isomers formed as components of
commercial hydrogenation increase the risk for coronary heart
disease or other health related conditions.
    Comments from medical associations, trade associations, a
consumer advocacy group, the Canadian government, and the EC
supported the proposed definition, focusing particularly on two
considerations. First, trans fatty acids or unusual cis isomers
formed during commercial hydrogenation of unsaturated fats may
increase the risk of coronary heart disease, and, second,
monounsaturated fats, as defined in the proposal, may reduce
blood total and LDL-cholesterol and reduce the risk of coronary
heart disease. Comments also suggested that the inclusion of
trans fatty acids may mislead consumers, who perceive
monounsaturates as healthful or at least as not harmful.
    FDA concludes that there is no need to amend the proposed
definition of ``monounsaturated fat.'' The comments received, the
scientific reports they discuss (Refs. 12 and 43), and the
concerns addressed in the preceding discussion of trans fatty
acids establish that to include trans isomers in the definition
of ``monounsaturated fat'' would be misleading and will not
assist consumers in maintaining healthy dietary practices. The
agency is not willing to include trans isomers in the label
definition of ``monounsaturated fat'' until there is further
consensus based on publicly available, well-designed, and
well-conducted studies. However, as more data concerning the
action and safety of trans fatty acids become available, ther its
decisions to define monounsaturates as the usual
cis-monounsaturated fatty acids.
    28. One comment also objected to the proposed definition of
``monounsaturated fat'' because it would require manufacturers to
conduct a further analysis of lipids to differentiate between cis
and trans isomers. The comment argued that this extra analysis
was not justified by available scientific data and would cause a
financial burden.
    The agency disagrees with the comment. The declaration of
poly- and mono-unsaturates is voluntary. Therefore, an analysis
of unsaturated isomeric forms is only required if the
manufacturer chooses to declare poly- or mono-unsaturates or to
make fatty acid or cholesterol claims. In such cases, given the
controversy on the effect of trans fatty acids, the additional
analysis is necessary to ensure that the declaration or claims
are not misleading.
    29. One comment suggested that the agency should include
stearic acid in the definition of monounsaturated fat by virtue
of its effects on blood total and LDL-cholesterol. The comment
stated that scientific data suggests that stearic acid does
notincrease blood LDL-cholesterol, and that it is rapidly
converted to oleic acid, an unsaturated fat that does not raise
blood total and LDL-cholesterol levels.
    FDA does not agree that stearic acid should be included in
the definition of monounsaturated fat. Chemically, stearic acid
is a saturated fat, and the agency, therefore, finds that it
would be inappropriate to include it with monounsaturated
fats.The agency has acknowledged above that some studies and some
consensus statements suggest that stearic acid does not increase
LDL-cholesterol relative to other saturated fats. However,
stearic acid does increase LDL-cholesterol relative to
monounsaturates and polyunsaturates (Refs. 12 and 43).
Accordingly, the agency is not including stearic acid in the
definition of monounsaturated fat.
4. General Issues Related to Declaration of Fats and Fatty Acids
    a. Calculation of fatty acids as triglycerides.
    30. A comment was received that disagreed with proposed101.9
(c)(4)(i) and (c)(4)(ii), which would require saturated fat and
unsaturated fat content to be calculated as triglycerides. The
comment noted that values in current data bases are reported as
the free fatty acids.
    Current 101.25(c)(2)(ii) requires that fatty acids be
calculated as triglycerides. This requirement dates back to the
initial nutrition labeling regulation promulgated in 1974. This
requirement was a result of comments from industry at that time.
    To provide consumers with nutrition information that can
readily be used for comparison to available nutrient data bases,
FDA agrees that saturated and poly- and mono-unsaturated fat
should be declared as free fatty acids instead of as
triglycerides. As a consequence of the change in method of
reporting, slightly lower values for the various fatty acid
declarations will appear on the label because the weight of the
glycerol molecule in triglycerides is not included when free
fatty acids are declared. Also, fatty acids from mono- and
di-glycerides used as a source of fat in many products will be
included using this revised means of reporting fatty acids.
Accordingly, FDA is amending 101.9(c)(2)(i) for saturated fat
and 101.9(c)(2) (ii) and (iii) for polyunsaturated and
monounsaturated fat, respectively, to remove the requirement that
the fatty acids be ``calculated as triglycerides.''
    b. Increments for declaring fats and fatty acids.The
mandatory nutrition labeling proposal retained the current
requirement for the declaration of fat in g and added, as a
requirement, the amount of saturated fatty acids in g (55 FR
29487 at 29495). In the supplementary proposal, FDA proposed to
changeted fat, and monounsaturated fat in 0.5 (\1/2\)-g
increments. The agency made this change in the proposed
provisions to make the increments in which these nutrients are
declared more consistent with the levels at which these
substances will have nutritional significance. FDA believed the
proposed change would consequently provide consumers with more
precise information and a greater ability to discriminate among
products. In this context, a level of less than 0.25 g per
serving was established as the level at which saturated and
unsaturated fat content would be expressed as zero.
    31. Over 25 responses concerning fat increments were received
in response to the request for comments. Almost twice the number
of comments disagreed with the proposal to declare total fat and
fatty acids in 0.5-g increments as those who agreed with the
proposal to do so. The rationale given by essentially all who
disagreed with the proposed change was the lack of analytical
methods that are adequate and sensitive enough to provide data to
that degree of precision. Several comments recommended that the
fat content of foods containing 3 or less g fat be declared in
0.5 (\1/2\)-g increments, and the fat content of foods containing
more than 3 g be declared in whole g increments. These comments
suggested that the precision of 0.5 g increments forfat
declarations is less important for higher fat foods.
Additionally, these comments stated that the variability of some
fat assays warrants whole-g increments, especially for moderate
and high levels of fat. It should be noted, however, that
severalcomments stated that methodology does exist to support the
0.5-g increment declaration. One comment noted the desirability
of keeping all macronutrients, including fat, in whole-g
increments. Several comments cited the cost of assaying to the
0.5 (\1/2\)-g level of precision as a reason for retaining the
whole-g increment declarations for these nutrients.
    FDA has given careful consideration to the comments. The
agency recognizes that labeling requirements must not only convey
desired nutrition information for the consumer but must also be
enforceable. Because of concerns about analytical
precision,variability, and the effect of product matrices on the
methods necessary to quantify total fat, saturated fat, and poly-
and mono-unsaturated fat declarations in 0.5-g increments, FDA
has concluded that such precision is not necessary for amounts of
fat above 3 g per serving of food. Consequently, the agency is
modifying 101.9(c)(2) and (c)(2)(i) through (c)(2)(iii) to
require that levels below 3.0 g per serving be declared in 0.5
(\1/2\)-g increments and levels above 3.0 g be declared in g
increments.
    The agency disagrees that cost, although a factor, is a
sufficient reason in and of itself to retain the current whole-g
increments for total fat, saturated fat, and poly- and
mono-unsaturated fats. The public health benefits attributed to
decreasing dietary intakes of fat (Refs. 2, 3, 4, and 47) justify
the use of 0.5-g increments to allow consumers to differentiate
between products containing low levels of fats.
    32. A few comments urged that fats not be declared in 0.5
(\1/2\)-g increments to improve the legibility of the label.
    The agency is concerned about the legibility of the label.
However, because of the public health significance of dietary
intake of fats, FDA believes it is important to provide the
increased precision at low levels of fat. Inasmuch as legibility
ismore dependent upon factors such as type size and color
contrast than the addition of a decimal point and digit, FDA
urges manufacturers to consider the readability factor and use
great care to ensure that the information is legible.
    33. Two comments requested that the agency permit the
declaration of total fat and saturated fat in tenths of a g.
    FDA does not agree. It is not possible to require the
declaration of total fat and saturated fat in tenths of g
increments because this degree of precision cannot be reliably
obtained in all foods with available methodology.
    34. Comments stated that the change from whole-g increments
would be confusing and cumbersome to consumers. One comment
requested that the agency adopt a consistent rule for all
macronutrients by rounding values to the nearest g.
    FDA does not agree that the use of different increments for
different nutrients will be confusing and cumbersome. These final
rules allow for calories to be declared to the nearest 5 or 10
calorie increment depending on amount, for fats to the nearest
0.5 (\1/2\) or whole g, for cholesterol to the nearest 5
milligram (mg) amount, for carbohydrates and protein to the
nearest g and for sodium to the nearest 5 or 10 mg increment. The
rationale for each of these increments was explained when the
increments were proposed.
    c. Amounts of fatty acids to be rounded to zero.
    35. A few comments disagreed with 0.25 g as the cut-off level
at which fatty acids could be declared at zero. The primary
reason given for disagreeing with the 0.25 g cut-off was that the
analytical methods are not sensitive and precise enough to detect
that level with any degree of reliability. One of the comments
noted that FDA had, in its proposal on serving sizes, referenced
consumer complaints about fractional numbers. The comment felt it
was contradictory to introduce a potentially confusing
requirement for the proposed 0.25 g cut-off. Other comments
stated it would be confusing to consumers if, because of the
rounding requirements, a product containing either 0.4 g or 0.45
g total fat and 0.3 g saturated fat is declared as ``0'' g total
fat and 0.5 g saturated fat.
    The agency is persuaded that the level of 0.25 g as the
cutoff for a zero declaration for saturated, polyunsaturated, and
monounsaturated fat content implies unwarranted precision. The
ability to distinguish 0.24 g as zero and 0.26 g as a 0.5 g
increment is presently unsubstantiated. Therefore, FDA is
amending 101.9(c)(2)(i) through (c)(2)(iii), to require that
when a serving contains less than 0.5 g of saturated fats,
polyunsaturated fats, or monounsaturated fat, the content of the
fatty acids will be expressed as zero.
5. Cholesterol
    36. The majority of comments agreed with the proposal for the
mandatory declaration of cholesterol content. A few comments
disagreed stating that dietary cholesterol does not play a
significant role in the etiology of atherosclerotic vascular
disease. Some comments stated that the declaration of cholesterol
would mislead consumers into believing that a food free of, or
low in, cholesterol would be effective in lowering serum
cholesterol levels no matter how much saturated fat or total fat
it contained. These comments suggested that declarations of
cholesterol content should be either voluntary or not permitted.
    FDA disagrees that the declaration of cholesterol should be
voluntary or not permitted. The declaration of cholesterol
content is required by section 403(q)(1)(D) of the act. While
section 403(q)(2)(B) of the act allows the Secretary to delete
nutrient information that is not necessary to assist consumers in
maintaining healthy dietary practices, FDA does not believe that
this is the case for cholesterol. There is a strong scientific
consensus that high dietary intakes of total fat, saturated fatty
acids, and cholesterol are associated with an increased risk of
atherosclerotic cardiovascular disease, most notably with
elevations in blood LDL-cholesterol and increased risk of
coronary heart disease (Refs. 2, 3, 4, and 30).
    Further, numerous controlled experiments, in both animals and
humans, verify that dietary saturated fats and cholesterol
elevate blood LDL-cholesterol. For this reason, current
recommendations suggest limiting cholesterol to 300 mg per day as
a meanses that the declaration of cholesterol is warranted and
will be beneficial to many individuals in the general population
in the monitoring of their cholesterol intake. Therefore, no
changes have been made in 101.9(c)(5), redesignated as
101.9(c)(3).
    37. A few comments requested the mandatory declaration of a
cholesterol-saturated fat index. This index provides a single
number for individual foods that describes their cholesterol and
saturated fat content. The index indicates the potential of a
given food, diet, or menu to raise blood cholesterol levels.
    The agency is not persuaded that the declaration of a
cholesterol-saturated fat index on the nutrition label is
warranted. There currently exists no consensus on the efficacy of
this index. Therefore, FDA believes that the declarations of
cholesterol and saturated fat, as required by the 1990
amendments, are sufficient for those who wish to moderate their
intake of these nutrients.
    38. Several comments disagreed with the proposal to declare
``not a significant source of cholesterol'' if cholesterol is
present at less than 2 mg per serving. One comment suggested that
the label declare zero cholesterol only if the product is
virtually devoid of cholesterol. The comments stated that it is
misleading to have even minute quantities of a food component in
a product when the label declares that the product is free of
that component.
    The agency is not persuaded by these arguments. As discussed
in the July 19, 1990, tentative final rule on cholesterol (55 FR
29460 at 29461), FDA purposely selected a value, less than 2 mg
of cholesterol per serving, that is dietarily insignificant yet
that can be detected with reasonable analytical reliability. A
quantitative declaration other than zero would not necessarily be
more correct because methodological limitations do not generally
permit precise quantification of cholesterol content within the
95 percent confidence level below 2 mg amounts. It is also
extremely unlikely that sufficient quantities of foods containing
less than 2 mg of cholesterol per serving would be consumed on a
daily basis to represent a significant level of cholesterol
intake.
    39. A few comments requested that foods having less than 5 mg
of cholesterol per serving be permitted to indicate ``not a
significant source of cholesterol'' so that skim milk, at 4 mg
cholesterol per cup, and similar foods could use the statement.
Proposed 101.9(c)(5) only allows its use on labels of foods
containing less than 2 mg of cholesterol per serving.
    FDA disagrees. The agency believes that the statement ``not a
significant source of cholesterol'' is only appropriate on foods
that contribute truly insignificant amounts of cholesterol to the
diet. In the companion document on nutrient content claims
published elsewhere in this issue of the Federal Register, the
agency has determined that foods that contain less than 2 mg of
cholesterol per serving are dietarily insignificant sources, and
that foods that contain larger amounts, both individually and
collectively, contribute significantly to a person's daily
cholesterol intake. Therefore, FDA is taking no action on the
basis of these comments.
D. Sodium
    40. The majority of comments supported the proposal for the
mandatory declaration of sodium. A few comments requested
alternate methods of declaring sodium, such as a sodium to
potassium ratio and a sodium balance system.
    The agency has no data, nor were any submitted, that
demonstrate that these alternative methods would more effectively
present sodium content. Accordingly, FDA has not revised the
provision for the declaration of sodium content.
    41. A few comments questioned potential beneficial effects of
sodium restriction in nonhypertensive populations. The comments
suggested that there is still debate within the scientific
community as to whether it is appropriate for the general
populatnts stated that control of sodium intake is only relevant
for those segments of the population that are sodium sensitive.
These comments requested that the agency not permit the
declaration of sodium, or that it make the declaration of sodium
voluntary.
    The agency disagrees. Section 403(q)(1)(D) of the act
requires the declaration of sodium. While section 403(q)(2)(B) of
the act allows the Secretary to delete nutrient information that
is not necessary to assist consumers in maintaining healthy
dietary practices, the bulk of the accumulated evidence strongly
supports the prevailing consensus that it would be prudent for
the general population to reduce sodium consumption, even though
not all people display increased blood pressure in response to
high sodium intakes.
    The Surgeon General's report (Ref. 2) asserts the need for
moderation in sodium consumption, not only because there is a
benefit to persons whose blood pressures do rise with sodium
intake, but also because there is no biological marker for
individual sodium sensitivity. Additionally, the report notes
that there is no apparent harm from moderate sodium restrictions
(Ref. 2). Accordingly, 101.9(c)(9), redesignated as101.9(c)(4),
will continue to require sodium declaration.
    42. One comment from a national manufacturer of carbonated
soft drink syrups explained that these syrups must be mixed with
local water supplies, and that, therefore, the final products
vary in sodium content. The comment suggested that the sodium in
these products be declared as an average value, such as ``less
than 35 mg,'' even though the product contains slightly more, or
slightly less, than that amount. Further, the comment suggested
that the manufacturer be allowed to make a claim, such as ``very
low sodium,'' based on that range.
    Data on the sodium content of the United States water
supplies were previously submitted, reviewed, and discussed inthe
April 18, 1984, final rule on the declaration of sodium content
(49 FR 15524). FDA's evaluation of these data revealed that a
single label would accurately reflect the sodium content of all
but 10 percent of soft drink products bottled in the United
States, and that a second label could apply to the remainder
without severely overstating the sodium content.
    Further, the agency stated that the manufacturer had the
option of using a single nationwide label with the highest
possible sodium level declared. This approach would result in the
sodium content being overstated by about 50 mg for a majority of
products. While 50 mg is not an insignificant amount of sodium,
it represents a relatively small portion of the daily sodium
intake for all but those persons on extremely restricted sodium
diets. Even if sodium were declared based on the highest level of
sodium found in any source of water, all regular and diet soft
drinks would fall into the ``low sodium'' category.
    FDA is not persuaded by the comment to the July 19, 1990 or
November 27, 1991, proposals that an average value representing a
range of sodium levels, such as ``less than 35 mg,'' is
appropriate for these products. Sodium content may be
underrepresented by this method.
    Inasmuch as the declaration of sodium is required by section
403(q)(1)(D) of the act, and no new data were presented with the
comment, the agency is denying the request that a range of sodium
content be declared on the nutrition label or be allowed as a
basis to support a nutritional claim.

E. Potassium

    43. Several comments supported the proposal for the voluntary
declaration of potassium. One comment, however, requested that
the agency not allow any declaration of potassium content. The
comment suggested that the general population is unaware ofthe
dietary role of potassium, and any declaration of potassium
content would only serve to confuse the consumer. No data were
provided to support this argument.
    FDA is not persuaded that it should not permit the voluntary
declaration of potassium content. As discussed in the mandatory
nutrition labeling proposal (55 FR 29500 at 29501), beneficial
effects of potassium intake relative to reducing mortality from
stroke have been reported. Data from animal studies suggest that
dietary potassium may lower blood pressure and the risk for heart
disease and may also protect against vascular damage and stroke
(Ref. 3).
    In addition, epidemiological evidence for humans suggests
that diets with high levels of potassium--but also low levels of
sodium--may be beneficial in lowering blood pressure (Ref. 3).
Moreover, the IOM report concluded that even though deficits or
excesses of potassium intake do not pose public health concerns,
the voluntary declaration of potassium would be beneficial to
consumers (Ref. 1). Based on the foregoing evidence, FDA
concludes that the declaration of potassium in nutrition labeling
mayassist in maintaining healthy dietary practices. Accordingly,
in 101.9(c)(10), redesignated as 101.9(c)(5), FDA will allow
potassium to be declared in nutrition labeling on a voluntary
basis.
    44. Comments from several health and professional
associations, consumers, consumer groups, and universities
supported mandatory declaration of potassium content. The
comments stated that this information is potentially helpful to
persons with kidney disease. Others referred to epidemiological
evidence of a positive association among high potassium intake,
low sodium intake, and lower blood pressure.
    Although potassium has been acknowledged as a potential
public health issue (Refs. 1 and 49), no specific, quantitative
recommendations have been made by national consensus reports.
Accordingly, potassium does not meet FDA's criteria for inclusion
as a mandatory element of nutrition labeling, as discussed in
FDA's mandatory nutrition labeling proposal (55 FR 29487 at 29493
and 29500).
    Until such time as quantitative recommendations are made, the
agency does not believe there are sufficient grounds to require
labeling of potassium content. Therefore, FDA is continuing to
permit potassium content labeling in nutrition labeling on a
voluntary basis in 101.9(c)(5).
    45. One comment suggested that the declaration of potassium
content should be mandatory only if magnesium is not required as
a reference nutrient. The comment stated that potassium and
magnesium are abundant in whole grain cereals, legumes, nuts, and
other unprocessed foods. Further, the comment suggested that if
magnesium is required, potassium should be voluntary.
    The agency is not persuaded that the mandatory declaration of
magnesium is warranted, or that the declaration of potassium
should be mandatory because the declaration of magnesium is not
required. Magnesium is not a nutrient for which there are
significant public health concerns (Ref. 23).
    Further, while magnesium and potassium are found together in
many foods, using magnesium as a reference nutrient for potassium
in food labeling is questionable because there are many fruits
and some vegetables that are excellent sources of potassium but
poor to moderate sources of magnesium (Ref. 23). The agency does
not believe that a mandatory declaration of magnesium content is
warranted at this time.
    46. One comment suggested that information on potassium
should be available from manufacturers' toll-free telephone
numbers. The comment explained that consumers who must monitor
their potassium intake, such as renal dialysis patients, often
have difficulty determining how much potassium is in a product.
The comment suggested that manufacturer's toll-free numbers would
provide easy access to more detailed nutrient contentinformation.
    While FDA encourages manufacturers to make additional
information available to consumers, this request is beyond the
authority of the agency. Toll-free telephone numbers for product
information may or may not be supplied according to the
preference o

F. Total Carbohydrate, Dietary Fiber, Sugars,Sugar Alcohol, and
Other Carbohydrate

1. Total Carbohydrate
    47. Comments from trade associations, manufacturers,
professional societies, and another federal agency recommended
that FDA reconsider proposed 101.9(c)(6) which excludes dietary
fiber from total carbohydrate. As noted in several comments,
dietary fiber traditionally has been included as part of the
carbohydrate content of food, is considered part of carbohydrate
in current nutrition labeling regulations, is included within
total carbohydrate for nutritional labeling by Canada, and is
included in the Atwater method of determining ``carbohydrate by
difference.'' Other comments pointed out that excluding dietary
fiber from total carbohydrate is consistent with definitions for
labeling used by Codex Alimentarius Commission and the European
Community (EC), which include only metabolized carbohydrate. A
comment remarked that to exclude dietary fiber from total
carbohydrate is inconsistent with all major data bases and U.S.
publications on food composition and is different from the way
carbohydrate has been presented to the consumer in nutrition
labeling for the past two decades. This comment suggested that
inconsistency in definitions will contribute to consumer
confusion.
    In the 1990 mandatory nutrition labeling proposal, FDA
proposed mandatory declaration of total digestible carbohydrate,
which excluded dietary fiber, the nondigestible portion of
carbohydrate. Several comments noted that while the intent of
this definition for total carbohydrate was to include only energy
yielding components, in fact there is evidence that fermentation
of dietary fiber yields available energy. Comments noted that
dietary fiber content was accounted for in deriving both the
general energy factor of 4 calories per g of carbohydrate and the
specific Atwater factors for calculation of energy value of
carbohydrate in foods. The comments stated that total
carbohydrate (excluding dietary fiber) content as defined in
proposed 101.9(c)(6) is not appropriate for calculating calories
from carbohydrate as proposed in 101.9(c)(3). As a result, two
different values for ``total carbohydrate'' would be required to
comply with nutritional labeling: (1) Total carbohydrate
(excluding dietary fiber) for the content declaration, and (2)
total carbohydrate (including dietary fiber) for calorie
calculation.
    In the mandatory nutritional labeling proposal, FDA stated
that the reason for declaration of carbohydrate content was, in
part, to allow consumers to determine the percentage of calories
from carbohydrate (55 FR 29487 at 29497). Several comments argued
that departing from the established use of the term
``carbohydrate'' (i.e., including dietary fiber) used in
determining carbohydrate calories by the Atwater method, will be
confusing and thereby detract from the value to consumers of the
caloric information. Several comments suggested that in
separating dietary fiber from ``energy yielding'' components of
carbohydrate, FDA's logic was faulty for two reasons. First,
carbohydrate fractions are not clearly delineated as digestible
or nondigestiblefractions. Rather, there exists a continuum of
digestibility among carbohydrate substances. Second, dietary
fiber is appropriately included in total carbohydrate for
calculation of energy content with use of Atwater factors.
    Many comments noted that, except for lignin, substances
comprising dietary fiber are carbohydrates. Comments pointed out
that dietary guidelines (Ref. 4) urge increased consumption of
types of foods rich in both dietary fiber and complex
carbohydrate and stated that separating these carbohydrate
components in nutrition labeling will mislead consumers as to the
nature of dietary fiber.
    FDA is persuaded by the comments that the separation of
dietary fiber from carbohydrate is inconsistent with established
methods of reporting food composition and confuses the issue of
calculating energy content. Further, the agency agrees that the
sunderstanding of label information and its application to
dietary recommendations that link dietary fiber and complex
carbohydrate together in advising increased consumption of foods
high in both. Accordingly, FDA is modifying 101.9(c)(6) to
include dietary fiber in the declaration of total carbohydrates.
This action results in the inclusion of both digestible and
nondigestible carbohydrates under total carbohydrates.
    Section 101.9(c)(6) is also modified to state that total
carbohydrate content is to be calculated by subtracting the sum
of crude protein, total fat, moisture, and ash from the total
weight of food. Additionally, since total carbohydrate now
includes dietary fiber, the paragraphs relating to dietary fiber
are redesignated under 101.9(c)(6)(i) rather than
under101.9(c)(7). Consequently, the remaining paragraphs
within101.9(c) are renumbered.
2. Dietary Fiber
    48. Comments from consumers, health professionals, health
promotional organizations, and manufacturers agreed that
declaration of dietary fiber should be mandatory. Other
manufacturers, trade associations, and a university food science
department disagreed and urged voluntary, rather than mandatory,
declaration of dietary fiber. The arguments against required
listing of dietary fiber included: (1) Analytical methods for
dietary fiber in foods are not yet routine, are expensive, and
lack precision in some types of foods; (2) mandatory declaration
imposes an unnecessary analytical burden on producers of foods
that are not significant sources of fiber; and (3) mandatory
declaration will likely encourage fiber supplementation of foods.
    The agency does not agree that the specified methods for
fiber analysis are difficult and expensive. The operations
involved and equipment required for the methods are standard in
analytical laboratories. The agency recognizes that the official
AOAC method for dietary fiber analysis is relatively recent.
However, as a validated method, it should be included in current
nutrition labeling regulations.
    In regard to the analytical burden on producers of foods with
insignificant amounts of fiber, the agency advises that if there
is adequate and reliable reason to believe that fiber is not
present, there is no need to analyze for it; it can be declared
as ``0''. Additionally, 101.9(c)(6)(i) provides for expression
of dietary fiber in 1 g increments in recognition of the
precision of analytical methods. For foods that contain less than
1 g of dietary fiber per serving, manufacturers may chooseto
state ``contains less than 1 g'' or to omit dietary fiber from
the list of nutrients and to state at the bottom of the nutrition
label ``Not a significant source of dietary fiber.''
    There have always been concerns that nutrition labeling will
encourage the supplementation (i.e., fortification) of foods. In
part for this reason, FDA published a policy statement on the
addition of nutrients to food on January 25, 1980 (45 FR 6314).
The statement was issued to promote the rational addition of
nutrients to foods to preserve a balance of nutrients in the diet
of American consumers. In the statement, FDA established
guidelines in 104.20, which the agency urges manufacturers to
follow if they elect to add nutrients to a food.
    FDA intends to continue to monitor the marketplace through
the Food Labeling and Packaging Survey, consumer and industry
complaints, and other means to determine if inappropriate
fortification is occurring. If the agency finds that there is a
problem with inappropriate fortification of foods with dietary
fiber or any other nutrients, it will take steps to ensure that
overfortification does not result in the imbalance of essential
nutrients in the diet of American consumers or the presence of
excessi
    Thus the agency is not persuaded that there is a compelling
cause not to require declaration of dietary fiber in nutrition
labeling. Section 403(q)(1)(D) of the act requires dietary fiber
to be included in nutrition labeling. Section 403(q)(2)(B) of the
act allows a required nutrient to be deleted if the Secretary
determines that the nutrient is not necessary to assist consumers
in maintaining healthy dietary practices, but no information
contained in the comments would lead to such a conclusion. In
fact, most comments supported the usefulness of mandatory
declaration of dietary fiber. Accordingly, 101.9(c)(6)(i) will
require the declaration of dietary fiber in nutrition labeling.
    This action represents the final disposition of two petitions
regarding label declaration of carbohydrates, including dietary
fiber. One petition from the Kellogg Co. dated May 14, 1978
(Docket No. 78P - 0091), requested, in part, permission to list
under ``carbohydrate'' the amounts of ``starches and related
carbohydrates'' and ``sucrose and other sugars.'' The other
petition from the Center for Science in the Public Interest dated
June 1, 1987 (Docket No. 87P - 0194/CP) requested that dietary
fiber be a mandatory component of nutrition labeling, and that
regulatory letters be sent to all manufacturers making misleading
claims about fiber content.
    49. Comments from nutritionists representing state
cooperative extension services and from one manufacturer
cautioned that declaration of soluble and insoluble subcomponents
of dietary fiber should be prohibited because the methodology for
separating soluble from insoluble fiber is inadequate, and
because there is no scientific agreement as to the health effects
of the subgroups of dietary fiber.
    The agency advises that analytical methods for the
measurement of soluble and insoluble dietary fiber are now part
of an official AOAC method for dietary fiber (Ref. 50). While
experience with these methods is limited, they will allow for
accurate separation of these subcomponents.
    In regard to scientific agreement as to the health effects of
soluble or insoluble fiber, FDA has evaluated the health effects
of the dietary fiber subgroups and has concluded that there is
sufficient scientific agreement to issue a final rule permitting
health claims relating to the effects of intake of soluble
dietary fiber on heart disease. This decision is discussed in a
companion document entitled ``Food Labeling; Health Claims;
Dietary Fiber and Cardiovascular Disease'' published elsewhere in
this issue of the Federal Register.
    Accordingly, 101.9(c)(7)(i), redesignated
as101.9(c)(6)(i)(A) and (c)(6)(i)(B), will continue to allow the
voluntary declaration of soluble and insoluble dietary fiber in
nutrition labeling, except that when a claim is made about either
type of fiber, label declaration of that type of fiber will be
required. To clarify that soluble and insoluble fiber are to be
indented under dietary fiber rather than under total
carbohydrate, FDA has modified 101.9(c)(6)(i)(A) and
(c)(6)(i)(B) to state ``indented under dietary fiber.'' FDA has
also modified these two sections to remove the requirement that
whenever one type of fiber is declared, the other type must also
be declared. Because total dietary fiber is a mandatory component
of nutrition labeling, the amount of an undeclared subcomponent
(i.e., soluble or unsoluble fiber) can be calculated simply by
subtracting the amount of the declared subcomponent from the
amount of dietary fiber. This change will minimize space
requirements caused by the voluntary declaration of additional
nutrients.
3. Sugars
    a. Definition of sugars.
    50. Comments from consumers, consumer interest groups, state
governments, trade associations, food retailers, and a
manufacturer concurred with the agency's proposed definition for
sugars as the sum of all free mono- and oligo-saccharides through
fourydrate sweeteners added to foods, were cited as reasons for
support for the proposed definition. Comments from many
consumers, state governments, and a health promotion association
stated that information on content of both sugars and of sugar
derivatives is important to assist consumers to moderate intake
of sugars and to assist diabetics in maintaining healthy dietary
practices. Consumer interest groups argued that underreporting of
the sugars content in products rich in corn syrups is an
appropriate justification for an expanded definition for sugars.
A comment noted that the agency has a precedent for considering
sugar alcohols as sugars in 100.130(d)(4), which states that
``sugar-free'' type statements cannot be made on labels of diet
beverages containing ``sorbitol, mannitol, or other hexitols.''
    Other comments from a wide variety of manufacturers, trade
associations, foreign and state governments, professional
associations, and a Federal agency objected to the proposed
sugars definition. Most of these comments recommended that the
sugars definition be limited to monosaccharides and
disaccharides. One argument for limiting the sugars definition to
mono- and di-saccharides is that this is the traditional and
widely accepted use of the term ``sugars.'' They pointed out that
it is also thedefinition of the term in the IOM report
``Nutrition Labeling: Issues and Directions for the 1990s'' (Ref.
1). Many comments noted that for conformity with international
regulatory definitions for nutrition labeling (EC, Codex
Alimentarius Commission,and Canada) sugars should be defined as
mono- and di-saccharides.
    Another argument, brought forth in comments, for limiting the
sugars definition to mono- and di-saccharides is that there are
no compelling health or nutritional reasons for includingtri- and
tetra-saccharides as ``sugars.'' The comments pointed out that
the 1986 ``Report From FDA's Sugars Task Force'' (Ref. 51)
concluded that the only public health concern from sugars
consumption in the United States is the promotion of dental
caries. The IOM report (Ref. 1) concurred with this conclusion.
The comments argued that, in the absence of a clear relationship
between number of saccharide units and carcinogenicity, the
proposal to include tri- and tetra-saccharides within sugars is
not relevant to the public health concern of dental caries.
    Several comments questioned the agency's logic in including
tri- and tetra-saccharides with sugars. FDA had stated in the
1990 mandatory nutrition labeling proposal (55 FR 29487 at 29497)
that the intent of including tri- and tetra-saccharides as sugars
was to preclude potential underdeclaration of the sugars content
of foods containing corn syrups. Several comments noted that
mono- and di-saccharides are logically grouped in that they are
sweet, naturally occurring, and rapidly absorbed, but that these
characteristics are, for the most part, not in common with tri-
and tetra-saccharides. Comments also noted that most corn syrup
used in sweetening is in the form of high fructose corn syrup,
which is composed of 95 percent monosaccharides, and that high
fructose corn syrup accounts for two thirds of total U.S. corn
syrup consumption. Comments noted that corn syrups with greater
proportions of higher saccharides are used for technical purposes
other than sweetness. Thus, the comments argued that
underestimation of simple sugars from corn syrups is not of
sufficient importance to warrant imposing a unique sugars
definition for labeling purposes that would differ from common
usage of the term.
    Many comments objected to the proposed sugars definition on
methodological grounds, in that they claimed that the proposed
definition is not compatible with standardized analytical methods
for measuring sugars. The comments acknowledged that validate
foods of oligosaccharides larger than disaccharides is difficult,
costly, and inaccurate. The comments asserted that the lack of
validated analytical methodology appropriate for the definition
would result in compliance difficulties and inaccurate
information on the label.
    FDA has evaluated all comments in favor of the proposed
expanded sugars definition and those opposed to this definition.
FDA is persuaded that compliance with nutrition labeling will be
impeded by adopting a definition for sugars that is not supported
by validated analytical methods. FDA is also persuaded that the
usefulness of nutrition labeling will be hindered by adopting a
definition that is inconsistent with commonly accepted use, and
with the international use of the term.
    FDA finds that these factors outweigh any public health
benefit from including tri- and tetra-saccharides in the
definition of ``sugars'' for nutrition labeling purposes. The
public health concern associated with sugars consumption is the
promotionof dental caries. While simple sugars are the most
cariogenic carbohydrates, all fermentable carbohydrates,
including starches, are capable of promoting dental caries.
Factors such as the characteristics of the food that contains the
sugar (e.g., stickiness), the frequency of consumption, and the
sequence in a meal, appear to be as important in the etiology of
dental caries as the sugars themselves (Refs. 2 and 3). As such,
the inclusion of tri- and tetra-saccharides with sugars would not
improve the ability of the label to assist consumers in
maintaining healthy dietary practices with respect to dental
health.
    Therefore, the agency is modifying the definition of
``sugars'' in 101.9(c)(6)(ii)(A), redesignated as
101.9(c)(6)(ii), to include only free monosaccharides and
disaccharides.
    51. Several comments recommended that lactose be specifically
excluded from the sugars definition for nutrition labeling. These
comments asserted that the listing of lactose with sugars in
nutrition labeling may mislead some consumers who may equate the
lactose sugar content of dairy products with ``empty calories''
of products high in added sugar. The comments expressed a fear
that dietary guidelines to moderate sugars consumption may lead
some consumers to forego the important nutritional benefit of
dairy products if lactose is included in sugars content. Comments
also noted that intestinal digestion of lactose is inefficient.
As such, the digestion and absorption of lactose more closely
resembles complex carbohydrate than simple sugars. Furthermore,
comments argued that lactose is not sweet nor used as a sweetener
and could logically be separate from sugars used as sweeteners.
    FDA disagrees with the comments. As discussed in the
preceding comment, the agency has been persuaded of the need to
define ``sugars'' for nutrition labeling purposes to be
consistent with standard analytical methodologies and in
conformity with the traditional usage of the term. Lactose, a
di-saccharide, is clearly a sugar by conventional standards and
is identified with all other mono- and di-saccharides in routine
analytical procedures. The nutritional significance of the sugars
content of certain types of foods, such as lactose in dairy
products and natural sugars in fruit, and the importance of such
foods as sources for other important nutrients, needs to be
addressed through the consumer education program discussed below.
    Accordingly, the agency is not making the recommended change
to exclude lactose in the definition of sugars.
    52. Several comments suggested alternative definitions for
``sugars'' based upon physiological characteristics rather than
the number of saccharide units. Among these alternatives were
suggestions for definitions based on digestibility, caloric value
    FDA finds that such alternative approaches are not feasible
from a compliance standpoint because validated analytical methods
to quantitate sugars defined in these ways do not exist. In
addition, use of any of these definitions for sugars would be
unique to U.S. nutrition labeling and would thus likely impede
foreign trade. Moreover, because these definitions do not
correspond to the commonly recognized meaning of the term, the
resulting labeling information would be of limited usefulness.
    53. FDA received comments that suggested alternative
terminology for the ``sugars'' component of carbohydrate. The
agency's longstanding use of ``sugar'' as synonymous with sucrose
in ingredient labeling was cited as evidence of the need for an
alternative term. Several comments felt that FDA's distinction
between ``sugar'' and ``sugars'' would not be clearly understood
by consumers. Alternative terms suggested included ``sweeteners''
and ``simple carbohydrates.''
    FDA considered these comments but has concluded that it is
best to maintain the proposed terminology. The agency advises
that the term ``sweeteners'' would logically include the
noncarbohydrate intense sweeteners, which would not be
appropriately declared as a part of carbohydrate content. In
addition, the term appears to apply more to added sugars than to
total sugars and, therefore, would cause compliance problems
because it is not possible, in most foods, to differentiate
between added and naturally present sugars.
    ``Simple carbohydrates'' may have been a good term for the
originally proposed definition (i.e., mono-, di-,    tri-, and
tetra-saccharides). However, the agency finds it is too broad a
term to encompass only the traditional sugars (i.e., mono- and
di-saccharides).
    b. Total sugars versus added sugars.
    54. Some comments recommended mandating declaration of added
sugars only rather than total sugars. The comments noted that
consumers need to be made aware of added sugars because dietary
recommendations urge use of sugars in moderation, while at the
same time recommending increased consumption of fruits which are
sources of naturally occurring sugars. Other comments recommended
either mandatory or voluntary declaration of both added and
naturally occurring sugars. One comment suggested that added
sugars be required in addition to total sugars in foods
containing more than 2 g of added sugar.
    The agency is not persuaded that there is a need for
mandatory disclosure of added sugars in place of, or in addition
to, total sugars. There is no scientific evidence that the body
makes any physiological distinction between added sugar molecules
and those naturally occurring in a food. In addition, the agency
believes that it should not promulgate regulations that it cannot
enforce. When a product is sampled for compliance, laboratory
analysis yields a value for total sugars. For most foods, as
stated above, it is not possible to differentiate between added
and naturally occurring sugars. Accordingly, the agency would not
be able to determine the accuracy of a label declaration of added
sugars.
    Furthermore, declaration of only added sugars may
significantly underrepresent the sugars content of many foods
that are high in naturally occurring sugars. For example, in some
fruits canned in heavy syrup, added sugars may represent only
about 50percent of total sugars. Disclosure of only the added
sugars could be misleading to consumers who are concerned with
total sugar intake. Therefore, the agency is retaining the
provision of 101.9(c)(ii)(A), redesignated as 101.9(c)(ii), to
declare total sugar content, e.g., that added as well as that
naturally present.
    While FDA is not distinguishing, on the nutrition label,
between added and naturally present sugars, the agency does
intend to include information about this distinction in the
consumer education program that it is preparing. This information
will hng versus added sugars, and (3) appreciate the important
role in the total daily diet of foods, such as fruits and dairy
products, with naturally occurring sugars.
    c. Mandatory declaration of sugars.
    55. In the 1991 supplementary proposal, FDA requested
specific comments as to the utility, appropriateness, and
feasibility of requiring declaration of sugars content,
particularly as such declaration relates to, and is supported by,
public health goals (56 FR 60366 at 60369). Comments received
were relatively evenly divided on the issue of whether the
declaration of sugars should be mandatory in nutrition labeling.
In general, consumers and health professionals and their
associations supported mandatory declaration of sugars. Several
state attorneys general and a few industry groups also agreed
that consumers have a right to know the amount of sugars present.
Comments argued that section 403(q)(2)(B) of the act only allows
FDA to delete sugars as a mandatory component of nutrition
labeling if such information ``is not necessary to assist
consumers in maintaining healthy dietary practices,'' and that
such information is vital to this end. Comments from many
consumers, state governments, anda health promotion association
stated that information on sugars content is important to
diabetics in assisting them to maintain healthy dietary practices
and to consumers in general in selecting diets that will moderate
the intake of sugars. One comment urged mandatory declaration of
sugars as a way to inform consumers of the content of new foods
that are being marketed as ``low fat'' and ``fat free'' in which
fats are being replaced by sugars.
    Most industry groups as well as a few health professional
associations and the IOM report (Ref. 1) recommended allowing
sugars declaration to be voluntary. They argued that dietary
guidance recommendations have not specified quantitative goals
for sugar consumption, and that sugar declaration should not be
required until a definition has been recognized by scientific
communities that reflects physiological effects. They also
pointed out that data bases do not generally contain information
on sugars composition, so a substantial investment of time and
money is needed for analysis and data base update. The comments
argued that such an expenditure would be inappropriate for a
nutrient of little public health concern. There was also a
concern expressed that because total sugars would be declared
rather than only added sugars, consumers would be confused by the
amount of sugars in fruits and reduce their consumption of these
foods. Despite these concerns, industry generally conceded that
if sugars information is needed, requiring sugars declaration in
the nutrition label is a better approach than mandating grouping
of sweeteners in the ingredient statement, as the agency proposed
in a document on ingredient labeling (56 FR 28593, June 21,
1991). Final action on the issue of grouping sweeteners in the
ingredient statement is addressed in the final rule on
declaration of ingredients published elsewhere in this issue of
the Federal Register.
    FDA is persuaded that mandatory declaration of sugars is of
great interest to consumers, and that it will assist consumers in
planning diets that conform to current dietary guidelines for
Americans to avoid too much sugars (Ref. 4). As discussed above,
FDA is modifying its proposed definition of ``sugars'' to be in
conformity with general usage and international definitions for
this term. The use of this definition will minimize the costs
associated with necessary laboratory analyses and update of data
bases.
    Therefore, FDA is requiring in 101.9(c)(6)(ii) that
declaration of sugars be included in nutrition labeling.
4. Sugar Alcohol
    56. Comments from a wide variety of manufacturers, trade
associations, foreign and State governments, professional
associations, and a Federal agency were opposed to inclusion of
saccharide derivatives, specifically sugar alcohols, within the
proposedn. The agency's proposed definition included in its
coverage saccharide derivatives that have sweetening,
nutritional, and metabolic effects similar to simple sugars. The
comments stated that sugar alcohols are inappropriately included
with sugars because sugar alcohols have many different chemical
and physiological properties than sugars. Comments noted that it
is these differences that motivated the development of uses for
these substances and makes them useful as sugar substitutes.
Comments pointed out that a salient distinction between sugar
alcohols and sugars lies in their digestion and absorption, which
is slower for sugar alcohols. Also, intestinal absorption of
monosaccharide sugar alcohols occurs only by passive diffusion,
not by active or facilitated monosaccharide absorptive
mechanisms. As a result, significant portions of ingested sugar
alcohols remain unabsorbed and pass into the colon, where they
are fermented, similar to fiber and complex carbohydrate. Thus,
the caloric value,insulin response, and glycemic index for some
sugar alcohols are less than for sugars.
    Several comments also claimed that sugar alcohols have
reduced cariogenic potential compared to sucrose or other sugars.
The comments noted that FDA proposed in 101.13(o)(8) in the
document on the general principles for nutrient content claims to
permit chewing gums sweetened with sugar alcohols to be labeled
as ``sugar free'' or ``sugarless'' as a means of indicating that
these products do not promote tooth decay. The comments argued
that declaring sugar alcohols as sugars would deny manufacturers
the means to promote the reduced cariogenic potential of other
sugar alcohol sweetened products relative to sugar containing
products.
    Comments also noted that international regulatory definitions
for nutrition labeling (EC, Codex Alimentarius Commission, and
Canada) exclude sugar alcohols and provide for a separate
declaration of sugar alcohols under carbohydrates. As a result,
the comments stated that a definition for sugars that includes
sugar alcohols for U.S. nutrition labeling could be seen as an
obstruction to international trade. The IOM report (Ref. 1)
recommended that sugar alcohols not be grouped with sugars in
ingredient labeling. Some comments argued that in the absence of
any quantitative dietary guidelines concerning sugar alcohols, it
is inappropriate to require any declaration of sugar alcohols in
nutrition labeling.
    FDA is persuaded that sugar alcohols have metabolic effects
different than sugars, have a history of being considered to be
sugar substitutes rather than as sugars, and have a role in
contributing to dental health. FDA also acknowledges that the
proposal to define sugar alcohols as sugars for nutrition
labeling purposes is inconsistent with the nutrition labeling
practices of other countries. Thus, FDA is modifying
101.9(c)(6)(ii) to remove sugar alcohols from the definition of
``sugars'' fornutrition labeling. The agency is doing so in
recognition of their usefulness as sugar substitutes in reducing
the cariogenic potential of foods.
    However, FDA continues to believe that the content of
nutritive carbohydrate sweeteners used as sugar substitutes is of
interest and importance to consumers. Therefore, FDA is retaining
101.9(c)(6)(ii)(B), which provides for the voluntary declaration
of sugar alcohols except when a claim is made on the label or in
the labeling about sugar alcohol or sugars (e.g., ``sugar free'')
and sugar alcohols are present in the food, in which case their
declaration is mandatory. Because sugar alcohols will no longer
be a subcomponent of sugars, FDA is redesignating
101.9(c)(6)(ii)(B) as 101.9(c)(6)(iii).
    Removing sugar alcohol from the definition of sugars
necessitates a change in the definition of sugar alcohol. The
proposed definition included a criterion that sugar alcohols
``meet the definition of sugars as described in paragraph
(c)(6)(ii)(A).''1A101.9(c)(6)(iii) to use a chemical definition,
namely that sugar alcohols be defined as ``saccharide derivatives
in which a ketone or aldehyde group is replaced by a hydroxyl
group, and whose use in food is listed by FDA (e.g., mannitol) or
is GRAS (e.g., xylitol, sorbitol).''
    57. Comments from trade associations and manufacturers stated
that the term ``sugar alcohol'' is potentially confusing in that
consumers may assume such components contain a sugar and ethyl
alcohol. The comments requested that the term ``polyol,'' which
has been recognized by the EC, be used in lieu of ``sugar
alcohol.'' Another comment from the Canadian government included
a copy of their nutrition labeling regulations which allow for
declaration of the specific sugar alcohols by name (i.e.,
sorbitol, mannitol, and xylitol).
    FDA advises that the term ``polyol,'' a contraction of
``polyalcohol'' or of ``polyhydric alcohol,'' is neither uniquely
descriptive of the alcohol derivatives of saccharides used as
sugar substitutes, nor is it a term that FDA expects consumers to
recognize or understand. While the agency recognizes that it is a
term that may be used voluntarily on labeling in the EC, it is
unlikely that American consumers will have any concept of what it
represents. As such, the agency considers the term ``polyol'' to
be potentially more confusing to consumers than would be ``sugar
alcohol.''
    Despite this fact, FDA acknowledges that many consumers also
may not be familiar with the term ``sugar alcohol.'' Thus, FDA
has decided to adopt the approach used by the Canadian
Government, which allows manufacturers to use the specific name
of the sugar alcohol in the nutrition label. The names of sugar
alcohols that are listed or GRAS for use in food, (e.g., sorbitol
184.1835, mannitol 180.25, and xylitol 172.395) are currently
used in ingredient statements on labels of food packages and,
hence, should be recognized by many consumers.
    The primary disadvantage to this option is the introduction
of the name of an ingredient into the nutrition label. While FDA
is generally opposed to such a result, the agency concludes that
the arguments opposed to the term ``sugar alcohol'' and the
desire to harmonize with Canadian labeling regulations are more
compelling in this instance than the need to maintain a clear
separation between the nutrition label and ingredient list.
However, to avoid cluttering the nutrition label and confusing
consumers, if more than one sugar alcohol is used in a food,
101.9(c)(6)(iii) provides that the term ``sugar alcohol,'' and
not the names of the ingredients, must be used in the nutrition
label.
    Accordingly, FDA is modifying 101.9(c)(6)(iii) to specify
the continued use of the term ``sugar alcohol'' or,
alternatively, if only one sugar alcohol is present in the food,
the name of the specific sugar alcohol present in the food may be
used.
5. Other Carbohydrate
    a. Definition.
    58. In the supplementary proposal, FDA noted that the term
``complex carbohydrate'' has not been clearly or consistently
defined, and that consensus reports that have associated
increased consumption of dietary complex carbohydrate with health
benefits have not attempted to define this food component. The
agency solicited suggestions on appropriate chemical definitions
and analytical methodology for complex carbohydrate (56 FR 60366
at 60369). Many comments from trade associations, food
manufacturers, professional societies, and state and foreign
governments expressed opposition to the agency's proposed
definition for the term complex carbohydrate as the sum of
dextrins and starches that contain ten or more saccharide units
(56 FR 60366 at 60388). A majority of these comments also
recommended as an alternative that ``complex carbohydrate'' be
defined as the difference between total carbohydrate and sugars.
Comments that argued for changing the definition pointed to the
lack of existing analyte comments raised concerns about the
feasibility of compliance and the economic burden of developing
methods and data bases. These comments also pointed out that
complex carbohydrate content defined as the difference between
total carbohydrate and sugars could readily be calculated.
    Another criticism of the proposed complex carbohydrate
definition was that the cutoff at 10 saccharide units is
arbitrary. These comments noted that there are no known
nutritional or physiological differences, nor a methodological
justification, tomake a distinction between polysaccharides
smaller than 10 saccharide units and those with 10 or more
saccharide units.Several comments were concerned that there is
the potential for consumer confusion regarding total carbohydrate
because neither of the subcomponents for total carbohydrate
included the 5 to 9 saccharide unit polysaccharides.
    Several comments suggested that the commonly accepted usage
of ``complex carbohydrate'' includes all carbohydrates larger
than disaccharides. Other comments suggested that complex
carbohydrate should be defined as all digestible polysaccharides
(e.g., dextrins, starch, and glycogen) rather than on the basis
of the number of saccharide units. Comments emphasized that while
there is not a consensus on a precise definition for ``complex
carbohydrate,'' the agency's proposed definition is not commonly
recognized, nor is it consistent with the use of the term in the
IOM report (Ref. 1).
    One comment from a State government recommended that to avert
undue emphasis on complex carbohydrate substances added to foods
and to avoid the potential for misleading claims about complex
carbohydrates, the term ``other carbohydrate'' should be used
rather than ``complex carbohydrate.''
    The agency noted in the supplementary proposal (56 FR 60366
at 60368) that identification of a specific benefit for complex
carbohydrate is confounded by the fact that diets high in complex
carbohydrate are usually mixed diets that contain significant
amounts of cereal grains, fruits, and vegetables which are high
in fiber, vitamins, and minerals and low in fats (Ref. 2). Thus
the extent to which complex carbohydrate provides health benefits
separate from those provided by fiber, vitamins, minerals, and
reduced fat is unclear. FDA has evaluated comments concerning the
complex carbohydrate definition and concludes that there is not
sufficient consensus on the meaning of the term to justify
adopting a specific definition for ``complex carbohydrate.''
    In response to the comments that suggested defining this term
as ``digestible polysaccharides,'' FDA advises that carbohydrate
digestibility is not clear cut. Some soluble dietary fiber is
relatively digestible, whereas some oligosaccharides are
relatively nondigestible. At this time there is not a consensus
regarding the most reliable methods for determining carbohydrate
digestibility nor for distinguishing energy derived from
intestinal digestion from that derived from colonic fermentation.
Asa result, the agency feels that it is inappropriate to base a
regulatory definition upon digestibility.
    FDA, therefore, is modifying the definition it proposed for
``complex carbohydrates''(101.9(c)(6)(i), redesignated
as101.9(c)(6)(iv)) to provide that it is the difference between
total carbohydrate and the sum of dietary fiber and sugars or, if
sugar alcohol is declared, the difference between total
carbohydrate and the sum of dietary fiber, sugars, and sugar
alcohol. This modified definition accommodates quantification of
the remaining carbohydrates by calculation rather than by
requiring additional laboratory analysis, and it resolves
concerns that the defined components of total carbohydrate were
not inclusive of all carbohydrates.
    In addition, because there is no consensus on a clear
definition for the term ``complex carbohydrate'' as it relates to
physiological effects, health benefits, or dietary guidelines,
the agency concurs with the recommendation from a state
governmentthat the term ``other carbohydrate'' be used rather
than ``complex carbohydrate.'' The agency recognizes that the new
definition will include many substances added to processed foods
for technical purposes, such as for texture modification or as
bulking agents. To declare these substances as complex
carbohydrates would be misleading. The intent of dietary
recommendations to increase the consumption of complex
carbohydrates and dietary fiber is to select diets with plenty of
fruits, vegetables, and grain products, not foods that have
complex carbohydrates as added texturizers or bulking agents.
Accordingly, FDA is modifying 101.9(c)(6)(iv) to change the
terminology from ``complex carbohydrate'' to ``other
carbohydrate.'' In addition, FDA is modifying 101.9(g)(4) and
(g)(6) to reflect this change in terminology.
    Finally, because ``other carbohydrate'' will be calculated as
that amount of carbohydrate remaining after subtraction of the
amount of dietary fiber, sugars, and sugar alcohols (when
declared) from total carbohydrate, it is logical to rearrange the
subcomponents of total carbohydrate to place ``other
carbohydrate'' at the bottom of the list. This reordering should
help to reduce any potential confusion over the meaning of the
term ``other carbohydrate.'' Accordingly, dietary fiber is
designated as101.9(c)(6)(i), sugars as 101.9(c)(6)(ii), sugar
alcohol as101.9(c)(6)(iii), and other carbohydrates as
101.9(c)(6)(iv).
    b. Voluntary declaration of ``other carbohydrate''.
    59. In the supplementary proposal, FDA requested specific
comment on the utility, appropriateness, and feasibility of
mandatory declaration of complex carbohydrate content,
particularly as it relates to, and is supported by, public health
goals (56 FR 60366 at 60369). Based on the comments and
information that it received in response to the supplementary
proposal, the agency said it would decide, under section
403(q)(2) of the act, whether to include complex carbohydrate in
the required list of nutrients in nutrition labeling. Several
comments from consumers, health professionals, a manufacturer,
and state governments supported mandatory listing of complex
carbohydrates on the grounds that this information will be
helpful to persons attempting to follow dietary recommendations.
However, a much larger number of comments from health
professional associations, academia, manufacturers, trade
associations, and foreign governments supported voluntary listing
of complex carbohydrates. The overriding factors cited in these
comments were the lack of an accepted definition for ``complex
carbohydrates'' and the lack of reliable analytical methods for
determining amounts present. Comments also stated that dietary
recommendations do not specify amountsof complex carbohydrates to
be consumed; therefore quantitative information in nutrition
labeling is not necessary to assist consumers in maintaining
healthy dietary practices. Additionally, comments noted that the
IOM report (Ref. 1) recommended voluntary listing of complex
carbohydrate. Comments also pointed out that currently available
data bases do not contain information on complex carbohydrates,
and that there would be an inherent variability in amounts
present in minimally processed foods.
    In light of these comments, the agency's decision to drop the
term ``complex carbohydrate'' because of the lack of a consensus
on the meaning of the term, and the lack of methods for reliably
determining the amounts present, FDA has reassessed the
requirement in section 403(q)(1)(D) of the act to declare complex
carbohydrates. Section 403(q)(2)(B) of the act allows the
Secretary to determine whether information relating to nutrients
specified in section 403(q)(1)(C), (q)(1)(D), (q)(1)(E), or
(q)(2) a specific definition for ``complex carbohydrates,'' it is
not possible to include quantitative information in the nutrition
label that would assist consumers in maintaining healthy dietary
practices. Therefore, under the provisions of section
403(q)(2)(B) of the act, FDA is deleting the requirement for the
listing ``complex carbohydrate'' in the nutrition label and is
allowing for declaration of ``other carbohydrate'' on a voluntary
basis.
    When ``other carbohydrate'' is omitted from the label, the
declared subcomponents of total carbohydrate (i.e., dietary fiber
and sugars) will not add up to the value for total carbohydrate
in most foods. Consumer education programs should inform
interested persons that other forms of carbohydrate beyond those
declared on the label are in the food product. This situation is
analogous to the fat category where the sum of saturated,
polyunsaturated, and monounsaturated fatty acids often do not add
up to 100 percent of the value for total fat because trans fatty
acids are not included in the definition of the fatty acids but
are included in the value for total fat.

G. Protein

1. Quantitative Protein Content
    60. Several food manufacturers agreed with the proposed
provision requiring that if the protein in foods represented or
purported to be for adults and children 4 or more years of age
has a protein digestibility-corrected amino acid score (PDCAAS)
of less than 20 percent, and if foods represented or purported
for children below 4 years have a protein quality value less than
40 percent of casein, the protein content statement must be
modified by an adjacent statement, ``not a significant source of
protein,'' regardless of the actual amount of protein present.
However, other food manufacturers objected to this provision of
the proposal. These comments argued that the statement has little
value in terms of the total dietary protein intake, and thatthere
is no evidence of protein malnutrition in this country. These
comments argued that, therefore, the statement is unnecessary.
One food manufacturer stated that the statement should only be
required if a claim is made. Another comment stated thatthe
declaration of the percent of the RDI for protein should be
required instead of the statement.
    FDA disagrees with the comments that state the statement is
unnecessary. Information on protein quantity alone can be
misleading on foods that are of low protein quality. As stated in
the supplementary proposal, dietary protein serves as a sourceof
essential and nonessential amino acids, the building blocks of
body protein. Because excess amino acids are not stored in the
body, humans need a constant supply of good quality dietary
protein to support growth and development. The determination of
the quality of a protein is dependent upon the proportion and
availability of essential amino acids (i.e., those amino acids
that the human body cannot manufacture but must obtain through
the diet) as well as the quantity of protein present. Foods that
contain proteins that are low in one or more of the essential
amino acids are known as incomplete proteins and are lower
quality proteins than those that contain all the essential amino
acids in sufficient quantities to support growth and development.
    The agency believes that nutrition labeling must inform
consumers when the quality of the protein is below minimum
specified levels. The majority of comments supported this
position. Although the agency agrees that protein deficiency is
not commonin the United States, protein quality is still of
concern for certain segments of the population, such as the very
young and the elderly. Accordingly, the agency concludes that
nutrition labeling must allow consumers to readily identify foods
with partic
    Nonetheless, in the case of foods for adults and children
over 1 year of age, the agency agrees with the comment that the
percent of the reference value for protein (discussed below) is a
satisfactory alternative to the statement, ``not a significant
source of protein,'' to allow consumers to readily identify foods
of low protein quality. However, as discussed in the final rule
entitled ``Food Labeling; Reference Daily Intakes and Daily
Reference Values'' published elsewhere in this issue of the
Federal Register, the label reference value for protein for
adults and children 4 or more years of age has been established
as a DRV rather than an RDI. As discussed in that document, this
change to a DRV is necessary because the agency is no longer
basing the label reference value for protein for this group on
the Recommended Dietary Allowances for protein. Rather, they are
now being based on percent of calories. However, because FDA did
not propose DRV's for infants, children less than 4 years of age,
pregnant women, and lactating women, the protein label reference
values for these groups remain as RDI's.
    Accordingly, the agency is amending 101.9(c)(8),
redesignated as 101.9(c)(7), to permit the optional declaration
of percent of the DRV or RDI for protein, as appropriate,
expressed as ``Percent Daily Value,'' in lieu of the statement
``not asignificant source of protein'' when the food is
represented or purported to be for use by adults and children 4
or more years of age and the protein quality value is a PDCAAS of
less than 20 expressed as a percent, or when the food is
represented or purported to be for use by children under 4 years
of age and the protein quality value is a PDCAAS of less than 40
expressed as a percent.
    FDA is not requiring declaration of percent DRV or RDI for
protein instead of the subject statement, as requested in the
comment, because of cost considerations. If a manufacturer is
aware that the protein in a particular food product represented
or purported to be for adults and children 4 or more years of age
has a PDCAAS of less than 20 percent, or that the protein in a
food represented or purported for children below 4 years has a
protein quality value less than 40 percent of casein, and the
manufacturer does not want to go to the expense of determining
the precise percent of the label reference value present in the
food, the agency has no objection to the use of the statement
``not a significant source of protein.''
    In conjunction with this change, FDA is making a parallel
modification in proposed 101.9(c)(8)(i), redesignated as
101.9(c)(7)(i), by deleting the prohibition on the declaration
of percent of the RDI for protein on foods represented or
purported to be for use by adults and children 4 or more years of
age with a PDCAAS of less than 20 percent, or on foods
represented or purported to be for use by children under 4 years
of age with a protein quality value of less than 40 percent of
the reference standard. That prohibition is no longer necessary
because the PDCAAS method for assessing protein quality is more
exact in measuring the protein quality for humans one year of age
and above than the protein efficiency ratio (PER) which was
previously used for all age groups. Because the PER is being
retained to measure protein quality for infant foods, FDA has
retained this prohibition for declarations on foods represented
or purported to be for use by infants with a protein quality
value of lessthan 40 percent of the reference standard of casein.
    61. One comment requested that the food-specific conversion
factors used by AOAC, and permitted in proposed 101.9(c)(8), to
convert amounts of nitrogen to protein content should be allowed
in calculating the PDCAAS whenever such factors are available.
The comment stated that in some cases (e.g., peanut butter) the
amino acid score used in calculating the PDCAAS is artificially
reduced when a conversion factor of 6.25 must to be used to
calculate protein content rather than a food-specific conve
    FDA agrees that there is an inconsistency in
proposed101.9(c)(8) pertaining to the factors for converting g
of nitrogen to g of protein when calculating protein content and
when calculating the PDCAAS. While the method for calculating the
PDCAAS described in the Report of the Joint Federation of
Agriculture Organization and the World Health Organization
(FAO/WHO) Consultation (Ref. 8) specifies a conversion factor of
6.25 (i.e., g of nitrogen x 6.25 = g of protein), the agency
finds it appropriate to use more specific conversion factors for
those foods where the official AOAC procedures require them.
Therefore, to allow for consistent methods of calculating g of
protein, the agency is modifying 101.9(c)(7)(ii) to state that
food-specific conversion factors required by the AOAC are to be
used when calculating the PDCAAS.
2. Protein Content as a Percentage of the RDI/DRV
    62. Several comments opposed the voluntary (in the absence of
a claim) declaration of protein as percent of the RDI on the
labels of foods intended for adults and children 4 or more years
of age. The comments contended that mandatory declaration of
protein as percent RDI for all groups would provide the consumer
with information on how amino acid and protein needs are met and
would assist consumers in appreciating that protein is an
important part of the diet.
    FDA disagrees with these comments. In the preamble of the
mandatory nutrition labeling proposal, FDA stated that current
evidence suggests that the diet typically consumed in the United
States provides for an adequate protein intake of
sufficientlyhigh biological quality to meet the nutritional needs
of adults and children 4 or more years of age (55 FR 29487 at
29499). Because protein intakes generally are adequate and not a
public health concern for this population group, FDA finds that
the additional costs associated with determination of the PDCAAS,
which are necessary to calculate the percent of the DRV for
protein, are not warranted on foods for this group unless protein
claims are made. Therefore, while declaration of the quantitative
amount of protein will continue to be required on all foods,
101.9(c)(7)(i) allows voluntary declaration of the percent of
the DRV for protein, expressed as ``Percent Daily Value,'' for
foods intended for adults and children 4 or more years of age
unless a protein claim is made for the product.
    63. Two baby food manufacturers suggested that the protein
content expressed as percent of the RDI for protein should be
voluntary for all foods, including those for infants and children
less than four years of age, unless the food is infant formula or
a protein claim is made. The comments stated that data show that
breast or cow milk and formula are the main contributors of
protein during the first 18 months, and that other foods are not
sole sources of protein for infants above 4 months. One
manufacturer provided survey data on the protein intake of
children 2 to 18 months of age. The comments also stated that
recent evidence shows that the protein intake of children 1 to 4
years of age is 100 percent of the RDI, that nutrition
information expressed as percent of the RDI would not be helpful
to the parents, and that the requirement is burdensome. Other
comments supported mandatory declaration of protein content
expressed as percent of the RDI for children less than 4 years of
age.
    FDA rejects the suggestion that protein content expressed as
the percent of the RDI should be voluntary for foods specifically
intended for infants and children under 4 years of age. As noted
in the preamble of the mandatory nutrition labeling proposal,
mandatory declaration of the percent RDI is warranted for this
age group because of the importance of protein quality in diets
derived from a limited number of foods (55 FR 29487 at 29499).
FDA acknowledges that breast or cow milk and formula are
thcontribution to total protein intake. For example, at 6 to 7
months of age, infants are receiving approximately one-third of
the total protein intake from baby foods (Ref. 52).
    The agency recognizes that required declaration of the
percent of the RDI for protein for foods for infants and children
less than 4 years of age presents a burden to manufacturers.
However, protein nutriture is critical during this period of life
which is marked by rapid growth and development. Both protein
quantity and quality are major factors in the utilization of
protein. Because of the importance of adequate high quality
protein in the diets of infants and young children, FDA considers
thedeclaration of percent of the RDI for protein necessary.
Moreover, with the information on digestibility the agency is
providing in appendix B (see comment 66 in this document),
declaring the percent of the RDI for protein should not be overly
costly or difficult.
    64. Several comments suggested the use of a system similar to
the current approach of expressing the percent of the U.S. RDA
for protein. Recommendations were made for the use of a single
RDI or two RDI's (i.e., an RDI for proteins of high qualityand
another RDI for those of low quality) to calculate the percent
RDI as long as the food is not intended for infants and toddlers.
Three baby food manufacturers favored establishment of specific
low and high protein quality-based RDI's to calculate the percent
RDI for foods intended for infants.
    FDA disagrees with the use of a system similar to the current
approach of expressing protein as percent U.S. RDA. The use of
breakpoints, as found in the existing regulation, creates
artificial differences in apparent protein nutritive values of
some foods when significant differences do not exist.
3. Protein Quality
    65. One comment questioned FDA's authority to change the
proposed protein quality methodology. The comment contended that
the 1990 amendments did not require a change in methodology, and
that the proposal must be reevaluated pursuant to President
Bush's directive in his State of the Union address on January 28,
1992, and set forth in his memorandum on Reducing the Burden of
Government Regulations (Ref. 53).
    FDA stated in the supplementary proposal that while not
directed to do so by the 1990 amendments, it was proposing to
modify the approach for determination of protein quality in the
mandatory nutrition labeling proposal. The agency did so in
response to a citizen's petition submitted by Protein
Technologies International Inc. (Docket No. 90P - 0052),
requesting that the agency accept an amino acid method that is
corrected for digestibility as an alternative method for
evaluating protein quality. FDA tentatively decided that the
petition had merit, and that some of the concepts in the petition
should be integrated into the rulemaking since protein quality is
an important part of nutrition labeling. The agency has the
authority under sections 201(n) of the act (21 U.S.C. 321(n), 403
(a) and (q) of the act, and 701(a) of the act (21 U.S.C. 371(a))
to modify the original proposed protein quality methodology to
reflect expanding scientific knowledge. This final rule
represents the final disposition of the subject petition.
    66. Several comments commended the agency for acceptance of
the PDCAAS method for assessing the protein quality of foods for
regulatory purposes. Comments stated that the PDCAAS method was
entirely appropriate and consistent with the FAO/WHO Consultation
on protein quality evaluation (Ref. 8). Other comments from food
manufacturers and a trade association conditionally supported the
PDCAAS method. Several comments recommended that PDCAAS not be
used as the sole method for measuring protein qualitspectrum of
foods. Two comments stated that compliance problems necessitate a
transition period of 2 to 5 years to ease the logistical and
economic burdens. Several comments supported the method but
recommended that the current PER method also be permitted.
    One manufacturer recommended: (1) Manufacturers be permitted
to use calculated PDCAAS values for common foods and food
ingredients; (2) that FDA issue a list of estimated digestibility
values and PDCAAS values for major foodstuffs and ingredients
before issuing the final regulation; and (3) that FDA convene an
expert group to produce a data base on digestibilities and PDCAAS
values and to make provision to update such a list.
    Another manufacturer requested that FDA allow any valid
methodology for determining protein quality for adults and
children more than 1 year of age.
    FDA disagrees with the recommendations that the PDCAAS method
not be used as the sole method until more technical and practical
knowledge is gained on its application to a broad spectrum of
foods and disagrees that a transition period is needed. FDA
advises that since most food products in the market place are
intended for adults and children above 4 years of age, on which
the declaration of percent DRV of protein is voluntary, a delay
in the implementation of the PDCAAS as the sole method is not
necessary.
    The agency also rejects the recommendation that the PER
method continue to be permitted for foods for adults and children
1 or more years of age as an option to the PDCAAS method for the
following reasons: (1) The PDCAAS is based on human amino acid
requirements and, therefore, is inherently more appropriate for
evaluating the protein content of foods intended for human
consumption than the PER which is based on the amino acid
requirements of the rat (Ref. 8), (2) the PDCAAS method is
recommended for regulatory purposes by a recognized international
organization experienced in establishing such standards (Ref. 8),
and (3) values obtained by the two methods differ so that their
simultaneous use on different foods would not allow for
comparisonof food products.
    FDA considered the recommendation that manufacturers be
permitted to calculate PDCAAS values. The two pieces of
information that are needed for this calculation are the amino
acid content and the digestibility of the food. FDA has concluded
that current representative amino acid data bases on raw and
processed food products are not sufficiently reliable to allow
for calculated PDCAAS values. Current data bases often lack
information on key essential amino acids, and the information
that is therewas often obtained using methodology that is now
outdated. In addition, food processing, i.e., the chemical,
biological, or physical treatment of foods, can reduce the amino
acid content of the food, so that only data from a food that
underwent similarprocessing or treatment should be utilized in
calculating the PDCAAS. In time, the agency believes it will be
possible to calculate PDCAAS values using representative amino
acid data backed by periodic analytical spot checks, but, at the
current time, more and better data are needed.
    FDA does agree, however, that a data base on digestibility
values could be of assistance in implementing the PDCAAS method
and in reducing the expense of implementing this new methodology
by eliminating the need for a bioassay. Therefore, FDA is
providing a limited data base on published true digestibility
values (determined using humans and rats) of commonly used foods
and food ingredients, which manufacturers may use to calculate
the PDCAAS for food products. The agency has decided not to
publish the digestibility values in the Code of Federal
Regulation at this time because the values are interim and
subject to change on a frequent basis. The data base is being
published in Appendix B to this document and is also available
from the Division oistration, 200 C St. SW., Washington, DC
20204.
    Appendix B lists foods from nine major food groups. In the
development of this data base, the agency examined scientific
data that included reports by national and international
organizations, review articles and other scientific articles. In
examining the data, FDA first considered true digestibility
values of protein foods obtained using adult subjects, followed
by data using the rat as an animal model. The agency did not
consider digestibility data obtained using in vitro methods or
other animal species. Comparative reviews of digestibility of
some protein using humans and the rat model suggest that the true
digestibility of a variety of foods is similar in humans and
rats.
    There are gaps in knowledge of the digestibility of protein
in common food sources. Therefore, the data in Appendix B of this
document are tentative. FDA believes that with the implementation
of this regulation, better data will be forthcoming, and that, in
due course, it will be able to revise the data base. The agency
encourages industry to submit additional data to enable FDA to
expand the assortment of foods included in the data base and to
update current data.
    FDA concludes that it would be premature to convene an expert
group to develop a data base on digestibilities and PDCAAS
values. There is a need to allow time for the compilation of
reliable data based on digestibility and amino acid analyses
obtained by the methods specified in this regulation. FDA will
reconsider the idea of convening of an expert group on protein
quality as such data become available.
    The agency advises that manufacturers are not precluded from
using other analytical methods for their own quality control
purposes as long as they assure themselves that such unofficial
methods compare adequately with the official methods. For
compliance purposes the methods specified in the regulation will
be used by FDA.
    67. Several comments recommended that the proposed new method
(PDCAAS) for the evaluation of protein quality be eliminated from
the regulation. Some comments stated that the PDCAAS method will
not provide flexibility and will be unnecessarily burdensome and
expensive, because it requires that digestibility and amino acid
analysis be performed on every product for which a declaration of
the percent of the RDI for protein is made. One comment stated
that foods are often reformulated, creating an ongoing cost.
Several comments expressed concern that, because of the costs,
the PDCAAS could have unintended negative effects on the
competitive position of smaller companies and on the willingness
of manufacturers to provide complete nutrition information to the
consumer.
    A few comments argued that for some foods, the PDCAAS will
result in lower values being declared for the percent of the RDI
than current methodologies using the PER, and that this will
effect the ability of the foods to make claims about protein
content. Another manufacturer opposed the change to new
methodology and commented that the PDCAAS methodology should be
reviewed and scrutinized by the AOAC before application.
    FDA does not agree that the PDCAAS should be eliminated. FDA
wishes to clarify that declaration of the percent DRV for protein
(which uses the PDCAAS method) is voluntary for foods intended
for adults and children 4 or more years of age unless a protein
claim is made for the product. Therefore, for this age group, the
burden and expense of the PDCAAS method are voluntarily assumed
by the manufacturer.
    FDA acknowledged in the preceding comment that values
obtained for percent of label reference value differ when
calculated using the PDCAAS rather than the PER. However, the
PDCAAS, based on human requirements, is inherently more
appropriate for assel more accurately describe the role of the
protein product in meeting human nutrition requirements.
    FDA advises that the analytical methodologies for amino acid
analyses involved in the calculation of the PDCAAS method have
undergone collaborative studies and have been published in the
Journal of the Association of Official Analytical Chemists.
    68. One comment expressed uncertainty about the proposed
amino acid scoring pattern used in calculating the PDCAAS and
stated that the WHO/FAO recommended further research to confirm
the currently accepted values of preschool children.
    FDA acknowledges that the WHO/FAO Consultation (Ref. 8)
recommended further research on the proposed scoring pattern to
confirm and reinforce the existing information. The Consultation
concluded, however, that the proposed scoring pattern is
robustand represents the best available estimate of indispensable
amino acids for this age group. Because of the high protein
requirements of the preschool age group for adequate growth and
development, protein foods and diets with an amino acid pattern
thateffectively meets the needs of the preschool child will
adequately meet the needs of older children and adults, whereas
the reverse may not be true (Ref. 10). Therefore, FDA concludes
that the proposed amino acid scoring pattern for preschool age
children is at present the most suitable pattern for use in the
evaluation of dietary protein quality for all age groups, except
infants.
    69. Comments agreed that the amino acid pattern for 1 to 4
year old children should be the same as the amino acid reference
pattern for 2 to 5 year old children when calculating the PDCAAS.
According to the data presented in the comments: (1) There is
little difference in the portion of protein and amino acids
needed for maintenance and growth between the two age groups; (2)
there is no sound nutrition rationale for using 70 percent of
casein as the reference standard for 1 to 3 year old children as
recommended by the Codex Committee on Nutrition and Foods for
Special Dietary Uses (Ref. 10a); and (3) there is no evidence
that the pattern of intake of amino acids for 1 to 3 year old
children differs from, or that the pattern is inadequate compared
to, the pattern for 2 to 5 year old children. The comments also
confirmed that there is sufficient overlap between the age groups
to render one standard adequate.
    In the preamble to the supplementary proposal, FDA
specifically requested comments on the inconsistency between the
FAO/WHO and the Codex Committee on Nutrition and Foods for
Special Dietary Uses standards for the protein quality of foods
intended for children 1 to 3 years old. The data presented in the
comments (Refs. 54 and 55) supported the agency's tentative
conclusion to use the amino acid scoring pattern for preschool 2
to 5 year old children for determining the PDCAAS of foods
intended for children over 1 year of age. Therefore, the agency
is maintaining the requirement in 101.9(c)(7) that the PDCAAS be
used to measure protein quality in foods for children above 1
year of age.
    However, the agency inappropriately left a parenthetical
notation in proposed 101.9(c)(8) that indicated that casein was
to be used as the reference standard for determining the PDCAAS
for children greater than 1 but less than 4 years of age. Because
by definition the PDCAAS uses an amino acid scoring pattern based
on human requirements as the standard, the agency has modified
proposed 101.9(c)(8), redesignated as101.9(c)(7), to remove the
reference to casein for that age group.
    70. Two comments disagreed with retaining the PER method and
the casein standard for assessing protein quality for infants.
The comments asserted that the requirement was not consistent
with the FAO/WHO Consultation recommendation for the use of the
    FDA acknowledges that the FAO/WHO Consultation (Ref. 8)
recommended that the amino acid composition of human breast milk
should be the basis of the scoring pattern to evaluate protein
quality in foods for infants under the age of one. However, in
the same document, the Consultation stated that further data on
the amino acid profile of human breast milk using standardized
methods of analysis are required to confirm the pattern for
calculating the chemical score of infant formulas (Ref. 8).
    Because of the uncertainties expressed in the FAO/WHO report
(Ref. 8) and the inconsistencies in reported amino acid patterns
of human breast milk (Ref. 56), the agency finds that it is
premature to use the FAO/WHO reference pattern for infants,
especially since this population group relies on relatively few
foods for nutrients. Until further data become available, the
safer course is to continue to use the current PER method using
casein as a standard. When more data become available, FDA
wouldbe willing to reconsider this position in response to a
petition.
    71. A few comments stated that the use of the PDCAAS will
understate the biological quality of vegetable proteins consumed
in a mixed diet. Another comment requested that FDA provide
manufacturers with ways to communicate the complementary nature
of different proteins from different sources.
    FDA agrees that use of the PDCAAS does not indicate the value
of individual proteins consumed as part of a mixed diet. However,
this is true with any method used to measure quality of proteins
in individual foods. The calculation of the corrected amount of
protein of a food does not take into account the complementary
potential of the food in a mixed diet, i.e., how a food rich in a
particular essential amino acid can ``complement'' a food low in
that amino acid to result in a total diet that provides
sufficient amounts of the amino acid. What the method does is
allow for a greater awareness of the value of protein sources
when consumed alone.
    While FDA acknowledges that more consumer education would be
helpful on the complementary effects of individual foods in mixed
diets, providing such information is beyond the scope of
nutrition labeling. Space limitations within the nutrition
labelgenerally prevent the addition of information to communicate
the complementary nature of different proteins. However, FDA
advises that the regulation does provide in 101.9(e) for the
voluntary inclusion of a second column to declare the nutrient
content of common combinations of foods (e.g., milk and cereal,
peanut butter and bread). It would be possible to declare in this
column the percent of the DRV or RDI for protein, as appropriate,
for the combination of foods. Also, the manufacturer may include
nonmisleading statements about the complementary nature of
protein sources in materials outside the nutrition label.
    72. Several comments expressed concern over the amino acid
analytical methodology and urged that high performance liquid
chromatography (HPLC) technology be incorporated into the
methodology, and that hydrolysis time be tailored for specific
foods. One comment suggested that FDA appraise the use of plasma
aminograms as indicators of protein quality.
    FDA agrees that the HPLC technology should be incorporated
into the suitable methodology for amino acid analyses. In the
preamble of the supplementary proposal, the agency stated that
the analytical methodology for PDCAAS is described in the
Reportof the Joint FAO/WHO Consultation, section 5.4.1 (Ref. 8).
The analytical methodology includes HPLC and provides flexibility
in the hydrolysis of specific foods.
    The agency has evaluated the merits of using plasma
aminograms for protein quality evaluation. FDA believes that the
method is not appropriate for this purpose. Current methodologies
using plasma amino acids for predicting the protein quality of
foo.
    73. One comment requested information on how to implement the
PDCAAS method and on whether commercial testing laboratories have
the necessary capabilities to determine the PDCAAS value.
    FDA advises that the methods for determining a food product's
PDCAAS is found in ``Protein Quality Evaluation, Report of a
Joint FAO/WHO Expert Consultation'' which is being incorporated
by reference into the final rule. As stated in101.9(c)(7)(ii),
this report is available from the Division of Nutrition, Center
for Food Safety and Applied Nutrition (HFF - 260), Food and Drug
Administration, 200 C St. SW., Washington, DC 20204, or is
available for inspection at the Office of the Federal Register,
800 North Capitol St. NW., suite 700, Washington, DC. To assist
persons in using this report, FDA has modified 101.9(c)(7)(ii)
to add the specific sections of the FAO/WHO report in which the
methodology is found. These sectionsare 5.4.1, 7.2.1, and 8.00.
    For those foods for which the digestibility factors are known
and found in FDA's interim data base, commercial testing
laboratories will be able to calculate the PDCAAS after running
an amino acid analysis as described in the FAO/WHO report,
section 8.00 (Ref. 8). The equipment necessary for amino acid
analyses is commonly used by commercial laboratories and should
be widely available.
    For those foods for which the digestibility factor is not
known, digestibility values must be determined in laboratories
according to methods in the FAO/WHO Report, sections 7.2.1 and
8.00 (Ref. 8).
    74. A comment noted that the agency had not specified
increments for reporting the ``Percent Daily Value'' for protein
as had been done in proposed 101.9(c)(11)(iii) for reporting the
``Percent Daily Value'' for vitamins and minerals.
    FDA acknowledges the oversight and is
modifying101.9(c)(7)(i) by specifying that the ``Percent Daily
Value'' for protein is to be declared to the nearest whole
percent as it is for fat and carbohydrate in 101.9(d)(7)(ii).
4. Conclusion
    In all other respects, 101.9(c)(8), redesignated
as101.9(c)(7), remains unchanged except that 101.9(c)(7)(iii)
has been reserved in this document. That paragraph is included in
the companion document entitled ``Food Labeling; Reference Daily
Intakes and Daily Reference Values'' published elsewhere in this
issue of the Federal Register in which the DRV for protein for
adults and children over 4 years of age and the RDI for protein
for infants, children less than 4 years of age, pregnant women,
and lactating women are established.

H. Vitamins and Minerals

    75. Retaining the requirement for mandatory listing of
vitamin A, vitamin C, calcium, and iron on the food label was
supported by a large number of comments representing a broad
spectrum of consumers and consumer organizations, public health
organizations, health care professionals, industry
representatives, and trade associations. These comments agreed
with the rationale stated in the proposal for continuing the
mandatory declaration of these nutrients in nutrition labeling.
    There were, however, some comments that did not support the
mandatory listing of these nutrients. Some comments suggested
that vitamin A and vitamin C should not be mandatory but should
be allowed on the food label on a voluntary basis. One comment
questioned whether inadequate intake of these vitamins is a
public health issue, noting that some milk is fortified with
vitamin A and stating the belief that consumers are aware that
citrus fruits are sources of vitamin C. One comment noted that
the IOM report (Ref. 1) recommends that vitamin A and vitamin C
should be allowed on the food label rather than required.
Additionally, a few comments recommended voluntary rather than
mandatory declaration of calcium and iron.
    In view of the strong support for the mandatory listing of
vitamin A, vitamin C, calcium, and iron on the food label, and in
the absence of strong opposition to the agency's proposal to
require the listing of these nutrients on the label, FDA is
notpersuaded that voluntary listing of these nutrients is
desirable and in the interest of the public health. While the IOM
report does suggest that vitamin A and vitamin C could be
allowed, rather than required, on the food label, it identifies
vitamin A and vitamin C as potential public health issues and
states that certain subpopulations are still at risk for
deficiencies of these vitamins (Ref. 1). The report states that
inadequate dietary intake of vitamin A is found in children under
5 years of age, and that two segments of the population are at
risk for vitamin C deficiency (infants fed cow's milk exclusively
and elderly individuals on inadequate diets).
    FDA continues to believe that public health concerns exist
for vitamin A in these at-risk groups and for the general public.
While fortification of certain foods, such as low fat and skim
milk, has helped to improve intakes of this vitamin among healthy
persons consuming a balanced diet, the inclusion of adequate
vitamin A in the diet still requires care and effort on the part
of a consumer in selecting good food sources of this vitamin.
Vitamin A is found in a relatively limited number of foods within
the food supply, and these foods must be selectively chosen by
consumers on a regular basis to ensure adequate intake.
    FDA also continues to believe, as supported by numerous
comments, that vitamin C is a nutrient with public health
significance, in that, even with fortification efforts and
greater year-round availability of citrus fruits and dark green
vegetables, certain subpopulations are considered at risk (see 55
FR 29487 at 29501).
    In the case of calcium and iron, these minerals are
identified as public health issues in the IOM report (Ref. 1) and
by numerous other sources, including the Surgeon General's report
(Ref. 2), Diet and Health (Ref. 3), and the report on
``NutritionMonitoring in the United States'' (Ref. 49).
    Therefore, 101.9(c)(11)(ii), redesignated
as101.9(c)(8)(ii), requires vitamin A, vitamin C, calcium, and
iron as mandatory elements of nutrition labeling.
    76. Many comments were received from persons at risk of iron
overload, particularly hemochromatotics, supporting mandatory
labeling of iron and requesting that the food label declare both
added and naturally occurring iron.
    FDA has carefully considered these comments. The agency
recognizes that a segment of the population is at risk of iron
overload. In deciding whether the declaration of a nutrient or
component on the food label should be mandatory, however, the
agency must consider the broad public health significance of its
action. Inadequate intakes of dietary iron are responsible for
the most prevalent form of iron deficiency in the United States.
Iron deficiency remains a risk for certain segments of the U.S.
population, notably young children, adolescents, women of
childbearing age, and pregnant women, especially those with low
incomes (Refs. 2, 3, 23, and 49). Thus, public health concerns
relative to iron, as stated in the National Nutrition Goals for
the Year 2000 (Ref. 47), center on the prevention of iron
deficiency and support increased dietary intake of iron among
children 1 to 2 years of age, women 20 to 44 years of age, and
low-income pregnant women. The agency believes that the listing
of ironon the food label aids the consumer in making individual
food selections in structuring the total diet, and that this
total diet has significant effects on health.
    However, as discussed in comment 54 of this document
concerning added sugars, the agency has taken the position that
it should not attempt to regulate actions that it cannot enforce.
Because available laboratory analytical methods do not
differentiat
    Therefore, the agency is denying the request that
manufacturers declare on their labels separate quantitative
amounts of added and naturally occurring iron.
    For the segment of the U.S. population at risk of iron
overload, the agency notes that the food label will provide
quantitative declaration of iron and vitamin C content of a food,
as well as a listing of ingredients (including iron compounds if
iron is added to the food). As absorption of nonheme iron may be
enhanced by consumption of vitamin C containing foods, those at
risk of iron overload can decrease their simultaneous consumption
of foods containing iron and vitamin C by using the information
on the iron and vitamin C content of foods found on the food
label.
    77. The majority of comments that the agency received from
consumers, health care professionals, public health agencies,
universities, industry, and trade associations agreed with FDA's
proposal to allow thiamin, riboflavin, and niacin to be
listedvoluntarily unless a claim is made, or unless these
nutrients are added to a food. Most comments based their position
on the decline of public health concern for deficiencies of these
vitamins over the past 20 years.
    On the other hand, there were some comments that advocated
continued mandatory listing of these vitamins on the food label.
Several comments expressed the opinion that mandatory inclusion
of thiamin, riboflavin, and niacin within nutrition labeling
contributed to the reduction of the incidence of deficiencies of
these vitamins in the United States. One comment stated that
these vitamins continue to be important to a significant portion
of the U.S. population, that listing these vitamins on the label
provides information on the nutritional properties of a food, and
that the 1990 amendments direct FDA to mandate declaration of any
vitamin that the agency deems to be important for the maintenance
of healthy dietary practices.
    FDA does not agree that the listing of thiamin, riboflavin,
and niacin on the food label has been the major cause of the
declining incidence of deficiencies of these vitamins. Rather,
the agency believes that the variety and abundance of the food
supply and the enrichment of many standardized foods with these
vitamins are the primary factors responsible for reducing the
occurrence of deficiencies of these vitamins (Ref. 57).
    FDA acknowledges that these vitamins continue to be important
nutrients, and that listing these vitamins on the label provides
information on the nutritional properties of a food. However, the
agency notes that while the 1990 amendments direct the agency to
include in the nutrition label information that will assist
consumers in maintaining healthy dietary practices, not all
information related to maintaining healthy dietary practices can
be included on the food label. If all such information were
included, all essential nutrients would be declared on the
nutrition label. Not only would space constraints not allow for
this, but the large amount of information would interfere with
consumers' abilities to use the information of the greatest
public health significance (see discussion in response to comment
3 of this document). Such a result would be contrary to the
intent of Congress (Ref. 16).
    For this reason, FDA developed criteria in its mandatory
nutrition labeling proposal to assist it in determining which
nutrients to require in nutrition labeling (55 FR 29487 at
29493). These criteria specify that nutrients should be required
when quantitative intake recommendations have been made in
scientific consensus documents, and when the nutrient is of
particular public health significance. Based on the preponderance
of comments that agreed with FDA's assessment that thiamin,
riboflavin, and niacin are no longer of particular public health
significance, FDA has decided to provide in101.9(c)(11)(ii),
redesignated as 101.9(c)(8)(ii), for the voluntary declaration
of thiamin, riboflavin, and niacin.
    78. Several comments requested that FDA clarify whether
thiamin, riboflavin, and niacin are required to be listed on the
nutrition label of a product made with enriched flour, a
standardized food, if no claim is made about these enrichment
nutrients other than their listing in the ingredient statement as
part of enriched flour. Similarly, another comment suggested that
FDA explicitly state, as in current 101.9(h)(7), that labeling
of voluntary nutrients will not become mandatory if present ina
food product as part of an enriched ingredient that has a
standard of identity. The comment also requested that nutrients
added strictly for a technological effect not be required to be
declared in nutrition labeling, in a similar fashion to current
101.9(h)(6).
    Proposed 101.9(c)(11)(ii) stated that vitamins and minerals
(other than vitamin A, vitamin C, calcium, and iron which must be
declared) need only be declared in the nutrition label when they
are added as a nutrient supplement, or when a claim ismade about
them. FDA's intent in this section, which is redesignated as
101.9(c)(8)(ii), was that when a food product is made with
enriched flour as an ingredient, but the label does not make an
``enriched'' claim or use ``enriched'' in the name ofthe food,
the nutrition label need not declare the enrichment nutrients.
If, however, the product is made with unenriched flour and
supplemented with nutrients as ingredients to achieve the
equivalent of a product made with enriched flour, the product's
label must list the enrichment nutrients in the nutrition label.
Information on the amount of the enrichment nutrients is also
required if an ``enriched'' claim is made on the label, or if
``enriched'' is used in the name of the food. Section
101.9(c)(8)(ii) is modified to clarify this requirement.
    FDA agrees with the comment that nutrients that are not
required to be declared in the nutrition label and are added to a
food strictly for a technological effect need not be declared if
the nutrient is declared solely in the ingredient statement and
is otherwise not referred to on the label or in labeling or
advertising. This provision, similar to current 101.9(h)(6), is
added to 101.9(c)(8)(ii).
    79. Several comments stated that listing of other vitamins
and minerals should be required, even without a claim, such as
vitamin D, magnesium, and phosphorus. One comment supported the
listing of all vitamins (even those absent from the food or food
product).
    FDA notes from these comments consumer interest in a variety
of nutrients but points out that not all nutrient information
related to maintaining healthy dietary practices can be included
on the food label. As discussed in comment 3 of this document,
the agency must be selective with regard to the information that
it requires to be listed on the label. Thus, it emphasizes
nutrients or components of particular public health significance.
FDA does not believe that all vitamins and minerals are ofequal
public health significance, a view that is supported by the IOM
and Nutrition Monitoring reports (Refs. 1 and 49). The agency is
also aware that space limitations on the food label require that
it use discretion in deciding which nutrients it requires to be
listed there.
    FDA does not agree that vitamin D, magnesium, or phosphorus
are of particular public health significance in the United
States. Because the human requirement for vitamin D can be met
with sufficient exposure to sunlight, and because milk and other
foods are fortified with vitamin D, deficiencies in this vitamin
are very rare (Ref. 23). Magnesium and phosphorus are cited in
both the Nutrition Monitoring and IOM reports as food components
that are not currently public health issues (Refs. 1 and 49).
FDA, therefore, is not requiring mandatory listing of vitamin D,
magnesium, or phosphorus in the nutrition label.
    80. Some comments suggested not requiring any vitamins or
minerals on the food label unless claims are made, or the
nutrient is added to the food, in order to minimize the space
requirements of nutrition labeling.
    As discussed in the supplementary proposal (56 FR 60366 at
60368), FDA interprets section 403(q)(1)(E) of the act to require
the inclusion of vitamins and minerals currently required
in101.9(c)(7)(iii) if the Secretary (or FDA, by delegation)
determines that such information will assist consumers in
maintaining healthy dietary practices. For the reasons discussed
above, FDA, has determined that vitamin A, vitamin C, calcium,
andiron meet the criterion in section 403(q)(1)(E) of the act
and, therefore, must continue to be required elements of
nutrition labels. The 1990 amendments did not provide for
nutrients to be omitted to save space. Therefore,101.9(c)(8)(ii)
continues to require declaration of vitamin A, vitamin C,
calcium, and iron.
    81. One comment requested that the final regulations allow
for the voluntary identification of foods that are important
sources of beta-carotene, either as a subset of vitamin A or
through an independent designation. The comment stated that
beta-carotene may reduce the risk of chronic diseases and appears
to have its own independent biological functions in addition to
serving as a source of vitamin A. The comment also noted that
``Recommended Dietary Allowances'' (10th ed.) states that ``For
food products containing large quantities of carotenoids, it
would be advisable in nutrition labeling to distinguish between
retinol, which in large amounts is toxic, and carotenoids, which
are not'' (Ref. 23).
    The agency has carefully reviewed the relationship of
beta-carotene to cancer in the companion document entitled ``Food
Labeling; Health Claims and Label Statements; Antioxidant
Vitamins and Cancer,'' published elsewhere in this issue of the
Federal Register. Based on that review and the stated
recommendations in ``Recommended Dietary Allowances,'' FDA has
concluded that there should be a method within nutrition labeling
to voluntarily distinguish the amount of beta-carotene present in
foodproducts. Accordingly, FDA is adding101.9(c)(8)(vi), which
states that the percent of vitamin A that is present as
beta-carotene may be declared to the nearest 10-percent increment
immediately adjacent to or beneath the nutrient name (e.g.,
``Vitamin A (90 percent as beta-carotene),''see example in
appendix C).
    82. A few comments objected to FDA's proposed deletion of the
synonyms vitamin B1 and B2 for thiamin and riboflavin,
respectively. The comments argued that many consumers continue to
use these terms and understand them better than the
``scary-sounding chemical'' names. Similar comments were received
in response to proposed 101.36 Nutrition labeling of dietary
supplements of vitamins and minerals.
    Based on the comments, the agency has reassessed its position
on this issue. FDA believes that for consistency the chemical
name of the nutrient (i.e., thiamin and riboflavin) must always
be given when the nutrient is declared in nutrition labeling, or
when claims are made (e.g., ``high in thiamin''). However, the
agency will not object to the voluntary parenthetical listing of
``vitamin B1'' or ``B1'' following ``thiamin'' and ``vitamin B2''
or ``B2'' following ``riboflavin.'' Accordingly, FDA has modified
101.9(c)(11)(v), redesignated as 101.9(c)(8)(v), by adding
vitamin B1 and vitamin B2 as synonyms for thiamin and riboflavin,
respectively. While FDA believes a similar change is appropriate
in proposed 101.36, given the requirements of the DS Act, FDA is
taking no action with respect to dietary supplements, and thus is
not acting on proposed 101.36, at this time.
    83. One comment suggested that other synonyms be allowed,
namely ``pyridoxine'' as a synonym for vitamin B6 and
beta-carotene for vitamin A when the vitamin A is solely
beta-carotene.
    FDA rejects this comment. Pyridoxine is only one of three
different forms of vitamin B6 (Ref. 23). In addition, the agency
believes that it would be a more difficult term for consumers to
use and understand. In regard to vitamin A, FDA believes new
101.9(c)(8)(vi) (see comment 81 of this document) is a
preferable course because in most foods, beta-carotene is only a
fraction of the total vitamin A content.
    84. FDA received a few comments that addressed increments for
those nutrients that are expressed as a percent of a reference
standard. One comment proposed a more complex incremental scheme
than that in the proposed rules, suggesting 1-percent increments
up to and including the 5-percent level, 2-percent increments
from 6 percent up to and including the 12-percent level,
5-percent increments from 15 percent up to and including 50
percent, and 10-percent increments above the 50 percent level.
Another comment suggested increments of 5 percent up to the
50-percent level and 10-percent increments thereafter. This
comment suggested that FDA not permit the use of 2-percent
increments because the necessary measurements are not accurate
enough to allow for such small increments.
    FDA proposed to maintain the current increments for vitamins
and minerals, i.e., percentages are expressed in 2-percent
increments up to and including the 10-percent level, 5-percent
increments above 10 percent and up to and including the
50-percentlevel, and 10-percent increments above the 50-percent
level. FDA considered both the comment suggesting a more complex
incremental system and the comment suggesting omission of the
2-percent increments. Neither suggestion provided sufficient
justification for the change. One appears to believe that an
accuracy of 1 percent of the RDI is necessary, the other that an
accuracy at 5 percent of the RDI is sufficient. Inasmuch as the
agency has experienced no problems with the increments that have
been in use since the early 1970's, and given that so few
comments addressed this issue, the agency sees no need to modify
the incremental scheme in proposed 101.9(c)(11)(iii),
redesignated as 101.9(c)(8)(iii).
    85. One comment noted that regulations specified that all
nutrients except vitamins and minerals were to be expressed to
the ``nearest'' unit or increment (e.g., total carbohydrates are
to be expressed to the nearest g, and sodium to the nearest 5-mg
increment between 5 and 140 mg of sodium and to the nearest 10-mg
increment above 140 mg of sodium). The comment asked for
direction on reporting amounts of vitamins and minerals.
    To clarify the regulations and promote consistency, FDA is
modifying 101.9(c)(8)(iii) to specify that vitamins and minerals
are to be expressed to the nearest 2-percent increment up to and
including the 10-percent level, the nearest 5-percent increment
above 10 percent and up to and including the 50-percent level,
and the nearest 10-percent increment above the 50-percent level.
    86. A comment objected to the provision in proposed
101.9(c)(11)(iii) that allows vitamins and minerals that are not
present to be represented by an asterisk that refers to a
statement ``Contains less than 2 percent of the Daily Value of
this (these) nutrient (nutrients).'' The comment stated that
consumers might be misled into thinking that small amounts of the
vitamin or mineral are present when they are not.
    FDA considered this comment and has concluded that the
flexibility the use of the asterisk provides in allowing
manufacturers to reduce the space needed for nutrition labeling
outweighs any slight misunderstanding about the amount of a
vitamin or mineral present in a food that might result. Amounts
of either zero or less than 2 percent of the RDI (declared as
Percent Daily Value) for these nutrients are physiologically
insignificant.
    The RDI's are provided in 101.9(c)(8)(iv) which has been
reserved in this document. That paragraph is included in the
companion document entitled ``Food Labeling; Reference Daily
Intakes and Daily Reference Values'' (hereinafter referred to as
ths FDA from promulgating regulations based upon recommended
daily allowances of vitamins and minerals other than the U.S.
RDA's currently specified in 101.9(c)(7)(iv) until November 8,
1993, 101.9(c)(8)(iv) includes values for only one age group
(i.e., adults and children 4 or more years of age) rather than
the 5 proposed groups (i.e., adults and children 4 or more years
of age, children less than 4 years of age, infants, pregnant
women, and lactating women). FDA intends to adopt in accordance
with section 203 of the DS Act, appropriate RDI's for all groups.
Therefore, FDA has adopted the references to such groups in
101.9 (a)(4), (c)(8)(i), (e), and (1) of this final rule, even
though such values do not exist at this time. In the meantime,
suggested RDI values for other age groups, which, to be
consistent with the DS Act, are based on the 1968 RDA's are
presented as guidance in the preamble of the RDI/DRV final rule.

IV. Analytical Procedures

A. General Issues

    87. Several comments asserted that FDA should explicitly
state the methods to be used for the analysis of various
nutrients. Some comments expressed the opinion that the agency
should not mandate listing of any nutrient when there are serious
issues with the reliability of the analytical method. Dietary
fiber was specifically cited as one example. One comment added
that FDA must specify the method of analysis for analytes not
available in the AOAC.
    The agency acknowledges the concern expressed in the
comments. In the mandatory nutrition labeling proposal (55 FR
29487 at 29498), FDA discussed the analytical methodologies for
sugars and dietary fiber. The agency noted that in the 17 years
since the promulgation of 101.9, it had acquired substantial
experience under the regulation, and techniques for analyzing
foods for their nutrient content have greatly improved. The
agency considers it is inadvisable to explicitly state a method
for a particular nutrient. The applicability of a specific method
to products of different matrices varies. As noted in several
comments, values for some nutrients, such as fat, are dependent
upon the procedure used. If a specific method is cited, it
maygive the erroneous impression that other methods that are more
appropriate to the matrix or that utilize newer techniques could
not, or would not, be acceptable. It is FDA's policy and practice
that any method used to support a nutrient declaration value
requires appropriate validation if it has not been collaborated
for that nutrient in a specific matrix. Validation procedures are
a necessary component of sound analytical technique and are
frequently used even with official, collaborated methods.
    The agency agrees that no nutrient should be a required
component in nutrition labeling if there is no satisfactory
analytical method for determining its level in a food. In fact,
this view was a major factor in the agency's decision not to
requiredeclaration of complex carbohydrates. FDA believes that
there is adequate methodology to assay for the nutrients that it
has made mandatory elements of the nutrition label, even, as
explained below, for dietary fiber.
    Analysis is not needed for nutrients, however, where reliable
databases or scientific knowledge establish that a nutrient is
not present in the product. For example, there is no need to
analyze for cholesterol in fruits and vegetables or for dietary
fiber in seafood. Costs associated with nutrition labeling will
be contained by not analyzing for a nutrient where there is no
reasonable expectation that the nutrient occurs in the food.
    88. Some comments noted that analytical variability--which
ranges from 1 percent to as high as 20 percent according to one
comment--may be a function of the method selected and its
inherent variability, the laboratory performing the analysis, the
levelarbohydrates or vitamin C in potatoes. These comments
observed that the nutrient levels may differ between the time of
harvesting and processing, as well as after a period of storage.
One comment recommended that FDA allow flexibility in selecting
analytical methodology such that there would be a broadened range
of methods used to generate nutritional information.
    FDA advises that manufacturers are free to use methods of
their choice for ascertaining the quantity to declare on the
label as well as for screening purposes as part of their quality
control procedures. However, when questions arise as to the
validity of the data, the agency will utilize the methods of the
AOAC or other validated procedures.
    Given the analytical problems in determining values for
complex carbohydrates, the agency has deleted the requirement for
declaring complex carbohydrates and is eliminating the term
``complex carbohydrate'' from the nutrition label. As discussed
above, it is using instead the term ``other carbohydrate.'' The
term ``other carbohydrate'' is defined as the difference between
total carbohydrates and the sum of dietary fiber, sugars, and
sugar alcohols (when declared). Because a specific method of
analysis is no longer required for complex carbohydrates, the
concern about measuring this food component in potatoes has been
addressed. In regard to the concern about the analysis for
vitamin C in potatoes, FDA advises that in this situation vitamin
C is a naturally occurring, or Class II, nutrient. Thus, the
declaration is in compliance if the nutrient is present at a
level of 80 percent or more of the declared label value. It
should be noted that current regulations and 101.9(g)(6) permit
reasonable excesses within current good manufacturing practice
for both vitamin C and other carbohydrate.
    As more nutritional analyses are performed in support of
label values, more methodologies will be validated. As a result,
the number of methods that manufacturers may use in determining
the amount of a nutrient will increase. Moreover, products that
heretofore had not been labeled with nutrition information will
now be subjected to testing. These new matrices will create new
challenges for both the food industry and the agency. However,
these challenges should not impede the development of full,
accurate nutrition information on food labels. The agency is
committed to working with industry to provide valid nutrition
label information that will promote selection of healthier diets
by U.S. consumers.
    89. Some comments suggested that FDA work with trade
associations and industry on the analytical techniques required
to prepare nutrition labels. One comment recommended that
designations be made as to which food matrices are appropriate
for existing methods and which ones are not.
    FDA agrees it should be actively involved in the review of
suitable methods to be used in the implementation of mandatory
nutrition labeling. The AOAC Task Force on Nutrient Labeling
Methods was established early in 1992 by AOAC for the purpose of
assisting its membership in meeting the requirements of the
agency's regulations. The agency worked closely with the Task
Force, participating in meetings as well as in evaluating
appropriate methods for various matrices. The Nutrient Labeling
Task Force Report on Analytes for Nutritional Labeling is
available from the agency or AOAC. The report lists the methods
that are adequate for various nutrients and various matrices. As
pointed out in the official AOAC publication, The Referee (Ref.
58), not all analyte/matrix combinations in the report have been
fully collaboratively studied, however.
    In this context, it should also be noted that 101.9(g)(2) of
these final rules allows for the use of other reliable and
appropriate analytical procedures if no AOAC method is available
or appropriate. Sources of such methods include FDA's ``Lipihods
may be found in ``Approved Methods of the American Association of
Cereal Chemists'' and ``Official Methods and Recommended
Practices of the American Oil Chemists Society.''
    The method of analysis used must be suitable to achieve the
purpose for which it is used. For example, the method used to
quantify vitamin C for nutrition labeling must be able to
determine whether ascorbic acid or isoascorbic acid is present in
the food. Isoascorbic acid and sometimes ascorbic acid are used
as antioxidants in food processing. Only ascorbic acid, however,
is an active form of vitamin C and considered in the
determination of vitamin C content of the food. Thus, the method
must be able to distinguish ascorbic acid from isoascorbic acid.
    90. FDA received several comments regarding the use of the
Official Methods of Analysis of the AOAC. One comment stated that
the latest edition of this reference should be cited to avoid
obsolescence when new editions are issued.
    FDA does not have authority to not reference a particular
edition of the Official Methods. The Office of the Federal
Register requires that each statement of incorporation by
reference into the Code of Federal Regulations contain specific
information, including the date and edition of the publication.
Accordingly, FDA has not modified 101.9(g)(2).
    91. A comment supported a policy whereby FDA would verify
laboratory analysis results on file at a firm to substantiate the
nutrition label information in lieu of doing nutrient analysis
from a limited sample of products. The comment expressed the
opinion that FDA should be required to perform additional
sampling and testing and to consider the statistical variation
inherent in test procedures before initiating a legal action,
such as a seizure.
    The agency disagrees with the comment. FDA is a law
enforcement agency, and its mission is consumer protection. To
support a misbranding charge for inaccurate nutrient content
information, FDA must have accurate, reliable, and objective data
to present in a court of law. To obtain that information, FDA
relies upon the work performed by its trained employees because
it does not have legal authority in most instances to inspect a
food manufacturing firm's records.
    The practice of performing nutrient analysis from a composite
of 12 subsamples is well established. Compositing the contents of
the twelve containers yields a numerical result essentially
equivalent to what would be obtained if each container were
analyzed, and the results averaged. Thus, the composite value is
considered to be the same as the average of a sample of twelve
containers. As noted in 101.9(g)(4) and (g)(5), FDA will not
take regulatory action based on a determination of a nutrient
value that fails to meet appropriate levels by a factor inherent
in the variability generally recognized for the analytical method
used on that food at the level involved.

B. Calories

    92. Comments stated that the regulations should clarify how
calories are to be calculated. Several comments recommended
adding ``caloric content may be determined by the Atwater
method'' to proposed 101.9(c)(3). Some comments objected to the
use of the specific Atwater food factors published in ``USDA
Handbook 74'' which have not been updated since 1955. Another
comment noted that if a food item is a commodity-type product for
which a specific Atwater factor is available, the caloric content
for these products should be required to be calculated using the
specific Atwater factors.
    Several comments disagreed with the proposal to subtract
dietary fiber from the amount of carbohydrate before applying the
general factor of 4 (i.e., 4 calories per g of carbohydrate).
These comments contended that the general factor is intended
toapply to total carbohydrate including fiber. Because the
gastrointestinal effects of dietary fiber were taken into account
in the derivation of the general factors, these comments did not
consider it to be legitimate to exclude fiber from carbohydrate
cont when calculating caloric content.
    One comment suggested that calories be calculated from
carbohydrate-plus-dietary fiber if the general factor of 4
calories per g of carbohydrate is used. Alternatively, the
comment suggested that calories be calculated from available
carbohydrate if the general factor of 3.75 calories per g of
carbohydrate is used. The factor of 3.75 calories per g for
carbohydrate is used by the United Kingdom for calculation of
available carbohydrate energy (Ref. 61).
    One comment suggested that both total dietary fiber and other
nondigestible carbohydrate should be subtracted from the total
carbohydrate content before calculating calories contributed by
carbohydrates. As noted in the comments, many new food
ingredients such as reduced-calorie fats, fat substitutes, and
modified carbohydrates have been developed in recent years. Some
of these ingredients have caloric values substantially less than
the general factors of 4, 4, and 9 for protein, carbohydrate,and
fat, respectively. Comments requested specific allowances for
ingredients used as reduced calorie replacements for conventional
ingredients to permit methods for calculating the available
calories other than use of the general factors.
    The agency recognizes that confusion may exist about methods
for calculating caloric content because of the proposed changes
in how total carbohydrate content has been defined in
101.9(c)(6) and because of the changes in the treatment of
dietaryfiber. Therefore, the agency is modifying 101.9(c)(1) to
clarify how caloric content is calculated by providing five
options for calculating the energy value of foods in
101.9(c)(1)(i).
    The first option, which is set forth in 101.9(c)(1)(i)(A),
is the use of specific Atwater factors that are found in Table 13
in ``Energy Value of Foods--Basis and Derivation'' by A. L.
Merrill and B. K. Watt, USDA Handbook No. 74 (1955). FDA
disagrees with the comment that suggested requiring the use of
specific Atwater food factors for those foods for which such
factors exist. The agency does not believe that there is any need
to limit a manufacturer's flexibility in selecting a method for
determining caloric content. Current regulations do not require
the use of specific Atwater food factors, and no data were
presented to support a change in current practices.
    The second and third options utilize the general factors of
4, 4, and 9 calories per g for protein, carbohydrate, and fat,
respectively. In 101.9(c)(1)(i)(B), which provides for
calculating calories by general factors, dietary fiber is
includedin total carbohydrate. FDA also recognizes, however, that
doing so can result in significant error for the caloric value of
some foods because of the relatively low energy value of dietary
fiber. Adjustments for dietary fiber content are therefore
appropriate for nutrition labeling of some foods.
    However, because some soluble dietary fiber can make a
significant contribution to a food's energy value (Ref. 61), FDA
does not consider it appropriate to allow an absolute exclusion
of all dietary fiber from caloric calculation. Recognizing that
there can be significant levels of available energy in some
soluble fiber, and that official AOAC methods for dietary fiber
now provide for separation of soluble and insoluble fiber, the
agency considers it appropriate to permit exclusion of the
insoluble component of dietary fiber alone from calculation of
carbohydrate calories. Accordingly, FDA has added
101.9(c)(1)(i)(C) in the final rule to permit calculation of
caloric contribution from the carbohydrate portion of food by
multiplying carbohydrate content minus insoluble dietary fiber
content by the general factor of 4 calories per g.
    In addition, 101.9(c)(1)(i)(D) permits manufacturers or
users of soluble dietary fiber additives or other food additive
substances with reduced available energy to petition for use of
alternative energy factors in nutrition labeling through
established procedures for food additive or GRAS petitions.
Soluble dietary fiber substances are frequently added to foods to
replace fully caloric nutrients in formulating reduced calorie
foods. In such cases, the burden for establishing the actual
energyvalue of the food is appropriately with the manufacturer.
    The calculation of the caloric contribution of novel fats and
carbohydrates has been discussed in section III. of this
document. The agency has stated that it will consider
digestibility of new products on a case-by-case basis as
requested. In support of this action, the agency requests that
manufacturers who wish to declare adjusted values for the energy
contribution of a substance, based on reduced digestibility,
submit information on digestibility of the substance, analytical
assay procedures for the substance, and data on interference with
required methods of analysis. As stated in section III. of this
document, this information should be included in a food additive
petition or a petition for affirmation that the use of a
substance is GRAS. The agency will then publish the specific
digestibility coefficients in 21 CFR part 172 for food additives
and in 21 CFR part 184 for GRAS substances. These coefficients
can be utilized in determining the caloric value of specific food
ingredients.
    Other procedures may be required for particular foods and
will be addressed by other appropriate means. FDA is allowing for
this contingency in 101.9(c)(1)(i)(D) by adding ``or other
means, as appropriate.'' For example, in the voluntary nutrition
labeling program for raw fish, data were presented to FDA
supporting a value of fat and calories for the fish ``orange
roughy'' that omits a portion of the total fat since more than 90
percent of the fat in the product is in a wax ester that is
notmetabolized (Ref. 62). FDA published these corrected values
for available fat and calories in Appendix B ``Nutrition Labeling
Provided by FDA for the 20 Most Frequently Consumed Fish'' (57 FR
8175; March 6, 1992).
    To afford even more flexibility in determining caloric
content, FDA is including 101.9(c)(1)(i)(E), to provide for the
use of bomb calorimetry. The agency notes that the caloric value
so obtained must be corrected for nonmetabolizable protein by
subtracting 1.25 calories per g of protein to correct for
incomplete digestibility, as discussed in Energy Value of Foods,
Basis and Derivation, ``USDA Handbook No. 74'' (Ref. 63). The
caloric value determined by bomb calorimetry may give a higher
value than the other allowed methods. However, because it would
produce an over-estimation of the caloric content of the food,
FDA would not consider it to be disadvantageous to the consumer.
A primary consideration in selecting which method to use must be
the accuracy of the declaration of the caloric content in light
of the agency's compliance criteria in 101.9(g).
    The agency is aware that some manufacturers have developed
their own specific factors for conventional food ingredients that
they use in calculating the caloric content of their products.
FDA views this practice as analogous to using data bases to
determine nutrient label values, in that the manufacturer assumes
the responsibility for ensuring that the values obtained are
consistent with those obtained analytically by FDA. As such, the
agency does not believe it needs to provide for this option in
101.9(c)(1)(i).
    In summary, the agency is amending 101.9(c)(1) to permit
five optional methods for calculation of caloric content of
foods: (1) Specific Atwater food factors (i.e., the Atwater
method) given in Table 13, ``Energy Value of Foods--Basis and
Derivation,'' A.L. Merrill and B.K. Watt, USDA Handbook No. 74
(1955), (2) general factors of 4, 4, and 9 calories per g of
protein, total carbohydrate including dietary fiber, and total
fat, respectively, as described in USDA Handbook No. 74, (3)
general facy, as discussed in USDA Handbook No. 74, except that
insoluble dietary fiber content may be subtracted from total
carbohydrate content before calculating the caloric contribution
of the carbohydrate portion of the food; (4) specific factors for
particular food ingredients approved by FDA through incorporation
in 21 CFR parts 172 or 184 or other means, as appropriate, or (5)
bomb calorimetry data after subtraction of 1.25 calories per g
protein to correct for incomplete digestibility, as described
inUSDA Handbook No. 74, p. 10.
    By providing for these varied means of calculating caloric
content, FDA is giving manufacturers flexibility in how they
determine calorie content in a variety of foods, both
conventional foods and new foods developed to meet changing
marketing strategies.
    93. A recommendation was made in one comment that products
with a negligible amount of dietary fiber (suggested as less than
2.5 percent) should not be required to have dietary fiber
analysis for determination of caloric content.
    The agency advises that because revised 101.9(c)(6) now
includes dietary fiber in total carbohydrate content, separate
analysis for dietary fiber is no longer required for calculation
of either carbohydrate content or calories from carbohydrate.
Therefore, the concern expressed in this comment has been
addressed.
    94. Several comments asked for clarification on the
discussion in the mandatory nutrition labeling proposal (55 FR
29487 at 29493 and 29503) of the possible caloric contribution of
macronutrient substitutes or other ingredients such as certain
typesof soluble fibers or gums. While one comment agreed with the
agency's position that manufacturers of these ingredients should
be asked to provide evidence that these substances do not
contribute to the energy value of food, another comment found the
concern unwarranted and opposed the use of any correction factors
for calories from soluble fibers (e.g., gums). One comment noted
that more research is needed in this area.
    In the mandatory nutrition labeling proposal, the agency
expressed concern that available energy of soluble dietary fiber
food additives (e.g., gums) would not be included under the
agency's proposed method for caloric calculation, which excluded
energy contribution of all dietary fiber. Innovations in food
technology have resulted in reduced calorie foods that utilize
various soluble dietary fibers and other modified carbohydrates,
proteins, and fats for technical effects that allow reduction of
total fat content. The agency considered it inappropriate to
automatically assign a zero energy value to all soluble dietary
fiber additives when some of these substances may have available
energy. Likewise, some modified carbohydrate additives may have
less available energy than the 4 calorie per g assigned by the
general energy factor for carbohydrate but still have available
energy. The agency has determined that petitions regarding
specific caloric values for these types of food ingredients are
appropriate. FDA's new policy is discussed in comment 92 of this
document.

C. Fats, Fatty Acids, and Cholesterol

    95. As discussed in section III. C. of this document,
comments raised many questions about analytical procedures to be
used to measure fat and their reliability. In addition, several
comments expressed concern regarding the adequacy of methods for
measuring cholesterol. One comment cited a published article on a
method for measuring cholesterol that is undergoing collaborative
study under the auspices of the AOAC.
    The agency believes that its new definition for total fat in
101.9(c)(2) (i.e., total lipid fatty acids expressed as
triglycerides) will help to clarify what analytical procedures
are to be used by clarifying what compounds are to be included in
t total lipid fatty acids in a variety of product matrices. A
recent publication of the AOAC, The Referee (Ref. 32), contains a
compilation of these methods. Other reliable and appropriate
methods are also cited in comment 89 of this document.
    FDA notes that issues exist about the reliability of methods
for measuring low levels of fat. As discussed in a recent
article, fat determinations are reliable down to concentrations
of 1 to 5 g per 100 g, provided a large enough test portion is
taken to obtain at least 50 mg of weighable residue (Ref. 64).
The premise is that accuracy generally increases when larger
amounts are used for analysis so that there is always a minimum
quantity of extracted fat available for weighing.
    Although official analytical methodologies for determining
cholesterol content are somewhat limited at the present time, the
agency is pleased to note that comments indicate that this is an
area of active research. FDA, as a member of the AOAC Task Force
on Nutrient Labeling Methods, looks forward to the development of
additional collaborated methods for a range of matrices.

D. Dietary Fiber

    In proposed 101.9(c)(3) of the mandatory nutrition labeling
proposal and proposed 101.9(c)(7)(ii) of the supplementary
nutrition labeling proposal, FDA specified that total, soluble,
and insoluble dietary fiber content are to be determined by the
method ``Total Dietary Fiber in Foods, Enzymatic Gravimetric
Method.''
    96. One comment noted that satisfactory analytical procedures
for measuring dietary fiber are available and cited the American
Association of Cereal Chemists Method No. 32 - 21 and the
proposed AOAC method. This comment stated these methodologies
were at least as accurate as certain other sanctioned procedures.
It acknowledged that research should continue, however, to
improve the utility and standardization of analytical methods for
fiber. Another comment noted that the precision of the proposed
method may cause difficulties at low levels, typical of that
found in some fruits and vegetables (more than 1 percent to 5
percent) and especially when fat is present in the sample. The
comment stated that because of questions concerning the accuracy
of methods for measuring dietary fiber, companies may elect not
to declare low levels of fiber in their products. The comment
stated that there is a more accurate method for use in these
situations.
    Two comments from the meat industry expressed concern that
the proposed method for fiber had only been evaluated on cereals,
grains, and breads. They questioned the applicability of the
method to other types of products. According to another industry
comment, currently approved methods for analyzing dietary fiber
seriously underestimate dietary fiber content of high moisture
foods, which leads to inaccurate and misleading label
information. This comment said that a current analytical method
for high moisture products is unavailable. A comment expressed
the hope that the tests adopted to measure dietary fiber would
not falsely exclude low molecular weight bulking agents, such as
polydextrose.
    The few comments that addressed analytical methods for
soluble and insoluble fibers were split on whether available
methods are adequate or inadequate. One comment to the
supplementary proposal noted that the agency's cited method does
not measure soluble fiber directly. The comment said that the
method measures total dietary fiber and insoluble dietary fiber,
then calculates soluble dietary fiber as the difference between
the two.
    With the currently available AOAC methods for dietary fiber
and its components, FDA believes that suitable methodology exists
for the analysis of dietary fiber for nutrition labeling
purposes. Since the issuance of the supplementary proposal, two
addble and insoluble dietary fiber. The concern expressed by the
comments as to the availability of validated methods for
measuring dietary fiber is therefore alleviated. Because methods
for measuring dietary fiber are now included in the Official
Methods of Analysis of the AOAC, 101.9(c)(7)(ii) of the
supplementary proposal, which described dietary fiber
methodology, and 101.9(g)(2), which directs compliance by
official AOAC methods, are redundant. As such, 101.9(c)(7)(ii)
has not been included in the final rule.
    The enzymatic-gravimetric method (AOAC 15, 985.29) cited in
proposed 101.9(c)(7)(ii), is valid for high-moisture foods and
those with fat present in the product. The method specifies
drying conditions, as well as defatting procedures, that are to
be performed before analysis for total dietary fiber. If drying
conditions are a part of the analysis, analytical results must
incorporate the loss on drying to obtain the total dietary fiber
content of the ``as received'' product. Likewise, any loss of
weight from fat or sugar removal must also be compensated for in
the calculations.
    Regarding the comment on the appropriateness of including low
molecular weight bulking agents in dietary fiber, in the absence
of a consensus on a chemical definition for dietary fiber, the
available analytical methods dictate classification of
suchingredients. Some manufactured low molecular weight,
carbohydrate based, bulking agents, such as polydextrose, do not
analyze as dietary fiber in the official AOAC methods for
measuring dietary fiber. Such food ingredients, while not
reported as dietary fiber in nutrition labeling, would be
included in total carbohydrate and reported as ``other
carbohydrate.''
    97. A few comments recommended that dietary fiber be listed
in 0.5-g increments, believing it to be a meaningful quantity to
declare. Other comments concurred with the agency's proposal of
whole-g declaration for dietary fiber on the basis that 0.5g
requires greater analytical precision than is possible for
measuring dietary fiber.
    The agency disagrees with the first comment. No data were
presented to support a change to 0.5-g increments. Therefore, FDA
continues to believe that the precision of the analytical
methodology for determining quantitative amounts of dietary
fiberdoes not allow for accuracy to the 0.5 g level. Accordingly,
101.9(c)(6)(i) will require that dietary fiber be expressed to
the nearest g.
    98. One comment recommended use of the word ``fiber'' in lieu
of ``dietary fiber.'' The comment stated that consistency with
the 1990 amendments was not needed and was far less important
than using terms that consumers understood. The comment
alsocontended that insertion of the word ``dietary'' into each
term of fiber content would clutter the label.
    The agency disagrees with the comment. No data were presented
to support the contention that the term ``dietary fiber'' would
confuse consumers. FDA believes that it is important to
distinguish between dietary fiber and crude fiber to ensure
thatthere is no question as to what fiber components are
declared.
    99. One comment took exception to the agency's citation of
USDA Handbook 74 (1955) as the reference for the subtraction of
dietary fiber in the calculation of total carbohydrate at
101.9(c)(6) of the supplementary nutrition labeling proposal. In
the cited reference, dietary fiber is not a part of the
calculation. The comment noted that, as a defined, analyzable
entity, dietary fiber was unknown in 1955.
    FDA acknowledges the accuracy of the comment in regard to the
concept of what dietary fiber was in 1955. As noted previously,
the agency has modified 101.9(c)(6) so that dietary fiber is no
longer subtracted from the weight of the total food in t

E. Sugars and Other Carbohydrate

    In the mandatory nutrition labeling proposal, FDA discussed
the analytical methodologies for sugars and dietary fiber (55 FR
29487 at 29498). The agency acknowledged in the supplementary
proposal (56 FR 60366 at 60369) that analytical problems werea
concern for the mandatory declaration of sugars and complex
carbohydrate.
    100. Essentially all of the comments stated that current
methodology is inadequate to determine the levels of sugars and
complex carbohydrate, as defined. Other comments described the
methodology as unavailable, costly, difficult, and imprecise. One
comment noted that there are no current analytical standards for
measuring complex carbohydrates as defined in FDA's proposed
rule. There were no comments that provided references for
available, validated analytical methodology for these food
components. Another comment noted that assay techniques for the
quantitative determination of polysaccharides of 10 and higher
saccharide units are beyond the reasonable capabilities of many
in the food industry.
    One comment included an evaluation of two liquid
chromatographic procedures for monosaccharides through
pentasaccharides and delineated disadvantages of each. The
technique of high performance ion chromatography was identified
by the comment as thetechnique that could provide the most
accurate values for sugars. This technique has not however, been
studied collaboratively by the AOAC.
    Concerns expressed in regard to the analytical determination
of sugars and complex carbohydrates have been alleviated by the
revision of the definition of ``sugars'' to include only the sum
of mono- and disaccharides and of ``other carbohydrate'' asthe
difference between total carbohydrates and the sum of dietary
fiber, sugars, and sugar alcohols (when declared). There is
established methodology in the Official Methods of Analysis of
the AOAC for the determination of mono- and disaccharides in
several food groups with a high degree of confidence. Continued
analytical work will be necessary to validate methodology for a
wider, more diverse food supply.
    The agency will use the technique of HPLC in monitoring
compliance with label statements concerning sugars content. The
agency's use of this technique will not preclude the use of
emerging technologies such as high performance ion chromatography
orsupercritical fluid chromatography as they are developed and
validated.

F. Vitamins

    101. A comment stated that the agency should list those
carotene fractions that can be included in the declaration of
vitamin A for labeling purposes. It noted that a variety of HPLC
methods for vitamin A and carotene are available and currently
inuse by industry and FDA. The comment also stated that
analytical reagents required for AOAC official methods for
determining vitamin A content are no longer available.
    In its RDI/DRV proposal, the agency proposed that vitamin A
content is to be expressed in retinol equivalents (55 FR 29476 at
29485). One retinol equivalent was established to be equivalent
to 1 microgram (æg) retinol or 6 æg beta-carotene. The
nomenclature for vitamin A as retinol equivalents was carried
forward in the supplementary proposal at 101.9(c)(11)(iv). FDA
is aware of literature data where alpha-carotene is present in
some carrots in significant amounts. To account for this and
other carotene fractions, the agency also recognizes the National
Academy of Sciences' definition of retinol equivalents as 12 æg
of provitamin A carotenoids other than beta-carotene (Ref. 23).
    As noted above, the agency worked closely and actively with
the AOAC Task Force on Nutrient Labeling Methods to judge the
adequacy of AOAC methods to meet nutrition labeling needs. The
decreased availability of the analytical reagents for some method
vitamin A content in margarine is no longer readily available.
However, FDA advises that the Nutrient Surveillance Branch (HFF -
266), Center for Food Safety and Applied Nutrition, FDA, can
provide limited quantities of the dye upon request to the address
listed at the beginning of this document.

V. Format

A. Legal Authority for an Improved Nutrition Label Format

    Congress clearly intended that nutrition information be
presented to the public in a manner that facilitates
understanding of the information and that assists consumers in
maintaining healthy dietary practices. This fact is evidenced by
at least two provisions of the 1990 amendments. Section 403(q)(1)
of the act, which was added by the 1990 amendments, states:

    The Secretary may by regulation require any information
required to be placed on the label or labeling by this
subparagraph or subparagraph (2)(A) (section 403(q)(1) or (2)(A))
to be highlighted on the label or labeling by larger type, bold
type, orcontrasting color if the Secretary determines that such
highlighting will assist consumers in maintaining healthy dietary
practices.

In addition, section 2(b)(1)(A) of the 1990 amendments states
that the implementing regulations shall:

    * * * require the required information to be conveyed to the
public in a manner which enables the public to readily observe
and comprehend such information and to understand its relative
significance in the context of a total daily diet.

    Consistent with the authority vested in the Secretary (and
FDA, by delegation) to determine if specific label information
will assist consumers in maintaining healthy dietary practices,
the House report accompanying the 1990 amendments directs FDA to
consider a variety of format options, including: ``information
about the recommended daily intake, the use of descriptive terms
such as `high,' `medium,' and `low' or use of universal symbols
to indicate desirable or undesirable levels of particular
nutrients.'' The report goes on to state: ``While the bill does
not mandate any particular approach, it does require the
Secretary to specify requirements that would permit the consumer
to understand the nutrition information pertaining to a
particularfood in relation to recommended dietary information''
(Ref. 16).

B. The Role of the Nutrition Label

    The 1990 amendments provide several descriptions of the role
of the nutrition label. Section 403(q)(1) of the act, which was
added by the 1990 amendments, uses the language ``assist
consumers in maintaining healthy dietary practices.'' Section
2(b)(1)(A) of the 1990 amendments uses the language ``enables the
public to readily observe and comprehend such information and to
understand its relative significance in the context of a total
daily diet.'' In the format proposal, FDA requested comment about
how the nutrition label can best assume the information role
mandated by the 1990 amendments.
    102. A number of comments from food manufacturers, trade
associations, health promotion organizations, and consumer groups
identified more than one role for the nutrition label in
assisting consumers in maintaining healthy dietary practices. One
illustrative comment from a health professional organization
described two different roles of the food label as: (1) Helping
consumers choose appropriate foods and (2) helping consumers to
understand the ``importance of diet and proper dietary
behaviorsto a healthy life.'' Similarly, a comment from a trade
association made the distinction between the food label
``contributing to the consumer's understanding of the relative
significance of the food in the context of a total daily diet''
and providing``guidance on how to use information in the food
label to make appropriate food choices.'' Many comments made
similar distinctions between the food label helping to place the
particular product in the context of a daily diet and the food
label providinges. A number of comments from industry questioned
whether the act mandated an explicit educational role for the
nutrition label to provide guidance to consumers on how to
maintain healthy dietary practices.
    Many comments argued that the nutrition label cannot by
itself provide all the information important to maintaining
healthy dietary practices but reached different conclusions about
the relevance of this limitation for the nutrition label format.
Anumber of comments, particularly from industry, pointed out that
because of the limited space available on the food label, the
nutrition label cannot be expected to adequately convey all the
information consumers need to understand the importance of
nutrition information in maintaining healthy dietary practices.
These comments concluded that the role of the nutrition label
should be limited to providing factual, product-specific
information, and that the broader dietary guidance role should be
reserved to off-label activities of public and private nutrition
education programs. These comments asserted that these programs
will have sufficient time and space to inform consumers about the
concepts of flexibility and personal choice necessary to maintain
healthy dietary practices.
    Other comments, primarily from consumer organizations and
health professional groups, acknowledged the necessity of
off-label consumer education to help consumers understand how to
use the nutrition information to maintain healthy dietary
practices but saw the nutrition label as a useful food selection
tool that needs to be integrated with off-label educational
programs.
    FDA agrees that the nutrition label can and should help
consumers make informed food choices, and that it can also
contribute to helping consumers maintain healthy dietary
practices. The two roles are by no means inconsistent. To help
consumers make appropriate food choices contributes undoubtedly
to maintaining healthy dietary practices. Among those choices are
choices that will assist the consumer in maintaining healthy
dietary practices. Maintaining healthy dietary practices,
however, is a larger and more complex goal than informing food
choices, and one that requires motivation and knowledge of how to
combine and balance the many different kinds of foods and eating
occasions that constitute a total diet. The 1990 amendments
require the agency to take both senses of the possible role of
the nutrition label into account in evaluating alternative
formats for the nutrition label. However, the agency also agrees
that the mandated role of the nutrition label to assist consumers
in maintaininghealthy dietary practices does not encompass an
explicit educational role for the nutrition label to provide
dietary guidance to consumers.
    The agency believes that the nutrition label format needs to
give first consideration to helping consumers make informed food
choices by enabling them to both comprehend the nutritional value
of the food and to understand its relative significance in the
context of the total daily diet as called for in section
2(b)(1)(A) of the 1990 amendments.
    The agency's view is that the basic format elements that best
serve the mandated role of the nutrition label must be identified
and justified on the basis of consumer research. Therefore, the
implications of format elements for the use of the nutrition
label in assisting consumers to understand the nutritional value
of the food and to understand the food in the context of the
total daily diet were extensively examined in the agency's format
research.

C. Need for Consumer Research

    Section 2(b)(1)(A) of the 1990 amendments specifies criteria
for an acceptable format for nutrition label information. The
operative terms in this section, ``readily observe and
comprehend'' and ``understand its relative significance in the
context ofderstanding can only be measured objectively in terms
of behavior (i.e., in terms of how well consumers use a format
for a specific task). Formats and format elements can be assessed
subjectively by asking consumers or experts to judge the
usefulness of various formats. Behavior-based performance
measures, however, rather than subjective judgment, are generally
accepted as the more reliable and valid way to evaluate the
consequences of information displays on consumer perception and
understanding.
    Major scientific groups (Refs. 1, 65, and 66) urged FDA to
subject possible nutrition label formats to consumer testing to
objectively determine which formats can be used most effectively
by consumers. FDA has placed considerable emphasis on the
importance of consumer research in developing a new format for
the nutrition label because of this advice and because the
techniques of consumer research (surveys, focus groups,
experiments, and preference polls) provide the best and perhaps
the only possible bases for evaluating alternative nutrition
label formats against the consumer perception and understanding
criteria specified in the 1990 amendments.
    103. A number of comments argued that virtually any nutrition
label format, even the current format, can serve to help
consumers put foods in the context of a total daily diet
depending on the knowledge and understanding of the person
reading the label. To the same point, many comments recommended
nutrition education activities to supplement the public's
understanding of label information. Some comments suggested that
nutrition education activities can be an alternative to including
one or more information elements, such as a listing of DRV's for
certain macronutrients, on the nutrition label.
    FDA agrees that each person's knowledge is the necessary
context for understanding label information, and that nutrition
education activities can be an important complement to the
public's understanding of label information. FDA disagrees with
the implication sometimes drawn from these facts that FDA is
thereby relieved from the burden of adopting a format based in
part on the available evidence about what kind of format does the
best job at achieving the objectives of the 1990 amendments.
Although various considerations bear on the selection of a final
nutrition label format, FDA believes that an essential criterion
is how well a format conveys information that Congress expected
would be provided by the nutrition label. Congress expected that
such information would allow people to decide whether, based on
the nutrition content of the food, they would want to buy the
food (Refs. 67 and 68) and to understand the relative
significance of the food in the context of the daily diet
(section 2(b)(1)(A) of the 1990 amendments). FDA has sought to
measure, and has sought other information that measures, the
ability of various formats to achieve these objectives.

D. Consumer Research Submitted as Comments or Referenced in
Comments to the Format Proposal

1. Background
    The agency reviewed a number of qualitative studies (i.e.,
five focus groups, seven preference polls) and quantitative
studies (i.e., five surveys, seven experiments,) that were
submitted as comments or referenced in comments to the format
proposal. Consumer research studies about format issues were
conducted by FDA, food industry groups, individual food
companies, consumer groups, public health organizations, health
professionals, and academic researchers.
    Much of this work was done in response to FDA requests for
additional information, and became available only in comments
submitted in response to such requests. For example, FDA
published an advance notice of proposed rulemaking in the Federal
Register of August 8, 1989 (54 FR 32610), soliciting public
comment on a wide range of food labeling issues, including: (1)
Whether to revise requirements for nutrition labeling and (2)
whether to change the nutrition label format. FDA subsequentlyng,
the last of which was held in Atlanta, Georgia on December 13,
1989. This last public hearing focused on the nutrition label
format.
    Additionally, in the Federal Register of May 20, 1991 (56 FR
23072), FDA published a notice announcing the availability of a
report on research on alternative nutrition label formats that
had been conducted by the agency and inviting comments on the
report. The comments on this notice were used in the design and
execution of subsequent consumer research conducted by the
agency. In the Federal Register of July 1, 1991 (56 FR 29963),
FDA announced a plan for a cooperative pilot programwith industry
to test alternative nutrition label formats that led to several
industry sponsored nutrition label format studies. In the Federal
Register of April 2, 1992 (57 FR 11277), FDA gave notice of a
meeting for industry, including small businesses, at which the
agency presented the results of its research studies related to
the format and design of the nutrition label, so that comments to
the proposed format rule (57 FR 32058) could be as informed as
possible.
    In a number of instances, FDA staff provided materials,
information, support, and consultation on technical aspects of
study design and label format to researchers. In addition, FDA
received many comments from the general public in response to
articles in newspapers and newsletters that solicited consumer
opinions in the form of informal polls based on examples of
possible nutrition label formats provided by FDA. Table 1
presents a summary of the various research studies received in
response to the format proposal and the format research conducted
by FDA.

Table 1.--Research Studies Submitted as Comments or Referenced in
Comments to Docket Number 91N - 0162: Food Labeling: Format of
Nutrition Label; Proposal

A. Experimental Studies Submitted

1. Frito-Lay Study (Ref. 74)

    a. Design: Between subjects; five format cells.
    b. Subjects: Central location test; adults, age 18+ who
purchased and/or ate salty snacks in the past 4 weeks, one site,
N=750.
    c. Formats tested: Same as FDA Study 1 on actual product.
    d. Key dependent measures: (1) Scale based on seven
questions, three number-of-serving type and four dietary-judgment
type; (2) rating of single format based on helpfulness, ease of
use and adequacy of information.
    e. Assessment/comments: All subjects saw same product. Well
controlled study.

2. GMA/NFPA Industry Study (Ref. 71)

    a. Design: Between subjects; seven format cells.
    b. Subjects: Shopping mall intercept/central location test;
adults 18+ who did at least half of household food shopping,
quota controls on age, income, education and race; 36 sites,
N=5,600.
    c. Formats tested: Same as FDA Study 2 on realistic product
mockups.
    d. Key dependent measures: (1) Product comparison task
identical to FDA task, (2) four-product comparison task with
specific nutrient probes, (3) dietary judgment task with specific
nutrient probes, ``if you were trying to get more/limit
(NUTRIENT) in your diet, how would you feel about eating this
(FOOD)?''; (4) rating of single format based on adequacy of
information and ease of use, (5) self-report of whether subjects
knew how to use the DRV information.
    e. Assessment/comments: Each subject worked with only one
format executed in a variety of ways. Products are confounded
with tasks. Format executions are inconsistent across products.
Percent DRV formats are sometimes executed with 1, 3, or 4
columndisplays depending on product, while other formats have
either one- or two-column displays. This complex execution for
Percent DRV formats may explain why they show poor product
comparison performance and are rated more negatively than other
formats. Expo

3. FDA Format Study 1 (Ref. 69)

    a. Design: Repeated measures within subjects; subjects
assigned to one row of a 5(formats) X 5(products) Greco-Latin
Square.
    b. Subjects: Shopping mall intercept/central location test,
adults 18+ who did at least half of household food shopping,
quota controls on age, income, education and race; eight sites in
seven states, N=1,560.
    c. Formats tested: Five formats (see format proposal (57 FR
32058)).
    d. Key dependent measures: (1) Product comparison task,
measured both accuracy and time; (2) preference rating for most
liked/least liked format among the five seen in the study, and
reasons for choices.
    e. Assessment/comments: Formats presented as two dimensional
nutrition labels of realistic size but not on packages.

4. FDA Format Study 2 (Ref. 70)

    a. Design: Repeated measures within subjects; subjects
assigned to one row of one of three 4(formats) X 4(products)
Greco-Latin Squares.
    b. Subjects: Shopping mall intercept/central location test,
adults 18 + who did at least half of household food shopping, no
quota controls; 8 sites, N=1,232.
    c. Formats tested: seven formats (see format proposal (57 FR
32058)).
    d. Key dependent measures: (1) Product comparison task,
measured both accuracy and time; (2) judgments of front panel
nutrition claims; (3) judgments of nutrients that need to be
balanced in the diet after eating product; (4) product
healthfulness ratings before and after seeing nutrition label;
(5) estimate of how many servings of product needed to meet daily
requirement; (6) preference for most liked/least liked format out
of the four seen, with stated reasons for choices.
    e. Assessment/comments: Formats presented as two dimensional
nutrition labels of realistic size but not on packages. All
formats not tested on product comparison task. FDA Study 1 data
used to impute product comparison performance for Control and
Adjective formats. Percent DV/With DRV used as proxy for Percent
DV/Without DRV on product comparison task.

5. Geiger Study (Ref. 72)

    a. Design: Repeated measures within subject; subjects
assigned to one of two format sets of either five or six formats.
    b. Subjects: Shopping mall intercept/central location test,
one site, eligibility requirements not specified, N=243.
    c. Formats Tested: 11 formats including versions of Control,
Control/DRV, Percent DV/With DRV, Percent DV/Without DRV and
versions with adjectives, bar graphs and various combinations of
these design features.
    d. Dependent measures: (1) Reading accuracy; (2)
number-of-serving type questions; (3) perceived usefulness of
various formats based on a conjoint measurement
procedure-equivalent to preference for large choice set.
    e. Assessment/comments: All formats executed on same product.
Learning effects across repeated measures may confound format
effects on performance measures--the same information is
available on all formats. Correct answers to number-of-serving
type questions are not clearly defined.

6. Byrd-Bredbenner (Ref. 73)

    a. Design: Repeated measures within subject; seven format
cells.
    b. Subjects: Supermarket intercept, 15 sites in same
geographic area, food shoppers 18+, age and education quota
controls, health and nutrition-related workers excluded, N=309.
    c. Formats tested: seven formats including versions of
Percent DV, Adjective, DRV Listing and Control in various
combinations.
    d. Key dependent measures: (1) Hybrid scale consisting of
number-of-serving type questions and product comparison
questions; (2) scale consisting of product comparison questions;
(3) preference ratings of most helpful/least helpful with stated
reasons.
    e. Assessment/comments: Products confounded with formats.
Order of format presentation partially confounded with amount of
information in format.

7. Burton (Ref. 75)

    a. Design: Between subjects, 4(formats) X 3(reference values:
none/daily/meal) X 2(high/low nutrient values).
    b. Subjects: Recruited for a university sponsored project by
letter, cross-section of adults, n=500.
    c. Formats tested: Versions of Control, Adjective and Percent
DV with and without different versions of a DRV listing.
    d. Key dependent measures: (1) Product ratings: bad-good, not
nutritious/very nutritious, purchase intentions; (2)
number-or-serving type measure; (3) rank ordering of formats on
quantity and quality of information.
    e. Assessment/comments: Well controlled study.

B. Survey Studies Submitted

    1. AHA Quantitative Study (Ref. 87)
    a. Design: Central location test, details unspecified.
    b. Subjects: N=405.
    c. Relevant format topics: Frequency of reading back of food
labels, magnitude estimation of amount of fat in product,
awareness of calorie base for fat, knowledge of how to adjust Fat
DRV if person eats less than 2,350, likelihood of using
information on food label to help reduce fat intake.
    d. Assessment/comments: Most of the survey is devoted to
issues related to use of the word ``healthy'' on food labels.

2. CSPI Study (Ref. 95)

    a. Design: National probability sample of telephone
households.
    b. Subjects: N=2,008 adults, assigned to one of eight
versions of magnitude estimation question.
    c. Relevant format topics: Magnitude estimation of fat
amounts in product.
    d. Assessment/comments: Fat is the only nutrient considered.

3. American Meat Institute/Roper Study (Ref. 96)

    a. Design: Multistage, stratified national probability sample
of households, in home interviews.
    b. Subjects: N=2,000, males who shared food shopping
responsibility equally with other people in the household were
selected when possible, otherwise any food shopper available from
household was selected.
    c. Relevant format topics: Attitudes and behavior regarding
food labels, understanding of ``RDA'' and ``DRV.''
    d. Assessment/comments: A comprehensive survey on food
labeling issues.

4. Kellogg Study (Ref. 97)

    a. Design: Not a probability sample, subjects call toll-free
number for some product-specific reason.
    b. Subjects: N=272, unknown characteristics.
    c. Relevant format topics: Understanding of DRV, knowledge of
how to adjust DRV for varying calorie needs, rated helpfulness of
DRV information.
    d. Assessment/comments: Sample characteristics are unknown.

5. GMA/NFPA Industry Study (Ref. 71)

    a. Design: Central location test (see description above).
Questions that were asked before subjects saw any food label
formats or questions that did not involve use of formats are
considered survey questions.
    b. Subjects: See description above.
    c. Relevant format topics: Frequency of reading food labels,
frequency of various uses of food label information,
understanding of DRV concept.
    d. Assessment/comments: Large sample, detailed questions
about possible label uses, DRV questions are asked after
respondents have been exposed to particular food label formats,
subjects exposed to Control format are not asked DRV questions,
only 4,790 respondents are asked the DRV questions.

6. National Consumers League (Ref. 98)

    a. Design: National probability sample of telephone
households.
    b. Subjects: N=1,139, 1,007 who read nutrition labels at
least sometimes completed full questionnaire.
    c. Relevant format topics: Frequency of reading food labels,
reasons for reading food labels.
    d. Assessment/comments: Most of the survey is devoted to
issues related to use of the word ``healthy'' on food labels.

C. Focus Group Studies Submitted

    1. FDA Study 1 (Ref. 85).
    2. FDA Study 2 (Ref. 86).
    3. Geiger (Ref. 72).
    4. AHA Study 1 (Ref. 87).
    5. AHA Study 2 (Ref. 87).

D. Informal Preference Polls

    1. COSTCO Newsletter (Ref. 88).
    2. Washington Post (Ref. 89).
    3. Nutrition Action Newsletter (Ref. 90).
    4. USA Today (Ref. 91).
    5. Daily Herald (Ref. 92).
    6. Atlanta Constitution (Ref. 92).
    7. New York Times (Ref. 94).
    The studies vary greatly in the issues addressed,
methodology, sampling, types of nutrition formats studied, types
of evaluation measures used to assess formats, and degree of
control used in the research. Many of the consumer studies
submitted or referenced in comments about nutrition label formats
were based on recently conducted research studies and on
interpretations that had not yet appeared in the scientific
literature.
    FDA considers the findings of research studies submitted in
comments to constitute an important separate class of comments
for purposes of evaluating various nutrition label formats.
Research findings based on specific measures need to be
considered as distinct from conclusions based on combining
findings across several different measures. Research findings
also need to be considered in the context of a body of similar
research to evaluate consistency in the pattern of effects across
studies (e.g., reliability) and consistency in the identification
of important controlling factors (e.g., validity). Comments
offering conclusions based on research findings are discussed
below in the relevant sections. In this section, the research
findings themselves are discussed in terms of methodologies used,
types of evaluation measures, consistency of effects across
studies, the strength of effects, and implications for the design
of the nutrition label format. The agency believes that to
clarify its reasons for decisions about the nutrition label
format that rely on research findings, it is necessary to
articulate its understanding of the relevance, reliability, and
relative significance of the various research findings.
    To facilitate discussion of research findings, FDA considers
it useful to distinguish among three primary types of evaluation
measures used to assess nutrition label formats: Performance
measures based on label use tasks, consumer preference judgments
of various formats, and questions about consumer understanding of
selected elements of possible nutrition labels, such as Daily
Values (DV's) (called in the format proposal DRV's). Each type of
evaluation measure has a different relevance to the selection of
an improved format for the nutrition label.
2. Performance Measures Based on Specific Label Use Tasks
    As a rule, different tasks and performance measures have been
used to evaluate how well a format meets the different primary
performance objectives specified in the 1990 amendments. These
objectives as discussed in section V.B. of this document are: (1)
To enable consumers to readily observe the nutrition information,
(2) to enable consumers to comprehend the nutrition content of
the particular product, and (3) to enable consumers to understand
the relative significance of product nutrition information in the
context of a total daily diet. (Objectives (1) and (2) are
closely linked for testing purposes and will be frequently
discussed together in this document). In the research reviewed by
FDA, measures to evaluate formats in relation to these objectives
have appeared only in experimental studies, probably because this
type of measure requires a substantial degree of control over the
conditions under which such measurements are taken.
    For performance measures based on specific label use tasks,
respondents are asked to perform a task using the information
from a nutrition label. The task is constructed so that a
performance measure can be defined (e.g., speed, accuracy,
likelihood of giving appropriate response), indicating the degree
to which the respondent can readily observe and comprehend
product label nutrition information or ``understand its relative
significance in the context of a total daily diet.''
    a. Product comparison tasks. The type of tasks most commonly
used to evaluate formats with respect to the objective of
enabling consumers to readily observe and comprehend product
nutrition information were product comparison tasks. These tasks
presented respondents with two or more product labels
simultaneously and asked them to engage in a relatively simple
information search (e.g., find differences between the products,
identify which product is higher or lower in a certain nutrient)
wherewers were scored correct/ incorrect and timed.
    104. The product comparison type of task was employed by the
two FDA format studies (Refs. 69 and 70) and by three other
studies submitted in comments to the format proposal (Refs. 71,
72, and 73). One study (Ref. 72) simply asked respondents to read
certain information from a product label and scored whether they
gave the correct answer.
    Performance levels on product comparison tasks were high,
with most of these studies finding accuracy levels of 70 - 90
percent correct.
    The product comparison type of task tended to produce
consistent format effects. The most consistent finding,
replicated in all studies, was that simple formats that have
clean, nonredundant displays of nutrient information per serving
worked best in this kind of task. Because it has the least amount
of information, the current format performed well on product
comparison tasks. But several studies (Refs. 70, 72, and 73)
found that other ways to display nutrient information per
serving, using either g/mg amounts or percent DV declarations,
were equally effective when the format was executed with a clean
and uncluttered appearance.
    Multiple column nutrient information per serving displays
were much more difficult than single column displays for
consumers to use for product comparisons. Several studies (Refs.
69, 70, and 71) found that product comparison performance dropped
sharply for labels using the ``as packaged/as prepared'' dual
declaration format. Both the major industry format study and
FDA's first experimental format study found that declaring
nutrient amounts per serving in adjacent columns of g/mg amounts
and percent DV led consumers to make more mistakes and to take
longer on the product comparison type of task. FDA's second
experimental study, however, showed that when g/mg nutrient
amount information was placed immediately next to the nutrient
name in an unordered array, and percents were placed in a column
array, the adverse effects on product comparison performance
disappeared.
    Most studies found that the addition to the label of a
listing of the DV's for some or all nutrients did not greatly
affect the ability of consumers to use the nutrition label for
product comparison purposes. Similarly, the use of a highlighting
orgrouping scheme on the nutrition label neither impaired or
improved respondents' performance of product comparison tasks.
    The use of adjectives on the nutrition label did appear to
cause respondents to miss nutrient differences between products
when the adjectives used to describe the nutrient for each
product were the same. Several studies (Refs. 69, 71, and 73)
found that formats using adjectives did not perform as well as
formats without adjectives on product comparison tasks.
    FDA considers this product comparison type of performance
measure to be a valid and reliable indicator of how well a given
format's information can be readily observed and comprehended.The
major conclusions that FDA drew from this research are that: (1)
Clean, uncluttered nutrition label formats work best, (2) dual
column declarations of nutrition information per serving make it
harder for consumers to readily observe and comprehend nutrition
information, and (3) adjective formats lead consumers to miss
quantitative differences between products when different nutrient
levels are characterized by the same adjective. The formats that
FDA tested (Refs. 69 and 70) that were effective on product
comparison tasks included the CONTROL, CONTROL/DRV, PERCENT,
PERCENT/DRV, GROUPING, and HIGHLIGHTING.
    b. Dietary judgment tasks.A different type of task was used
to evaluate formats with respect to the second performance
criterion, enabling consumers to understand the relative
significance of product nutrition information in the context of a
total daily diet. Most of the dietaryct (e.g., how likely they
would be to eat the product if they were trying to limit/increase
a specific nutrient in their diet; what nutrients they should try
to cut back on/get more of in other foods they eat that day after
eating three servings of the product; whether they consider a
high/low nutrient claim for the product to be correct/incorrect).
Respondent dietary judgments were then scored as
correct/incorrect or appropriate/ inappropriate.
    105. This type of dietary judgment task was used in the
second FDA format study (Ref. 70) and two industry studies (Refs.
71 and 74). Measures based on dietary judgment tasks produced
consistent format effects. The most consistent finding,
replicated by all these studies, was that providing adjectives to
describe nutrient levels per serving or declaring nutrient levels
per serving as percentages of the nutrient DV helped respondents
to make correct and appropriate dietary judgments relative to
formats where nutrient levels were declared in g/mg amounts.
These studies found that other nutrition label format design
elements, such as inclusion of a listing of DV's, highlighting,
or grouping nutrients on the label, did not improve performance
on these types of dietary judgment tasks.
    A consistent finding across two studies (Refs. 70 and 71) was
that percent DV declaration had a moderating effect on dietary
judgments relative to formats without percent DV declaration.
Respondents were less likely to consider a given nutrient level
unacceptably high or to say that they would avoid the food
entirely because of its nutrient levels when the amount per
serving information was presented as percent DV.
    Performance levels on this type of dietary judgment task
ranged from 45 to 80 percent correct across studies.
    FDA considers that how well consumers are able to judge the
magnitude of specific nutrient levels as measured by this type of
dietary judgment task is a valid and reliable indicator of how
effectively a given format helps consumers to understand
thesignificance of product nutrition information in the context
of a total daily diet. The major conclusions that FDA drew from
this research are that: (1) The declaration of nutrient amount
information as percentages of DV or the placement of adjectival
descriptors next to the nutrient amount information are both
effective ways to help consumers understand the significance of
product nutrition information in the context of a total daily
diet, (2) percent DV declarations moderate dietary judgments
about a food, and (3) other format elements, such as a list of
DRV's for important macronutrients, highlighting, or grouping
nutrients according to dietary guidelines, do not help consumers
make better dietary judgments. The formats that FDA tested that
were effective on this kind of task were the PERCENT DV, PERCENT
DV/DRV, and ADJECTIVE.
    106. Several comments on the format proposal recognized that
the ADJECTIVE and PERCENT DV formats have benefits to consumers
for dietary judgment types of tasks. They argued that both of
these format design elements correct a common tendency to
misjudge the magnitude of specific nutrient levels when they are
given in g/mg amounts. The comments asserted that consumers make
such errors because they tend to use a single numeric standard to
estimate magnitude.
    The comments stated that when nutrient amounts per serving
are declared in g/mg amounts, the numbers appearing on the label
cannot be used as guides to estimate the relative magnitude of
nutrient levels in the product. These comments continued: A
reasonable reference standard for one nutrient is quite different
from the reasonable reference standard for another nutrient. For
example, 5 g is a high level for saturated fat, but 100 mg is a
low level for sodium. Because the nutrients on the label
vadgments.
    Percent DV declarations, by contrast, display nutrient
amounts per serving in comparable units, and this type of display
facilitates the appropriate use of a common numeric reference
standard for all nutrients. Adjectival descriptors also serve to
translate nutrient levels into comparable units by describing
disparate nutrient amounts per serving in easily understood
ordinary language terms.
    FDA agrees that this reasoning provides a plausible
explanation of the research finding that percent DV declarations
and adjectives help consumers make more appropriate magnitude
estimates of nutrient levels in a product.
    c. Number-of-servings calculation tasks.Some studies employed
a one-product task where respondents were asked to estimate how
many servings of the product were needed to meet a daily
requirement for a certain nutrient. FDA considers this
number-of-servings type task to be relevant to consumers'
abilities to use product information for meal planning and
quantitative dietary management purposes that are properly
considered part of placing the product in the context of a total
daily diet.
    107. This type of performance measure was used in the second
FDA format study (Ref. 70) and in four other studies submitted in
comments to the format proposal (Refs. 71, 72, 73, and 75). The
number-of-servings type of performance measure did not show
consistent format effects across studies, possibly because of
variations in stimuli and procedures between studies.
    One well controlled study (Ref. 70) found that the current
nutrition label format was the worst format on this type of
performance task, but a less well-controlled study (Ref. 72)
found that the current nutrition label was among the best formats
on the number-of-servings type of task. One study (Ref. 70) that
asked for a numerical answer found that PERCENT DV formats
performed almost as well as the best formats on this measure, but
another study (Ref. 71) that asked the respondent to articulate
the computation process found that PERCENT DV formats were much
worse than the best performing formats.
    In all studies, performance levels on the number-of-servings
calculation task were noticeably lower than for the kinds of
tasks discussed above. Performance levels ranged from 10 to 50
percent correct across studies. All studies found that this type
of task was highly sensitive to respondents' education levels and
arithmetic sophistication.
    FDA considers this type of performance measure to be relevant
to how well a given format serves to place product information
into the context of the total daily diet, particularly with
respect to the role played by including a listing of the DV's
onthe label. However, a lack of consistent results across studies
and low levels of consumer competence to perform the required
computations suggests that this measure be considered of
secondary importance for evaluating nutrition label formats.
    d. Single product rating tasks.Some studies showed
respondents a product label in a given format and asked them to
rate the product on healthfulness or to rate purchase intentions
toward the product. The measure compared ratings made after
seeing front panel information consisting of nutrient or health
claims and ratings made after being exposed to product
information on the nutrition label.
    108. This type of measure was used in the second FDA format
study (Ref. 70) and in one other study submitted as a comment
(Ref. 75). Both studies showed a consistent effect: Consumers
relied more heavily on back panel than front panel nutrition
information when making general judgments about a product. They
become more negative toward the product after seeing the back
panel nutrition information relative to an initial impression
based on front panel information alone. Neither study found that
the
    Because it appeared insensitive to format effects, FDA does
not consider this type of single product rating comparison to be
an important consideration for evaluating how well a format meets
the criteria specified by the 1990 amendments.
    e. Measures based on two or more types of tasks.Some studies
defined scales that combined more than one type of performance
measure, such as product comparison questions and
number-of-servings questions. Where possible, FDA considered
these scales to represent only one of the composite measures, the
one they most resembled, so that they could be discussed in the
appropriate sections above. Such measures were evaluated by
comparing results with other measures from the same study and
with measures from other studies which utilized similar elements.
    109. One study (Ref. 73) submitted as a comment to the format
proposal reported results for a scale based on product comparison
questions and number-of-servings questions (discussed in section
V.D.2.a. and V.D.2.c. of this document). In this instance, the
results showed that format effects on the scale were quite
different from format effects on a different scale in the same
study which was clearly made up of product comparison questions.
Therefore, FDA considered this scale to be an example of the
number-of-servings type of performance measure.
    110. Another study (Ref. 74) reported results for a scale
based both on number-of-servings type questions and dietary
judgment type questions. Examination of the results for this
scale showed that the performance findings most resembled
findings from other studies based on dietary judgment questions.
Therefore, FDA considered this scale to be an example of a
dietary judgment performance measure.
3. Preference Judgment Measures
    In the research reviewed by FDA, consumer preference
judgments of various formats were primarily choice measures based
on a direct or implied question to respondents about which of
some given set of possible examples of nutrition label formats
wouldbe most helpful, most useful, or would work best for
consumers. Measures of this type occurred in all research
modalities and were often the principal measures in focus group
and informal preference poll research. Because respondents were
typically asked to express a relative preference, the set of
choices presented to respondents influenced the selection process
and thereby constitutes an important limitation on the validity
of this type of measure.
    Preference measures are not of the same order as behavioral
measures, which address how well a given format performs in a
given label use situation. Stated preferences for formats reflect
a respondent's implicit theory about what kind of format
generally works best. Judgment in these instances is abstracted
from any particular product or any particular label use
situation.
    An extensive scientific literature review suggests that
untested theories about the amount and type of information that
are most useful to consumers are sometimes wrong (Refs. 76, 77,
78, 79, and 80). In particular, studies of preference for
nutrition information have generally shown that consumers prefer
the largest amount of information offered (Refs. 81, 82, and 83)
but perform best with limited amounts of information specifically
related to the task (Ref. 84).
    a. Experimental studies.111. Both FDA format studies and
three other studies submitted as comments on the format proposal
employed relative choice measures of format preference based on
choosing a most preferred or least preferred format from the set
of formats being evaluated in the study (Refs. 69, 70, 72, 73,
and 75). Direct comparisons between studies are difficult because
no two studies used exactly the same choice set of formats.
    Despite differences between studies, there were basic
consistencies in the pattern of preference results across
studies. In every case, respondents tended to prefer the format
with the most information in the choice set and tended to dislike
formatsset. The addition of a listing of DV's to the nutrition
label for some or all of the nutrients was seen as more
informative, and was always highly preferred, over alternatives
lacking a listing of DV's. All of the studies that asked
respondents to give reasons for their selection of a certain
format (Refs. 69, 70, and 73) found that providing more
information was one of the most common reasons given.
    Other format features in addition to a DV list that were
viewed positively relative to formats without such features were
adjectives, bar graphs, highlighting, and, to a lesser extent,
grouping and declaring nutrient amount per serving as percent ofa
DV.
    All studies that included both performance and relative
preference measures (Refs. 69, 70, 72, 73, and 75) found little
or no consistency in the pattern of format results across
performance and preference measures. For studies that included
performance measures of the product comparison type (Refs. 69,
70, 72, and 73), the common finding was an inverse ordering
between formats that were preferred and formats that performed
well on this type of task.
    Two experimental studies (Refs. 71 and 74) varied formats
between subjects such that each subject saw a single format and
rated only that format. One study (Ref. 74) asked respondents to
rate the helpfulness, ease of use, and adequacy of information of
the single format. The other study (Ref. 71) asked respondents to
rate ease of use and adequacy of information. Neither study found
that respondents gave the highest ratings to the format with the
most information. A format similar to the currentformat that did
not include a listing of DV's for some nutrients was among the
highest rated formats in both studies.
    In one study, respondents expressed suspicion toward formats
using adjectives (which provided relatively more information),
apparently because they felt the company was deciding how and
when the adjectives were used. In the other study,
respondentswere more negative toward formats using percent DV
declarations. However, in the latter study, the PERCENT DV
formats were executed with extra columns of information, so that
a single nutrition label had as many as four numeric columns.
Respondents inthis study considered the PERCENT DV with DV list
format to provide more information than was desirable.
    FDA is convinced by these results that consumer preferences
for various nutrition label formats were very sensitive both to
the set of formats the respondent was asked to compare and to the
particular methodology used to measure preference. Moreover,
preferences did not correspond to objective measures of format
performance. This lack of correspondence raises serious questions
about the underlying validity of such measures, even though
respondents were asked to base preferences on which formats they
thought would work best. Given these methodological problems and
the apparent lack of validity, FDA considers preference measures
to be of secondary importance for decisions about the nutrition
label format.
    b. Focus group studies.Research using focus group discussions
about nutrition label issues elicited preferences for various
kinds of format design elements by showing the group examples of
different formats and asking them to discuss their reactions.
    112. The two FDA focus group studies and three other focus
group studies submitted as comments on the format proposal
discussed the groups' reactions to various format elements (Refs.
72, 85, 86, and 87). In every study, respondents indicated
strongly that they would like more information on the nutrition
label, particularly with respect to helping them understand
whether given nutrient levels could be considered high or low. A
listing of DV's for some or all nutrients was always among the
most prnt of calorie declaration for macronutrients. Other
features, such as adjectives or pie charts, received some
favorable mentions, but fewer than the above features.
    Respondents in focus group discussions often stated they
would like to see a simpler and easier to use label than the
current nutrition label. One focus group study (Ref. 85) asked
respondents to consider in detail how they might use certain
formatfeatures and found that pie charts and bar graphs were seen
to be hard to use. Formats using adjectives were sometimes
criticized because of suspicion about who decided how and when
the adjectives were to be used.
    FDA considers the focus group preference results to be
consistent with the preference results of experimental studies.
    c. Informal preference polls.Many comments from the general
public were generated by articles in newspapers and newsletters
that solicited consumer opinions in the form of informal polls
based on examples of possible nutrition label formats. FDA
considers such articles to be informal preference polls and
therefore a form of research. FDA recognizes limitations on the
validity of such research: respondents are highly self selected,
no background information about respondents is available,
responses are influenced by the accompanying news article, and
responses depend on the choice set of formats given in the
article. FDA has tried to identify the actual articles and the
choice sets of formats presented to readers in interpreting these
comments.
    113. FDA identified seven informal preference polls that
generated comments on the format proposal (Refs. 88, 89, 90, 91,
92, 93, and 94). One informal preference poll conducted by a
consumer buying club in its newsletter (Ref. 88) asked
consumersto rate their preferences toward three formats taken
from FDA's research formats: ADJECTIVE, HIGHLIGHTING, and
GROUPING. Over 400 responses were received. Seventy percent of
the responses favored the ADJECTIVE format.
    A midwest newspaper (Ref. 92) published examples of all seven
formats used in FDA's format study 2 and asked readers to
indicate which one they preferred. Approximately 100 responses
were received. Sixty-five percent of the responses favored the
ADJECTIVE format.
    A consumer group newsletter (Ref. 90) published an example of
a recommended format that included adjectives and a listing of
DV's for macronutrients and asked readers to respond to FDA in
support of the recommended format. Approximately 130 responses
were received in support of such a format.
    A nationally distributed newspaper and a regional newspaper
(Refs. 91 and 93) published an example of a graphically enhanced
PERCENT DV with DRV format (Appendix C from the format proposal).
Approximately 40 responses were received. Sixty-five percent of
the responses disapproved of the published format.
    A major eastern newspaper (Ref. 89) published examples of
four formats taken from the format proposal; PERCENT DV with DRV
(Appendix C), CONTROL with DRV Ranges (Appendix E), CONTROL with
Sex-Specific DRV (Appendix E), and CONTROL with Dietary Guidance
(Appendix F). It asked readers to respond to FDA with their
preferences, and approximately 450 responses were received. Two
formats (CONTROL with DRV ranges and CONTROL with Sex-Specific
DRV) were most preferred overall, each by approximately 35
percent of respondents.
    FDA considers the results of informal opinion polls to be
consistent with the preference results observed in experimental
studies and focus groups. Most consumers say they prefer the
format with the most information out of the set of formats they
are asked to evaluate. However, FDA is not convinced that formats
that have more information are necessarily the formats that best
meet the criteria specified in section 2(b)(1)(A) of the 1990
amendments.
    4. Measures of consumer understanding.Some of the research
submitted or referenced in comments to the nutrition label format
proposal consisted of survey questions about consumer
understanding of various elements of proposed nutrition labels
(Refs. 71, 87, 95, 96, 97, and 98). Some of these questions
addressed topics such as whether consumers use nutrition labels
and, if so, for what purposes. Other questions addressed the
concept of a DV: how consumers understand it, whether they can
use it, how they might use it, or whether they are aware of it. A
third type of question about consumer use of format elements was
how consumers assign magnitude estimates to nutrient levels.
    FDA considers this kind of research about format elements to
provide an important context for the decision about an improved
nutrition label format. Although it does not directly address the
format objectives specified by the 1990 amendments, this research
does provide some insight on how consumers understand and use the
nutrition label.
    a. Survey questions about consumer use of nutrition
labels.114. Four surveys submitted as comments on the format
proposal (Refs. 71, 87, 96, and 98) asked questions about how
often respondents read nutrition labels and ingredient
information on food packages. These studies consistently found
that approximately70 to 80 percent of consumers report that they
read this information almost always, often, or sometimes. These
figures are consistent with a number of other surveys (Refs. 99
and 100) that asked similar questions.
    In several studies submitted as comments, consumers were
asked about specific purposes for reading nutrition labels. One
study (Ref. 71), which asked 5,600 respondents detailed questions
about possible label uses, found that the most common purposes
for reading nutrition labels were: To calculate how high or low
the product is in certain nutrients, to get a general idea of
nutritional content, to compare different types of food products,
and to help determine brand choices. The least common purposes
for reading nutrition labels were to help in meal planning or to
figure out how much of the product you should eat.
    Other submitted studies reported results consistent with
these findings. Specifically, one study (Ref. 96) found that only
7 percent of those who read nutrition labels did so ``to help in
planning a specific meal.'' Another study (Ref. 87) found that 83
percent of respondents would be very or somewhat likely to use
information on the food label to help reduce fat intake.
    FDA considers the results of these questions about consumer
uses of nutrition labels to show that consumers are already using
nutrition labels for purposes that are consistent with the format
objectives of the 1990 amendments. Indeed, the two most common
types of reported uses: (1) To evaluate nutrition characteristics
of single products and (2) to assist in making choices between
products, correspond well to the two primary criteria specified
for formats in the 1990 amendments. The agency believes that the
introduction of a revised nutrition label and accompanying
educational activities will have a significant impact on use of
the nutrition label for these purposes in the future.
    b. Questions about Daily Values.115. FDA received a number of
studies as comments on the format proposal that asked questions
related to consumer understanding of the concept of DV's. One
study (Ref. 96) reported that 22 percent of respondents said that
they were familiar with the term ``Daily Reference Value'' or
``DRV'' compared with 65 percent who said they were familiar with
the term ``Recommended Daily Allowance'' or ``RDA''. Two studies
(Refs. 87 and 97) found that only about half of respondents could
correctly identify (i.e., read from the label) the DRV for a
specific nutrient.
    One study (Ref. 71) found that approximately two-thirds of
all respondents considered the DRV to be appropriate for
``everyone'' or ``most people.'' The same study found that 71
percent of respondents considered the DRV to apply to them
personally.mately two-thirds of all respondents stated that they
understood that a DRV based on 2,350 calories would be high for a
person who ate less than a 2,350 calorie diet.
    One study (Ref. 87) showed respondents a label with a listing
of DRV information and a footnote stating that DRV's were based
on a 2,350 calorie diet. It found that more than half of all
respondents could not correctly answer a question about the
number of calories on which the DRV was based.
    FDA considers results of questions about consumer use and
understanding of DRV's to be tentative and likely to change
because the public's exposure to the concept has been very
limited, and educational activities to explain the concept have
not beenundertaken. The experimental format research (see section
V.D.2. of this document) did not find that listing the DRV's on
the nutrition label had much effect in a positive or negative
direction on label uses that required evaluation or comparison of
specific products, although it did improve calculation of number
of servings needed to meet a daily requirement. None of this
research, however, evaluated the impact of listing the DRV's on
the food label on consumers' overall dietary management behavior,
either alone or in conjunction with possible education
initiatives.
    FDA concludes that in the absence of reliable guidance from
research findings, it has to rely on other comments to evaluate
the potential value of listing the DRV's on the food label as a
guide to better overall dietary management behavior.
    c. Magnitude estimation of specific nutrient amounts.116. Two
studies submitted as comments addressed the issue of consumers'
ability to make correct magnitude judgments about the level of a
nutrient when told the amount. One study (Ref. 87) found that
over one-half of all respondents considered 13 g of fat to be a
large amount of fat.
    A more detailed study (Ref. 95) asked respondents to estimate
whether a given amount of fat in a product would be considered a
low, medium, or high amount of fat. The amount of fat was
systematically manipulated to determine how respondents assigned
magnitudes across a range of values (7, 13, 20, and 33 g of fat).
At the lowest level (7 g of fat per serving), approximately 20
percent of respondents considered the product to be high in fat.
At the highest level (33 g of fat per serving), approximately 50
percent of respondents considered the product to be high in fat.
The same magnitude estimation results were found when the amount
of fat was expressed as a percentage of the DRV for fat. When
amount of fat was expressed as a percentage of theDRV for fat,
however, respondents were slightly less likely to give a ``don't
know/can't tell'' answer than they were when fat amounts were
expressed in g. Also, respondents were more likely to give a
``medium'' answer when the level of fat was expressed as 50
percent of the DRV instead of 33 g.
    FDA considers these findings to show that consumers estimate
nutrient level magnitudes of fat in a reasonable manner. However,
the agency also concludes that a tendency exists for some
consumers to see low fat levels as too high and other consumersto
see high fat levels as less than high. More research is necessary
to determine whether these results might be due to response
biases inherent in the particular kinds of questions being used,
or whether they reflect the different attitudes toward fatin the
general population. FDA considers these findings to be consistent
with the results of magnitude estimation measures used in
experimental studies (see section V.D.2. of this document). FDA
is convinced that an important consideration for decisions about
the nutrition label format is whether the format helps consumers
make appropriate magnitude estimations of nutrient levels in the
product.

E. Criteria to Use in Judging Nutrition Labeling Format

    Section 2(b)(1)(A) of the 1990 amendments specifies the
requirements that an appropriate nutrition label format must meet
(see Section V.C. of this document), but it does not specify how
to weight these requirements with respect to various possible
label uses or how to weight the various measures intended to
evaluate alternate formats against the requirements. The 1990
amendments also do not specify how to balance the benefits of a
revised nutrition label against the practical limitations of
small package sizes and the interests of many consumers,
particularly older and less educated consumers, to have a highly
legible label. In the format proposal, FDA requested comment on
the criteria to use in judging nutrition label formats.
    117. Most comments strongly supported the view expressed in
the format proposal that a simple, uncluttered nutrition label is
highly desirable. Comments from consumer groups and health
professional organizations emphasized the benefits of a simpleand
uncluttered label for older and less educated consumers. Comments
from food manufacturers and industry associations emphasized in
addition that a simple, uncluttered format would allow greater
flexibility to accommodate packaging constraints. Consumer
research conducted by industry and by FDA demonstrated that
simpler, less cluttered label formats help consumers to make
comparisons between products.
    FDA is convinced by the research results and these comments
that a simple and uncluttered format is the best way for
information on the nutrition label to be ``readily observed and
comprehended,'' as called for by the 1990 amendments.
Accordingly, FDA is taking the steps discussed below to minimize
the amount of information and the number of columns used on
thenutrition label.
    118. A number of comments from food manufacturers, consumer
groups and health professional groups called for consistent label
formats for both FDA and USDA regulated food products. The
comments identified many benefits of having a uniform format for
all food products including: (1) Making it easier for consumers
to compare different kinds of products, (2) making it easier for
consumers to become familiar with, and to learn how to use, the
new labels, and (3) reducing the likelihood of consumer confusion
because of apparent inconsistencies between different food
labels.
    FDA agrees that consistency between FDA and USDA regulated
food labels should be an important consideration in decisions
about the nutrition label format.
    119. A number of comments from food manufacturers, consumer
groups, and health professional groups argued that decisions
about the nutrition label format should be informed by consumer
testing, and that the agency should not propose formats that have
not been tested. For the most part, these comments were directed
at three label formats included in the format proposal that
presented more elaborate listings of DRV's and more extensive
educational footnotes than any of the formats included in
FDA'sprevious nutrition label format research.
    The agency agrees with the comments that emphasized the
importance of consumer research in informing decisions about the
nutrition label format. However, the agency is satisfied that
most of the format elements that have been suggested for a
revised nutrition label format have been sufficiently tested to
permit research-based conclusions about their effects on consumer
comprehension and label use behavior. The agency's view is that
format elements that were less well tested, such as those
suggested by the three formats described above, do not introduce
sufficiently novel elements to the nutrition label to require
independent testing. Information about the performance
characteristics of more cluttered labels, listings of DV's, and
elaborate footnotes is already available from extant research and
can be extrapolated to estimate the performance characteristics
of these particular formats as well.
    120. The agency received a number of comments about the
relative importance that should be assigned to product comparison
versus dietary judgment measures of format performance in making
decisions about nutrition label format. Many comments, primarily
from food manufacturers and trade associations, argued that
enabling consumers to compare the nutritional characteristics of
food products is the fundamental use for the nutrition label and
concluded that label formats should be evaluated mainly on this
basis. Other comments, primarily from consumer groups and health
professional organizations, gave more emphasis to the importance
of the food label for helping consumers to make dietary judgments
about the nutritional value of the food product thatinvolve
placing the product in the context of a total daily diet. These
comments concluded that decisions about a nutrition label format
need to take account of both product comparison measures and
dietary judgment measures. The research on the reported frequency
of different kinds of nutrition label uses showed that comparing
products and assessing nutritional value are the two most
important consumer uses of the nutrition label and are considered
about equally important by consumers.
    FDA is convinced by the research and by these comments that
decisions about a nutrition label format should consider both
types of label uses and evaluation measures rather than only one.
Use of the nutrition label to compare products is dependent on
the consumer's ability to comprehend the nutrition information,
and use of the nutrition label to assess nutritional value is
dependent on the consumer's ability both to comprehend the
information and to understand its significance in the context of
the total daily diet. Accordingly, FDA has considered these
primary nutrition label uses in making decisions about the
nutrition label format.
    121. One comment from a health professional argued that
consumer preferences for nutrition label formats should be
considered as important as the ability of a format to achieve the
format objectives specified in the 1990 amendments because a
formatthat is more preferred will be more likely to be used by
consumers.
    FDA is not aware of any data that support the assertion that
a more preferred label format will be more likely to be read. The
agency's view is that people read the nutrition label because
they are interested in what it says, not because they havean
impulse to read the label based on its appearance. Actual ease of
use, that is, the ease with which a consumer can extract needed
information from the nutrition label, rather than preference for
a format, is likely to influence the probability of reading a
nutrition label. The consumer research shows that consumer
preference for different label formats is, if anything,
negatively related to actual ease of use (see sections V.D.2. and
V.D.3. of this document). Therefore, FDA does not agree that
preference should be considered as important as performance
criteria for decisions about nutrition label formats.
    122. One comment from a consumer organization argued that
label uses should be weighted according to the likelihood that
consumers engage in such uses. The comment recommended that less
importance be given to label uses that assume that consumers will
add up their daily totals of fat, saturated fat, or other
nutrients because relatively few consumers are likely to engage
in such difficult and burdensome monitoring. The comment
suggested that many more people are interested in making
qualitative judgments about individual foods, such as ``is this
food high or low in fat?'' and recommended that dietary judgment
measures assessing this aspect of label use be given the most
weight in decisions about the nutrition label format.
    Consistent with the comment, the consumer research did not
show quantitative monitoring of dietary intake to be a common
label use behavior. However, it also did not show that making
qualitative judgments about a food is the only important use of
thd be an important, but not overriding, consideration in making
a decision about the nutrition label format. Other evidence shows
that consumers use the nutrition label to compare products and to
assess a product's nutritional value. Accordingly, FDA has
considered facilitating qualitative judgments as one of the
dietary judgment factors important for evaluating the various
proposed formats.

F. FDA's Tentative View

    In the format proposal, FDA presented its tentative
conclusions about the elements that it will include in the final
nutrition format and requested comments about them. The agency
listed the following four elements as those that were likely to
be included in the final nutrition format:
    (1) The information must be presented in a manner that is
simple and minimizes clutter.
    (2) The information must be presented in tabular fashion,
although perhaps enhanced by other graphic devices to provide
rapid access to, and greater visibility of, key nutrition
information.
    (3) The nutrition information display must include either a
listing of the quantitative amount of each nutrient, in absolute
terms (e.g., g), or a listing of the amount as a percent of the
proposed RDI or DRV, or both.
    (4) Nutrient information must be linked to the dietary
guidance that is considered important to public health.
    123. Comments mentioning the first three elements were
unanimously supportive. Comments mentioning the fourth element
were generally supportive, although a number of comments argued
either that the nutrition label cannot or should not be the
primary vehicle for providing general dietary recommendations, or
that educational materials should not appear on the food label at
all.
    The agency disagrees with statements that the nutrition label
should not play a role in educating consumers. FDA is convinced
that the nutrition label is an important source of basic
information for consumers, and that the 1990 amendments require
that the label facilitate consumer education. The agency's view
of the educational role of the nutrition label is elaborated in
section V.B. of this document.
    However, the agency does agree that the nutrition label
cannot be the primary vehicle for providing general dietary
recommendations. Accordingly, as discussed later in this
document, FDA, USDA, health professional organizations, and the
food industry are developing a comprehensive consumer education
program that will ease the transition to the revised nutrition
label and help consumers to use the label to make well-informed
dietary choices.
    FDA points out that under the act, the requirement that
nutrition information be linked to dietary recommendations need
not require presentation of dietary guidance on the label. The
House report that accompanied the 1990 amendments states,
``Whilethe bill does not mandate any particular approach, it does
require the Secretary to specify requirements that would permit
the consumer to understand the nutrition information pertaining
to a particular food in relation to recommended dietary
information'' (Ref. 16). The declaration of nutrient amounts as
percent DV provides such information. For the nutrient in the
food for which a DV has been established, the percent DV advises
the consumer how much of the recommended intake of that nutrient
is provided by the food. Seen in this way, a requirement that
nutrition information be linked to dietary guidance plays a
greater role in describing the food than in presenting
educational material.

G. The Format and Format Elements

    FDA received approximately 1,000 responses to the format
proposal and to a public meeting, notice of which was given in
the Federal Register of July 23, 1992. Responses were received
from consumers, health professionals, trade and retail associs of
particular concern to the individual or organization commenting,
but a large number included a reference to the specific formats
favored or opposed.
1. Titles and Terminology
    a. Title for the nutrition label. 124. A number of comments
addressed the issue of the title for the nutrition label. The
majority of comments supported retaining the current label
heading ``Nutrition Information per Serving.'' Comments suggested
that consumers are familiar with this heading, and that the title
is descriptive of the information that follows. One comment
opposed the introduction of any new title because new terms are
confusing. Another comment expressed concern that new titles have
not been proposed or tested. Other comments suggested such terms
as ``Nutrition Information,'' ``Nutrient Information,'' and
``Nutrient Facts.''
    FDA acknowledges that the current title is descriptive and
familiar to consumers. However, the agency also notes that the
current title requires more space than several alternatives that
are equally descriptive. The agency has concluded that modifying
one of the suggested alternatives to ``Nutrition Facts'' yields a
term that will clearly describe the information declared on the
nutrition label. This more succinct term also allows the title of
the nutrition label to use a larger typeface in the same space so
that the nutrition label will be more readily noticed, and thus,
more readily observed by consumers. Accordingly, in 101.9(d),
the agency is requiring that the term ``Nutrition Facts'' instead
of ``Nutrition Information per Serving,'' be presented as the
heading of the nutrition information.
    b. Terminology for subcomponents of nutrients. In its format
proposal (57 FR 32070 at 32071), FDA solicited comment on certain
format elements not addressed by research studies. The agency
requested comment on what terminology and graphic elements would
most effectively distinguish subcomponents of nutrients from the
declaration of the total amount of the nutrient and improve their
visibility in the nutrition label display.
    i. Subcomponents of fat and of carbohydrate. 125. The
majority of comments supported the use of the terms ``total fat''
and ``total carbohydrate.'' Many comments suggested using
indentation of subcomponents as a graphic means to further
distinguish subcomponents because it is a commonly used technique
that would be easily understood by most consumers. A few comments
suggested bolding and highlighting of the broader classification
to further distinguish subcomponents of fat and carbohydrate.
Other comments suggested using such terms as ``includes,''
``including,'' ``of which,'' and ``which includes'' before the
subcomponent to further establish that the subcomponent is a part
of a broader classification.
    Section 403(q)(1) of the act specifies that nutrition
labeling shall include information on several nutrients,
including total fat and total carbohydrates. In order to be
consistent with the terminology used in the 1990 amendments, the
supplementary proposal (56 FR 60366 at 60387 and 60388) included
provision for listing ``total fat'' and ``total carbohydrate'' as
mandatory elements of the nutrition label. Given the statutory
derivation of this terminology, the support for its use in the
comments, the fact that the terminology reflects the broad
category of nutrient, and the lack of opposition to the use of
this terminology, the agency is retaining the provisions for the
declaration of total fat and total carbohydrate based in
101.9(c)(2) and (c)(6) and, by reference, in 101.9(d)(7).
    The agency agrees that indentation of subcomponents along
with the use of the term ``total'' before the major
classification provides effective means of establishing separate
and recognizable subcomponent status. The agency is not providing
for the uy is persuaded by the comments that the use of
indentation of subcomponents is sufficient to clearly distinguish
the subcomponents of total fat and total carbohydrate because it
is a commonly used and well understood graphic device. Therefore,
the agency is requiring the indentation of saturated,
polyunsaturated, and monounsaturated fatty acids in
101.9(c)(2)(i) through (c)(2)(iii), respectively, and the
indentation of dietary fiber, sugars, sugar alcohol, and other
carbohydrates in 101.9(c)(6)(i) through (c)(6)(iv),
respectively, when such nutrients are declared. In addition, as
explained in section V.H.1. of this preamble, the broader
classifications must be highlighted by boldface print as provided
in 101.9(d)(1)(iv).
    ii. Calories and calories from fat. 126. The plurality of
comments supported using the term ``total'' preceding or
following ``calories'' to denote that it includes the calories
from fat (i.e., ``total calories'' or ``calories, total''). Some
comments suggested that a potential for confusion exists because
``calories from fat'' must be declared on the nutrition label,
and consumers may be unaware that they are included in the larger
category ``calories.'' These comments expressed concern that
consumers would mistakenly add calories from fat to the larger
classification declared simply as ``calories.'' Additionally,
several comments suggested indenting ``calories from fat'' to
further distinguish it from ``total calories.''
    The agency is persuaded by the comments that the term
``total'' preceding or following ``calories'' will better enable
consumers to understand that it is the larger classification of
which the subcategory ``calories from fat'' is a part. The
agencynotes that it is requiring the term ``total'' for the other
larger classifications, total fat and total carbohydrate. A label
that has the term ``total'' preceding two of the three larger
classifications may have the potential to confuse consumers
withregard to the third. However, the agency also notes that the
term ``calories'' has fewer words, and therefore requires less
space and minimizes clutter. Furthermore, consumers have been
seeing the term ``calories'' on labels to designate total
calories,and, unlike the other nutrient subcomponents, the
subcomponent ``calories from fat'' designates subcomponent status
by its structure. Therefore, in 101.9(c)(1), FDA is providing
for the use of the terms ``total calories;'' ``calories, total;''
or ``calories.'' In addition, in 101.9(c)(1)(ii), the agency is
requiring that the subcategory ``calories from fat'' be indented
for consistency with other nutrient subcomponents when it is
listed in a column under the total calorie information.
    c. Terminology for Daily Reference Value. In its mandatory
nutrition labeling proposal (55 FR 29487) FDA asked for comments
concerning an appropriate single new term to be used to refer to
all the reference values in the nutrition label. On its own, FDA
arrived at ``Daily Value (DV)'' as a possibility for use as this
single term. FDA used this term in the research that it conducted
on formats. Most consumers correctly interpreted the general
meaning of the term. However, during probing infocus group
discussions conducted by the agency, several consumers commented
that the word ``value'' may connote something of worth and
suggested that another term might be appropriate for food
labeling purposes. In its supplementary proposal
publishedNovember 27, 1991 (56 FR 60366 at 60371), the agency
reiterated its request for comment on, and suggestions for,
appropriate terminology to be used to refer to the entire set of
reference values.
    127. A number of comments responded to the issue of
terminology for a single term to denote all label reference
values. Two comments stated that the word ``value'' may give the
impression that these levels are goals to be achieved rather than
pointsmmended Daily Standard,'' ``Reference Value,'' ``Daily
Amount,'' ``Reference Daily Intake,'' and ``Recommended Daily
Intake.'' However, no general agreement emerged from the
comments, and no research data were submitted in support of
suggested alternatives for the term ``Daily Value.''
    One comment stated that the term ``reference'' has little
meaning for most consumers, while a few others said that the use
of the term precludes persons assuming that the value is a goal.
Another comment stated that the term ``standard'' avoids
theconfusion of having to differentiate between minimum and
maximum intakes. One comment suggested that the term ``U.S. RDA''
be retained to denote all label reference values. Many other
comments requested retention of the U.S. RDA's; however, those
comments appeared to be referring to retention of the current
numerical values for the U.S. RDA, not the terminology to be used
on the label.
    FDA disagrees that the term ``U.S. RDA'' should be retained.
The term was developed in 1972 when label reference values for
all nutrients listed on the label were derived from the
Recommended Dietary Allowances (Ref. 23). The term was developed
to suggest the link between the Recommended Dietary Allowance and
the label reference values developed by the agency. However, the
reference values for a number of the nutrients that are to be
included in the nutrition label, under the final rule on DRV's
and RDI's, published elsewhere in this issue of the Federal
Register, are not based on a Recommended Dietary Allowance value
because the National Academy of Sciences has not established
Recommended Dietary Allowances for these nutrients. It therefore
would be inaccurate and misleading to retain the term ``U.S.
RDA.''
    Further, the agency believes that terms that use the words
``recommended,'' ``requirement,'' or ``need'' would be misleading
to consumers and would complicate nutrition education efforts.
Some of the reference values that FDA is adopting are intended to
guide consumers relative to maximum intakes (for example,
saturated fat), while others are intended to serve as a basis for
planning general diets to meet nutrient requirements (for
example, vitamin C) or as minimum intakes (for example,
potassium). It would be incorrect to imply that FDA
``recommends'' that consumers consume the maximum intake level
for total fats, or that such levels are ``required'' or
``needed.'' Also, FDA cannot agree that the term ``standard'' is
appropriate. While the comment argued that this term does not
suggest a minimum or a maximum, the agency believes that it
commonly implies a level to be achieved or surpassed, and for
which it is undesirable to fall below. Thus, it may connote a
minimum level for many consumers.
    Moreover, the term ``daily intake'' suggests a requirement or
prescriptive need for individuals, rather than a general
reference point. Furthermore, the agency is concerned that if the
term were used, it could become a source of confusion in
information and educational materials on nutrition because
``daily intake'' for nutrients is used to mean current
consumption levels, rather than reference intakes based on
dietary recommendations. For example, the current daily intake of
fat is estimated to be 95 g per day based on food consumption
surveys. However, the agency's DRV for fat is 65 g for a 2,000
calorie diet and is based on dietary guidance.
    After reviewing the comments carefully, the agency concludes
that it is appropriate to retain the proposed term ``Daily
Value.'' FDA research has shown that the term is generally
understood by consumers as a point of reference, and no
appropriate or well-supported alternatives have been suggested to
the agency. FDA acknowledges that two comments suggested that the
word ``value'' may be indicative of a goal. However, no data were
submitted to support this suggestion, and no other comments
objectedence values on the nutrition label and is providing for
its inclusion in 101.9(d)(6).
    To preclude any confusion, the agency points out that the
Daily Values are a specific, regulatorily established set of
reference values that have been derived based on dietary guidance
and, for certain nutrients, on the assumption of a 2,000 calorie
per day diet (see the document on RDI's and DRV's published
elsewhere in this issue of the Federal Register). FDA recognizes
that alternate daily caloric requirements (e.g., 2,500 calories)
produce alternate recommended values for those nutrients with
dietary recommendations that are based on calorie requirements,
and that these alternate values can be considered ``daily
values'' for people consuming the given calorie level. However,
the recommended values for various calorie intake levels other
than 2,000 calories per day should not be confused with the Daily
Values, specifically the DRV's that FDA is establishing by
regulation (see 101.9(c)(7)(iii)) and that are referenced in
several of the regulations that FDA is adopting today (see, e.g.,
101.13).
2. PERCENT DV Format
    The majority of comments that supported the PERCENT DV or
PERCENT DV with DRV format were from consumer groups and health
promotion organizations, although several industry and other
types of organizations also supported the proposed format. The
majority of comments that opposed the PERCENT DV format were from
industry.
    128. The major argument given in support of the PERCENT DV
format was that the percent formats are easy to use and provide
clear information about how a food fits into a total daily diet.
FDA's research showing that the percent formats have superior
performance characteristics, particularly with regard to label
tasks related to dietary judgments, was sometimes cited. Some
comments argued that consumers are mainly interested in using the
nutrition label to make qualitative judgments about
specificfoods, such as whether the food is low or high in a
nutrient of interest. Many fewer people, it was argued, keep
running lists of nutrient amounts throughout the day. The
comments argued that the percent format facilitates this type of
qualitative judgment.
    Many of the comments opposed to the PERCENT DV format also
addressed the issue of consumers' ability to use the PERCENT DV
information, arguing that consumers would not be able to use
percent displays effectively. Specific arguments included that
the percent formats did not perform well in the Grocery
Manufacturers of America and the National Food Processors
Association (GMA/NFPA) industry study (Ref. 71), and that
consumers do not understand percents.
    FDA has carefully considered the arguments regarding percent
displays but finds no basis not to conclude that consumers will
be able to use PERCENT DV declarations more effectively than they
would any other format tested. The consumer research (seesection
V.D.2. of this document) supports the assertion that the PERCENT
DV format, with or without a listing of the seven macronutrient
reference DV's, improves consumers' abilities to make correct
dietary judgments about a food in the context of a total daily
diet. This result was replicated in three separate studies (Refs.
70, 71, and 74), two of them industry-sponsored, and on three
different dietary judgment tasks: judging the correctness of
nutrient claims about the product, identifying the nutrients in
the product that needed to be counterbalanced by changes in the
daily diet, and judging how much to eat of the given food if you
want to reduce intake of certain nutrients. In one
industry-sponsored study (Ref. 71), the PERCENT DV format helped
consumers judge how much to eat of a given food despite the fact
that PERCENT DV formats were executed with extra columns of
nutrient information per serving.
    As noted in section V.D.2. of this document, the percent DV
format element is one of only two format elements that have been
shown to improve consumer performance on dietary judgment tasks
(the other format element is the use of adjectives). In addition,
the PERCENT DV format, when executed without additional columns,
scored as well or better than any other format on all of the
other tasks measured in FDA's study. No evidence was submitted to
FDA showing that consumers cannot effectively use a PERCENT DV
format when it is appropriately executed.
    FDA studies (Refs. 69 and 70) found that for label use tasks
involving simple comparisons between products, PERCENT DV
declaration formats were best executed as single column displays
with g/mg amounts next to the nutrient name and not in a
column.Executed in this manner, no difference was found between
PERCENT DV formats and the CONTROL format on product comparison
tasks.
    The GMA/NFPA industry study (Ref. 71) found that when the
format was executed as two adjacent columns of numbers with
different units (g/mg amounts and percent DV declarations),
performance on simple comparison tasks was adversely affected.
This result is likely attributable to the additional columns
added to the format, particularly since the units differed, and
is not an inherent weakness of the PERCENT DV declaration formats
(see section V.D.2. of this document). FDA considers the
placement ofg/mg amounts in an unordered array next to nutrient
names to be a necessary feature of the PERCENT DV format because
it improves consumers' abilities to readily observe and
comprehend the percent information on the nutrition label as
demonstrated by FDAformat studies. Thus, use of this format is
consistent with section 2(b)(1)(A) of the 1990 amendments.
    The argument that people have difficulty in understanding
percents is not borne out by the consumer research. In the
nutrition label situation, a consistent system of percents is
used such that virtually all the nutrients on the label can be
declared in equivalent units, in this instance percent DV. A list
of nutrients declared in equivalent units has the unique property
that the list of values is self-anchoring, that is, values in the
list can serve as references for each other. A low value on the
list is likely to be a ``true'' low value, a high value on the
list is likely to be a ``true'' high value. This consistency is
not possible when the list contains nutrients declared in very
different units. Five g of saturated fat may be a ``true''high
value and 115 mg of sodium may be a ``true'' low value, but few
consumers see the number 5 as high and the number 115 as low
according to FDA research. Percent DV declarations help consumers
because they overcome the problems associated with declaring
nutrients in nonequivalent units (see comment 106 of this
document).
    Gram/milligram formats with a list of DV's give consumers the
numbers they would need to calculate percentages and thus to
transform the amounts to equivalent numbers. However, research,
including FDA's format research, has consistently shown that most
consumers are unwilling or unable to transform data provided on
labels (Refs. 70 and 101). Available evidence shows that
providing consumers with raw data is not effective. Providing
them with data in the form needed to make judgments, e.g., in
consistent percentages, is effective.
    Consumers have been seeing vitamin and mineral levels
expressed as percent of U.S. RDA on food products for about 20
years. Few know what the U.S. RDA's are for specific nutrients or
even know what units the U.S. RDA's are in. No arguments have
been raised that percents in this context are difficult to use or
hard to understand. The presentation of macronutrient data in
percents is a logical extension of the system that consumers have
been using with apparent success for years.
    Therefore, FDA is requiring in 101.9(d)(7)(ii) that
nutrition information per serving be declared as percent of the
DV in the primary columnar display on the nutrition label.
    129. Many of the industry supporters of the PERCENT DV format
cited the relatively small space requirements for the format,
particularly if the DV listing is not required.
    FDA agrees that the PERCENT DV format without a DV listing
requires little additional space relative to the CONTROL format.
A strength of the PERCENT DV format not shared by any other
format except ADJECTIVE is that consumers can use it equally well
for most label use tasks with or without the reference DV
listing. For this reason, the agency is not requiring that the
reference DV list be displayed as such. Rather, it is displayed
as part of an example of recommended nutrient amounts for
different calorie intake levels, and the normal placement is not
beside the Percent DV information but beneath it.
    In addition, the calorie-specific daily value list may be
omitted in simplified formats and on small and intermediate sized
packages (101.9(f)(5) and (j)(13), respectively). In contrast,
labels declaring amounts of nutrients only in g/mg units require
consumers to compare the reference DV list with the amount
declarations in order to make dietary judgments. Thus, for such
labels, the presentation of the reference DV list adjacent to the
declaration of amounts is necessary for most label use tasks.
    Accordingly, 101.9(d)(9) provides that daily values for
2,000 and 2,500 calorie diets be placed in columns beneath the
vitamin and mineral information. However, if space is not
adequate beneath the vitamin and mineral information,
101.9(d)(11) provides that the calorie-specific daily value
information may be placed to the right of the Percent DV
information. In addition, 101.9(f)(5) allows the calorie
specific daily value information to be omitted from labels of
products that qualify for the simplified format, and
101.9(j)(13)(ii)(C) allows it to be omitted from packages with
40 or less square inches of label surface available to bear
labeling.
    130. A number of comments argued against the PERCENT DV
format because of poor legibility of the basic format. They
argued that legibility will be lower because the absolute amount
declarations are hidden and are likely to be hard for consumers
tofind and because two numbers are required for each nutrient.
    FDA disagrees that the basic format has poor legibility. The
agency's research showed that consumers are easily able to use
the PERCENT DV format displayed with amounts by weight in
parentheses next to the nutrient name (see section V.D.2. of this
document). Most consumers will not need to use the amounts by
weight. The format prominently and clearly displays the one piece
of nutrient information that will be most easily used and
understood by the general population. The amounts by weight are
provided for consumers who find it easier to use them, such as
individuals who manage their diets using g/mg amounts.
    131. Other comments argued that consumers will be confused
because they will have to learn a new type of declaration, and
those consumers used to the amount by weight declarations may
mistakenly use the percentages as absolute amounts.
    FDA disagrees with this argument. Evidence from consumer
research shows that consumers generally are not able to
effectively use the current format for some important label uses,
such as placing a food in the context of their total daily diet
(see section V.D.2.b. of this document). In contrast, research
shows that consumers are able to use percent displays for all of
the label uses tested, including those tasks related to dietary
judgments, such as placing the food in the context of the total
daily diet. As consumer education reaches more people, and as
consumers become more familiar with the percent display format,
its effective use will increase. In addition, under
101.9(d)(7)(ii), as explained in section V.H.1. of this
document, the symbol for percent (i.e., ``%'') must be used after
each number. Therefore, consumers are not likely to use the
percentages as absolute weight amounts.
    Many of the comments opposed to the use of the PERCENT DV
formats did not acknowledge that these formats provide g/mg
amount information on the nutrition label. FDA included amounts
by weight to meet the needs of consumers who had come to rely on
such information. An appropriate balance must be achieved between
how much and how prominently information can be presented on the
label. The relative numbers of people likely to use different
information is an important consideration in achieving
thisbalance. Few people currently engage in the kind of dietary
management that requires keeping daily running sums of particular
nutrients, such as assumed by some of the comments opposed to
PERCENT DV formats.
    132. Several comments stated that PERCENT DV formats are
misleading because they provide inappropriate dietary guidance or
offer no guidance to those consumers whose daily requirements
differ from the DV. Concern was expressed that consumers will
believe the numbers apply to them personally.
    The agency disagrees that PERCENT DV declarations are
misleading because they provide inappropriate dietary guidance. A
major advantage of a percent unit is that it communicates the
relative magnitude of the nutrient level in a food without the
consumer having to be concerned about the absolute level or units
of the underlying scale being used. Knowledge of quantitative
dietary goals for specific nutrients is not inherent in, or
necessary for, accurate magnitude assessments of the nutrient
levelsin the food. The DV base of the percent does not have to
exactly fit each individual's needs in order for the percent to
accurately reflect the relative magnitude of the nutrient level
in the product.
    FDA considers estimation of the relative magnitude of
nutrient levels in the food to be central to the placement of a
food in the context of the total daily diet. FDA's research and
other research submitted as comments to the format proposal
showed that consumers were able to use PERCENT DV formats to
assess high/low levels of nutrients more effectively than any
other format (see section V.D.2.b. of this document).Therefore,
for purposes of placing the food in the context of a total daily
diet, a label use for which consumers have no need to adjust the
scale for individual variations, the declaration of nutrient
amounts as percent DV cannot be considered misleading or
inappropriate dietary guidance.
    Although, for the reasons described above, detailed knowledge
of the DV's and their relation to an individual's diet is not
necessary for using a PERCENT DV format to make product
comparisons or dietary judgments about the product, it is useful
for other dietary management purposes. Information about how
daily values vary by calorie needs will help those people who so
desire to estimate their own personal daily values and will help
them to differentiate the concept of a reference Daily Value used
for labeling and regulatory purposed from personally appropriate
dietary guidance.
    Therefore, to decrease the likelihood of consumer
misunderstanding, the agency is requiring in 101.9(d)(9)(i) that
a footnote accompany the percent DV declarations stating that
these declarations are based on a 2,000 calorie diet, and that
personal needs vary depending on an individual's calorie intake.
In addition, to assist consumers in estimating their own
quantitative dietary needs relative to the reference DV's, the
footnote will display daily values of total fat, saturated fat,
cholesterol, sodium, total carbohydrate, and dietary fiber based
on 2,000 calories and 2,500 calories. By providing a concrete
example of how individual dietary needs may vary depending on
calorie intake level, the footnote will help people to place
their personal dietary needs with respect to the reference Daily
Values and to avoid any misunderstanding as to whether the
reference DV's are dietary guidance meant for them.
    133. Several comments argued that PERCENT DV formats are
mathematically misleading because they are calculated against an
implied range if the terms ``or less'' and ``or more'' are used,
or because consumers will not be able to reconstruct the percents
from the absolute amounts and the DV numbers because of the
rounding rules for percents of macronutrients.
    FDA disagrees that consumers will see qualifying terms such
as ``or less'' and ``or more'' as constituting a range from which
a percent cannot be calculated. These terms were included on the
PERCENT DV/DRV format in the second FDA label format study, and
no subject expressed confusion resulting from their presence
(Ref. 70).
    FDA agrees that the rounding rules for percents for
macronutrients and sodium in proposed 101.9(c)(12) (56 FR 60366
at 60390) have the potential to cause consumer confusion when
used with the PERCENT DV format. The agency notes that the amount
by weight declarations for these nutrients have already been
rounded, and that additional rounding of the percents may lead to
an undesirable degree of inaccuracy, depending on the specific
percent.
    Therefore, the agency is requiring in 101.9(d)(7)(ii) that
percent declarations for macronutrients, sodium, and potassium in
the PERCENT DV/DV format be calculated by using as the numerator
the actual amount of the specified nutrient before rounding for
label declaration. The resulting proportion will be transformed
to a percentage and rounded to the nearest whole percent.
    134. Some comments argued that the PERCENT DV format should
not be selected because the lack of DRV's for some nutrients will
result in blanks in the principal numeric column. The comments
argued that such blanks will leave consumers with no information
about the level of some nutrients and will be confusing to them.
However, none of the comments that supported the PERCENT DV
format suggested that the lack of DV values for some nutrients
was a disadvantage of the format.
    Several of the comments that discussed the lack of reference
values in the context of whether the DV list should be required
on the label provided arguments that apply to all uses of the DV
information and thus also apply to the PERCENT DV format. These
comments argued that it is more beneficial for consumers to have
the values for some nutrients than to have the values for no
nutrients.
    The agency disagrees that blanks in the principal numeric
column resulting from the lack of DV's for some nutrients is
sufficient reason to reject the PERCENT DV format. The g/mg
amounts will be listed for nutrients that have no DV, so that
some information will be presented for these nutrients. Since a
reference value has not been set for these nutrients, none of the
alternative formats would give additional information to help the
consumer evaluate the food with respect to nutrients that lack
aDV. For example, no value will appear in the DV listing for the
nutrients, so comparison of the amount in the product with the
DV, as might be done with the CONTROL/DV format, would not be
possible. No scheme for assigning adjectives or for highlighting
would be able to include nutrients without a DV, so that formats
using these elements would not present more information about
such nutrients than the PERCENT DV format. Since no other format
overcomes the gap in information that results from lack of DV's
for some nutrients, the lack of DV's for specific nutrients
cannot be seen a reason to reject the PERCENT DV format. The
agency agrees with the comments that argued that presenting DV
related information for some nutrients is more beneficial to
consumers than withholding such information about all nutrients.
    135. Some comments argued that PERCENT DV formats are
calculated against a base that will change as scientific
knowledge about nutrition changes, just as dietary guidance
changes as knowledge increases, and that, therefore, a PERCENT DV
declaration should not be required.
    These comments address the issue of putting on the label
dietary information that will change over time with increasing
knowledge. The underlying assumption of these comments is that
percent DV declarations will communicate quantitative dietary
goalsfor specific nutrients, but, as discussed above (see comment
132 of this document), FDA's view is that percent DV's are not
likely to be used for this purpose. U.S. RDA's have been subject
to change in the same sense, but this fact has not prevented
their successful use on nutrition labels as a basis for declaring
nutrient amounts as percentages. Therefore, the agency disagrees
that the possibility of change is a substantial reason to avoid
percent declarations on the nutrition label.
    136. A number of comments argued that PERCENT DV formats
encourage good/bad food judgments.
    The agency does not agree with this argument. Both FDA and
industry research found that PERCENT DV declarations tend to
produce the most accurate judgments about whether products are
high or low in various nutrients. The g/mg formats were more
likely to lead to extreme and inappropriate dietary judgments
than PERCENT DV declarations (section V.D.2.b. of this document
and Ref. 102).
    137. One comment expressed the view that FDA does not have
the legal authority to require percentages, since the 1990
amendments only require the declaration of amounts. Others argued
that the 1990 amendments do not mandate that FDA change the
current format.
    FDA disagrees with these comments. As discussed above,
section 2(b)(1)(A) of the 1990 amendments requires that the
nutrition information be conveyed in a manner that enables the
public to understand the relative significance of the nutrition
information in the context of the total daily diet. Moreover, the
legislative history states that this provision requires the
Secretary to specify requirements that permit the consumer to
understand the nutrition information about a particular food in
relation to recommended dietary information (Ref. 16, p. 18).
Expressing the level of a nutrient in the food as a percent of a
reference amount (the DV) is the simplest and most
straightforward way of permitting the consumer to understand the
amount of a nutrient in the context of the total daily diet.
Thus, the 1990 amendments provide clear authority to require
percentages. Moreover, given the requirements of the 1990
amendments, and particularly the requirement in section
2(b)(1)(A) of the 1990 amendments, revision of the current format
is unavoidable and necessary.
    138. A comment from a foreign government stated that PERCENT
DV information is country-specific because the DRV information on
which it is based varies by country, and mandatory inclusion of
percent DV information on a label would make it difficult to
achieve equivalence in nutrition labeling requirements between
the United States and other countries. The comment noted that
their free trade agreement with the United States requires that
the two countries work toward equivalent requirements on
nutrition labeling. The comment pointed out that Codex guidelines
provide for supplementary nutrition information only on a
voluntary basis.
    The agency supports efforts toward international
harmonization of food labeling. However, the 1990 amendments
direct FDA to require a number of format elements that are not in
harmony with international food labeling. The agency believes
that it has been directed to require a format that will enable
consumers to choose appropriate foods and to place the food
within the context of their total daily diet, without the
constraints of meeting international guidelines.
3. The DV List on the Label
    a. Including the DV list on the label. A number of comments
from industry, consumer groups, and health promotion
organizations addressed the issue of whether the DV list should
be required, optional, or not permitted on the nutrition label.
    139. The major arguments supporting mandatory inclusion of
the DV list on the label, made primarily by consumer groups and
health professional organizations, were: (1) That the DV's must
be listed for people to estimate how their needs may vary frome
restrictive diet than represented in the DV's, (2) that consumers
need the DV information on the label because they have to become
comfortable and familiar with the DV concept in order for them to
use the new nutrition label to place the food in the context of
their daily diet, to put nutrient content information in
perspective, or to provide a frame of reference for decision
making, and (3) that consumers need the information because
quantitative dietary goals are necessary in order to encourage
and help consumers understand proper dietary practices.
    The major argument against inclusion of the DV list on the
label, made primarily by food manufacturers and food industry
associations, was that consumers will misinterpret the DV's as
dietary recommendations for their personal dietary needs, which
will lead to the DV's providing inappropriate dietary guidance.
Comments argued that DV's are unacceptable for dietary guidance
because they are population based reference values for an
``average'' consumer that do not take account of individual
differences such as sex, weight, activity level, and other
factors influencing personal dietary needs.
    Many comments opposed to requiring DV's argued that a listing
of DV's on the nutrition label provides no product-specific
information to consumers, and that mandating the listing on all
labels requires repeating the same information on millions of
food labels. One comment likened the requirement of placing the
list of DV's on food labels to a requirement that banks provide
addition and subtraction tables to their customers in each and
every monthly statement. Many of these comments argued that
inclusion of a list of DV's on the nutrition label will
significantly increase the space requirements of the nutrition
label, and that the increased space needs will make it extremely
difficult for small packages to comply with nutrition labeling
requirements.
    Many comments opposed the mandatory inclusion of the DV's on
the nutrition label because it will clutter the label and thereby
decrease consumers' ability to readily observe and comprehend the
nutrition information on the label. A number of comments,
particularly from industry, supported optional inclusion of the
DV list. The arguments for making the listing of DV's optional
were similar to those for opposing it.
    The agency finds merit in the argument that presenting the DV
list on the label may potentially mislead consumers by giving
undue prominence to values intended as references only and not as
dietary guidance for individuals. The consumer research (see
section V.D.4.b. of this document) showed that consumers were
likely to interpret a single list of values labeled as ``Daily
Values'' as personally applicable. At the same time, the agency
agrees with the comments that argued that consumers should be
able to assess how their personal dietary needs, which vary by
factors such as age, sex, and activity level, may differ from the
reference DV's used on the label. After extensive consideration,
the agency is convinced that the best solution to these
conflicting requirements is not to list the reference DV's
identified as such as part of the primary information, but to
provide a footnote as specified in 101.9(d)(9)(i) that gives
individualized daily values of total fat, saturated fat,
cholesterol,sodium, total carbohydrate, and dietary fiber based
on 2,000 calories and 2,500 calories.
    Without a prominent display of the list of reference DV's for
macronutrients on the label, the likelihood that consumers will
misunderstand the reference DV's as personally relevant dietary
guidance is greatly reduced. At the same time, a concrete example
of how recommended nutrient amounts vary depending on an
individual's needs will help consumers to understand how their
own dietary needs stand with respect to the reference Daily
Values.
    The agency believes that the information in the footnote will
accomplish virtually all the benefits that comments identified
would follow from including the list of reference DV's on the
label. By enabling consumers to evaluate the appropriatenessof
the percent DV's for their personal needs, the information in the
footnote will serve to increase consumer confidence in the
nutrition label and lead to more effective use. For consumers who
want to practice quantitative dietary regulation that involves
setting intake targets for certain nutrients and keeping a
running tally of intake of these nutrients, the information in
the footnote will provide maximum flexibility in the use of the
nutrition label. The percent DV's on the label can be adjustedfor
different personal needs or an individual's caloric intake either
by working with the percentages (such as having a target value of
120 percent for 2,400 calorie diet and a target of 90 percent for
an 1,800 calorie diet) or by working from absolute values derived
from the calorie-based daily values in the footnote. The footnote
will yield these benefits without implying that a specific
reference DV is the appropriate target for every consumer.
    However, FDA also agrees with comments that point out that
inclusion of DV-related information on the nutrition label, such
as that in the required footnote, imposes significant costs in
terms of label space without providing product-specific
information. Considering the appropriate balance, FDA is
convinced that the agency should be flexible in requiring the
footnote on product labels, particularly since the benefits of
having such a listing are not relative to the specific food that
carries the information, and that the information will be
available to consumers if it appears on a significant percentage
of food labels.
    Therefore, the information specified in 101.9 (d)(9) and
(d)(10) may be omitted from small and intermediate sized packages
as provided for by 101.9(j)(13), and from products that qualify
for a simplified format as provided for in 101.9(f)(5).
    140. Comments also addressed the placement on the label of
information intended as context to help people more effectively
use the nutrition information of the label. In the proposal, this
information was a listing of the reference DV's. The agency has
considered these comments in deciding the issue of the placement
of the information in the footnote that FDA is requiring instead
of a listing of the reference DV's. Several comments suggested
that the daily value information should be required tobe listed
in a column beside the percent DV information, not in a footnote.
Some comments agreed that placement in a footnote is sometimes
necessary and suggested that FDA require a listing on separate
lines rather than in a string. Others also recommended that
placement in tabular form be required. Many of the industry
comments stated that, in order to accommodate daily value
information on many packages, flexibility in placement is
essential.
    Because the PERCENT DV formats do not require consumers to
use information about the reference DV values to perform
product-related dietary management tasks, the agency believes
that allowing some flexibility in placement of the
calorie-specific daily value information and excluding small and
intermediate sized packages and products with simplified labels
from the requirement to provide the footnote information will not
compromise the effectiveness of the format (see comment 128 of
this document).As long as the information appears on a
substantial percentage of food packages, it will be readily
available to consumers. FDA recognizes that the added information
requires increased label space and agrees that manufacturers
should have flexibility to place it so that they can use
available label space efficiently. Thus, in101.9(d)(11), FDA is
providing that the footnote information may be placed to the
right of the percent DV information when there is not adequate
space to place it beneath that information.
    b. Lack of reference values. In its final rule on RDI's and
DRV's published elsewhere in this issue of the Federal Register,
FDA has established DRV's for total fat, saturated fat,
cholesterol, sodium, total carbohydrate, dietary fiber,potassium,
and protein. However, the agency has not established DRV's for
polyunsaturated fat, monounsaturated fat, other carbohydrates,
sugars, sugar alcohol, soluble fiber, and insoluble fiber.
Formats that include these nutrients will show missing values
under the percent DV and the DV columns. In the format proposal
(57 FR 32059 at 32070), FDA requested comment as to whether the
missing values will cause consumer confusion and, if so, whether
it would be helpful to place an entry in the column stating that
a reference value is not available.
    141. The comments were divided on whether the lack of
reference values for some nutrients would be confusing to
consumers. Several comments stated that appropriate educational
efforts would reduce consumer confusion, and that the potential
confusion possibly caused by missing values does not outweigh the
usefulness of providing the percent DV and DV information for
nutrients for which such values exist. Several comments suggested
that it was most appropriate to merely leave the entry blank,
citing concerns about label clutter and the need to keep the
label simple. Other comments suggested that the format include an
entry of some type to indicate that a reference value has not
been established. One comment pointed out that it is current
practice for some food and supplement labels to state that a U.S.
RDA has not been established for some nutrients.
    As noted in comment 134, the agency agrees that presenting DV
information for some nutrients is more beneficial to consumers
than withholding such information about all nutrients. The agency
is concerned about space limitations on food labels and label
clutter. The label format presented in this final rule contains
considerably more information than is required by the existing
label, and comparisons to the current practice of stating that
U.S. RDA's have not been established for some nutrients may
overlook the increased information required on the nutrition
label. Given the fact that nutrition labeling has been extended
to virtually all foods regulated by the agency and the concern
that too much information on the nutrition label may overwhelm
consumers, FDA finds no basis to conclude that statements that a
reference value is not available for the particular component
will add clarity to the label. Therefore, FDA is not providing
for the use of statements regarding the lack of established
reference values.
    c. Use of qualifying terms in presentation ofcalorie-specific
recommended nutrient amounts. The agency proposed (57 FR 32058 at
32070) to require the use of the qualifying terms ``or less'' and
``or more'' in conjunction with the proposed DRV list. Comments
about qualifying terms are relevant to the presentation of
calorie-specific daily values, as provided for by 101.9(d)(8).
While the agency did not specifically discuss the nutrients for
which each of these qualifying terms were appropriate in the
proposal, the examples presented in the appendices made it clear
that the agency's intent was to use these terms in conjunction
with those nutrients for which current dietary guidance specifies
an ``open-ended'' decrease or increase in consumption. Therefore,
because recommendations for total fat, saturated fat,
cholesterol, and sodium intake are stated in specific amounts or
less (Refs. 2, 3, and 4), the agency used the qualifying term
``or less'' with these nutrients. On the other hand, the
recommendation for carbohydrate is stated as 55 percent or more
(Ref. 3). Thus the agency used the qualifier ``or more'' with
this nutrient. FDA included such qualifiers in its research.
    142. A few comments opposed the use of the qualifying terms
because of the interest in reducing label clutter or because
their use conveys a message that a food should be avoided if it
contains high amounts of a nutrient qualified by ``or less.''
Sese of the qualifying terms if a range of values was used rather
than a single value. One comment considered the qualifying terms
unnecessary if FDA adopts a 2,000 calorie base.
    The majority of comments supported the use of qualifying
terms and suggested that such terms convey to consumers the
notion of a variable target intake rather than a prescriptive
intake. Some comments supported ``or less'' and ``or more.''
Others stated that ``less than'' was preferable to ``or less,''
and one stated that ``no more than'' and ``no less than'' were
preferable to ``or less'' and ``or more.'' These comments argued
that the recommendation for saturated fat intake was less than 10
percent of calories from saturated fat, and therefore the use of
``less than''as a qualifier in general was more appropriate. No
comment presented data concerning consumer use and interpretation
of qualifier terms. A comment suggested dropping ``or more'' for
carbohydrate, regardless of calorie base, as it is in conflict
with the Dietary Guidelines for Americans (Ref. 4), which
recommend the use of sugars only in moderation.
    FDA agrees that the use of qualifying terms assists consumers
in appropriate interpretation of the daily value information and
may help to preclude too literal an interpretation of the values.
Moreover, since no single caloric level can be specificfor all
individuals, the agency concludes that qualifying terms are
appropriate regardless of the caloric level used. Furthermore,
the agency is convinced that regardless of which term is
selected, the qualifiers should be used consistently to avoid
consumer confusion.
    FDA acknowledges that, while Diet and Health (Ref. 3)
recommends that 55 percent or more of calories be consumed as
carbohydrate, the Dietary Guidelines (Ref. 4) recommend the use
of sugars only in moderation. The label format will list sugars
as asubset of carbohydrate. The agency is persuaded that the use
of the qualifying term ``or more'' or ``more than'' with
carbohydrate has the potential to be misleading to consumers
given that carbohydrate includes sugars. The use of this term may
be particularly confusing to consumers when the source of
carbohydrate in a food is primarily sugars. Therefore, FDA will
not provide for the use on the label of the qualifier ``more
than'' or ``or more'' with carbohydrate.
    FDA finds merit in the term ``or less'' because this term is
presented after the quantitative value and thus does not
interfere with the consumer's ability to locate the quantitative
values (especially when the daily values are presented in a
column). However, the agency believes that the term ``less than''
conveys a less specific target and thus meets the concerns of
many comments that asserted that consumers need to be alerted to
the fact that recommended amounts vary greatly from individual
toindividual. The agency also acknowledges that the qualifying
term ``less than'' is more consistent with the recommendation for
saturated fat.
    Therefore, FDA is persuaded that qualifying terms should be
included when daily values are presented, and that the qualifying
term should be ``less than.'' The agency has included this
requirement in 101.9(d)(9)(i). For consistency and to avoid
consumer confusion, FDA will not provide for the use of the term
``or less.''
    143. One comment stated that the agency should allow the use
of the term ``or more'' with dietary fiber because such a
qualifier is consistent with current dietary guidance.
    FDA disagrees that it is appropriate to use the qualifying
term ``or more'' with dietary fiber. While there is relatively
little evidence that high fiber intake impedes mineral absorption
and bioavailability (Ref. 3), concerns about excessive
fiberconsumption have led to specific recommended ranges for
dietary fiber intake rather than open-ended recommendations. The
report from the Life Sciences Research Organization (Ref. 103),
which provides the basis for the DV for dietary fiber,
specifically pfor recommended dietary fiber intake (10 to 13 g
per 1,000 calories, or approximately 20 to 35 g per day) and is
not stated as 25 g or more. Therefore, the use of ``or more''
with fiber is not consistent with dietary recommendations, and
FDA will not provide for its use on the label to qualify dietary
fiber.
    d. Clarifying footnote for daily value caloric intake level.
In the format proposal (57 FR 32058 at 32071), FDA asked for
comment on the effectiveness of a footnote to convey to consumers
the need to modify the DV amounts to meet their nutritional needs
and for suggestions for alternative footnote statements. The
proposal included the following explanatory footnote in the
PERCENT DV with DRV graphic format: ``For a 2,350 calorie diet.
Your Daily Value may be higher or lower depending on your calorie
intake.''
    Comment was requested on the following three alternative
footnotes listed in the proposal (57 FR 32058 at 32071):
    (1) Based on a 2,300 calorie diet. Fewer calories are
recommended for women and young children.
    (2) As part of a 2,400 calorie diet. Many young children and
women over 50 need 2,000 calories or less. For a 2,000 calorie
diet the Daily Value would be less than 65 g Fat, less than 20 g
Saturated Fat, less than 275 g Carbohydrate, and 25 g Fiber
(Sodium and Cholesterol do not change).
    (3) A 2,000 calorie diet is for women over 50 and young
children. Most teenagers, sedentary men, active and very active
persons, and pregnant and breastfeeding women need more calories.
    144. Comments were received from manufacturers, health
promotion organizations, State governments, trade associations,
universities, and consumer advocate organizations. The majority
of comments supported the requirement of a footnote to clarify
the calorie base for the daily value listing.
    The explanatory footnote in the PERCENT DV with DRV graphic
format was specifically supported by five of the comments. This
footnote stated ``For a 2,350 calorie diet. Your Daily Value may
be higher or lower, depending on your calorie intake.'' However,
one manufacturer objected to this footnote on the basis that it
was ambiguous and ineffective and did not provide the necessary
information.
    Most comments stated that it is important for the consumer to
understand that the DV may need to be adjusted because it is
based on the number of calories consumed, and recommended calorie
consumption depends on various factors, such as physical activity
level, age, sex, weight, height, and metabolism. Two comments,
although opposing inclusion of the DV, argued that an explanatory
footnote should be included if the DV is included.
    Two comments objected to all of the alternatives. One
comment, from a consumer advocacy organization, asserted that the
third alternative listed above would create more confusion by
attempting to identify every segment of the population.
    Two comments, one from a health professional organization and
the other from a food manufacturer, stated a preference for using
the footnote in the format shown in appendix F (57 FR 32058 at
32089) as the footnote clarifying the DV list. This footnote
summarizes the Dietary Guidelines and includes statements such as
``Eat a wide variety of foods.'' However, it does not include a
reference to the DV's or the caloric level on which they are
based because the format that the footnote appears on does not
list DV's. (The format in appendix F of the format proposal is
discussed in section V.G.11. of this document). One of the
comments recommending this footnote stated that the footnote
should be prefaced with a statement about DV's varying with
calorie needs. The following was suggested: ``Your calorie, fat,
carbohydrate, fiber, and protein intake will vary based on age,
height, weight, metabolism and activity level.''
    A trade association opposed the footnote included in appendix
F of the format proposal as the clarifying footnote for the DV
list, stating that it added to the clutter, and that it did not
provide any relevant information to the consumer because toomany
calculations would be required to use the information in the
footnote. Two comments suggested that the statement in the
footnote of appendix F, ``choose a diet low in fat (30% or
less),'' be clarified to reference the total caloric intake for
the day. One comment requested that footnote information be
consistent with dietary guidance and the nutrition label by
changing ``Use sugar * * * in moderation'' to ``Use sugars * * *
in moderation.''
    Various other suggestions were made for the appropriate
wording of a footnote. Some comments were concerned with brevity
in the interest of space. Other comments emphasized clarity,
offering longer footnotes. Several comments were concerned
withconciseness. Some variations arose from the preference for a
range rather than a single number for the DV.
    The various footnotes offered in the comments stated that an
individual's DV depends on calorie needs, listed individual
characteristics that affect calorie needs, or simply provided the
calorie base of the DV on the label.
    Three comments from manufacturers suggested that footnotes
should be optional. Another comment suggested a voluntary program
of explanatory information provided in footnotes, with a
stipulation that the footnote become mandatory if 70 percent of
packages were not including it by 1996. Two comments objected to
inclusion of a footnote on the basis that the label cannot be a
source of dietary guidance. Others were concerned about the space
used by the footnote.
    Two comments addressed other possible footnotes to the
nutrition information. Inclusion of dietary guidance for special
dietary needs was suggested by a manufacturer. A consumer
advocacy organization suggested the use of footnotes to explain
the use of adjectives such as ``high'' on the basis that FDA is
wrong in assuming that people will relate ``high'' to the idea of
limiting consumption elsewhere. The following statement was
suggested: ``People eating this food may need to limit the fat
(or other nutrient) that they consume from other foods.''
    The agency notes that the consumer research suggests that
many consumers do not notice footnotes. One survey asked
respondents how many calories the DRV's were based on, while the
respondents viewed a label with a footnote providing the
information. Over half of the respondents could not give the
answer (Ref. 87). Another study provided subjects with a
nutrition label that had a footnote stating the caloric level
base (2,350 calories) of the DRV list. Over 70 percent of the
subjects stated that the DRV's applied to them (Ref. 71), even
though, according to this report, this caloric base should apply
to only 10 percent of the population. However, in other surveys,
when respondents were asked directly whether DRV's based on a
specific calorie level applied to them, most recognized that
adjustment would be needed and were able to give the correct
direction of adjustment (Refs. 87 and 97). These results show
that although most consumers do not notice footnotes, those who
are given the information (and by inference, those who do read
the footnote) are able to interpret it appropriately.
    FDA, in section V.B. of this document, has addressed comments
regarding whether the nutrition label is subject to a requirement
to provide general dietary guidance to consumers. The agency
concluded that the first consideration for the nutrition label
must be to help consumers make informed food choices by enabling
them to both comprehend the nutritional value of the food and to
understand its relative significance in the context of the total
daily diet. Thus, general dietary guidance is not to be provided
as part of the nutrition label. If a particular label has space,
however, general dietary guidance may be included outside of the
nutrition label, as discussed further below.
    The agency does not agree that a footnote should be placed on
the label urging people, after eating the food, to limit
nutrients in which the food is high. Such information depends on
the use of adjectives with the PERCENT DV format, which the
agency is not allowing, as discussed in section H.4. of this
document.
    FDA agrees with the majority of comments that a footnote is
necessary to help consumers determine how their individual
dietary needs compare with the reference DV's used on the label
and to prevent the possible misunderstanding of the applicability
of the reference DV's (see comment 132 of this document). At a
minimum, an acceptable footnote must specify the calorie level
used for the reference DV's so that consumers have some basis to
evaluate possible differences between their dietary needs and the
reference DV's used on the label. Also, a statement that an
individual's daily values vary according to calorie needs is
essential when calorie-specific daily values are presented. FDA
has provided for a footnote that includes this information in
101.9(d)(9)(i).
    The agency believes that many other pieces of information
mentioned in comments, such as the information presented in
Appendices E and F in the format proposal, may be appropriately
included on the food label to give useful context to the
nutrition information. However, this information may not be
included within the nutrition label itself because such
additional information would require significant additional space
to present and therefore would detract from the readily
identifiable image of the nutrition label. Specifically, it would
be appropriate to list typical calorie intakes for men, women,
and children and to summarize dietary guidance on the food label.
Furthermore, if the manufacturer is willing to supply copies of
the Dietary Guidelines for Americans upon request, a statement of
the availability of such information would be appropriate.
Examples include the following:
    (1) Typical intakes for women are 1,600 to 2,200 calories,
for men 2,000 to 3,000 calories, and for children (ages 4 to 14)
1,800 to 2,500 calories.
    (2) Use this nutrition information to help you plan your
total daily diet. The Dietary Guidelines recommend that
Americans:
     Eat a wide variety of foods
     Choose a diet with plenty of vegetables, fruits, and grain
products
     Choose a diet low in fat (30 percent of calories or less),
saturated fat (less than 10 percent of calories), and cholesterol
     Use sugars, salt, and sodium in moderation
    (3) For more complete information, a copy of the ``Dietary
Guidelines for Americans'' may be obtained from (manufacturer's
name and address).
    e. Footnote listing the caloric conversion factors.145. A
number of comments addressed the inclusion of caloric conversion
factors for fat, carbohydrate, and protein on the label to help
consumers use the nutrition information to apply the recommended
Dietary Guidelines for Americans. Several comments agreed that
stating the caloric value per gram of fat, carbohydrate, and
protein would help consumers better understand and use the
nutrition information on the label. Many other comments objected
to the inclusion of caloric conversion factors because ofspace
considerations and because of reservations about how many people
would be able and likely to use such information.
    The agency is persuaded by comments from nutrition education
experts that the public will benefit from having the caloric
conversion factors on the label. FDA recognizes, as discussed
above, that 9, 4, and 4 calories per gram for fat, carbohydrate,
and protein, respectively, are general factors that may not apply
to all foods. However, they are applicable to the majority of
foods, and therefore, inclusion of these factors will be useful
as a general guide. Moreover, FDA finds that any concerns about
4. CONTROL Format; Expression of Absolute Amounts in
Grams/Milligrams
    The majority of comments that opposed the PERCENT DV formats
supported the CONTROL format. A few comments supported the
current format, rejecting the revised list of nutrients, the new
order in which nutrients are declared, and the PERCENT DV
display. FDA has responded to the comments opposing the revision
in the required list of nutrients and the order of nutrients
earlier in this document (section III.A.2. of this document). A
majority of the industry comments supported the CONTROL format
withoutthe DRV list. Some consumer groups and health
organizations also supported the CONTROL format; however, they
recommended that the DRV list be included.
    146. Most comments in favor of the CONTROL format stated that
research has not consistently shown that any other format has
better performance characteristics on label use tasks than the
CONTROL format.
    The agency disagrees with this argument. Both FDA's and
industry's research found that the simplest label formats with
the smallest amount of information and the least number of
columns had the best performance for label use tasks that require
only simple comparisons or identifying differences between
products. Because it has the least amount of information, the
CONTROL format performs well on this kind of task. FDA's research
suggests that with certain placement of the information, some
other formats, including the percent formats, that provide more
information can perform as well as the CONTROL format on these
tasks. The industry study demonstrates that these other formats
can also be designed in ways that lead to poorer performance on
simple comparison tasks (e.g., by adding more columns to the
display).
    Both FDA's and industry's research also shows that for label
use tasks that require consumers to make dietary judgments about
the product, such as whether the food is high or low in certain
nutrients or how the food fits into a daily diet, the best
performing formats are those that include either the PERCENT DV
declaration or adjectives. Other design elements such as listing
reference DV's, grouping nutrients, or highlighting nutrients do
not appear to improve performance on these types of
dietaryjudgment tasks. The CONTROL format is among the poorest
performers on tasks that require dietary judgment.
    Being able to comprehend the nutrition information and to
understand its relative significance in the context of a total
daily diet means, at least in part, that consumers must be able
to make accurate high/low judgments about the food. PERCENT DV
and ADJECTIVE displays present high/low information directly. The
g/mg formats (such as the CONTROL format) require that the
consumer calculate percentages to get the information. The
CONTROL format requires, in addition, that the consumer know the
recommended amount for each nutrient. Research results show that
consumers do not know the recommended amounts for nutrients, that
many are not able to make such calculations, and that many are
not willing to make the large number of calculations that would
be required to include all of the listed nutrients in the
judgment (see comments 105, 106, and 107 of this document).
    147. Other comments supported the CONTROL format because it
is uncluttered, because consumers are used to it, and because it
is more consistent with dietary guidance, which is given in terms
of g/mg amounts, than is the PERCENT DV format.
    The agency agrees that simplicity and lack of clutter are
important criteria in selecting a format. However, enough
effective information must be presented to make the nutrition
label useful. Therefore, the selection of a required nutrition
label cannot be based simply on which one has the least amount of
information.
    Some of the arguments about consumer familiarity with a
format were addressed in section V.G. 2 above. The agency noted
that evidence from consumer research shows that consumers are not
able to effectively use the current format for some important
laity with it does not have important benefits. In contrast,
research shows that consumers are able to effectively use the
PERCENT DV format, even though the format is new to them. The
agency also noted above that g/mg amounts will continue to appear
onthe nutrition label for use by consumers who have come to rely
on nutrition information presented this way.
    148. A large number of comments were opposed to the CONTROL
format because it does not meet the criterion in the 1990
amendments of enabling consumers to understand the significance
of the nutrition information in the context of a total daily
diet. This argument was sometimes stated in conjunction with
FDA's research finding that the CONTROL format had poor
performance characteristics, particularly with regard to the
dietary judgment tasks.
    FDA agrees with this argument. A summary of research findings
related to the CONTROL format appears in section V.D.2. of this
document. For all the reasons discussed in this section, FDA
concludes that the CONTROL format is not adequate to meet the
criteria of the 1990 amendments.
5. HIGHLIGHTING Format
    Highlighting was discussed in the format proposal both as a
separate format and as a format enhancement. Most comments
regarding the use of HIGHLIGHTING dealt with it as a format
enhancement, and these comments are discussed in a later section.
    149. The comments that discussed HIGHLIGHTING as a format
were from industry health professional organizations, and
consumer advocacy organizations. Most comments were opposed to
the format. Many of these comments discussed the HIGHLIGHTING,
ADJECTIVE, and GROUPING formats together. The comments argued
that the HIGHLIGHTING format is inadequate and misleading because
it gives undue emphasis to desirable components, thus tending to
obscure the levels of undesirable components. In addition, the
comments stated that the HIGHLIGHTING, ADJECTIVE, and GROUPING
formats have no satisfactory means of communicating the level of
components that do not have a DV, such as complex carbohydrates
and sugars. Some comments argued that a modified HIGHLIGHTING
format that flagged both desirable and undesirable components of
a product should not be selected because extensive consumer
testing would have to be conducted to determine whether people
are able to distinguish between the two types of flags. Other
comments argued that the HIGHLIGHTING, ADJECTIVE, and GROUPING
formats foster good-bad food messages.
    Several comments from professional organizations argued that
the HIGHLIGHTING format is redundant because nutrient content
claims can be made on the front of the package. These comments
stated that if anything is highlighted, it should be undesirable
components to balance the front panel. Other comments argued that
this format did not score well in consumer research and did not
improve consumer comprehension of the label. One comment noted
that international harmonization is problematic with
HIGHLIGHTING, ADJECTIVE, and GROUPING formats because in Canada,
such information is generally required to be grouped together and
given equal prominence, whereas these formats include some form
of emphasis in one or more parts of the nutrition label.
    A supporting comment argued that the HIGHLIGHTING format is
best because it is straightforward, easy to understand, and
information can be quickly gleaned from it. The agency is
persuaded by the comments and the research that the HIGHLIGHTING
format should not be selected. FDA notes that this format has
most of the disadvantages of the CONTROL/DV format (of which it
is a variant), and it has several additional limitations. The
format did not score well in consumer research on measures that
involved putting the food into the context of a total daily diet.
In addition, it emphasizes desirable features of products, which
may already be emphasized by front panel statements and which may
tend to obscure the levels of less desirable components. TherING
format.
6. ADJECTIVE Format
    Issues regarding the use of adjectives to describe nutrient
levels arose in three contexts: support or opposition to the
ADJECTIVE format itself; mandatory use of adjectives with another
format, particularly the PERCENT DV with DRV format; and
voluntary use of adjectives as a format enhancement. Adjectives
as a format enhancement are discussed in section V.H.2. of this
document.
    150. Some comments argued that adjectives are inherently
value-laden and would communicate a good-bad food perception.
    The agency does not agree with this argument. As noted above
in the discussion of this argument for the PERCENT DV format
(section V.G.2. of this document), both FDA and industry research
found that the ADJECTIVE and the PERCENT DV declarations tended
to produce the most accurate judgments about whether products are
high or low in various nutrients (Refs. 70 and 71). The g/mg
formats were more likely than ADJECTIVE formats to lead to
extreme and inappropriate dietary judgments, such as responses
that a food was high in a nutrient in which it was actually low,
or that a food should be avoided altogether because of a
particular nutrient level. The agency is not requiring the
ADJECTIVE format for other reasons.
    151. Several comments argued that a complete scheme for
assigning adjectives to all nutrients required to be listed on
the label does not exist. They argued that because DRV's have not
been established for all nutrients, including sugars and
polyunsaturated fats, an acceptable scheme would be time
consuming to develop.
    The agency agrees that a complete scheme for assigning
adjectives to all nutrients does not currently exist, and that
the lack of DV's for some nutrients would complicate the
development of such a scheme. However, as explained in Comment
134, all of the alternative formats except the CONTROL format
share the limitation that DV's have not been set for some
nutrients. Because the limitation is constant for almost all
formats, it cannot be seen as a disadvantage unique to one
format. The agency believes that providing DV information for the
nutrients that have DV's is more beneficial than withholding it
for all nutrients because it is unavailable for some.
Nonetheless, the agency is not requiring the ADJECTIVE format for
reasons stated in comment152 of this document. Therefore, the
issue raised by these comments need not be addressed further.
    152. Several comments opposed the ADJECTIVE format because it
would be confusing to consumers. One comment argued that the
format provides information on whether a nutrient is high,
medium, or low, but not whether it is a desirable or
undesirablenutrient. Some comments argued that the format is too
cluttered and directive. Some comments noted that the ADJECTIVE
format showed a number of weaknesses in the consumer research,
particularly a tendency for consumers to fail to differentiate
betweenproducts when different nutrient levels were described by
the same adjective. The comments noted that wide ranges, as
proposed for the category ``medium,'' would be misleading to
consumers who did not attend to the nutrient values.
    Several comments supported the ADJECTIVE format, arguing that
the format is easy to read and does not require math calculations
or working with numbers at all. One stated that it would be
easier for the elderly and visually impaired to use. Other
comments supported it because it was preferred by consumers in
the research.
    FDA is not requiring the ADJECTIVE format for the following
reasons. The agency agrees that the ADJECTIVE format showed
weakness on an important label use task, the product comparison
task that required detecting differences between nutrients.
Theagency also agrees with the comments that argued that the wide
range for some of the adjective categories may be misleading to
consumers who use the label in certain ways. The agency
acknowledges that the ADJECTIVE format was the most preferred in
some seference measures must be interpreted cautiously and cannot
be used as a definitive criterion, for the reasons discussed in
section V.D.3. of this document. The agency further notes that
none of the studies provided evidence that the ADJECTIVE format
is easier for elderly consumers to use.
7. GROUPING Format
    Grouping by whether dietary guidelines recommend choosing a
diet high or low in specific nutrients was tested in FDA's Study
2 (Ref. 70). This format element did not generate many comments,
and the comments about it were frequently included in statements
about the HIGHLIGHTING or ADJECTIVE format. Most of the comments
were opposed to GROUPING.
    153. One argument against the GROUPING format was that it is
too value-laden, lending itself to a good - bad food message.
Another comment argued that the GROUPING format does not provide
meaningful information related to the particular product. Other
related comments argued that the GROUPING format did not have
good performance characteristics in research, and that subjects
reported that they found it too prescriptive. Some comments
argued that it would be confusing to consumers in general, and
one comment argued that it would be especially confusing to
consumers with diabetes.
    The agency agrees with these comments. FDA's research showed
that the GROUPING format did not perform well on the dietary
management tasks and did not offer any significant advantages
over other formats (Ref. 70). In addition, although the format
was not strongly disliked, many subjects who disliked it reported
that they found it too prescriptive. This complaint is consistent
with the complaints of many of the comments.
    The agency has decided not to require that nutrients be
listed under the GROUPING format headings for the reasons
discussed in the paragraphs above and below.
    154. A few comments argued that the GROUPING format would be
a challenge to implement because adequate consensus does not
exist on where to place some subcomponents, such as
polyunsaturated fats. In addition, comments challenged the format
becauseits recommendations are not entirely consistent with those
of the dietary guidelines. For example, the dietary guidelines
recommend moderate intake of some nutrients, such as sodium, but
the GROUPING format recommends low intake.
    The agency agrees that the placement of some nutrients and
nutrient subcomponents is problematic under the GROUPING format.
This problem in placing all nutrients is one of the reasons the
agency has decided against the GROUPING format.
    155. Several comments in support of the GROUPING format
argued that it provides nutrition education by stating which
nutrients should be eaten in greater and lesser amounts. A few
comments argued that the proposed order of nutrients on the label
tends to group them into those targeted for lower and higher
intakes, so that the GROUPING format is unnecessary with the new
nutrient order.
    FDA agrees that the intent of the grouping format is to
provide general dietary guidance. However, the fact that the
format did not offer significant advantages over other formats on
any performance measure considered in the consumer research shows
that dietary guidance as offered in this format did not benefit
consumers. The proposed new order of nutrients uses the widely
accepted design of placing first the elements of greatest
importance and is intended to accomplish some of the goals of the
GROUPING format. The GROUPING format's failure to convey the
intended dietary guidance, as measured in the consumer research
(Ref. 70), is one of the reasons the agency has decided against
this format.
8. Modified Grouping Format
    A few comments mentioned the Modified Grouping format in
which the order of the nutrients changed according to the amount
in the product.
    156. Almost all comments were opposed to the Modified
Grouping format. The major argument against it was that it would
reduce consistency and increase confusion among consumers.
Comments stated that using this type of format is especially
difficult for older people, who have a particular need for
nutrition information. As the population ages, larger numbers of
consumers will have difficulty with such a format.
    The agency agrees with this argument and notes further that
available research shows that with advancing age, consumers have
increasing difficulty extracting relevant information from
displays in which the order of nutrients vary (Ref. 104).
    The agency is not requiring the Modified Grouping format
because it has no reason to believe that this format would meet
the requirements of the 1990 amendments for the reasons stated
above. The agency further notes that consistency of placement
ofnutrition information is a principle that has guided the
development of the new format because such consistency has been
shown to help consumers, as noted above.
9. CONTROL Format With DV Ranges
    In its format proposal (57 FR 32058 at 32072), the agency
discussed several alternative formats to those tested by the
agency. For those reference values based on caloric intake, one
alternative was the use of a range of DV's based on a caloric
intake range instead of a single caloric intake value (Appendix E
in the format proposal). The agency requested comment on this
alternative.
    157. Comments were evenly divided concerning the use of
ranges for DV's. A number of comments, primarily from food
industry representatives, supported the use of a range for the
DV's because a range could assist consumers in realizing that
nutritional needs vary with individuals, and ranges are easier
for consumers to work with than single DV values. Others
supported the use of a range because the use of a specific
reference value would cause consumers to conclude that the values
applied directly to them as individuals. Several comments
suggested specific caloric ranges to be used (including 1600 to
2800, 1600 to 2400, and 1500 to 2800).
    A number of comments from a variety of groups, including
consumer advocates and the food industry, argued against the use
of ranges. Reasons for opposing the use of ranges included
concerns that ranges would be confusing to consumers, that they
would overwhelm consumers, that they are too broad to be
meaningful, that they use more label space than single values,
that consumers would not be able to calculate their reference
value from a range, and that they have not been evaluated in
appropriately designed studies to determine if they would be more
effective and less misleading than a single value. One comment
cited research conducted for the purposes of developing a dietary
guidance graphic (Ref. 105) that showed that consumers
experienced difficulties using a range of values relative to
dietary guidance.
    FDA has carefully considered these comments and concludes
that there is not sufficient support, nor a substantial
rationale, for providing reference values as a range. The agency
notes that no comment contained research or other data to
substantiate the utility or appropriateness of ranges. No
evidence shows that consumers do in fact find ranges easier to
work with, and no data suggest that ranges are less likely than
single values to confuse or mislead consumers. In fact, the
agency has reviewedthe literature on how people assign magnitude
to numbers (e.g., Ref. 120). This literature concludes that in
order to estimate magnitude, people generally have to answer the
question, ``compared with what?'' usually invoking a norm or
reference standard as a context for comparison. The DV is
intended to be such a reference standard. When expressed as a
range, the value of the DV as a norm against which the level of
the nutrient can be understood is compromised because the norm
cannot be easily identif information (section 2(b)(1)(A) of the
1990 amendments). Ranges apparently have the opposite effect.
    The agency is also concerned that the use of ranges would
mislead some consumers to believe that the consumption of a
nutrient at any level within the declared range is appropriate
for them. For consumers whose calorie intake is at the middle or
low end of the range, however, the label could induce consumption
of nutrients such as fat or saturated fat in excess of the
dietary guidelines, which would adversely affect public health.
    For these reasons, FDA has rejected the presentation of
reference values as ranges. The argument that consumers need
assistance to realize that nutritional needs vary with
individuals has been addressed by requiring daily value
information for 2,000 and 2,500 calorie diets.
10. CONTROL Format With Sex-specific Daily Values
    158. A few comments supported reference values based on
gender (Appendix E in the format proposal). One comment suggested
that gender specific reference values were appropriate because
women have different nutritional requirements than men.
Othercomments opposed the use of separate reference values for
men and women. The primary concerns were the issues of space,
readability, and clutter on the label. One comment suggested that
the format presented too much information for the consumer to
process. The same comment opposed gender-based reference values
because their use did not recognize that some active women are
more like men in terms of their calorie need, while some older
men have calorie needs more like those suggested for women.
Several comments argued that gender is only one factor to
consider in determining an individuals's dietary intake, and
therefore its presentation on the label has the potential of
inappropriately emphasizing one factor.
    While the agency acknowledges that women in general have
different nutritional needs than men, FDA notes that such
comparisons can be made for a variety of groups comprised of
persons 4 or more years of age. Thus, the agency agrees that the
use of gender specific reference values may inappropriately
emphasize only one factor in evaluating dietary intake. However,
the agency agrees that examples of recommended nutrient amounts
for different calorie level intakes may help consumers to
estimate their personal daily recommendations. Therefore, as
discussed in comment 139 of this document, FDA is requiring the
inclusion of recommended nutrient amount information for 2,000
calorie and for 2,500 calorie diets in the nutrition information.
11. CONTROL Format With dietary guidance
    159. A number of comments, primarily from industry, supported
the CONTROL Format with Dietary Guidance (Appendix F in the
format proposal). Several comments supported this format on the
basis that it was most like the current format and therefore
familiar to consumers. Two comments argued that FDA should select
the CONTROL Format with Dietary Guidance because USDA prefers it,
and harmonization between the two agencies is important.
Supporting comments argued that this format helps consumers to
put the food in the context of a total daily diet, reinforces the
dietary guidelines, and is simple and uncluttered. One comment
suggested that the caloric equivalents of the macronutrients may
enable consumers to better utilize the information provided.
    The major arguments against CONTROL with Dietary Guidance
were that the information is too vague to be effective and adds
clutter to the label. Some of these comments noted that the
footnote discusses foods when the information on the label is
about nutrients, so that, except for vague information about fat,
no relevant information about recommended nutrient amounts is
available on this label. Several comments argued that it would
not be clear to consumers that the dietary guidance information
appbasis to assume that consumers will be able to relate the
dietary guidance to the nutrition information. Comments also
pointed out that many calculation steps and further instruction
would be required to apply the dietary guidance to the consumer's
daily diet.
    Some comments noted that the dietary guidance footnote would
be problematic for meat products because it recommends a diet
high in vegetables, fruits, and grain products, which might imply
to consumers that they should not eat meat.
    The agency does not agree that CONTROL with Dietary Guidance
format as shown in the format proposal is consistent with the
requirements, or effective for meeting the objectives, of the
1990 amendments. The addition of the dietary guideline and
calorie conversion information does not serve to put the levels
of nutrients in the food into the context of a daily diet.
However, the format includes information that helps consumers to
understand the significance of the nutrient levels in the food.
    Therefore, the agency is incorporating one of the elements
from this proposed format as a mandatory requirement of the
nutrition label format. Specifically, FDA recognizes that it will
be useful to some consumers to have the caloric conversion
factors on the label. The placement beside the nutrient names as
shown in Appendix F of the format proposal is not acceptable,
however, because the g/mg amounts will be placed beside the
nutrient names in the required format. Both pieces of information
on the same line would decrease or eliminate the spacing that
helps to make the format comprehensible. Therefore, the agency is
requiring that the caloric conversion factors be included on the
nutrition label as a footnote, as described in 101.9(d)(10).
12. New Formats Submitted as Comments
    160. One comment suggested a format for the nutrition label
quite different from any other format suggested and quite
different from any format that had been previously tested. Called
a graphical profile, the format expressed quantitative nutrition
information in terms of distance along a spoke radiating from a
central point, where each spoke represented a mandatory nutrient
component. The points on each spoke were connected with each
other to form a pattern that distinctively identified the
nutrient profile for the product. The comment claimed that the
format has a number of advantages, including: (1) Providing
consumers with easily remembered mental ``shapes'' of the
products that they wish to consume or avoid and (2) helping
consumers to place the food in the context of a total daily diet
by allowing for easy comparison between the shape of the nutrient
profile for the product and an ideal shape based on dietary
recommendations.
    FDA is impressed by the ingenuity of this format but is
convinced that such an innovative format for the nutrition label
cannot be required without extensive consumer testing. No
consumer research to support use of this format to accomplish the
requirements specified in the 1990 amendments was submitted.
Furthermore, FDA notes that the format encourages a comparison
between the specific food and the dietary guidelines, whereas the
recommended comparison is between the total daily diet and the
dietary guidelines. Therefore, FDA is not adopting the graphical
profile format. The agency is prepared to work with interested
persons to develop consumer research that would show the
usefulness and validity of this format.

H. Graphic Enhancements and Format Elements

    The agency received numerous comments concerning the various
format elements and graphic enhancements discussed and
illustrated in the format proposal.
1. Format Legibility
    161. Many comments, particularly from older and
vision-impaired consumers and from organizations and health care
professionals serving their needs, suggested that the legibility
of nutrition information should be improved through regulations
specifyie letters, minimum type spacing, and greater color
contrast between print material and background, such as black
lettering on white background. An alternative suggestion was for
FDA to specify minimum print size and color contrast.
    A number of comments from package designers and from
individual food companies pointed to the problems in requiring
graphic elements that add to the space requirements for the
nutrition label, given the limited size of many packages and the
competition for label space from other required or desirable
information (e.g., UPC information, storage and preparation
instructions, recipes). Other comments urged the agency to set
minimal and flexible standards for graphic requirements related
to readability that would allow manufacturers to accommodate the
wide variation in package sizes, package shapes, and current
graphic elements on packages.
    The agency recognizes that mandating graphic elements to
assure a desirable level of readability for the required
nutrition information has both advantages and disadvantages,
particularly when these elements may require more space. The
agency agrees that some flexibility in the mandated graphical
elements is necessary in order to accommodate the wide range of
packages on which the information will appear. However, FDA also
agrees with the comments that stated that the readability of the
nutritionlabel needs to be improved to help older and
vision-impaired consumers who otherwise would be effectively
denied access to nutrition information of food packages. The
agency points out that, as stated in the format proposal (section
V.B.), certain graphic techniques go directly to the requirements
in the 1990 amendments (section 2(b)(1)(A)) that the information
be presented in a way that enables consumers to readily observe
the information.
    With the aim of achieving minimal readability standards for
the required nutrition information, FDA has developed for use in
this document, and in the presentation of the new label, a format
design that incorporates many of the graphic elements suggested
by the comments to produce a more readable label. FDA agrees with
comments that argue that a consistent ``look'' to the required
nutrition information on food packages will help consumers to
find the information on the package and to recognize
theinformation for what it is--a profile of the nutrient content
of the food. Although FDA is providing for some flexibility in
label execution, companies are encouraged to use this label as a
model for designing labels for their packages (see section
101.9(d)). The specifications for this presentation of the
graphical elements are included in the Code of Federal
Regulations as appendix B to part 101.
    FDA understands that some flexibility in execution of the
nutrition label is necessary to accommodate various package sizes
and shapes and current graphic features that serve as brand
identifiers. For these reasons, a number of graphic
alternativesto the model label are permitted. For example,
101.9(d)(11) allows the footnote to be moved to the right of the
percent DV information if space is not adequate beneath the
vitamin and mineral information, and 101.9(d)(8) allows for a
vertical display of vitamin and mineral information when more
than four vitamins and minerals are declared.
    In addition, although the model label calls for dark or one
color type on a white or neutral color background, flexibility in
background and type color is allowed 101.9(d)(1)(i). FDA is
aware that some products traditionally use color as a brand
identifier and print nutrition information in white or neutral
color type on a darker color background. This type of graphic
technique, called ``reverse type,'' is known to have poorer
readability characteristics than regular type. FDA is not
prohibitiose described above, in those situations in which
reverse type will be significantly less expensive than the FDA
preferred alternative, or in other similar appropriate
circumstances. If reverse type is used, FDA expects that the
impairment in readability resulting from such a technique will be
compensated for by the use of other graphic techniques to improve
readability, such as increased type size. It will not be
acceptable to reduce the contrast between print and background,
whether by light letters on a light background or dark letters on
a dark background, to the point where readability of the label is
significantly degraded.
    Although the agency is committed to the flexible application
of graphic techniques to achieve an acceptable level of
readability for the required nutrition information, FDA considers
it necessary in order to ensure that the nutrition information is
conveyed in a manner that enables the public to readily observe
and comprehend such information to set minimal standards and
requirements for certain key graphic elements of the nutrition
label. Such requirements will prevent confusion about the
minimallevel of readability that is necessary for the nutrition
information. The key graphic elements that are specifically
required on all packages are set forth in 101.9(d) (1)(ii),
101.9(d)(1)(iii), and 101.9(d)(1)(iv). They consist principally
of type size and type style requirements, namely that a
consistent upper and lower case type style be used, that a
single, easy-to-read type style be used, that product information
be in at least 8 point type (the lower case ``o'' no smaller than
\1/16\th inch) with at least 2 points leading (i.e., space above
and below letters) and kerned (space between letters) no tighter
than {dieresis} setting, and that headings, certain nutrient
names, and percentage amounts be highlighted by bolding or other
form of highlighting.
    In addition, to preserve a readily identifiable image or
``look'' for the label, a number of other graphic elements are
required, as discussed in 101.9(d)(1)(i), (d)(1)(v), (d)(2), and
(d)(7)(ii). These sections require a hairline box to set off the
nutrition information, hairline rules in certain places within
the nutrition label, larger print size for the title, ``Nutrition
Facts,'' and display of the percent sign (%) after the numerical
value of the percent DV for each nutrient.
    FDA has been persuaded by comments and careful consideration
of alternatives that these requirements will benefit a
substantial number of consumers who currently have difficulty
reading nutrition information on food packages. FDA considers
this benefit to be worth the cost of the small increase in space
allotted to nutrition information that may be required for some
food packages. The agency notes that the previous regulations on
type size also mandated a minimum type size for the lower case
``o'' of \1/16\th inch but applied the same minimum type size
requirement to the upper case ``o'' as well, which resulted in
most manufacturers using all upper case type styles. The
practical effect of requiring upper and lower case type styles
and keepingthe same minimum type size requirement will be to
increase the minimum size of upper case letters by approximately
30 percent. To further compensate for the increased demands on
label space, FDA will allow the information in the footnote,
which unlike the product specific information is the same for all
products, to meet or exceed a 6 point type minimum type size
requirement. FDA considers that the requirements for upper and
lower case type styles, leading, and kerning will enhance the
readability of6 point type enough that it will not present
problems for most consumers.
2. Other Graphic Enhancements
    162. Several comments from retailers, manufacturers, graphics
designers, universities, and a nutrition professional group were
directed specifically at reverse printing as used in the graphic
adaptations of appendix C of the proposal. Two comments sould be
helpful to consumers and should be permitted on a voluntary
basis. However, the remaining comments addressed technical or
legibility problems. Several comments stated that the format
examples were not readable, that the reverse printing overwhelmed
the smaller type, and that ``stacked'' titles (i.e., with
components arranged vertically, e.g., placing ``servings'' on one
line and ``per container'' on the next line) were confusing. The
principal technical problem mentioned was that reverse printing
tends to fill in and become unreadable, depending on the printing
process used and available label area.
    One manufacturer of many products stated that only two out of
four printing processes used by the firm would be able to
implement the graphic adaptations of appendix C. Another
manufacturer stated that it would be able to print reverse
graphics only for large containers. A graphics design firm stated
that reverse printing adds significant costs. Another design firm
cited two technical barriers to reverse printing: Multi-screen
labels are difficult to hold in alignment and retain clarity, and
reverse printing cannot be applied to packages with light
backgrounds because the background must be dark for light,
reverse print, to show through. The comments stated that many
brand identification colors are light, and manufacturers object
to having tochange them, arguing that brand identification would
be lost.
    FDA agrees that reverse printing should not be required,
given the difficulties mentioned in the comments. The agency
finds convincing the arguments against the legibility of reverse
print discussed in the professional literature (Ref. 107).
    Because of the need for flexibility to place the nutrition
panel in variously sized panels, the agency does not object to
stacked titles.
    163. The majority of comments stated that other kinds of
graphic enhancement of nutrition information, such as
underlining, bolding, and using larger type size or contrasting
color, would encourage and assist consumers in using the
information. However, opinions were divided as to whether the
combination of enhancements illustrated in appendix C of the
proposal would be helpful. A number of comments criticized the
graphic adaptations as cluttered, jumbled, obtrusive, or
distracting from or overwhelming the smaller type. However,
except for a few suggestions that graphic enhancements be
entirely at the discretion of the manufacturer, the comments
favored some degree of standardization through FDA regulations.
Comments frequently stressed the need for uniformity of
appearance of labels across the food supply to facilitate
education efforts and consumer access to and use of the nutrition
information.
    Several comments from consumers, a graphics designer, and a
nutrition professional group stated that the number of different
font sizes on a label should be minimized to ensure legibility.
One comment cited a book in support of this view (Ref. 108).
Several comments urged FDA to keep the label uncluttered. Other
comments provided specific guidelines for maximizing the label's
usefulness. These guidelines generally involved removing as much
print as possible, keeping titles linear rather than stacking
them, and including only essential information.
    Several comments from industry and health education
organizations endorsed voluntary, judicious use of other graphic
enhancements such as spacing, indentation, use of upper and lower
case letters, and selection of type face and size. These
commentsgenerally opposed making such enhancements mandatory
until consumer research is conducted to ensure that they
effectively aid consumer comprehension. Other comments from
groups representing older readers and the vision impaired
provided research demonstro provide guidance for the selection of
graphic devices for further consideration. Two manufacturers
opposed graphic enhancements altogether as contrary to the
requirements of 1990 amendments for consistency in presentation
of information. One commentcharacterized the combination of
extreme bolding and close proximity of columns in the graphic
adaptation of the PERCENT DV with DRV format as diverting
attention from the quantitative values and stated that, with
respect to the objectives of the 1990 amendments, the format
constitutes near misbranding.
    Based on the research submitted in comments, the agency is
convinced that it can proceed to require certain graphic
enhancements. While some comments questioned the appropriateness
of requiring such enhancements at this time, other comments
submitted research that demonstrated that these enhancements are
effective and appropriate for creating a nutrition label that is
readily observable and comprehensible, as required by the 1990
amendments. The agency agrees that keeping the format uncluttered
is important and therefore has minimized clutter in the model
format. The agency has carefully considered which format
enhancements to combine, based on the comments and the research
presented.
    FDA is convinced that the specific elements mandated provide
a visually integrated image that will give the nutrition label a
uniformity of appearance across the various types of packages in
the market and will enhance consumer use of the information. For
example, an important element in the appearance of the nutrition
label is its pattern of bolding. In 101.9(d)(1)(iv), the agency
is requiring bolding of the heading ``Nutrition Facts'' which is
being employed as an identifying title, like a logo or trademark,
to distinguish the nutrition label from other information on the
package, as well as bolding of headings of certain nutrient names
and percentage amounts. The agency is convinced that this and the
other measures that it is requiring willserve to establish the
readily identifiable ``look'' that it is seeking and finds to be
necessary to achieve the relevant goals of the 1990 amendments.
    164. Comments were received from a consumer, a health care
provider, a State government agency, and two manufacturers
suggesting that industry be permitted to use graphical devices,
such as pie charts, to illustrate nutrient content claims.
Commentssuggested that uniformity of labeling could be maintained
by requiring that any supplementary graphics be placed outside
the nutrition label area. Other suggestions were that FDA permit
the voluntary inclusion on the label of information from
authoritative sources, such as the Dietary Guidelines for
Americans (Ref. 4) or Diet and Health (Ref. 3), to aid consumer
understanding of nutrition information in the broader context of
current dietary advice to the public.
    FDA has no objections to the use of graphic devices to
amplify or explain nutrition information, provided that the
illustrations are presented in a manner that is truthful and not
misleading, and that the devices are not placed within the label
areain which the nutrition information appears. The agency also
agrees that supplementary information outside the nutrition label
can help consumers better understand the characteristics of
individual foods in relation to the total diet. However, such
supplementary information must be consistent with the
requirements for nutrient content or health claims that are
established in companion documents published elsewhere in this
issue of the Federal Register. Manufacturers are also encouraged
to utilize other means to disseminate dietary guidance
information, such as incorporation of such materials in
promotional and print advertising materials or by means of shelf
talkers and placards at point-of-sale.
    165. Several industry comments requested that manufacturers
be given flexibility, either in the case of small packages or in
general, to declare vitamins and minerals and DRV's in either
tabular or linear arrangement for both full and simplified
formats. A manufacturer suggested permitting a linear array for
micronutrients present at levels of at least 2 percent of the DV.
    The agency agrees that manufacturers need flexibility in
accommodating the required nutrition information, particularly
for small packages and printable surface areas that are oddly
shaped or narrow. Consequently, FDA is providing options in the
display of a number of the types of information required. For
example, in 101.9(d)(ii), the agency is providing that the
information about calorie-specific daily values and caloric
conversion information may be placed beneath the vitamin and
mineral declarations or to right of the Percent DV column. In
101.9(d)(8), the agency is providing that the vitamin and
mineral declarations may be presented vertically in the Percent
DV column when more than four vitamins and minerals are declared
or horizontally beneath the macronutrients. In 101.9(f), the
agency is providing for optional use of a simplified format under
certain circumstances. In 101.9(j)(13)(i), the agency is
providing exemptions to the requirement to bear nutrition
labeling for small packages, and in 101.9(j)(13)(ii), the agency
is providing options for intermediate sized packages to minimize
the amount of label space that must be used for nutrition
labeling. For small or intermediate sized packages, manufacturers
may list nutrition information in a linear fashion, use certain
abbreviations, omit the calorie-specific recommended nutrient
amounts, or present the nutrition information on other label
panels (see sections V.G.3., V.H.1., and V.J. of this document).
3. Highlighting
    In the format proposal, the agency requested comments on the
feasibility of allowing highlighting as a voluntary graphic
enhancement of the principal format. Specifically, the agency
requested comments related to whether the use of voluntary
highlighting would confuse or assist consumers to observe and
comprehend label information. Many different possible schemes for
highlighting could be applied to the nutrition label. Many
comments addressed noncontingent highlighting, in which certain
material is highlighted regardless of any product
characteristics, such as highlighting certain nutrient names
(e.g., fat, sodium, cholesterol) or titles (e.g., Percent DV,
Amount per serving) on the nutrition panel. Other comments
addressed contingent highlighting, in which certain material is
highlighted only if the product has certain characteristics, such
as highlighting the nutrition information for fat on the label of
a product that meets FDA's criterion for low fat. In it's
research, the agency tested a contingent highlighting scheme that
highlighted nutrients whose levels in the food qualified for
adjectival descriptors (high or low depending on the nutrient)
that were consistent with dietary guidelines.
    Many different possible techniques of highlighting exist,
including boldface print, all capital letters, italic print,
larger print, reverse print, different colored print, and color
banding. In the format examples published in the format proposal,
only boldface print was used for highlighting.
    166. Many of the comments that discussed noncontingent
highlighting suggested that highlighting should be considered as
a format enhancement and should be used for column headings and
names of nutrients. Other comments argued that the agency should
require the highlighting of certain nutrients because of their
health significance regardless of the level in the product. The
nutrients most frequently mentioned in the comments in this
regard were those associated with chronic disease, such as
sodium, fat, and cholesterol. Most of these comments suggested
that boldface type and all capital letters are adequate to
achieve such highlighting.
    Several comments addressed the issue of whether noncontingent
nutrition label highlighting should be mandatory or voluntary.
Some manufacturers objected to any required highlighting because
of the increased cost and increased label space required. Other
comments argued that if highlighting is allowed at all, it should
be mandatory so that the benefits of highlighting would be
universally available to consumers, and so that labels would be
uniform. Some of these comments argued that uniformity of labels
is important to reduce consumer confusion.
    The agency agrees than mandatory highlighting imposes some
burden on manufacturers and needs to be justified on the basis of
obvious benefits to consumers. The agency also agrees that the
use of highlighting to enhance column headings and nutrient names
can increase the visual interest of the label and make it more
legible for some consumers.
    However, the agency is concerned that allowing too many
optional highlighting schemes will lead to less consistency
between labels, and that highlighting has the potential to
increase label clutter and consumer confusion. Therefore, in
101.9(d)(i)(iv) the agency is requiring mandatory highlighting
of the title of the nutrition panel, ``Nutrition Facts'',
headings (``Amount per serving'' and ``% Daily Value''), nutrient
names (``Calories'', ``Total Fat'' ``Cholesterol,'' ``sodium,''
``total Carbohydrate,'' and ``Protein'') and percentage amounts
for certain nutrients. The agency concludes that these
requirements, by establishing a specific ``look,'' strike an
appropriate balance between establishing a nutrition label that
is readily observable and one that only increases clutter and
confusion.
    The agency does not agree that nutrients associated with
chronic disease should be highlighted. The agency notes that the
revised order of the nutrients already calls attention to the
nutrients of major public health significance.
    167. Several comments supported or opposed the voluntary
highlighting of certain nutrients based on their level in the
product (contingent highlighting). Supporting comments, primarily
from food manufacturers or trade associations, argued that
allowing such highlighting would provide useful information to
consumers. Opposing comments, primarily from health
professionals, professional associations, and consumer advocate
or health promotion organizations, argued that allowing
contingent highlighting on a voluntary basis would likely lead to
inconsistent and possibly self-serving highlighting that would be
more likely to misinform than inform consumers.
    Several comments supported contingent highlighting but
recommended that it be subject to the definitions used in FDA's
research studies or to the requirements for nutrient content
claims, because highlighting is a nutrient content claim. Related
comments supported a link between a nutrient content claim on the
front panel and highlighting the relevant information on the
nutrition panel. For example, if a low fat claim is made on the
front panel, the nutrition information for fat would be
highlighted.
    Some comments recommended that highlighting of undesirable
rather than desirable aspects of foods be required because
manufacturers will emphasize the good qualities, and such
highlighting will provide a balance.In contrast, a trade
association commented that the highlighting of ``bad'' nutrients
should not be required because it would be misleading. Another
comment suggested that highlighting be allowed only on a
case-by-case basis so that both FDA and food manufacturers have
maximum flexibility.
    Some comments opposed highlighting and argued that attention
should be drawn to specific nutrient levels by nutrient content
claims rather than by highlighting. Other comments argued that
the highlighting of positive information only will
accentuatebenefits without including information about risks and
will lead consumers to ignore vital information on negative
aspects of certain products. Several comments from industry were
opposed to contingent highlighting because it would communicate a
good foodopriately addressed in educational materials.
    Other comments opposed highlighting because, they claimed,
FDA will not have adequate enforcement resources. Permitting
voluntary highlighting will open the door for inappropriate use
of highlighting, which will then require additional regulatory
intervention. Another comment, consistent with comments about
other forms of dietary guidance on the label, argued that
selective highlighting based on current dietary recommendations
will change over time, and that it is unwise to include shifting
format elements on the label.
    FDA is not persuaded by the comments that contingent
highlighting on the nutrition label will benefit consumers.
Consumer research did not find that contingent highlighting
increases effective use of the nutrition label for product
comparison or dietary judgment uses (see section V.D.2.a.and
V.D.2.b. of this document). There is no consensus among the
comments, and the consumer research does not support that
requiring a particular contingent highlighting scheme is
appropriate. FDA is persuaded by comments that voluntary
contingent highlighting can be applied inconsistently in a way
that would be potentially misleading to consumers. Among products
with similar nutrition profiles, some would highlight certain
nutrients and others would not. Consumers who read nutrition
labels could not depend on the fact that all labels of similar
products would look the same, and the differences could undermine
the credibility of the information on the nutrition label and
lead to consumer confusion.
    The inconsistency of labels that would result also leads the
agency to disagree that highlighting should be allowed on a
case-by-case basis. Consistent treatment of similar information
is important for effective use of the nutrition label by
consumers. FDA agrees that attention can be drawn to the levels
of specific nutrients in a food by nutrient content claims rather
than by highlighting information on the nutrition label. Because
of problems of inconsistent treatment of similar information and
the lack of demonstrated benefits in the research, FDA is not
allowing the use of contingent highlighting on the nutrition
label.
4. Adjectives
    168. A majority of the comments that addressed the issue of
adjectives were in favor of their use in some context. Several
consumer groups recommended that the PERCENT DV with DRV format
be enhanced by a footnote providing FDA's definition of high and
low for nutrients. The comments argued that adjectives help
consumers to make qualitative judgments about the food without
having to make calculations or evaluate percentages. Some
comments noted that the ADJECTIVE format was the most preferred
in FDA's research.
    In contrast, comments opposed to adjectives in the nutrition
label argued that they clutter the format and would be redundant
with PERCENT DV formats. Comments also expressed the concern that
the wide range of values that will fall in the medium category
may mislead consumers.
    The agency does not agree that there is a need to include
adjectives on the PERCENT DV with DRV format. The agency's
research showed that consumers are able to make accurate high/low
judgments using percents alone, and that when consumers relied on
adjectives (see section V.D.2.a. above), they tended to overlook
nutrient differences if the nutrients were described by the same
adjective. The additional words would add clutter to the label
without benefit to consumers. Therefore, FDA is not providingfor
the addition of adjectives to the PERCENT DV with DV format.

I. Dual Declarations

    A few comments raised format issues with respect to foods
that use dual declaration displays on the nutrition label (i.e.,
nutrition information declared both on as packaged/as prepared or
per serving/per 100 g bases). A number of issues related to thzes
that such displays raise special format issues, and these are
discussed here.
    169. One trade association comment urged that FDA consider
the label space requirements of dual declaration formats before
deciding to require a change in the current nutrition label
format. Several comments stated that the PERCENT DV with DRV
format would not leave room enough for both ``as packaged'' and
``as prepared'' columns on nutrition labels. One industry comment
stated that industry sponsored research has shown that the
PERCENT DV with DRV format becomes particularly complex and
cumbersomewhen used on food labels using dual declaration
displays.
    FDA agrees that dual declaration displays have much greater
space requirements than the single base display required for most
food products. FDA notes, however, that dual declaration displays
are voluntary, not mandatory. For this reason, FDA does not agree
that the requirements of dual declaration displays should be a
determining factor in the decision about the nutrition label
format. However, FDA does agree that the unique requirements of
dual declaration displays should be accommodated as much as
possible within the constraints of the format that is required
for all nutrition labels, and the agency therefore has created a
new section, 101.9(e), to specifically address the format of the
nutrition label when dual columns are utilized.
    FDA notes that several studies submitted as comments or
listed in the format proposal found that dual declaration
displays of any kind made the product nutrition label
significantly harder for consumers to use and understand (see
section V.D.2. of this document). The execution of dual
declaration displays for the PERCENT DV with DV format in the
GMA/NFPA industry study (Ref. 71) demonstrated the problem with
multiple column displays on the nutrition label. In that study, a
number of the dual declaration displays placed both g/mg and
PERCENT DV declarations in separate columns on the nutrition
label, resulting in as many as five columns of information when
DV's were listed. The display was complex and cumbersome, but it
is not the only, and certainly not the best, way to execute the
PERCENT DV format with dual declarations.
    FDA's execution of the PERCENT DV format in Study 2 (Ref. 70)
purposely arranged the percent declarations in two columns and
presented a single g/mg declaration per nutrient in a noncolumn
array next to the nutrient name so as not to intrude visually on
the columns of percents. With this execution, the PERCENT DV with
DRV format performed as well as other formats on labels using
dual declaration displays. The agency is convinced that dual
declaration executions of the PERCENT DV with DV format should
follow the pattern of minimizing the number of columns displaying
nutrient amount per serving information. Accordingly, in
101.9(e) the declaration of the required g/mg nutrient amounts
on dual declaration nutrition labels is required to follow the
same requirements as for single declaration nutrition labels,
which is to be in an unordered array next to the nutrient name.
Placement of optional g/mg amounts is discussed in comment 170 of
this document.
    170. A comment argued that declaring two g/mg amounts in
parentheses next to the nutrient name as proposed in the dual
declaration format example included in the format proposal looks
like ``matrix coordinates'' and is likely to be confusing to
manyconsumers. One comment suggested an alternative for
presenting dual declaration g amounts. The comment suggested that
only the g/mg amounts of the product as packaged be in the table,
and that a footnote provide the additional amounts in the second
declaration for the food.
    FDA agrees that even when presented in a noncolumn array, the
declaration of two g/mg amounts for each nutrient on the
nutrition label (e.g., one each for as packaged/as prepared or
per serving/per 100 g) in addition to two columns of percent DV
amounome consumers. Based on the research findings (see section
V.D.2. of this document), FDA is confident that a single g/mg
amount declared in a noncolumn array does not have a detrimental
effect on consumers' abilities to understand and use the
nutrition information on the label, but FDA did not test a format
with two g/mg amounts for each nutrient. FDA is convinced by this
comment that only a single g/mg amount (as packaged and according
to the label serving size based on reference amounts in
101.12(b)) for each nutrient should be required in dual
declaration labels. The second set of g/mg amounts may be
presented optionally next to the required g/mg values,
differentiated from them by a comma or other means.
Alternatively, the second set of g/mg amounts may be presented in
a footnote. When the second set of g/mg amounts is presented in a
footnote, either the total amounts or the additional amounts may
be declared. When the additional amounts are declared, only those
nutrients that are present in different amounts than the amounts
declared in the required   g/mg information may be listed. The
footnote must clearly state which amount is declared. The agency
has included this provision in 101.9(e)(3)(i).
    Examples of nutrition label formats for products using a dual
declaration display that conforms to the new regulations are
presented in appendix E.

J. Simplified Format

    Most comments from consumers, industry, and professional
organizations supported the concept of using a simplified format
stating that it is easier for consumers to understand, cuts down
on label clutter, and gives manufacturers flexibility in
preparation of labels. The comments and FDA's response to
concerns raised by the comments are summarized below.
1. Terminology
    171. A few comments commended FDA and USDA for attempting to
bring consistency to nutrition labeling regulations by allowing
for similar types of simplified formats but requested that the
two agencies use the same term rather than ``simplified'' format
in FDA regulations and ``abbreviated'' format in USDA
regulations.
    Both FDA and USDA are in agreement with these comments. In
accordance with the language in the 1990 amendments, they will
use the term ``simplified'' format. However, because the foods
regulated by each agency are different, the specific
regulationspertaining to the simplified format will differ
somewhat by agency.
2. Required Use and Criteria
    172. Several comments disagreed with FDA's interpretation of
that part of section 403(q)(5)(C) of the act which states ``the
Secretary shall require the amounts of such nutrients to be
stated in a simplified form * * *'' (emphasis added) tomean that
foods that qualify for the simplified format must use that
format, i.e., they may not choose to use the full nutrition
labeling format. These comments urged that use of the simplified
format be optional as, in some instances, consumers may be better
served by one uniform nutrition label, and manufacturers should
be given the flexibility to meet consumer preferences. In support
of such labeling being voluntary, the comment cited the following
legislative history:

    However, the bill provides that the Secretary may permit the
information to be included on the label or labeling in a
simplified form if a food contains insignificant amounts of more
than one-half of the nutrients required to be on the label.
(emphasis added)

(Ref. 16)

    FDA advises that the draft legislation was revised by the
Senate subsequent to the above House report to replace the word
``may'' with ``shall'' and to add that the form of the simplified
format was to be ``prescribed by the Secretary'' (136
Congressional Record, S. 16607 (Oct. 24, 1990)). The changes were
not explained in legislative history, so that the intent of
Congress is not clear. The agency acknowledges that it is
possible that Congress was merely trying to require that the
final regulations the format be used whenever a food met the
qualifying criteria. In fact, inasmuch as the intent of the 1990
amendments was to increase the amount of nutrient information
provided to consumers, it is not entirely consistent that the act
would require less information on certain foods.
    Based on reassessment of the statute and its legislative
history in response to the comments, FDA concludes that use of
the simplified format should be at the manufacturer's discretion,
whenever a food product meets the criteria of containing
insignificant amounts of half of the required nutrients.
Accordingly, FDA has modified 101.9(f)(1) by changing the word
``shall'' to ``may.''
    173. A comment from a consumer group suggested that FDA
require a different base than one-half of all 15 required
nutrients for determining if a product qualifies for the
simplified format because using all 15 nutrients results in
``double-dipping.'' The comment suggested that calories from fat
should not be included because total fat represents the same
nutrient, and that total carbohydrate should also not be included
because complex carbohydrate and sugars comprise total
carbohydrate. The commentstated that this procedure would result
in a base of 13 different nutrients, of which seven must be
present in insignificant amounts to qualify to use the simplified
format.
    FDA agrees that counting both ``total fat'' and ``calories
from fat'' for the purpose of determining whether a food
qualifies for use of the simplified format results in a double
count being given to the fat content of a product. The agency,
therefore, has deleted calories from fat from the qualifying
criteria. However, in the case of ``total carbohydrate,'' the
agency notes that 101.9(c)(6) is revised to delete ``complex
carbohydrate'' as an element of nutrition labeling. Although
``other carbohydrate'' replaced ``complex carbohydrate,'' the
declaration of ``other carbohydrate'' is not mandatory.
Therefore, the required subcomponents in 101.9(f) no longer
comprise the total amount of the component ``total
carbohydrate.'' Accordingly, ``total carbohydrate'' must continue
to be included among the nutrients used as a base for determining
whether a food qualifies for use of the simplified format.
    The deletion of ``calories from fat'' and ``complex
carbohydrate'' results in a base of only 13 nutrients. According
to section 403(q)(5)(C) of the act, the simplified format is to
be allowed if a food contains insignificant amounts of ``more
than one-half the nutrients required * * *.'' Therefore, it
follows that the simplified format may be used when a food
contains insignificant amounts of seven or more of the base
nutrients.
    As a result, and in accordance with the reordering of
nutrients in 101.9(c), FDA has modified 101.9(f) to state that
the nutrition information may be presented in a simplified format
``when a food product contains insignificant amounts of seven or
more of the following: Calories, total fat, saturated fat,
cholesterol, sodium, total carbohydrate, dietary fiber, sugars,
protein, vitamin A, vitamin C, calcium, and iron.''
    Although FDA has deleted ``calories from fat'' from the list
of nutrients in 101.9(f) used to determine when a product may
use the simplified format, ``calories from fat'' continues to be
a nutrient that must be declared under section 403(a) of the act
in nutrition labeling when present in more than insignificant
amounts. Therefore, FDA has modified 101.9(f)(3)(ii),
redesignated as 101.9(f)(2)(ii), to require declaration of
calories from fat in addition to any other nutrients identifiedin
101.9(f) that are present in more than insignificant amounts.
For the same reason, the agency has modified 101.9(f)(4) to
require calories from fat to be included in the statement ``not a
significant source of &llmdash;&llmdash;&llmdash;'' if it is
present in
    174. One comment requested that FDA confirm that eligibility
for use of the simplified format is not limited to those foods
listed in the supplementary proposal as examples of foods that
would use the simplified format (56 FR 60421 at 60474), but that
the use of a simplified format is determined on a
product-by-product basis.
    FDA advises that the interpretation in the comment is
correct. The determination that a food qualifies for the
simplified format is dependent on the amount of nutrients in that
food.
    175. Some industry comments requested that FDA provide
guidance on how the simplified format applies to foods for
children under two years of age as these products are exempted by
proposed 101.9(j)(4) from labeling of calories from fat,
saturated fat, and cholesterol, all of which are included in the
list in proposed 101.9(f)(1) of the 15 ``required nutrients.''
Comments questioned whether the stipulation of insignificant
amounts of eight or more required nutrients for the simplified
formatapplies to such foods and if it does apply, whether
calories from fat, saturated fat, and cholesterol are included in
the eight insignificant nutrients, even though they are not
required to be labeled.
    In developing the proposed rules, FDA did not consider the
application of regulations governing the use of the simplified
format to foods for children less than 2 years of age. Since
these foods have a required base of only 11 nutrients (i.e.,
calories, total fat, sodium, total carbohydrate, dietary fiber,
sugars, protein, vitamin A, vitamin C, calcium, and iron), it is
appropriate that they be allowed to use a simplified format when
more than one-half (i.e., 6) of these 11 nutrients are present
ininsignificant amounts. Section 101.9(f) has been modified to
include this provision for foods for children less than 2 years
of age.
3. Definition of ``insignificant amount''
    176. A few comments recommended changes in FDA's proposed
definition of ``insignificant amount'' of a nutrient as that
amount that allows a declaration of zero in nutrition labeling.
The term ``insignificant amount'' was used in section
403(q)(5)(C)of the act in reference to when a food would be
exempt from nutrition labeling (proposed 101.9(a)) and to when a
food would qualify for the simplified format (proposed
101.9(f)(1)). Comments on both uses of the term are discussed in
this sectionto ensure consistent use of the term.
    A few comments stated that the use of a mathematical base for
determining ``insignificant amounts'' does not consider the
actual need for the nutrients in the maintenance of good health,
and that because FDA proposed to define ``source'' as from 10 to
19 percent of the RDI or DRV, anything less than 10 percent
should be ``insignificant.'' Other comments recommended the level
of insignificance be changed to 0.5, 2, 5, or 8 percent of the
RDI or DRV for particular nutrients. Another comment noted that
defining ``insignificant amount'' as an amount less than 0.5 g of
carbohydrate and protein is in conflict with the definition of
``insignificant amount'' for calories as less than 5 calories,
given that 1 g of carbohydrate and protein furnishes only
4calories. A comment also stated that, as a practical matter,
consumers cannot reasonably be expected to differentiate between
0.5 g amounts.
    FDA did take maintenance of health into consideration when it
based its proposed definition of ``insignificant amount'' for
calories (including calories from fat), total fat, cholesterol,
sugars, and sodium on the amount defined as ``free'' under the
proposed nutrient content claims rule (see final rule entitled
``Food Labeling: Nutrient Content Claims, General Principles,
Petitions, Definition of Terms'' published elsewhere in this
issue of the Federal Register). (In addition, in the final rule
on nutrient content claims, FDA defined ``saturated fat free'' as
less than 0.5 g of saturated fat per serving.) For most
nutrients, FDA has determined the level that is dietetically
trivial or physiologically inconsequential (56 FR 60421 at 60433)
ande, for those nutrients for which a level of ``free'' has been
defined, FDA is denying the request to change the definition of
``insignificant amount.''
    For those macronutrients that are required to be included in
nutrition labeling but that do not have definitions of ``free''
levels (i.e., total carbohydrate, dietary fiber, and protein),
FDA has reconsidered the proposed amounts and, in accordance with
the comments, it is specifying in 101.9 (f)(1) and (j)(4) that
an insignificant amount of these nutrients is ``an amount that
allows a declaration of less than 1 g.'' Because 1 g of each of
these food components yields 4 calories, this amount iscloser to
the amount that will yield an ``insignificant'' amount of
calories. By doing this, differentiation of amounts of 0.5 g will
no longer be necessary.
    In the case of vitamins and minerals, which also do not have
definitions of ``free'' levels, FDA is not persuaded that amounts
less than the amount defined as a source of a nutrient (i.e.,
less than 10 percent of the RDI) in 101.54(c)(1) in the document
on nutrient content claims published elsewhere in this issue of
the Federal Register can be considered ``insignificant.'' In
fact, assuming that as many as 20 foods are consumed in a day
(Ref. 109), levels of 5 percent or more of the RDI per food would
be sufficient to assure that a person's daily requirements were
met. Therefore, FDA rejects the suggestions that amounts greater
than 2 percent but less than 10 percent of the RDI be considered
insignificant.
    177. One comment expressed concern about a potential
compliance problem with 101.9(f)(1) for firms who elect to
``round down'' Class II nutrients under 101.9(g)(4)(ii) and to
``round up'' calories, sugars, total fat, saturated fat,
cholesterol, or sodium under 101.9(g)(5). The comment stated
that by defining ``insignificant'' based on analytical
capabilities while at the same time requiring that the simplified
format be used, a firm could find itself in violation of
101.9(f)(1) by either claiming that it is required to use the
simplified format, or by asserting that it must use the complete
nutrition format.
    FDA does not believe its proposed definition of
``insignificant'' has any bearing on this concern because the
concern could exist at any defined level. However, the amendment
of 101.9(f)(1) discussed above, whereby the simplified format is
allowed, not required, on foods that meet the qualifying
criteria, resolves the comment's concerns for Class II nutrients.
For other nutrients, FDA advises that firms should determine
label values to be in compliance with 101.9(g) and then
determine, based on those values, whether or not a food qualifies
to use the simplified format.
4. Nutrients To Be Declared
    178. While most comments supported the required declaration
of ``core'' nutrients (i.e., calories, total fat, total
carbohydrate, protein, and sodium) in the simplified format, a
few comments requested that the proposed mandatory declaration of
the ``core'' nutrients be deleted. Comments from firms
manufacturing honey, chewing gum, and spices requested that FDA
adopt a more simplified format for foods that have very limited
nutritional value. These comments requested that only those
nutrients that are present at more than insignificant levels be
required to be declared. For example, chewing gum would declare
only calories and carbohydrates, and seasoned salt only sodium.
    FDA is not persuaded that it should drop the requirement for
declaration of the core nutrients. This core information is
essential to aid consumers in learning about the relative
nutritional qualities of all foods, and it allows them to judge
the consequences of the food selections they make. Most comments
supported this position. Also, as discussed above in this section
on format, consistency in presentation is a principle that has
guided the agency in developing the new format because such
consistet making the requested change.
    179. A few comments from the soft drink industry expressed
opposition to mandatory listing of sugars in the simplified
format stating that it is not consistent with the intent of the
law which is to enhance consumers' understanding of sound dietary
practices. The comments contended that mandatory declaration of
sugars places undue emphasis on a nutrient that does not warrant
such emphasis in light of its physiological impact. One comment
explained that the greatest concern posed by sugars is their
potential carcinogenicity which, considering the rapid passage of
soft drinks through the mouth, is significantly lower than other
sugar-containing foods.
    A comment from the honey industry also objected to required
declaration of sugars on honey products on the basis that it
``could mislead consumers into thinking that the honey had been
manufactured from what consumers most likely regard as
`sugar'--table sugar.''
    As discussed in section III.F.3. of this document, the agency
has concluded that the mandatory declaration of sugars content in
nutrition labeling is consistent with the law. In regard to the
simplified format, the 1990 amendments and its legislative
history give no direction on the content of the simplified label,
only that it be ``in a simplified form prescribed by the
Secretary.'' Based on the criteria Congress put on the use of the
simplified format, it is possible to infer that its purpose is
not to save space on the label nor to allow the declaration of
otherwise mandatory nutrients to be omitted, but rather to modify
the label by allowing nutrients not present in significant
amounts to be omitted. FDA does not believe that the legislation
allows it, or that there is any reason, to permit a nutrient that
is required to be declared in complete nutrition labeling to be
omitted from the simplified format when that nutrient is present
in more than insignificant amounts.
    In response to the comments from the honey industry, FDA
acknowledges that consumers must be made aware of the different
purposes of the ingredient statement and the nutrition label and
be taught how to use the information in each. Developing this
awareness will be a component of the consumer education program
discussed in section IX. of this document.
    The comment presented no data to show that consumers will be
misled by a declaration of sugars in the nutrition label. As
discussed in section III.F.3. of this document, FDA believes that
sugars should be a mandatory component of nutrition labeling
because it will assist consumers in planning diets that conform
to current dietary guidelines and is of great interest to
consumers. Therefore, FDA finds no basis not to require that
sugars be treated in the nutrition label of honey as they would
be in the nutrition label of any other product that is a natural
source of sugars, such as fruits.
5. Use of statement ``Not a Significant Source of''
    A number of comments were received that addressed the
requirement in 101.9(f)(4) that the simplified format include
the statement ``Not a significant source of
&llmdash;&llmdash;&llmdash;&llmdash;&llmdash;,'' with the blank
filled in with the name of the missing nutrient, whenadditional
nutrients are voluntarily added to the food or declared in the
simplified format.
    180. Many comments on this subject supported the proposal.
However, a few comments from consumers, health professional
associations, and industry suggested that all simplified labels
should include a statement identifying those nutrients present in
insignificant amounts, such as ``Not a significant source of
&llmdash;&llmdash;&llmdash;&llmdash;&llmdash;.'' One comment
stated that consumers may be misled by the missing information
unless the nutrients that are not present are identified. Two
other statements that were suggested were2\ of the nutrients
required for full nutrition labeling.''
    FDA is not persuaded that consumers will be confused by the
absence of certain nutrients on simplified labels. Most of the
foods that will be able to use this format are basic commodities
or simple foods (e.g., oil, butter, sugar, syrups, juices,
drinks) for which it is reasonable to expect that consumers will
know that the missing nutrients are not present in the food.
Therefore, in response to the Congressional intent that the label
be ``simplified,'' the agency is not making the suggested change.
    181. Other industry comments generally opposed requiring such
a statement when additional nutrients are voluntarily added to
the food or declared in the simplified format on the grounds that
it clutters the label with a long list of nutrients that are not
present. One comment stated such a requirement is discriminatory,
especially if the additional nutrient is declared because of a
nutrient content claim. Another comment suggested that simplified
labels that declare naturally occurring nutrients betreated
differently from those that declare added nutrients. Several
comments suggested that the statement ``Not a significant source
of other nutrients'' be used in lieu of the proposed statement as
this would provide consumers information without cluttering the
label.
    FDA disagrees with the comments. When nutrients are
voluntarily added to a food or voluntarily declared in the
nutrition label, or when a nutrient content claim is made on the
label, the food is being marketed as a significant source of
nutrients. In such cases, the food label would be in violation of
section 201(n) of the act unless consumers are advised about the
full nutritional profile of the food.
    FDA shares the concern about the space required by the list
of nutrients not present. However, the statement ``Not a
significant source of other nutrients'' is too broad and
therefore could be misleading on a large proportion of foods.
Even though the food may not contain significant amounts of the
nutrients required in 101.9(c), it may contain significant
amounts of other essential nutrients that are not required to be
declared in nutrition labeling. The language suggested by the
comment, however, asserts that the food is not a source of any
other nutrients. Thus, to determine whether such a statement is
true, it would be necessary to analyze for all known essential
nutrients. The agency believes that such a situation makes no
sense and therefore is not making the suggested change.
    182. One industry comment opposed the exemption of
standardized enriched foods from the required statement ``Not a
significant source of
&llmdash;&llmdash;&llmdash;&llmdash;&llmdash;,'' stating that
there is no basis for treating different food products (i.e.,
nonstandardized enriched foods) discriminatorily. Another comment
wanted FDA to state that the addition of a nutrient such as
vitamin C to a food, if required by a standard of identity or
another government standard (i.e., a purchase specification to
qualify for use in the Special Supplemental Food Program for
Women, Infants, and Children (WIC)) would not require the
statement ``Not a significant source of
&llmdash;&llmdash;&llmdash;&llmdash;&llmdash;.''
    FDA is persuaded that foods containing added vitamins and
minerals, whether under a food standard or not and whether
required by purchase specifications or not, should be treated
similarly. Therefore, FDA has modified proposed 101.9(f)(4) to
require that if any vitamins or minerals are declared as part of
the simplified format for any reason, the statement ``not a
significant source of &llmdash;&llmdash;&llmdash;&llmdash;''
shall be included at the bottom of the nutrition label. This
statement is also required if any additional naturally occurring
nutrients are voluntarily declared in the simplified format. To
clarify the regulation, the requirement that any added vitamins
and minerals must be declared as part of the simplified format is
removed from proposed 101.9is subdivided into 101.9 (f)(3) and
(f)(4).
6. Format for the Simplified Label
    183. Many comments from industry responding to the
supplementary proposal were opposed to requiring the DRV list in
the simplified format, arguing that such a required list would
considerably expand the simplified format and therefore defeat
its purpose.
    A few comments responding to the format proposal argued that
examples of simplified formats illustrated at appendix D of the
proposed rule were merely abbreviated versions of the nutrition
label format and not simplified formats as called for by the1990
amendments. These comments were particularly critical of the
inclusion of the listing of DRV's in the simplified format
because they argued that an abbreviated list of DRV's would
communicate incomplete, and therefore misleading, information
about a total daily diet. Other comments supported the examples
of simplified formats in the proposed rule on the grounds that
they eliminated unnecessary information but retained a consistent
appearance with the regular format.
    FDA agrees that an important consideration for the simplified
format is that it retain common elements with the regular format
to facilitate consumer understanding and use of the nutrition
information. FDA does not agree that section 403(q)(5)(C) ofthe
act requires a simplified format that is simpler in other
respects than being an abbreviated version of the regular format.
As discussed in comment 179 of this document, the 1990 amendments
and their legislative history give no direction on the content of
the simplified format. However, FDA agrees with the concern
expressed about the value of an abbreviated list of DRV's. After
careful consideration of the comments, the agency is convinced
that by declaring quantitative amounts as percent of Daily
Values, the simplified format will retain sufficient common
elements with the regular format to facilitate consumer use and
comprehension. The agency is also convinced that not requiring
the full footnote and calorie conversion information required in
öA101.9 (d)(8) and (d)(9) on the simplified format will not
sacrifice important objectives of the legislation because the
information is not specific to the particular food and is
available on a significant portion of the food supply. Therefore,
FDA is requiring in 101.9(f)(5) that a simplified format contain
only quantitative and Percent of Daily Value information in the
same format as required for full or dual nutrition labeling in
101.9 (d) and (e), respectively.
    184. Comments to the format proposal addressing the use of
the simplified PERCENT DV formats generally preferred the use of
columns rather than the in-line presentation. Comments stated
that the in-line presentations appear significantly more
difficult to use and make it difficult to distinguish the actual
quantitative amounts from the DRV's. One comment was received
from a consumer interest group opposing the line concept on the
grounds that it is difficult to read, confusing, and will allow a
company to hide the content of fat, sodium, or other undesirable
nutrients in the product. The comment maintained that if a line
format is allowed, it should only be permitted when no additional
voluntary disclosures are made. The comment stated that such
additions would make the nutrition information comparable in
length to the required format, and FDA has already determined
that the required format would not be legible in a line format.
However, several industry comments were received in support of
allowing the abbreviated nutrition information to be presented in
either vertical columns or lines because of the flexibility and
saving in space provided in this option.
    FDA agrees that, where label space is adequate, the
simplified label is best understood when the information is
presented in columns, particularly when additional nutrients are
voluntarily disclosed. However, as discussed in section VI.K. of
this docuabels of small or intermediate sized packages.
Therefore, in 101.9(f)(5) the agency is requiring that nutrition
labels on products qualifying for the simplified format present
the required information in the same format as is required in
101.9 (d) and (e) for other packaged foods, except that foods in
small and intermediate sized packages that come under
101.9(j)(13) are allowed by that section to present the
information in a linear fashion. Examples of simplified formats
are seen in appendix F.
    185. One industry comment said that while it supported the
simplified nutrition label format for sugar, this format, as
depicted in the proposal, may be confusing on labels of soft
drinks because consumers may conclude that the soft drink has 36
g carbohydrate and 36 g sugars. If sugars are required to be
listed, the format should provide for indentation that would
clarify that sugars are a subcategory of total carbohydrates.
    FDA agrees that the format must allow for subcomponents to be
indented under the primary component. Accordingly,
101.9(c)(6)(ii) specifies that sugars are to be indented under
total carbohydrate. A similar requirement is specified for each
subcomponent. In addition, 101.9(d)(1)(iv) requires the primary
component to be highlighted to further differentiate it from its
subcomponents.

VI. Exemptions and Special Conditions

A. Small Business

    The 1990 amendments granted an exemption from mandatory
nutrition labeling for small businesses. Under section
403(q)(5)(D) of the act, a small business is defined as a
business with less than $500,000 annual gross sales of food or
any commodity, ora business with annual gross sales of more than
$500,000 but less than $50,000 in food sales. The exemption does
not apply to those products that make nutrition claims or provide
nutrition information.
    186. Many comments from industry, trade associations, and
international organizations have stressed that the dollar
exemption limits in proposed 101.9(j)(1) that implement the 1990
amendments are too low. The comments note that the sum of
analytical, printing, and other costs of nutrition labeling are
prohibitively expensive for low volume products. Many small food
producers that exceed the $500,000/$50,000 sales limit report
that they will suffer a severe economic hardship if forced to
comply with the nutrition labeling rules. One comment stated that
without an increased exemption, 25 percent of food businesses in
Kansas would close. Retail bakery and confectionery trade groups
stated that the nature of their business dictates that they
offerhundreds of different products throughout the year, and that
limiting and standardizing product lines would cause a retail
operation to lose its character and appeal. Yet, the need to
nutrition label products would force such standardization. Other
consequences for small businesses that would not qualify for the
exemption that were identified in the comments included the loss
of a substantial portion of annual profits, loss of low volume
product lines, and small business failure.
    FDA has considered these comments and believes that there is
merit in many of the contentions they raise. To gain adequate
information on what to recommend as a reasonable and appropriate
adjustment to the 1990 amendments' standard, FDA participatedin a
series of public forums that had been scheduled by USDA to
discuss the small business issue. These forums were held in May,
1992, in Kansas City, MO; Atlanta, GA; and San Francisco, CA. In
a notice of the public forums (57 FR 19410, May 6, 1992), FDA
announced its participation in the forums and requested comment
on a number of issues, such as which option should be used to
amend the current statutory exemption--increasing the gross
annual sales exemption, providing an exemption based on the
number ompliance with various limits and the effect on the
percent of the diet bearing nutrition labeling.
    The agency has compiled the information it received. However,
at the current time the agency is constrained by the requirements
of section 403(q)(5)(D) of the act. Therefore, 101.9(j)(1)(i)
has not been changed. If Congress amends the statute, FDA will
implement the change as soon as possible thereafter.
    187. Comments have questioned FDA's interpretation of that
part of section 403(q)(5)(D) of the act that states ``If a person
offers food for sale * * * or has business done in sales to
consumers'' to mean that foods produced by small businesses
thatare exempt would have to bear nutrition labeling if they were
sold by a larger retailer who was not exempt. The comments stated
that this interpretation would have a devastating impact on many
small entrepreneurs who primarily sell their products through
larger retailers or department stores.
    In 101.9(j)(1)(ii), FDA proposed that this exemption applied
to any ``person who manufactures, packs, or distributes food for
ultimate sale to consumers at the retail level as well as any
person directly involved in the retail sale of foods to
consumers.'' The legislative history was not specific as to
whether the term ``retailer'' applied only to the small business
retailer/producer or to a larger retailer acting as a middle-man
in handling the sale of the items to the ultimate consumer (Ref.
16). The agency is convinced by the comments that its
interpretation would have unintended consequences on small
businesses and, therefore, is removing ``as well as any person
directly involved in the retail sale of foods to consumers'' from
101.9(j)(1)(ii). To further clarify which foods are covered by
the small business exemption and to streamline the regulations,
FDA is also deleting the remaining portion of 101.9(j)(1)(ii)
and revising 101.9(j)(1)(i) to state that ``Food offered for
saleby a manufacturer, packer, or distributor who has annual
gross sales made or business done in sales to consumers that is
not more than * * *.'' The agency's intent with this change is
that the exemption will apply to persons whose name appears on
the label as the manufacturer, packer, or distributor of the
product, regardless of who ultimately sells the product to the
consumer. As a consequence, 101.9(j)(1)(iii) is redesignated as
101.9(j)(1)(ii).

B. Ready-to-Eat Foods

    188. Comments stated that proposed 101.9 (j)(2) and (j)(3)
did not adequately track section 403(q)(5)(A)(i) and
(q)(5)(A)(ii) of the act, which both pertain to foods ready for
consumption but differ in that section 403(q)(5)(A)(i) of the act
addresses foods served for immediate human consumption and
section 403(q)(5)(A)(ii) addresses similar types of foods that
are sold ready for human consumption but not for immediate human
consumption and that are processed and prepared primarily on the
premises.
    The agency is persuaded that proposed 101.9 (j)(2) and
(j)(3) in its supplementary mandatory nutrition labeling proposal
(56 FR 60366) (which were based on proposed 101.9 (h)(2) and
(h)(3) in FDA's July 19, 1990, mandatory nutrition
labelingproposal) did not adequately implement the 1990
amendments. FDA is therefore revising these two sections as
discussed below in accordance with the 1990 amendments and in
response to comments.
1. Foods for Immediate Human Consumption
    In proposed 101.9(j)(2) of its supplementary proposal (56 FR
60366), FDA proposed to exempt ``food products provided by
restaurants or other food service facilities offering
restaurant-type services (e.g., delicatessens, bakeries, feeding
facilities in organizations such as schools, colleges, hospitals,
and transportation carriers (such as trains and airplanes)).''
While this list was not all-inclusive, it was intended to respond
to section 403(q)(5)(A)(i) of the act which directed the agency
tois served for immediate human consumption * * *.'' Examples of
congressional intent concerning the types of facilities covered
by section 403(q)(5)(A)(i) of the act are limited in the
legislative history to cafeterias and hospitals (Ref. 16).
    189. While many comments supported the exemption in
101.9(j)(2) for restaurants, several comments requested
clarification about the coverage of the proposed section. For
example, comments asked whether it covers retail confectioners,
ready-to-eat food carryouts, vending machines, and food delivery
systems such as meals-on-wheels programs or establishments such
as pizza-delivery companies. Comments also pointed out the great
diversity in the types of establishments in which food is served
for immediate human consumption in the United States. For
instance, comments stated that in addition to full-service
restaurants, many establishments such as delicatessens, bakeries,
candy stores, and convenience stores provide customers with
tables and chairsto sit and immediately consume foods purchased.
Others, whether for lack of space or for other reasons, do not
provide such facilities. For example, frequently food franchises
in shopping malls sell cookies or other snack foods expecting
customers to eat the foods while walking in the mall or while
sitting on benches located throughout the mall.
    Comments from a company producing sandwich and salad items in
a commissary for sale in vending machines requested to be
included under this exemption because the subject foods are sold
for immediate consumption, not for ``take-home'' use, and because
the foods are prepared in a commissary kitchen similar to a
restaurant/cafeteria kitchen, where foods are assembled by hand
and subject to individual product variations. The comment argued
that mandatory nutrition labeling would require standardizationof
menu items, thereby prohibiting common day-to-day variations in
the food items produced, and would require larger labels or
smaller type-size, both of which would be difficult or impossible
to read through the small glass door of a refrigerated vending
machine.
    Similarly, one comment pointed out that some foods sold in
convenience stores are intended for immediate human consumption
and compete directly with foods served by restaurants and
delicatessens. It stated that many stores have seating areas for
customers to use while eating foods purchased on-site, and that
in some states, such convenience stores must have restaurant
licenses. Foods sold range from self-service beverages to
prewrapped sandwiches, prepared off-site by vendors and offered
for sale in store display cases.
    FDA notes that section 403(q)(5)(A)(i) of the act addresses
restaurants and ``other establishments in which food is served
for immediate human consumption.'' To respond to the comments
stating the proposed rules did not adequately track the 1990
amendments, the agency is revising proposed 101.9(j)(2) to
include a new paragraph (ii) that states that the exemption is to
include food products served in ``other establishments in which
food is served for immediate human consumption.'' In addition,in
response to comments seeking clarification of the coverage of
such ``other establishments,'' and in recognition of the
diversity of food service operations in the United States, the
agency advises that while some enforcement decisions will need to
bemade on a case-by-case basis, for efficient enforcement of the
act, it is providing in 101.9(j)(2)(ii) that, in addition to
food service in hospitals and cafeterias, the agency considers
that this exemption applies to establishments such as bakeries,
delicatessens, and retail confectionery stores where there are
facilities for ``immediate consumption'' on the premises (i.e.,
tables or counters with chairs); to food service vendors such as
lunch wagons, mall cookie counters, vending machines, and
sidestores); and to food delivery systems or establishments where
ready-to-eat foods are delivered to homes or offices for
immediate consumption.
    FDA recognizes that some persons might consider that it is
inconsistent for the agency to exempt packaged foods sold in
vending machines from nutrition labeling but not from ingredient
labeling. However, the agency is convinced that such foods are
exempted from nutrition labeling by section 403(q)(5)(A)(i) of
the act because vending machines serve food for immediate
consumption, and there is no similar statutory exemption from
ingredient labeling.
    Regarding convenience stores, FDA agrees that some foods sold
in such stores bear many similarities to foods sold at
restaurants and delicatessens and should qualify for similar
exemptions. Because circumstances will vary greatly according to
the services a particular convenience store offers, it is not
possible to state precisely which foods do or do not have to
provide nutrition labeling. Rather, determinations will have to
be made on a case-by-case basis. However, 101.9(j)(2) generally
provides an exemption for foods of the type served in restaurants
or other establishments in which food is served for immediate
human consumption. Such foods might include beverages (both
self-service and those served by store personnel), frankfurters
in a roll, cold sandwiches, pizzas, and hand-packed ice cream
cones.
    190. Many comments requested that proposed 101.9(j)(2) be
amended to clearly exempt foods ``sold for sale or use'' in
restaurants or other establishments in which food is served for
immediate human consumption as specified in section
403(q)(5)(A)(i) of the act. They argued that the statutory
language indicates that food intended for use in restaurants is
exempt from mandatory nutrition labeling in the absence of
nutrient content or health claims. The comments pointed out that
proposed 101.9(j)(2) merely provided an exemption for foods
provided by restaurants and did not cover foods intended for sale
or use in restaurants.
    The agency agrees that the proposed regulations did not fully
implement section 403(q)(5)(A)(i) of the act that covers foods
sold for sale or use in restaurants or other such establishments.
As directed in the statute, this exemption applies to allfoods
sold in restaurants, including packaged products such as a
specialty house dressings made by the restaurant, or used in
restaurants, such as portion controlled packages (e.g.,
individual catsup or coffee whitener packages) for use only in
restaurants. If a manufacturer makes a product for sale only in
restaurants (e.g., a package of candy), that product need not be
nutrition labeled. However, if there is a reasonable possibility
that the product will be purchased directly by consumers in a
setting other than a restaurant or other establishment in which
it is served for immediate consumption, it must be nutrition
labeled (see Ref. 25). Accordingly, FDA has modified proposed
101.9(j)(2) to add a new paragraph (iii) that exempts foods sold
for sale or use only in restaurants or other establishments in
which food is served for immediate human consumption.
    191. A few comments requested that the second sentence of
proposed 101.9(j)(2) be revised to adequately implement section
403(q)(5)(F) of the act that exempts food ``which is sold by a
food distributor if the food distributor principally sells food
to restaurants or other establishments in which food is served
for immediate human consumption and does not manufacture,
process, or repackage the food it sells.'' The comments pointed
out that the second sentence in proposed 101.9(j)(2) would only
exempt ``foods sold to restaurants by distributors * * *'' which
is duplicative of that part of 403(q)(5)(A)(i) of the act that
stipulates an exemption for foods sold for sale or use in
restaurants and fails to include the broader exemption in
403(q)(nts in which food is served for immediate human
consumption and who do not manufacture, process, or repackage the
food they sell.
    The agency is persuaded that there is a need to revise the
second sentence of proposed 101.9(j)(2). As discussed in the
legislative history (Ref. 25), the food distributor that sells
principally to restaurants and other food service establishments
is exempted from mandatory nutrition labeling requirements as
long as the food distributor does not manufacture the product
sold to the consumer. However, the legislative history states:

    The manufacturer of such products would be responsible for
providing the nutrition information on the products if there is a
reasonable possibility that the product will be purchased
directly by consumers, even if the principal customers are
restaurants and other wholesale purchasers.* * * [T]he
distributor is not liable as long as the distributor does not
manufacture the product sold to the consumer.
(Ref. 25)

    In essence, this legislative history makes clear that section
403(q)(5)(F) of the act is intended to direct the agency to do
for foods sold to restaurants what it does for foods sold to
consumers; that is, to hold the manufacturer, not the seller,
responsible for nutrition labeling of foods. (The only exception
to this approach is the voluntary nutrition labeling program for
raw fruits, vegetables, and fish in which the retailer is to
provide the nutrition information.) This exemption would applyto
an independent distributor who principally distributes
institutional foods directly to restaurants and similar
establishments and does not manufacture, process, or repackage
the food it sells.
    Thus, under this exemption, such a distributor is not
responsible for nutrition labeling a product, even if it sells
the product in a so-called ``cash and carry'' store, unless it
manufactures, processes, or repackages the food for sale to
consumers. On the other hand, a manufacturer of institutional
size food products is responsible for nutritionally labeling
those products if there is a reasonable possibility that they
will be sold to consumers, for example, through such a mechanism
as a cash and carry store.
    Therefore, proposed 101.9(j)(2) is modified by adding
101.9(j)(2)(iv) to fully implement this exemption.
    192. One comment recommended that statements such as ``for
food service use'' or ``not labeled for retail sale'' be used as
one means of qualifying for the exemption or that such foods be
identified by the size of the package. The comment suggested that
such a rule would be of particular help for foods imported for
the food service trade.
    The legislative history quoted in the preceding comment makes
clear that nutrition labeling is required ``if there is a
reasonable possibility that the product will be purchased
directly by consumers * * *.'' Therefore, the agency does not
believe that a label statement can be used as the basis for this
exemption. The agency is concerned that, if permitted, a label
statement such as ``for food-service use'' would be used to claim
exemption for products that Congress intended to be nutrition
labeled. Therefore, rather than create the possibility for
potentially misleading labeling, FDA is denying this request.
    Imported foods that are in large packages that are obviously
not intended or packaged for sale to consumers would be
considered exempt under 101.9(j)(9) which deals with foods
shipped in bulk form that are to be processed, labeled, or
repacked at a subsequent site.
    193. Several comments opposed proposed 101.9(j)(2) because
it would exempt restaurants from mandatory nutrition labeling.
These comments urged that restaurants, particularly the regional
and national chain restaurants, be required to have nutritn
computer analysis of nutrient databases; and (3) presented in
alternative ways such as brochures, menu boards, posters, or tray
liners. A few large fast food restaurant chains requested
guidelines for voluntary nutrition labeling with flexibility
informat and content. They requested that restaurants be allowed
to use their own serving sizes, present information on an
as-served basis, and update information annually.
    In response to comments requesting that restaurants be
required to provide nutrition information, the agency points out
that section 403(q)(5)(A)(i) of the act specifically exempts
restaurants and other establishments in which food is served for
immediate human consumption from mandatory nutrition labeling,
unless a nutrient content claim or a health claim is made. The
requirements that pertain when claims are made are discussed
extensively in the final rules on the general requirements for
nutrient content claims and health claims that are published
elsewhere in this issue of the Federal Register.
    FDA is aware, however, of the consumer interest in knowing
the nutrient content of foods eaten away from home. In response
to that interest and to the comments from fast food chains, the
agency intends to work closely with all interested parties,
particularly those in the food-service sector, to develop
guidelines for presenting nutrition content information in a
restaurant setting in such a way that it will not inhibit the
flow of useful nutrition information (e.g., claims) to the
consumer, whileat the same time providing assurance of the
reasonable accuracy of the information, thus furthering the goal
of the 1990 amendments to aid consumers in maintaining healthy
dietary practices.
    194. One comment agreed that the 1990 amendments exempt
restaurants from mandatory nutrition labeling but requested that
they be regulated under sections 201 and 403(a) of the act. The
comment also requested that FDA clarify that the 1990 amendments
have no preemptive effect on state or local regulation of the
nutritional disclosures by restaurants.
    FDA advises that the exemptions in 101.9(j) in no way exempt
any foods from regulations promulgated under sections 201 and
403(a) of the act. In regard to State and local preemption, the
legislative history states that ``Because food sold in
restaurants is exempt from the nutrition labeling requirements of
section 403(q)(1) through (q)(4) of the act, the bill does not
preempt any State nutrition labeling requirements for
restaurants. If States do require such labeling in restaurants,
it is important that they make every effort to make those
requirements consistent with the requirements of the bill.''
(Ref. 110).
    2. Foods Not for Immediate Consumption. 195. Many comments
objected to proposed 101.9(j)(3) that allowed an exemption for
in-store delicatessen and bakery foods only when they were sold
from behind service counters. Comments pointed out that the 1990
amendments made no distinction for such foods when sold from
behind the counter rather than from a self-service display. They
stated that such a rule would be totally unworkable and would
adversely affect the bakery and deli departments. Such a rule,
according to the comments, would make it impossible to sell foods
that are sold from behind the service counter during the day, at
night, when no service clerks are available, or to assemble
sandwiches and salads for fast pickup during the lunch hour from
self-service counters, without nutrition labeling those foods. A
trade association reported that 21 percent of in-store bakeries'
sales come from bulk self-service units and 42 percent from
prepackaged self-service cases, and that, if compelled to
standardize and label their products, such bakeries would be
unable to continue in competition with wholesale bakery items.
    The comments argued that in-store delicatessens and bakeries
should be able to adhere to the same regulations as their
independent counterparts, with whom they compete for business.
They stated that in-store delicatessens and bakeries operate as
independent bakeries in that their accountability is separate
from the rest of the store.
    Other comments stated that the intent of Congress was that
foods similar to restaurant foods that are ready for immediate
consumption, and that are produced by retailers that offer
variable and nonstandardized products, should be exempt from
nutrition labeling to eliminate the substantial burdens that
would otherwise be imposed if such labeling were mandatory. Many
comments have pointed out that the average baker or confectioner
produces hundreds of different products each year, and that the
average sales per product are relatively low. Comments stated
that the precise selection of foods produced is frequently
modified according to changing preferences of customers, seasons,
holidays, ingredient availability, and the individuality of the
baker/confectioner. Comments argued that it is this ability of
the retailer to vary and customize food products that gives the
establishment its character and appeal, and that forced
standardization to allow for nutrition labeling would eliminate
product competitiveness by disallowing innovation and creativity.
This loss of competitiveness, in turn, would create a major
economic burden and thereby lessen consumer choice.
    The agency is persuaded that proposed 101.9(j)(3) did not
adequately implement section 403(q)(5)(A)(ii) of the act, could
result in economic harm to in-store delicatessens and bakeries,
and created an artificial demarcation between foods sold from
service versus self-service areas of the delicatessen or bakery.
Therefore, FDA believes that it is necessary to
revise101.9(j)(3) to more closely reflect the language of the
act.
    As stated above, section 403(q)(5)(A)(ii) of the act applies
to foods that are: (1) Similar to the type addressed in section
403(q)(5)(A)(i) of the act, (2) ready for human consumption, and
(3) offered for sale to consumers but not for immediatehuman
consumption. Accordingly, FDA is modifying 101.9(j)(3) by adding
paragraphs (i) through (iii) to reflect these three statutory
requirements. There were no specific concerns presented in
comments that suggest any problems with these requirements.
    Section 403(q)(5)(A)(ii) of the act also requires that the
foods to which it applies be processed and prepared primarily in
the retail establishment. The agency is codifying this
requirement in 101.9(j)(3)(iv). Comments were very divided on
this issue, particularly for bakery items. Some comments argued
that breads shipped to a retail store in a semi-finished
condition and baked-off just before retail sale would not meet
the criteria of ``prepared and processed primarily'' at the
retail store and should not be exempt. However, other comments
disagreed with this position, stating that frozen dough products
are further processed at the bakery by being proofed, shaped,
molded, filled, decorated, cut, assembled, customized, or
otherwise completed or further processed and should be exempt.
    The legislative history discusses this situation stating that
for bakeries ``simply thawing frozen bread would not be
sufficient; the bread would have to be baked on the premise''
(Ref. 16). While this statement appears to indicate that baking
(i.e., cooking at a high temperature in an oven) is sufficient to
qualify a food for this exemption, comments have argued that
``baking'' is not equivalent to heating but includes a number of
steps, such as ``selecting, weighing, and mixing ingredients,
fermentation, and shaping and forming the dough prior to actual
heating the product.''
    This example with bread illustrates the difficulty in
applying the criterion of ``processed and prepared primarily in a
retail establishment.'' Because of the wide variety of foods sold
in delicatessens, bakeries, retail confectionery stores, and oth
impossible to identify for each type of food sold the exact
amount of processing or preparation that would be needed to say
that that food was ``processed and prepared primarily'' on-site.
    In many establishments, foods such as bakery items or salads,
may be prepared entirely on-site; however, in other similar
establishments, much less of the processing and preparation of
these foods is actually done on the premises. Similar variations
are encountered with other foods, such as cheeses, which may only
need to be sliced and portioned, or puddings, which may be
purchased in cans and only need to be put in trays in the display
case for portioning. The characteristics that all of these foods
have in common is that they are ready-to-eat, they are the same
type of foods sold in restaurants, and they are portioned and
packaged on-site.
    Legislative guidance to assist the agency in defining what is
meant by ``processed and prepared primarily in a retail
establishment'' in addition to that cited above is scant. However
during Senate debate, one of the sponsors of the bill that became
the 1990 amendments stated that:

    This exemption recognizes that when food is processed and
prepared primarily on the premises and sold there, as in the
prepared food sections of supermarkets, nutrition labeling is not
appropriate. On the other hand, if the preparation or processing
of food is standardized and is accomplished primarily at another
establishment and the same food is then shipped to a retail food
store in a form that requires minimal or no further processing,
nutrition labeling can be easily accomplished and is required.

(Ref. 110).

    The agency interprets this legislative history to mean that
if the food arrives at a store in a form to be sold directly to
the consumer (i.e., it is ``standardized''), then nutrition
labeling must be required. However, if the food is not
standardized, i.e., it has to undergo processing or preparation,
including portioning, before being sold to the consumer, then
nutrition labeling is inappropriate and should not be required.
In the case of the examples cited above, FDA finds that nutrition
labeling would therefore not be required on bread that is shaped;
filled, decorated, assembled, or customized and baked (i.e.,
cooked at a high temperature) in the retail establishment. Cheese
that is sliced and portioned according to directions given by
theconsumer, and pudding that is portioned according to
directions given by the consumer, also need not be nutrition
labeled. In these examples, the food is not ``standardized'' in
the form that it is to be sold to consumers when it arrives at
the store.Similarly, candies sold in retail confectionery stores
that are selected by consumers to be part of a packaged
assortment are not ``standardized.''
    However, because of the great diversity of situations in
which foods are sold, it must be recognized that a decision
regarding exactly what foods do or do not require nutrition
labeling cannot be fully resolved by regulation. Circumstances at
the retail location must be the deciding factor.
    196. A few comments from the retail baking and confectionery
industries and from grocery stores requested that the exemption
for single-unit bakeries, delicatessens, and confectioneries
apply equally to multi-unit establishments that do most or all of
their preparation at a central facility or shop. Each type of
respondent attempted to limit such an exemption by describing
what it would encompass. For example, a comment from the retail
baking industry described multi-unit bakeries as being
owned,controlled, and operated by the same entity and stated that
finished products would be delivered unwrapped or in bulk
delivery boxes to each store or outlet. The confectionery
industry requested that the exemption cover satellite operations
operated by tfoods prepared in central kitchens for use in the
retailer's own stores. Reasons given for using central facilities
included ensuring quality control through a controlled
environment that promotes food safety and integrity and allowing
for economies ofscale. Comments stated that the average number of
bakeries operated by a multi-unit retail bakery was 2.4 in 1988,
and that many small independent confectioneries only operate one
additional outlet.
    FDA does not believe that the 1990 amendments allow for
exemptions beyond those discussed in the preceding comments. This
position is based on the final criterion given in section
403(q)(5)(A)(ii) of the act, which states that foods to which the
section applies shall not be offered for sale outside the retail
establishment in which they are primarily processed and prepared.
The agency is codifying this requirement in 101.9(j)(3)(v).
While foods that are fully prepared and portioned (i.e.,
``standardized'') at the central facility are required to bear
nutrition labeling, there may be some types of food products or
circumstances in which the portioning or packaging is not
standardized, and in which nutrition labeling would consequently
not be required (e.g., salads that are portioned and packaged
according to directions given by the consumer).
    FDA notes that the problems presented in most of the comments
on this aspect of this exemption have more to do with the size of
the businesses than whether there are good reasons not to require
nutrition labeling. FDA believes that the best way to deal with
most of these comments is through a change in the small business
exemption.

C. Foods of No Nutritional Significance

    To reflect the first sentence in section 403(q)(5)(C) of the
act, FDA proposed an exemption for foods of no nutritional
significance in 101.9(a). It proposed to include the other
exemptions in 101.9(j). To minimize any confusion that
thesedifferences in placement may cause, the agency has decided
to group all exemptions in one place in this final rule.
Accordingly, that part of proposed 101.9(a) that exempted foods
of no nutritional significance is redesignated as 101.9(j)(4).
    197. Comments from the coffee industry noted that, unlike
FDA's mandatory nutrition labeling proposal, the supplementary
proposal did not explicitly identify coffee as being
nutritionally insignificant. Thus, the comment requested
clarification in the final rule. The comments pointed out that
coffee is always consumed as a brew. An analogy was drawn
to101.45(b)(4) in the guidelines for voluntary nutrition
labeling of raw fruit, vegetables, and fish, which states that
nutrition information is to be based on the edible portion of the
food. Comments stated that the available nutrients in brewed or
plain instant coffee would meet the criteria for being
nutritionally insignificant.
    The agency agrees that only the edible portion of coffee
should be considered in determining the nutritional significance
of the product. Therefore, based on a review of available
nutritional data on a serving of coffee and on the revisions in
the levels that are significant, discussed in comment 176 of this
document, FDA has concluded that coffee beans, roasted ground
coffee, or dry plain (i.e., unsweetened) instant coffees contain
no nutrients at other than nutritionally insignificant levels. As
a result, these foods are exempt from mandatory nutrition
labeling. Unsweetened plain tea powders or tea leaves likewise
would be exempt.
    In response to comments requesting clarification of the
exempt status of coffee and tea, FDA has included in 101.9(j)(4)
a listing of coffee beans (whole or ground), tea leaves, and
unsweetened plain instant coffee and tea as examples of foods
thahe coffee or tea.
    198. The spice industry commented that FDA did not establish
a reference amount for spices, thereby implying that spices are
exempt from mandatory nutrition labeling. Comments requested that
the agency provide an explicit statement in the final
ruleregarding the exemption of spices, spice blends (e.g., curry
powder), and condiment-type dehydrated vegetables (e.g., dried
garlic) as well as flavor extracts and food colors, from the
nutrition labeling requirements.
    As discussed in the final rule on serving size published
elsewhere in this issue of the Federal Register, FDA has set a
reference amount of \1/4\ teaspoon for most spices and
condiment-type dehydrated vegetables. In reviewing the
nutritional data in Agriculture Handbook No. 8 - 2 and 8 - 11
(Refs. 111 and 112), the agency has found that, under FDA's
criteria for determining nutritional insignificance, the vast
majority of spices, spice blends, and condiment-type vegetables
are exempt from mandatory nutrition labeling. FDA found, however,
that one spice (paprika) and one spice blend (chili powder),
exceed the cutoff levels for one or two nutrients. Using the
appropriate rounding procedures, paprika is over the cutoff for
vitamin A (6 percent of the RDI), and chili powder is over the
cutoff for both vitamin A (4 percent of the RDI) and sodium (5
mg) per \1/4\ teaspoon serving. The levels at which these
nutrients are nutritionally insignificant (i.e., the amounts that
can be rounded to zero) are less than 2 percent of the RDI for
vitamin A and less than 5 mg for sodium. Therefore, under the
act, paprika and chili powder will have to be nutrition labeled
(see Ref. 16, p. 16: ``Foods such as certain spices, which have
insignificant amounts of most but not all nutrients, are covered
by the nutrition labeling requirements.''). Because not all
spices and spice blends are nutritionally insignificant, they are
not included as a category under 101.9(j)(4).
    Condiment-type dehydrated vegetables, flavor extracts, and
food colors do meet the criteria for foods of no nutritional
significance and, therefore, are exempt from mandatory nutrition
labeling. As with unsweetened coffee and tea, 101.9(j)(4) will
include these examples of nutritionally insignificant foods.
    199. One comment suggested that ``fun foods'' defined as
foods with empty calories (i.e., those with no nutrients other
than calories), such as plain sugar candies, gum, and carbonated
beverages, should be exempt from mandatory nutrition labeling
except for a declaration of calories and the statement ``no other
significant sources of nutrients.'' The comment argued that the
statement ``Contains less than 2 percent of the RDI'' for such
foods is deceptive and miseducates consumers.
    FDA advises that these types of foods would qualify
under101.9(f) for the simplified label and would only be
required to list the core nutrients, not the statement ``Contains
less than 2 percent of * * *.'' Moreover, Congress did not
provide for an exemption of such a category of foods in the
statute. Therefore, the agency is taking no action on this
comment.
    200. The pickle industry commented that, as a cost-saving
measure, only sodium content (as is permitted under current
regulations) should be required to be labeled on dill pickle
products, rather than the full simplified format. The comment
arguedthat, even though a serving of dill pickles also contains 1
g of carbohydrate, sodium is the only nutrient of any concern to
consumers.
    FDA rejects this comment. Section 403(q)(5)(C) of the act
exempts from nutrition labeling foods that contain insignificant
amounts of all of the nutrients required within nutrition
labeling. The same section also provides for a simplified form of
nu the suggested changes in the regulations.
    201. One trade association commented that bottled water
products have little or no nutritional value, and that such
products should be exempt from mandatory nutrition labeling. The
comment asserted that the following industry practices should be
permitted without triggering nutrition labeling obligations: (1)
Bottlers should be allowed to add back minerals as flavor
enhancers that are removed during purification and declare
``minerals added'' on the principal display panel; (2) bottlers
should be allowed to describe bottled water with natural or added
fluoride as ``fluoridated water;'' (3) bottlers should be allowed
to add sodium fluoride or add back trace minerals that may
contain sodium as an incidental additive and still be permitted
to claim ``sodium free'' on the label; (4) ``essence'' bottled
water products (i.e., those containing 1 percent or less of juice
or flavors) should be considered nutritionally insignificant; and
(5) bottled mineral water products should be permitted to have a
listing on the label of certain minerals, e.g., sodium,
bicarbonate, calcium, magnesium, and other trace minerals in mg
per liter in addition to a declaration of total dissolved solids
content (which some state laws currently require). The comment
argued that the EC Directive on Nutrition Labeling expressly
exempts mineral water and other waters from nutrition labeling,
and, for the sake of harmonization, FDA should do likewise.
    FDA points out that, separate from this rulemaking on
nutrition labeling to implement the 1990 amendments, the agency
is in the process of amending its regulations on bottled water,
partly in response to a petition from the trade association that
submitted the comment. The bottled water regulations will address
certain aspects of labeling apart from nutrition labeling, e.g.,
definitions, information about mineral content, and required
label statements. Under the 1990 amendments, Federal regulations
will preempt any State standards of identity that are not
identical to it (section 403(a)(1) of the act).
    A recent IOM report, ``Food Labeling: Toward National
Uniformity'' (Ref. 113), noted that many States have expressed
concern about the heightened potential for consumer confusion
because of the increased number of bottled water products on the
market and the aggressive marketing and advertising claims of
superiority made for them. Thus, FDA maintains its position that
nutrition information relating to food must be provided for all
products, including bottled and mineral water, that contain more
than insignificant amounts of any of the nutrients or food
components that are required to be listed, or whose label,
labeling, or advertising contains a nutrient content claim or any
other nutrition information in any context. For products that
qualify for the simplified format, if manufacturers voluntarily
declare nutrients allowable under 101.9(c) that are not among
the 14 required nutrients (e.g., potassium), the required
statement ``Not a significant source of
&llmdash;&llmdash;&llmdash;&llmdash;&llmdash;,'' must be used,
with the blank filled in with the name of any of the 14 required
nutrients or food components that are not present or are present
in insignificant amounts. Moreover, if a product is voluntarily
enriched or fortified with added vitamins or minerals, any such
nutrients must be declared using the simplified format and
followed by the above statement. Thus, a product labeled as
``bottled water, minerals added'' will have to bear nutrition
labeling.
    The agency considers the identity statement ``fluoridated
water'' misleading if the product is derived from a source
naturally containing fluoride. Use of the term ``fluoridated''
represents that fluoride has been added in the processing. Thus,
the
    Bottled water products containing juice or other flavors are
subject to the same nutrition labeling requirements as any other
food. If a product meets the criteria for no nutritional
significance, and no claims are made, then nutrition labeling is
not required. A ``sodium free'' declaration on bottled water or
on any other food label will trigger nutrition labeling, because
such a claim promotes the nutritional properties of the product.
    202. One comment stated that, to avoid varying
interpretation, FDA should clarify what it means by the term
``implicit'' as it applies to nutrient content claims or
information that will bar a food from an exemption from nutrition
labeling under the ``no nutritional significance'' provisions.
    A thorough discussion of implicit claims may be found in the
companion documents on nutrient content claims and health claims,
found elsewhere in this issue of the Federal Register.

D. Foods for Infants and Children Less Than 2 Years of Age

    In the mandatory nutrition labeling proposal (101.9(h)(4))
and in the supplementary proposal (101.9(j)(4)), the agency
proposed to require that foods, other than infant formula, that
are represented or that purport to be specifically for infants
and toddlers less than 2 years of age bear nutrition labeling,
except that such labeling shall not include information on the
number of calories from fat or the amount of saturated fat and
cholesterol present in the food.
    203. The comments supported this proposal. One comment,
noting that proposed 101.9(j)(4) refers to ``toddlers less than
2 years of age'' and other references in 101.9 refer to toddlers
as children less than 4 years of age, recommended that ``children
less than 4 years of age'' be used, or that the term ``toddler''
be clarified. Another comment pointed out the practical fact that
some foods used in the 1 to 2 year age bracket are also used by
some children up to 4 years of age as well as by adults who have
problems chewing food.
    FDA does not agree that these special labeling requirements
proposed for foods for infants (other than infant formula) and
toddlers less than 2 years of age should be extended to children
less than 4 years of age despite the fact that no other nutrition
labeling requirements use 2 years of age as a cutoff. The agency
does not believe there is scientific support to change the cutoff
to 4 years because dietary recommendations for very young
children are specific in citing 2 years of age as the age under
which dietary modifications are not appropriate. For instance,
the ``Report of the Expert Panel on Blood Cholesterol Levels in
Children and Adolescents'' of the National Cholesterol Education
Program (NCEP) states:

    The fast growth of infants requires an energy-dense diet with
a higher percentage of calories from fat than is needed by older
children. Based on current knowledge, it is inappropriate to
apply nutrient guidelines for fats, cholesterol, and caloriesto
children under 2 years of age.

(Ref. 114.)

and:

    As toddlers over 2 years of age begin to eat with the family,
they may safely make the transition to this [recommended] eating
pattern.

(Ref. 114.)

    However, FDA believes that some clarification is needed as to
the types of foods addressed in 101.9(j)(4) (which is
redesignated as 101.9(j)(5)(i)). The agency advises that the
infant and toddler foods to which the special labeling
requirements are intended to apply are the types of foods
represented in 101.12(b), Table 1 entitled ``Reference Amounts
Customarily Consumed: Infant and Toddler Foods'' in the rule
entitled ``Food Labeling; Serving Sizes'' published elsewhere in
this issue of the Federal Register. FDA notes, however, that in
its serving size reproposal (November 27, 1991, 56 FR 60394 at
60397), ``toddlers'' was interpreted to mean children 1 through 3
years of age. Therefore, the agency advises that no special
significar;'' rather it is the age category that is important. To
reduce the possibility of confusion, FDA is replacing the word
``toddler'' with ``children.'' The distinguishing characteristic
of foods to which the special labeling requirements in
101.9(j)(5)(i) apply is that they are specifically represented
or purported to be ``for use by infants and children less than 2
years of age.'' Foods represented or purported to be for use by
``children less than 4 years of age'' or by ``children 3 or more
years of age'' are not subject to the special labeling
requirements of 101.9(j)(5)(i) but should fully declare required
information on fats and cholesterol.
    With regard to the comment that these foods are sometimes
used by older children or adults, FDA acknowledges that this
occurs. The agency believes, however, that the represented use of
the product must be the deciding factor. Inasmuch as the foods to
which 101.9(j)(5)(i) applies are represented to be for use by
infants and children less than 2 years of age, the agency
considers the use of these types of foods by children over 2
years of age or by older persons to be not particularly relevant
indetermining how these foods should be labeled. Accordingly, FDA
has not made any change to 101.9(j)(5)(i) in response to this
comment.
    204. One comment stated that the word ``or'' was used
ambiguously in the proposed version of 101.9(j)(5)(i) so that it
was not clear whether ``calories from saturated fat'' or
``saturated fat content'' was prohibited. The comment also
suggested that information on calories from saturated and
unsaturated fat and the amount of unsaturated fat also should be
prohibited, and that 101.9(j)(5)(i) should be clarified by
enumerating those parts of 101.9(c) that are affected.
    FDA agrees that all information relating to fatty acids
should be prohibited on foods represented or purported to be for
use by infants and toddlers (i.e., children) less than 2 years of
age. In the proposed version of 101.9(j)(5)(i), FDA only
specified the fatty acid component that is required in nutrition
labeling.
    Therefore, to make the suggested change and to clear up any
confusion, 101.9(j)(5)(i) is modified to state: ``* * * such
labeling shall not include calories from fat ((c)(1)(ii) of this
section), calories from saturated fat ((c)(1)(iii)), saturated
fat ((c)(2)(i)), polyunsaturated fat (c)(2)(ii)), monounsaturated
fat ((c)(2)(iii)), and cholesterol ((c)(3)).''
    205. One comment suggested the additional exclusion of fiber
on labels of foods for infants, citing a statement made by the
American Academy of Pediatrics that fiber probably is not needed
in infants less that 1 year old.
    FDA, in reviewing the reference cited by the comment (Ref.
115), noted that on the same page as this cited statement is the
additional statement that more work needs to be done before any
firm recommendations can be made on dietary fiber in pediatric
nutrition.
    FDA is thus not convinced that dietary fiber should be
excluded from nutrition labels for foods intended for infants and
children less than 2 years of age. Most foods included in Table 1
of 101.12(b) in the final rule entitled ``Food Labeling; Serving
Sizes,'' published elsewhere in this issue of the Federal
Register, contain less than 1 g of fiber per 100 g of edible
portion (Ref. 116). Under usual circumstances, these levels would
seem to preclude the consumption of high-fiber, low-calorie diets
by infants or children under 2 years of age who consume such
foods. Also, because dietary fiber has a natural laxative effect,
the label declaration of fiber content may be useful information
to the purchasers of these foods.
    206. A comment to the format proposal objected to the
inclusion of DRV's on foods for infants and toddlers because
DRV's were not proposed for infants or children less than 4 years
of age, and labels on jars of baby food are too small to allow
for thelts and children 4 or more years of age are not
appropriate for infants and toddlers, and that there could be
serious health consequences if a parent tried to adapt an
infant's diet to the proposed DRV's.
    FDA agrees with the comment for the reasons presented
therein. In addition, the agency believes that it is
inappropriate and unnecessary to include the caloric conversion
information required by 101.9(d)(10) on foods intended for
children less than 4 years of age because DRV's for this group
have not been established and calculation related to such values
may be misleading. Accordingly, for foods for infants and
children less than 4 years of age, the agency is adding an
exemption in 101.9(j)(5)(ii) that excludes the declaration of
Percent Daily Values for nutrients other than vitamins and
minerals for which there are RDI's specifically established for
infants and children less than 4 years of age. The exemption also
applies to the footnoteand caloric conversion information. Except
for the omission of this information, which is otherwise required
in 101.9(d)(2)(ii), and the footnote and caloric conversion
information required in (d)(9) and (d)(10), the format of the
nutrition labels on such products should comply with the
requirements of 101.9 (d), (e), or (f), as appropriate. Examples
of labels for foods for children less than 4 years of age and
less than 2 years of age are given in appendix G.

E. Medical Foods

    207. All comments received supported this exemption. In
addition, several comments expressed support for the agency's
intention, stated in the supplementary proposal (56 FR 60366 at
60377), to develop specific regulations for medical foods in the
near future. Some comments suggested that nutrition labeling,
intended for use by the general population, does not provide the
kind of information needed by health care professionals or
patients selecting or using medical foods. The comments noted
that, in light of this exemption, there is little guidance for
labeling of medical foods, other than general food labeling
regulations, citing the need for labeling of nutrient content and
purported uses and adequate and appropriate directions for use.
In addition to the need for specific labeling requirements, some
comments identified the need for quality control and good
manufacturing practices specific for medical foods.
    Section 403(q)(5)(iv) of the act exempts medical foods from
nutrition labeling requirements. The agency agrees with the
comments that the exemption for medical foods from nutrition
labeling is appropriate considering that these products are not
intended for use by the general population but rather are
intended for use under the supervision of a physician for
specific dietary management of a disease or condition. However,
the agency also recognizes that the exemption creates a void in
terms of specific labeling regulations suitable for these
products. FDA believes, as noted in some comments, that the
proper labeling of the nutrient content and purported uses of
medical foods, perhaps in a different manner or in more detail
than is required for other, more traditional foods, and adequate
and appropriate directions for use, as well as assurances of the
quality of medical food products, are all of vital public health
interest. While these issues are beyond the scope of this
rulemaking, the agencyintends to develop regulations covering
these aspects of medical foods in a future Federal Register
document.
    208. The comments support incorporation into the nutrition
labeling regulations the definition of medical foods from section
5(b) of the Orphan Drug Act (21 U.S.C. 360ee(b)(3)). Section
403(q)(5)(iv) of the act incorporated this definition by
reference into the statute, and FDA in proposed 101.9(j)(7) to
incorporate the statutory definition of ``medical food'' into the
nutrition labeling regulations. Some clarification of this
definition was included in the preamble and codified sections of
th some guidance in regard to the intended use of a medical food.
However, several comments cited particular products and asked
whether the products would be regulated as medical foods.
    FDA considers the statutory definition of medical foods, from
section 5(b) of the Orphan Drug Act, to delineate the principal
characteristics of medical foods. Additional clarification of
this definition, contained in the preamble of the proposal, gives
guidance on some of the types of products that the term ``medical
foods'' pertains to by identifying a variety of foods that the
agency regards as medical foods and some that are not presently
regarded as medical foods. Criteria that product must meet to be
considered a medical food are stated in the preamble of the
proposal, as well as in proposed101.9(j)(7), redesignated as
101.9(j)(8) in the final rule. FDA believes that this definition
and the information clarifying the definition in the proposal are
reasonable guides for use by industry in determining the
characteristics of a medical food at present.
    However, following review of the comments generated by this
proposal, FDA acknowledges that further clarification of the
types of products that are considered to be medical foods by the
agency would be helpful to manufacturers. While these commentsgo
beyond the scope of this rulemaking, the agency intends to
address this issue in a future Federal Register document.
    209. One comment suggested that in proposed101.9(j)(7)(v),
the words ``* * * provided only to a patient receiving active and
ongoing medical supervision * * *'' be changed to read ``* * *
intended only for a patient receiving active and ongoing medical
supervision * * *.'' The comment stated that manufacturers can
label products in a manner that gives a clear indication of the
intended level of supervision, but that the word ``provided'' in
this section might require a distribution system beyond the
control of the manufacturer, restrictingavailability of medical
foods to prescription status or distribution through an
institution.
    The agency agrees with this recommended change for the
reasons stated in the comment and has modified new
101.9(j)(8)(v) accordingly.
    210. One comment suggested that the word ``seeks'' in
proposed 101.9(j)(7)(v) be changed to ``require.'' The comment
noted that while some patients receiving a medical food under the
supervision of a physician are capable of seeking ``medical care
on a recurring basis,'' others receiving a medical food under the
supervision of a physician are not able to actually ``seek
medical care'' on their own (e.g., a comatose patient).
    FDA agrees with the suggested change. The agency acknowledges
that a medical food, under the supervision of a physician, may be
consumed by, or be administered enterally to, some patients
capable of seeking medical care and may be administered enterally
to other patients who may be too ill to actively seek medical
care. In both instances, the patient may require a medical food
for the specific dietary management of a disease or condition for
which distinctive nutritional requirements, based on recognized
scientific principles, are established by medical evaluation. FDA
has modified 101.9(j)(8)(v) accordingly.

F. Foods Shipped in Bulk Form

    211. FDA received many comments that supported
proposed101.9(j)(8) that exempts foods shipped in bulk form. A
few comments sought clarification of this exemption, requesting
that new 101.9(j)(8) include a statement that flavors and other
food ingredients (as opposed to processed foods) shipped in bulk
form from one manufacturer to another for use in the manufacture
of other foods are exempt.
    FDA intended the term ``processed'' in 101.9(j)(8)
(redesignated as 101.9(j)(9)) to indicate that food ingredients
used in the manufacture of other foods were exempt, maintaining
the scope of current 101.9(h)(8). However, for further clari
requested to state:

    Food products shipped in bulk form that are not for
distribution to consumers in such form and that are for use
solely in the manufacture of other foods or that are to be
processed, labeled, or repacked at a site other than where
originally processed or packed.

(Emphasis added).

G. Foods for Institutional Food Service Use

    212. Several comments objected to proposed 101.9(j)(9) that
would require manufacturers or distributors of foods for
institutional food service use (i.e., for use by hospitals,
schools, prisons) to provide nutrition information required by
this section directly to the institutions on a current basis. The
comments stated that this requirement was in conflict with
section 403(q)(5)(A)(i) of the act, which exempts food that is
sold for sale or use in restaurants or other establishments in
which food is served for immediate human consumption. The act
does not differentiate between food served in institutional and
noninstitutional settings. In fact, the comments pointed out that
the legislative history specifies that similar food service
establishments include cafeterias and hospitals.
    FDA agrees with the comments and has deleted proposed
101.9(j)(9) to bring the final rule into compliance with the
1990 amendments. To clarify that institutional food service
establishments are included under the exemption for restaurants
and other establishments, FDA has added them as examples in
101.9(j)(2)(ii).
    However, the agency finds merit in other comments that
supported nutrition labeling of foods sold to restaurants and
other food service establishments in order to enable food service
operators to become more aware of the nutritional content of
foodsthey serve, to offer more healthful menu options, and to use
more accurate descriptors on their menus. The agency, therefore,
encourages manufacturers, packers, and distributors to make
nutrition information available to food service operators
whenever possible.

H. Single-Ingredient Packaged Fish Products

    213. Comments received from the fish industry objected to the
inconsistencies between the voluntary nutrition labeling program
for raw fish and the mandatory nutrition labeling program. They
pointed to the potential for confusion when raw fish underthe
voluntary program are labeled on an ``as consumed'' (i.e., ``as
proposed'') basis, and the same fish, when frozen and packaged by
a manufacturer, are labeled on an ``as packaged'' basis. They
also pointed to the inconsistency with the USDA proposal that
allows single-ingredient raw meat and poultry items, whether
frozen or unfrozen, to be under a voluntary program with
nutrition information reported on either an ``as packaged'' or
``as consumed'' basis.
    FDA agrees that consumers may be confused to find
inconsistent nutrition labeling on two packages of the identical
fish (e.g., fillet of flounder) when one is under the voluntary
program for raw fish, and the other is under the mandatory
program forfrozen packaged fish. According to the final rule for
the voluntary program (November 27, 1991, 56 FR 60880; corrected
at, March 6, 1992, 57 FR 8174), nutrition information for raw
fish is to be reported for a 3 ounce, cooked edible portion (see
appendix B, 57 FR 8174 at 8175). The final rule on serving size,
published elsewhere in this issue of the Federal Register,
provides that under the mandatory nutrition labeling program,
nutrition information for frozen packaged fish is to be reported
for that amount required to prepare 85 g (approximately 3 ounces)
of cooked fish (101.12 (b) and (c)) but is to be based on the
product ``as packaged'' (101.9(b)(9)).
    To reduce the inconsistencies in nutrition labeling between
raw versus frozen packaged single-ingredient fish, and between
single-ingredient fish versus single-ingredient meat and poultry,
FDA is adding a special labeling provision for fish in 101dient
fish to be labeled on a cooked (i.e., ``as prepared'') basis
consistent with the voluntary program for fish and with USDA's
rules for single-ingredient meat and poultry products. Packaged
fish that contain added ingredients such as water, salt, or
additives such as sodium tripolyphosphate are considered
multi-ingredient processed packaged fish products and must
continue to be labeled on an ``as packaged'' basis.
    However, in the companion document on nutrient content claims
published elsewhere in this issue of the Federal Register, claims
such as ``lean,'' ``extra lean,'' and ``low fat'' are based on as
packaged values. Therefore, single ingredient packaged fish
products that make such claims must provide nutrition information
on an ``as packaged'' basis.

I. Raw Fish in Voluntary Nutrition Labeling Program

    214. One comment objected to the manner in which FDA defined
``raw'' for the purpose of determining what fish products are
covered by the exemption in proposed 101.9(j)(10) that subjects
the food to the voluntary nutrition labeling program. The comment
stated that ``absent a definition in the NLEA, the term ``raw''
means ``uncooked'' regardless of whether or not the product is
frozen and, therefore, packaged frozen raw fish should not be
subject to mandatory nutrition labeling.
    FDA discussed its interpretation of the word ``raw'' as it
pertains to fish in its proposed rule (July 2, 1991, 56 FR 30468
at 30470) and final rule (56 FR 60880 at 60886) implementing the
voluntary nutrition labeling program for raw fruit, vegetables,
and fish. Lacking legislative guidance, the agency chose to draw
a practical line in terms of retail selling practices and program
implementation rather than one based on a strict definition of
the term ``raw.'' While the agency included in the voluntary
program those fish that are generally sold raw (i.e., not heat
treated), it also included thermally processed shelled or
unshelled lobster, crab, and shrimp. The intent was to allow for
voluntary nutrition labeling of fish that are generally sold
refrigerated, or on ice, or are alive at purchase in fish stores
or in the fresh fish section of grocery stores and that are not
packaged or are packaged by the retailer or by a packer. These
are the types of products for which mandatory nutrition labeling
is most impractical. In contrast, providing nutrition labeling
for raw, frozen fish that are packaged by a manufacturer (usually
in a box with a printed label and brand name) and sold in the
frozen food case of a grocery store is no more difficult for a
manufacturer than providing nutrition labeling of other packaged
foods. Thus, these products appropriately come under the
mandatory nutrition labeling program.
    The agency has made a similar distinction with frozen
packaged raw fruit and vegetables and has received no comment on
it. It is likely that the greater concern on the part of the fish
industry was a result of the inconsistency between nutrient
values to be declared in the voluntary versus mandatory programs
(i.e., nutrient values based on ``as prepared'' versus ``as
packaged'' levels, respectively). The agency believes the
exemption in new 101.9(j)(11) should eliminate this concern.
Accordingly, FDA sees no need to amend its interpretation of the
term ``raw.''
    The agency would like to clarify, however, a
misinterpretation of the above definition of raw fish that
appeared in comments. FDA considers raw shellfish in or out of
the shell to be under the voluntary program, whether they are
sold bagged in plastic containers or displayed loosely in trays
or bowls. In addition, pasteurized crab meat that is not
shelf-stable and is sold on ice or refrigerated would be included
under the voluntary program, whereas canned pasteurized crab meat
that is shelf-stable wofacturing plant impractical. However, the
processing of the canned product is standardized, and nutrition
labeling can be easily accomplished and is required.

J. Meat Products Regulated by FDA

    215. Several comments recommended that nutrition labeling of
game meat should be on a voluntary rather than mandatory basis.
One game meat association stated that because buffalo is a red
meat, it should be exempt from FDA regulations and should be
allowed the option of voluntary labeling under USDA guidelines.
The comment also requested that any required nutrition
information should be allowed to be displayed at the point of
purchase to reduce costs associated with nutrition labeling.
    A number of comments expressed concern that the cost of
analytical testing and nutrition labeling would be prohibitive
for the small game meat producer. A request was made that an
economic impact study be conducted of the effect of the proposed
regulations on the buffalo industry before any final rule is
issued. Comments suggested small business exemptions for
producers marketing less than from 100,000 to 150,000 pounds per
year per each product label. A few comments also requested that
introductory test market, seasonal, short run, and experimental
products should be exempt from nutrition labeling.
    FDA is responsible for the regulation of all meats not
covered by USDA under the Federal Meat Inspection Act and the
Poultry Products Inspection Act (e.g., deer, bison, rabbit, wild
turkey, or ostrich, hereinafter identified as ``game meats'').
Therefore, the law does not provide an option for such products
to be covered by USDA guidelines. However, FDA appreciates the
fact that game meat producers have had little, if any, experience
with nutrition labeling, and that analytical data base
information is scarce. Accordingly, the agency will give game
meats as much latitude as possible under the 1990 amendments.
    Because many game meat producers are small enterprises, it is
possible that some will fall under the current small business
exemption. Many of those that do not may do so in the future if a
legislative amendment is passed to increase the exemption.
However, if an amendment is not forthcoming, all nonexempt
producers must provide the required nutrition information when
the regulations become effective.
    While the statute does not allow FDA to include raw game
meats under the voluntary nutrition labeling program for raw
fruit, vegetables, or fish, for consistency among all animal
flesh products, single-ingredient game meat products (frozen or
unfrozen, packaged or unpackaged) will also be included
in101.9(j)(11) that permits the information to be declared on
either an ``as purchased'' or ``as prepared'' basis (see comment
213 of this document).
    Also, in response to a comment, FDA is adding 101.9(j)(12)
to the final regulations to allow nutrition information to be
provided in accordance with paragraph (a)(2) of this section
which allows the required information to be placed on
labeling,that is on signs, posters, tags, or in binders or
booklets displayed at the point-of-purchase. FDA believes that
this action will allow game meat producers to give first priority
to nutrient analyses and data collection and to update nutrient
declarations more frequently than would be possible if the
information were printed on food labels.
    216. One comment requested the use of a data base to reduce
the cost of nutrition labeling for game meat. It was noted that
the nutrient composition of buffalo meat varies widely according
to whether the animal was grain fed or range fed and according to
age at slaughter. Another comment recommended that nutrient
information for buffalo meat come from actual sample testing and
not computer composites. The comment requested that FDA/USDA ``do
the same complete nutritional study for the buffalo industr
    FDA acknowledges that there is limited nutrient data
available on game meats. The agency advises that it does not
conduct nutrient analyses for any commodities; however, it is
willing to work cooperatively with game meat producers to produce
a validnutrient data base. To this end, the Agriculture Research
Service of USDA has experience in working collaboratively with
industry in developing food composition data (Ref. 117).
    217. Many game meat processors requested exemption from
nutrition labeling for custom services. Custom processed meat
includes wild game or domestic stock that is butchered to the
specifications of the customer. The meat may have been sold to
the customer or brought in by the customer for butchering.
Comments stated that because the customer owns the animal at the
time of butchering, the nutritional aspect of the meat product is
the responsibility of the customer.
    Consistent with similar regulations being issued
simultaneously by USDA for nutrition labeling of meat and poultry
products, FDA is exempting custom processed fish and game meats
from mandatory nutrition labeling. This exemption is found in new
101.9(j)(11)(ii). Legal authority for this is that what is being
sold is not the food but the processing. Thus, the food is not
subject to section 403(q) of the act.

K. Small Packages

    218. A number of comments supported the small package
exemption proposed in 101.9(j)(11). While a few comments
supported the provision that nutrition labeling be provided for
foods in small packages at the point of purchase in accordance
with paragraph 101.9(a)(2), many other comments objected to this
requirement. Several of these comments objected on the grounds
that the 1990 amendments did not include a requirement for point
of purchase disclosure for small packages, or that point of
purchase displays of nutrition information would create
``unnecessary clutter'' and ``place an undue burden on
retailers'' to find space for the information. One comment
stressed the economic impact the proposal would have on
supermarkets, especially those with front-end operations and
checkout lanes where a wide variety of small package items are
offered for sale. The comment stated that such areas would have
to be reconfigured with fewer items available because of space
lost to signage and fewer inventorychanges made throughout the
year. A comment raised a question about who would be held
responsible if the information was not available at the point of
purchase. Comments recommended that manufacturers, not retailers,
should be responsible for nutrition information on all packaged
foods. A suggestion was also made that interested consumers could
refer to larger retail packages of the same product or could
write or call the manufacturer for the nutrition information by
using an address or telephone number given on the package label.
    FDA is persuaded by the comments that it is impracticable to
require point of purchase display of nutrition information for
small packages. However, because section 403 (q)(5)(B) of the act
states only that the nutrition labeling requirements shall not
apply to the label of the food, not the labeling as is included
in section 403(q)(5)(C) and (q)(5)(D), the agency concludes that
nutrition information about food in small packages must be
provided to consumers through alternative means. The agency
agrees with the comments that manufacturers should bear the
responsibility for nutrition labeling of packaged foods and finds
merit in the suggestion that manufacturers provide an address or
telephone number on the package for consumers to write or call
for nutrition information. FDA believes that almost all small
packages should be able to add a short phrase, such as ``For
nutrition information, call 1-800-123-4567'' to the label. In
fact, many packages currently give an address or telephone number
for con
    Therefore, FDA has modified 101.9(j)(11), redesignated as
101.9(j)(13)(i), to delete the requirement that foods in small
packages that bear no nutrition claims or other nutrition
information provide the required nutrition labeling in accordance
with 101.9(a)(2). The agency replaced it with a requirement that
the manufacturer clearly state on the package label where a
consumer may write or call to obtain the required nutrition
information. If a manufacturer finds that it is impracticableto
comply with even this requirement on a particular product, the
manufacturer should write to the agency in accordance with
101.9(g)(9) (see section VI.P.3. of this document).
    219. A few comments from health professional organizations
expressed the belief that the 12 square inch standard for ``small
packages'' was too large, and that consumers should have as much
information as possible about what they purchase and consume. One
comment stated that ``with the increase in fabricated foods and
single serving size packaging, [they were] convinced that
nutrition information can and should go on less space,'' adding
that by using an abbreviated format, nutrition labeling is
possible on smaller packages, down to 8 square inches.
    However, several other comments objected to the 12 square
inch definition for small packages, stating that it would not
allow enough space for all the required information on the label,
especially on a product with a lengthy ingredient list. The
comments stated that the 12 square inch standard for exempting
small packages was established years ago when much less
information was required on food labels (i.e., before mandatory
nutrition labeling). The comments also expressed concern that
attempting to include all of the required information in 12
square inches would result in a label that would not be legible,
making it difficult for sight-impaired or elderly persons to
read. Comments also said that such a presentation would
discourage use of the nutrition information, thereby undermining
the purposes of the 1990 amendments.
    Two manufacturers commented on the unique space problems
arising when more than one language is used on small packages
inasmuch as 101.15(c)(2) requires that if a language other than
English is used, all information on the label must be printed in
both English and the other language. One comment pointed to the
fact that the United States has become an increasingly bilingual
nation, making Spanish-language labeling a ``necessity in many
parts of the country.''
    Several comments requested a more flexible rule based on
``practically available space'' or ``usable surface space'' on
labels. One comment stated that the term ``surface area available
to bear labeling'' is newly coined and unfamiliar and likely tobe
confusing. The comment recommended that the exemption be couched
in terms of ``total square inches on the information and
alternate panels,'' which are familiar terms to manufacturers.
    Other suggestions included: (1) Using a 20 square inch
surface area, (2) excluding the principal display panel from the
12 square inch requirement, (3) excluding odd shaped parts of
packages from the total surface area available for nutritional
labeling, (4) allowing a linear (i.e., string) format for the
nutrition information, (5) making the nutrition profile optional,
(6) allowing for abbreviations of nutrients, (7) deleting the
requirement for declaration of ``Servings per container'' on
single-serving containers, and (8) allowing required nutrition
information to appear anywhere on the package expected to be read
by consumers rather than just on the information panel as
required by 101.2. In regard to the latter comment, one comment
suggested that 101.2 be modified to allow required information
to be placed on other label panels adjoining the principal
display panel or the information panel when there was
insufficient space on a single panel.
    A few comments stated that no manufacturer should be required
to change its existing label style or container size to
accommodate nutrition labeling. The comments urged that areas of
a package not traditionally used for labeling should be excluded
frosurface area (e.g., many companies do not use lids of jars,
necks of bottles, or bottoms of cans for labels). One comment
recommended that current company practices be grandfathered until
the company changes its packaging or container.
    The agency received additional comments regarding small
package limitations in response to the format proposal. Several
comments from manufacturers of smaller size products such as
candy rolls and bars, chewing gum, canned fish, and cookies
stated that such labels could accommodate only the CONTROL
format. Two comments suggested raising the minimum 12 square inch
requirement for ``small packages'' to 13 square inches.
    A number of comments addressed the inclusion of the DRV's on
the labels of small packages. These comments apply to inclusion
of the footnote providing calorie-specific recommended nutrient
amount information specified in 101.9(d)(8)(i).
    The majority of comments asserted that it would be difficult
to accommodate the DRV's without a relaxation of the minimum
requirement of 12 square inches of printable label space. Most of
those seeking relief suggested the option of listing DRV's
inlinear rather than column array over an intermediate range of
printable package area. Alternate upper limits suggested were 20
and 26 square inches or no more than 30 percent of printable
package area devoted to the nutrition label. One manufacturer
provided support for 20 square inches as a minimum area below
which DRV's could not be accommodated without violating minimum
type size or principal display panel size requirements. It
submitted executions of the proposed and alternate formats for
several existing products. One comment suggested several
principles to be followed by FDA in establishing a range within
which the DRV listing could be modified or deleted while
preserving legibility and remaining in conformance with existing
labeling requirements concerning type size and area devoted to
the principal display panel.
    FDA acknowledges the need to give consumers as much
information as possible. The agency is persuaded, however, that
with requirements for more nutrition-related information, it may
be difficult to get all of the required information on packages
thatjust meet or slightly exceed 12 square inches of surface area
available to bear labeling, particularly for products that do not
qualify for the simplified format. However, in light of the
exemption from nutrition labeling on the package label for
products with less than this amount of usable surface space,
provided that no nutrition claim is made (see preceding comment),
the agency believes that exempting a larger number of foods by
increasing the definition of ``small package'' size would
undermine the intent of the 1990 amendments.
    However, based on the comments, FDA has concluded that
justification exists for developing a graduated system that would
allow added flexibility for foods in an intermediate package size
group. To select the dimensions of such an intermediate sized
package, FDA reviewed comment suggestions, examined the space
requirements of the required label with the calorie-specific
daily values, and reviewed data on available label area for a
sample of packaged foods (Ref. 117a). The agency is rejecting
suggestions such as the use of only ``practically available
space'' or ``usable surface space'' or the exclusion of ``oddly
shaped parts of packages'' because there is a significant
potential for differences of opinion about what is ``practically
available,'' ``usable,'' or ``oddly shaped.'' The remaining
suggestions are to exclude the principal display panel, to use an
upper limit of 20 or 26 square inches of surface area available
to bear labeling, or to require that no more than 30 percent of
the surface area available to bear labeling be devoted to the
information panel.
    The agency believes that the suggestion to apply the 30
percent criterion to space requirements necessary to comply with
FDA regulations has merit. Based on current requirements (see
101.1(b) and (c)), the principal display panel can be considered
to cover 40 percent of the total surface area available to bear
labeling. On the assumption that no more than half of the
remaining 60 percent of the label should be required to be
devoted to FDA-required information (i.e., the nutrition label
and ingredient list), 30 percent of the total surface area would
be used for such information. This is consistent with the
comment.
    Based on the data examined, FDA believes an upper limit of 40
square inches of surface available to bear labeling is
appropriate to define an intermediate sized package. The smallest
legal sized execution of the format required under 101.9(d)
isapproximately 7 square inches. For many processed foods, the
addition of the ingredient list could bring the space needed for
presenting this FDA-required information to 11 square inches.
Using the 30 percent factor, this information could be
accommodated on packages with 37 square inches available to bear
labeling. In order to provide incentive to allow sufficient space
to make the label readily observable and easily comprehensible,
the agency has decided to round this number up to 40 square
inches. The agency is providing for this upper limit in
101.9(j)(13)(ii).
    The agency does not agree with the comment that the term
``surface area available to bear labeling'' is newly coined and
unfamiliar, inasmuch as it has been used in 101.2(c)(3)(i) for
17 years.
    FDA looked to the comments for suggestions of added
flexibility for the labeling of foods in intermediate sized
packages available to bear labeling. Suggestions in the comments
included: Allowing a linear (i.e., string) format for nutrition
information (including the DRV listing), making DRV's optional,
allowing for abbreviations, deleting the requirement for
declaration of ``Servings per container'' on single-serving
containers, and allowing required nutrition information to appear
in other placesthan those required by 101.2 (i.e., the
information panel). Dependent upon the circumstances of a
particular package size and shape, the agency is not opposed to
the use of any of these suggested methods. In addition, as
provided for in 101.9(g)(9), manufacturers may request special
allowances for provision of the required information on tags
affixed to the product according to 101.9(a)(2) as discussed in
section VI.P.3. of this document, Foods For Which Labeling Is
Impracticable.
    In regard to the request to delete the requirement for
declaration of ``Servings per container'' on single serving
containers, FDA finds that inasmuch as the declaration of
``Serving size'' on such products will specify that the serving
is the entire unit (e.g., 1 can or 1 bar), it would be needlessly
repetitive to state that there is one serving per container.
Therefore, FDA has modified 101.9(d)(3)(ii) that pertains to all
container sizes to state that ``Servings per container'' is not
required on single serving containers as defined in 101.9(b)(6).
    While the provisions being made to increase flexibility are
for the purpose of making it easier for manufacturers to place
mandatory nutrition labeling on packages of an intermediate size,
they may also be used on ``small packages'' (i.e., packages with
less than 12 square inches of surface area available to bear
labeling) whose labels are exempt under 101.9(j)(13)(i) when
manufacturers elect to provide a nutrition label on those foods.
    FDA is providing in 101.9(j)(13)(ii)(A) that the required
nutrition information may be presented in a tabular fashion when
the package shape or size cannot accommodate a column display on
any label panel. This form of presentation is currently used on
many foods in long rectangular or round packages, such as candy
bars and is shown in Appendix H. In addition, to facilitate the
provision of information on small packages, 101.9(j)(13)(ii)(A)
provides for the use of a tabular presentation on allble to bear
labeling, regardless of the package shape. Further, if the label
will not accommodate a tabular display, 101.9(j)(13)(ii)(A) also
provides that the required nutrition information may be presented
in a linear (i.e., string) fashion.
    In regard to abbreviations, one comment stated that the
design limits of their company's printers for labels to be
affixed to foods packaged in retail stores limited the
description of nutrients to 10 characters. While the agency is
concerned about the use of abbreviations and any possible
consumer confusion they may cause, FDA believes their use under
limited and controlled conditions may be preferable to
overcrowding within the nutrition label. Therefore, based on this
comment, the agency is providing the following abbreviations in
101.9(j)(13)(ii)(B) for those mandatory nutrients whose names
exceed 10 characters.
    Serving size: Serv. size
    Servings per container: Servings
    Calories from fat: Fat cal
    Saturated fat: Sat fat
    Cholesterol: Cholest
    Total carbohydrate: Total carb
    Dietary fiber: Fiber

Section 101.9(d)(9)(iv) allows these abbreviations to also be
used in a footnote within the nutrition label.
    As discussed above in section V. of this document on the
format of the nutrition label, FDA is providing in
101.9(j)(13)(ii)(C) that the footnote and caloric conversion
information required in 101.9(d)(9) and (d)(10) may be omitted
on intermediate sized packages. When the footnote required by
101.9(d)(8) is omitted, an alternate footnote must be used that
states: ``Percent Daily Values are based on a 2,000 calorie
diet.''
    The agency believes that concerns expressed in comments
requesting that the nutrition information be allowed to appear
elsewhere on the package rather than just on the information
panel as required by 101.2 (see 101.9(i)) are generally
addressed by 101.2(a)(1). This section states that if the
information panel is too small to accommodate the necessary
information or is otherwise unusable label space, e.g., folded
flaps or can ends, the panel adjoining to the right may be used.
However, in recognition of the increased need for this
flexibility in packages with less than 40 square inches available
to bear labeling, FDA is providing in 101.9(j)(13)(ii)(D) that
nutrition labeling on intermediate sized packages may appear on
any label panel.
    As a conforming change, 101.9(c), (d), and (i) have been
modified to reflect the provisions of 101.9(j)(13).
    In regard to the comments requesting an exemption or
postponement based on current company labeling practices, FDA
advises that Congress did not provide in the 1990 amendments for
any such actions. The agency recognizes the possible economic
burdensassociated with changing labeling practices and has tried
to incorporate sufficient flexibility to minimize the need for
such changes but has no authority to prevent them. FDA advises
that in 101.1 the agency stated that, in determining the area of
the principal display panel, tops, bottoms, flanges at tops and
bottoms of cans, and shoulders and necks of bottles and jars were
to be excluded. Therefore, it is reasonable to conclude that the
agency will not include these areas in determining the ``surface
area available to bear labeling.''
    220. A comment requested clarification as to whether
manufacturers of products that are sold in small packages that
qualify for the small package exemption are required to omit
nutrition information from the label and then present it through
other means as required in proposed 101.9(a)(2), or whether they
may attempt to provide the nutrition labeling.
    While comment 218 of this document addressed the underlying
concern in this comment about the mandatory inclusion of required
nutrition information in labeling at the point of purchase, FDA
does not view this or any other exemption under 101.9(j) as
prohibiting a manufacturer from including nutrition labeling on
the label of a food product. The agency encourages the inclusion
of nutrition information on the label of exempted products
whenever possible. To clarify the situation, 101.9(j)(13)(i) has
been modified to state that the new requirement for an address or
telephone number for consumer use in obtaining nutrition
information is to apply to products that qualify and use this
exemption.

L. Shell Eggs

    221. One manufacturer commented on the labeling of egg
cartons, stating that proposed 101.9(j)(12) allowing for the
presentation of the required nutrition information immediately
beneath the carton lid is as impractical for many egg cartons as
requiring its display on the upper surface of the lid because
both surfaces conform to the shape of the eggs. The comment
suggested that packaging of this kind may not be readily
imprinted at all. The comment further stated that eggs are a
largely homogeneous agricultural commodity, and eggs sold at
retail in their shells should all be treated alike with respect
to nutrition labeling, whether the eggs are in bulk, on trays
without cartons, or in cartons. The comment requested that eggs
be exempt under 21U.S.C. 343(q)(3) that allows the Secretary to
provide that nutrition labeling be displayed at the point of
purchase for foods received in bulk containers.
    FDA is persuaded by the comment that it may be impractical
for egg cartons that conform to the shape of the eggs to bear
nutrition labeling. Accordingly, FDA is modifying 101.9(j)(12)
(redesignated as 101.9(j)(14)) to allow the required nutrition
information to appear on the inside or the outside of the carton,
or on an insert that can be clearly seen when the carton is
opened. By doing this, FDA is greatly expanding the total surface
area available to bear labeling.
    FDA rejects the suggestion that, because some eggs are sold
in bulk, all eggs should be allowed to be labeled at the point of
purchase according to the exemption for bulk foods
(101.9(j)(9)). As discussed above, nutrition labeling for eggs
may appear on the egg carton or on a package insert. FDA
concludes that there is no need to modify 101.9(j)(14) to allow
for further special conditions for shell eggs packed in cartons.
If, in fact, a manufacturer finds it impossible to label a
particularegg carton or to include a package insert, it may
request a special allowance from the agency, as discussed in
comment 223 of this document.

M. Multi Unit Packages

    222. A few comments disagreed with the requirement in
proposed 101.9(j)(13)(iii) that each unit within a multipack
state ``this unit not labeled for retail sale.'' Comments stated
that this requirement is redundant because 101.9(j)(13)(i) and
(j)(13)(ii) adequately prevent the product from being sold
without nutrition labeling.
    The agency does not agree that the requirement is redundant.
Although multiunit containers may be enclosed and are not
intended to be separated from the retail package under normal
conditions of sale, occasionally the individual units are
separated from the multiunit container and purchased separately.
Proposed 101.9(j)(13)(i) and (j)(13)(ii), redesignated
as101.9(j)(15)(i) and (j)(15)(ii), state: ``The multiunit
retail food package labeling contains all nutrition information
in accordance with this section;'' and ``The unit containers are
securely enclosed within and not intended to be separated from
the retail package under conditions of retail sale.'' These
sections cannot guarantee that the units in a multiunit package
will not be separated; e.g., frozen juice bars, soft drink
bottles, and sticks of butter are sometimes separated from an
enclosed multiunit package by consumers prior to purchase at the
retail level. Therefore, FDA is not modifying the regulation.
    223. A soft drink trade association requested a provision in
the final rule to exempt from nutrition labeling glass bottles
with lithographed labeling that are marketed in multi-unit
packages. These bottles, the comment pointed out, are often
loosely packed rather than securely enclosed. The comment made
reference to the technical limitations of labeling glass by the
lithograph method, and the impracticality of placing nutrition
labeling on the individual bottles or ``unit containers.'' The
commentrequested that the agency clarify the proposal to ensure
the continued availability of lithographed bottle multiunit
packages and suggested that the nutrition labeling information
appear on the information panels of the multiunit retail package.
    The agency acknowledges that there will be some circumstances
in which strict adherence to the regulations (in this case the
requirement that units be securely enclosed in the retail
package) is not technologically feasible, or some other
circumstance makes it impracticable. Proposed 101.9(g)(8) would
have allowed for alternative means of compliance or additional
exemptions to deal with the situation when firms were unable to
develop adequate nutrient profiles. The agency concludes based on
this comment that this latitude should be available for
additional circumstances. Accordingly, FDA is modifying
101.9(g)(8), redesignated as 101.9(g)(9), to broaden its scope
by stating ``When it is not technologically feasible, or some
other circumstance makes it impracticable, for firms to comply
with the requirements of this section (e.g., to develop adequate
nutrient profiles to comply with paragraph (c)), * * *.''
    Additionally, FDA believes that actions taken to address
technological or other problems on a case-by-case basis do not
need to be established by regulation in response to a petition to
initiate rulemaking. Therefore, the agency is replacing
``establish by regulation'' with ``permit'' in 101.9(g)(9) and
is deleting the reference to a petition, stating instead that
firms in need of such special allowances shall make their request
in writing to the Food and Drug Administration, Office of
Nutrition and Food Sciences (HFF-200), 200 C St., SW.,
Washington, DC 20204. However, FDA concludes that no change is
necessary in 101.9(j)(15) in response to this comment.

N. Foods Sold from Bulk Containers

    224. A food retailer wrote in support of the requirement in
proposed 101.9(j)(14) that nutrition labeling information for
bulk foods be provided at the point of purchase. However, the
comment took exception to the agency's intention to include
within the requirement individually wrapped bulk food items such
as candies, arguing that the exemption for small packages should
apply to small individually wrapped food items that are sold in
bulk.
    FDA disagrees with this comment. The labels of individually
wrapped small food items, such as bite size pieces of candy, are
exempt from nutrition labeling under the small package exemption
(101.9(j)(13)) because of the lack of space needed to print the
required information. However, under section 403(q) of the act,
foods sold from bulk containers must be nutrition labeled whether
or not they are individually wrapped. Nutrition labeling can, and
should, be presented on the labeling of the bulk container or on
a counter card, sign, or other appropriate device as identified
in 101.100(a)(2). Moreover, as discussed above, the exemption
for small packages only applies to the label and not to a
product's labeling. The agency reiterates its position as stated
in the Federal Register of July 19, 1990 55 FR 29487 at 29505,
and 56 FR 60366 at 60379:

* * * Many foods, such as candies, cookies, and pasta, are
offered for sale from large containers such as barrels or bins.
FDA has traditionally required that these foods be labeled in
accordance with section 403(i)(2) of the act through the use of a
    Accordingly, no changes are being made to 101.9(j)(14),
redesignated as 101.9(j)(16).
    225. Several other comments were received in support of the
requirement in the proposed version of 101.9(j)(16) that
nutrition labeling information for bulk foods be provided at the
point of purchase. Two comments recommended that nutrition
information be provided in the form of brochures or ``tear-off''
sheets at the point of purchase, so that consumers can have the
information available at home.
    FDA agrees that tear-off sheets or brochures with the
required nutrition information would be useful to consumers and
encourages manufacturers to provide retailers with the required
nutrition information in such form. Section 403(q)(3) of the act
states: ``For food that is received in bulk containers at a
retail establishment, the Secretary may, by regulation, provide
that the nutrition information required * * * be displayed at the
point of sale.'' Thus, the statute does not specify the form in
which this information is to be provided. Accordingly, FDA has
not made the recommended change.
    226. A retail ice cream manufacturer requested that the
proposed version of 101.9(j)(16) be clarified so that scoops of
ice cream that are dispensed by store employees from bulk ice
cream containers are clearly not subject to the ``sold from bulk
containers'' requirement.
    FDA advises that it is not necessary to exempt ice cream from
the requirements of 101.9(j)(16). Ice cream that is dispensed by
store employees from bulk ice cream containers at an ice cream
store is for immediate human consumption and would therefore be
exempt from mandatory nutrition labeling under revised
101.9(j)(2)(ii).
    227. A retail grocery chain stated that popular bulk food
items sold from bins and barrels but packaged by clerks for
customer convenience should not be required to have nutrition
labeling on each package.
    FDA advises that 101.9(j)(16) allows food products sold from
bulk food containers to display the required nutrition
information ``either on the labeling of the bulk container
plainly in view or in accordance with provisions of paragraph
(a)(2) of this section.'' Section 101.9(a)(2) allows use of
counter cards, signs, tags affixed to the product, or other
appropriate devices. Accordingly, the containers such foods are
put into when sold to the consumer need not bear nutrition
labeling as long asthe required nutrition information is plainly
in view, regardless of whether it is the consumer or a store
employee that packages the product. However, if the foods are
packaged in an area that is off-limits to customers, and the
information is not plainly in view, the required nutrition
information must be available on the package label or in labeling
adjacent to the packages accordingly to the provisions of
101.9(a)(2).

O. Foods Used as the Sole Item of the Diet

    228. One professional organization and one consumer interest
group wrote in support of FDA's tentative decision to delete the
exemption in current 101.9(h)(3) for foods promoted as the sole
item in a diet (such as formulated weight-loss products) and to
have the same labeling requirements for those products as all
other foods. The consumer interest group stated that
``considering the minimal long-term benefit from these products
and the potential for harm from the unsupervised use of these
products, FDA should consider greater labeling requirements for
these products.''
    FDA intends to monitor the use and labeling of foods used as
the sole item of the diet and, as discussed in the supplementary
proposal (56 FR 60366 at 60378), will consider at a later date
whether there should be additional or different requirementsfor
the nutrition labeling of these products.

P. Other Requests for Exemption

1. Donated Foods
    229. Two comments from food banks requested an exemption from
mandatory nutrition labeling, citing that food banks are
nonprofit charitable organizations, and as such, it would be
``unreasonably costly and unduly burdensome for (food banks) to
be required to apply complete nutrition labeling to repacked food
products.'' The comments stated that the exemption is necessary
to ensure that mandatory nutrition labeling rules do not hamper
the ability of charitable organizations to receive and
distributefoods to needy individuals.
    Section 403(q)(1) of the act requires nutrition labeling on
food that ``is intended for human consumption and is offered for
sale.'' Accordingly, donated foods that are given without charge
to the ultimate consumer are not subject to mandatory nutrition
labeling. This provision of the 1990 amendments was not included
in the proposed implementing regulations. To correct this
omission, the agency is modifying 101.9(a) to state that
``Nutrition information relating to food shall be provided for
all products intended for human consumption and offered for sale
* * *.''
    230. A second request from these foods banks was that food
companies having inventories of foods not in compliance with new
labeling rules after the effective date of section 403(q) of the
act be permitted to donate those products to charitable
organizations.
    Section 10(a)(2) of the 1990 amendments states that the new
nutrition labeling requirements shall not apply to foods labeled
before the effective date. Therefore, companies will be able to
continue to sell all foods that are labeled in compliance with
current regulations before the effective date of section 403(q)
of the act, May 8, 1993. As a result, there should be no
inventories of labeled food that cannot be sold to consumers. The
agency wishes to state, however, that it has long been the
agency's policy that misbranded foods, such as those that have
been the subject of a seizure or recall, can be donated to
charitable organizations rather than being destroyed if they do
not present a safety concern, and the recipient is fully informed
as to the problem with the food (e.g., short weight).
    231. Two comments expressed concern that if donated foods are
exempted from nutrition labeling, the goals of nutrition labeling
will not be met for individuals who rely on such foods.
    In passing the 1990 amendments, Congress intended to require
that consumers have the necessary information at their disposal
to select diets that are consistent with dietary recommendations
aimed at improving the health status of Americans. However,by
requiring nutrition labeling only on foods offered for sale,
Congress limited the coverage of the nutrition labeling
requirements. Therefore, while the agency would encourage
nutrition labeling on any foods repackaged or relabeled by
charitable organizations, the statute does not require such
labeling.
    The agency is pleased to note, however, that in conversations
with the Food and Nutrition Service, USDA, which administers the
Food Distribution Program, the Food and Nutrition Service has
stated that it plans to incorporate nutrition labeling on all
foods that it distributes to individuals. The Food Distribution
Program purchases surplus foods from American markets and
distributes them to State agencies for further distribution to
individuals and eligible local outlets.
2. Exported Foods
    232. Comments from a trade association and a manufacturer
requested that products intended for export be exempt from U.S.
nutrition labeling regulations because they will necessarily be
required to comply with the importing country's labeling
criteria.
    FDA advises that under section 801(e) of the act, foods
intended for export will not be deemed misbranded under section
403 of the act under certain circumstances. Section 801(e) states
that:

A food, drug, device or cosmetic intended for export shall not be
deemed to be adulterated or misbranded under this Act if it:
    (A) Accords to the specifications of the foreign purchaser,
    (B) Is not in conflict with the laws of the country to which
it is intended for export,
    (C) Is labeled on the outside of the shipping package that it
is intended for export, and
    (D) Is not sold or offered for sale in domestic commerce.

Thus, if a company complies with the requirements of section
801(e) of the act, it need not be concerned about misbranding the
food by failing to comply with section 403(q) of the act.
3. Foods for Which Labeling Is Impracticable
    233. Two dairy companies requested that returnable glass milk
bottles be exempt from nutrition labeling because the total
surface area available for labeling is much less than 12 square
inches. The labeling surface is the closure on the top of the
bottle. If the label were placed on the side of the bottle it
would be impossible to recycle the bottle for milk use because of
problems with washing and disinfecting the bottle after each use.
The comments stated that the returnable glass bottle is important
for the environment, and that many of their customers purchase it
for that reason. They suggested the nutrition labeling for milk
in returnable glass bottles be placed on placards at the point of
purchase.
    Other comments requested special allowances for uniquely
shaped package containers (such as containers of honey in the
shape of a bear, individual juice containers in the shape of a
hand grenade, or cheese balls) or packaging materials that do not
allow for fine printing (e.g., styrofoam ice cream cups).
    FDA is willing to consider allowing the required nutrition
information for returnable glass milk bottles to be available in
labeling, as provided for in 101.9(a)(2). As discussed in
comment 223 of this document, 101.9(g)(9) of this final rule
allows that when it is not technologically feasible, or some
other circumstance makes it impracticable, for firms to comply
with the requirements of nutrition labeling, FDA may permit
alternative means of compliance or additional exemptions to deal
with the situation.
4. Foods Purchased Under Government Contract
    234. One trade association and one manufacturer requested
exemptions for products produced for Government contracts (e.g.,
the National School Lunch Program, military feeding operations),
using the reasoning that any products sold would be offered to
the final consumer as part of a total meal/diet, and nutrition
information on the meal must be supplied by the facility offering
the meal.
    FDA advises that products of the type discussed in the
comment that are sold for use in restaurants and institutional
food service operations are exempt under 101.9(j)(2)(iii). As
long as it is not reasonably possible that they will be sold
directly to consumers, they need not be nutrition labeled.
Therefore, no further exemption is necessary.

VII. Other Issues

A. Assortments of Food

    235. A few comments requested clarification on whether
assortments of foods, such as a box of assorted chocolates or
nuts, would have to bear nutrient information on each type of
chocolate (or nuts), or whether an average nutrient value would
suffice.
    The agency advises that in the preamble to the mandatory
nutrition labeling proposal (55 FR 29487 at 29505), it stated
that ``where assortments of food are packaged, firms will be
required to express nutrient content based on the package as a
whole (e.g., the entire product contents may be combined for a
nutrient analysis).'' FDA recognizes that the terms ``will be
required'' and ``may be combined'' appear inconsistent.
Therefore, to clarify the regulation and in accordance with the
agency's intentto offer flexibility in the labeling of
assortments of foods, FDA has modified 101.9(e)(1), recodified
as 101.9(h)(1), by deleting ``of the total product'' and adding
a new sentence that states that when separately packaged
ingredients or assortments of the same type of foods are intended
to be eaten at the same time, the nutrition information may be
specified for each component or as a composite value. In
developing a composite nutrient value, the entire product
contents would be combined for anutrient analysis.
    In addition, to clarify the term ``assortments of food'' FDA
has modified 101.9(h)(1) by adding ``assortments of the same
type of food'' and including the example of assorted nuts.
    236. A few comments addressed the labeling of variety packs
containing an assortment of individually packaged products (e.g.,
assorted ready-to-eat breakfast cereals or snack foods such as
corn chips, cheese puffs, and potato chips). A food manufacturer
marketing variety packages stated that they currently label each
of the single-serving packages placed in a multi-serving
container separately. The comment stated that the outer wrapping
is generally transparent, making extensive labeling on the outer
wrapping infeasible. Another comment suggested that the outer
label contain the statement ``Individual inner units carry
nutrition information'' where each of the single-serving packages
in the variety pack bears nutrition labeling. The comment also
stated that larger sizes of the individual packages of foods in
the variety pack are invariably available to consumers at the
same location, and the nutrition labels on those larger packages
may be reviewed if desired.
    FDA points out that a primary purpose of the 1990 amendments
is to allow consumers to maintain healthy dietary practices. To
do this, consumers must have access to nutrition information at
the point of purchase. In many situations, consumers can look at
the nutrition labels of larger packages of the individual foods
for nutrition information. However, the agency does not agree
that it is always possible to do so.
    With respect to the transparent nature of the outer wrapping,
FDA does not believe this makes labeling on that wrapping
infeasible. Many bakery products are packaged in transparent
wrappers and these products provide nutrition and other label
information.
    Inasmuch as many variety packs are currently printing the
required nutrition information for each of the products contained
in the variety pack in a table on the outer package, and because
the outer packages are generally large, the agency concludesthat
a special allowance is not required for variety packs.
Accordingly, FDA rejects the suggestion that the outer label
merely state that the individual units within the package provide
nutrition information.
    However, the agency has no objection to manufacturers
labeling only the individual inner packages if the information is
provided in such a way that consumers can clearly see it at the
time of purchase. Examples of this type of packaging can be
foundcurrently in the marketplace where nutrition labeling is
provided on the tops of single-serving packages of breakfast
cereals. Accordingly, FDA is adding a new paragraph 101.9(h)(2)
to specify that nutrition labeling of single-serving packages
within variety packs must be clearly visible at the point of
purchase. Proposed 101.9(e)(2) is redesignated as 101.9(h)(4).
    237. FDA received comments from companies that sell food
products by mail order, particularly varieties of foods and food
assortments that are marketed as gifts. The comments requested
special provisions in the regulations to provide some flexibility
for packaged gift assortments because these packages are
assembled from several thousand separately labeled food items,
many of which are similar, differing only in size or flavor, and
which are used in many different assortments. Because of the
uniquecharacteristics of the mail order gift food industry,
caused, in part, by rapidly changing selections of gift packages
offered, the comments contended that nutrition labeling would
have a devastating effect on the industry, unless alternative
means of compliance are allowed.
    The comments requested that a new paragraph be added under
proposed 101.9(e) for assortments of foods intended to be used
as gifts, allowing for nutrition information on such foods to:
(1) Be included on labeling, (2) be based on uniform serving s,''
(4) be calculated as averages for categories of foods having
similar dietary uses or similar significant nutritional
characteristics for characterizing nutrients, (5) be based on
calculations from nutrient data bases, and (6) omit foods meeting
the definition of ``small package'' in 101.9(j)(13) from
determinations of nutrient content. A subsequent comment on
behalf of the mail order gift food companies modified the last
provision to state that foods in small packages only be omitted
if they are not listed in promotional catalogues and are
``optical garnishes'' used to enhance the appearance of the gift
package, or bonus items included as a free gift or promotional
item.
    FDA is persuaded that special allowances are justified for
gift packages containing a variety of foods (e.g., cheese, jams,
and crackers packed together in one gift box) or of food
assortments (e.g., several different types of jam in one box).
Accordingly, the agency is adding a new paragraph 101.9(h)(3) to
address gift packages.
    New 101.9(h)(3)(i) allows the required nutrition information
to appear on the label or in labeling that is within or attached
to the outer gift package. This provision allows the information
to be consolidated in a single document that could accompany
several different gift food packages that contain the same
assortment of foods, although not necessarily in consistent size
packages, as are identified in the document. This action is in
recognition of the fact that the person who buys the gift package
is generally not the person who will use the information.
According to the comments, on average, 65 percent of company
sales are shipped to recipients other than the purchaser.
Moreover, many packages shipped to purchasers are subsequently
offered as gifts to other persons.
    The ``outer package'' is intended to mean the container
directly within which component items are packed. It does not
mean the shipping carton, unless component items are packed
directly within the shipping carton instead of being packed in a
separate inner container.
     Comments also have persuaded the agency that standardizing
the serving sizes for foods included in gift packages will
simplify the simultaneous presentation of information on a
variety of different types of foods by putting the information
for all products on a comparable weight basis and, thereby,
increase the likelihood that consumers will use and understand
the information. The comments requested that, where there is no
uniform household measure that is either a common multiple or
fraction of the quantity of an individual food in an assortment,
one ounce (fluid or solid as appropriate) be used as the standard
serving size. Rather than leaving open the possibility of the use
of any ``uniform household measure,'' however, FDA believes that
an allowable exemption from the serving size requirements would
be permissible only when all of the foods in a particular gift
package are not subject to the same reference amount customarily
consumed, as specified in 101.12(b).
    FDA has no objection to the suggestion of a one ounce serving
size for solid foods in such circumstances. The selection of one
ounce is acceptable based on the fact that it is the simplest
value for use in calculations, many of the foods are packaged in
multiples of one ounce, and it is the same as the reference
amounts customarily consumed for many of the types of foods used
in gift packages (e.g., many cheeses, crackers, and nuts)
specified in 101.12 in the companion document on serving
sizespublished elsewhere in this issue of the Federal Register.
    In the case of liquids, the agency believes a larger serving
size is needed because of the extra weight of water, and because
there are no reference amounts specified in 101.12 for liquids
at only one fluid ounce. Based on the reference amounts inerages.
These are the reference amounts in 101.12(b) for maple syrup and
for all beverages, respectively. The agency does not believe that
it is reasonable to collapse the number of categories of foods
any further than these three groups for the purpose of nutrition
labeling of gift packages. Therefore, in response to the comments
and in an effort to minimize the number of different serving
sizes required in the nutrition labeling of gift packages,
101.9(h)(3)(ii) allows for a serving size of 1ounce for all
solid foods, 2 fluid ounces for nonbeverage liquids, and 8 fluid
ounces for beverages where there is no uniform reference amount
customarily consumed for each individual food used in an
assortment or variety of foods within a gift package.
    However, the agency believes it would be misleading to allow
nutrient content or health claims based on these serving sizes
for foods packaged in gift packs where they differ from reference
amounts specified in 101.12(b) that are used as criteria for the
claims. Therefore, 101.9(h)(3)(ii) states that the reference
amounts customarily consumed that are listed in101.12 must
continue to be used for purposes of evaluating whether individual
foods in a gift package qualify for nutrient content claims or
health claims.
    Inasmuch as section 403(q)(1)(B) of the act requires that the
number of servings per container be included on the nutrition
label, FDA does not believe that it has the authority to permit
this information to be deleted. However in 101.9(h)(3)(iii), FDA
is allowing for the number of servings per container to be stated
in the nutrition label as ``varied'' in recognition of the fact
that each type of gift package will furnish a different number of
servings. This action is consistent with 101.9(b)(8) in the
companion document on serving size published elsewhere in this
issue of the Federal Register, which allows a manufacturer to
declare ``varied'' for the number of servings per container for
random weight products. The assembling of gift packs has a random
quality. FDA is persuaded that requiring more specific
information on labeling would necessitate a unique label or
labeling for each package, negating the usefulness of these
special provisions.
    Section 101.9(h)(3)(iv) provides that average, composite
nutrient values may be declared in nutrition labeling for
reasonable categories of foods having similar dietary uses and
similar significant nutritional characteristics. While the
comments requested that composite values be allowed for
reasonable categories of foods having similar dietary uses or
similar significant nutritional characteristics, FDA believes
that both criteria are necessary. Many forms of cheese and peanut
butter have similar dietary uses in that they are used to make
sandwiches or are eaten on crackers, yet they have far different
nutritional characteristics and should not be composited.
    The comments suggested, and FDA concurs, that companies
should submit to FDA their determinations of ``reasonable
categories'' for review and acceptance. FDA's decision on the
companies' determinations will be based, in large part, on
whether the values of the characterizing nutrients for foodsin
the category meet the compliance criteria set forth
in101.9(g)(3) through (g)(6). To that end, companies should also
submit a list of proposed characterizing nutrients for each
``reasonable category'' of foods.
    For example, assuming total calories, total fat, saturated
fat, and cholesterol are the categorizing nutrients for a group
of cheeses, each cheese's content of these 4 nutrients would have
to be no greater than 20 percent in excess of the declared values
in the nutrition label, in accordance with 101.9(g)(5), or
reasonably less than the declared values, in accordance with
101.9(g)(6). Nutrients other than the characterizing nutrients
could be stated as an average, or composite, for the catego
    While the comments requested that FDA specifically permit the
use of data bases for calculating the nutrition information for
foods in gift packages, the agency does not believe a separate
policy from that which the agency is establishing for other
packaged foods (see section VII.B.2. of this document) is
necessary or appropriate.
    Section 101.9(h)(3)(v) allows foods that meet the definition
for small packages under 101.9(j)(13)(i) that are included in a
gift package to be omitted in determining the nutrition
information if they are not specifically listed in a
promotionalcatalogue, and they are used in small quantities as
``optical garnishes'' to enhance the appearance of the gift
package or are included as a free gift or promotional item.
According to the comment, these items are used in very small
quantities and may vary greatly from package to package. On the
understanding that the ``optical garnishes'' are generally small
plain candies wrapped in bright colored paper, the agency
believes that the small amount used will make an insignificant
nutrient contribution tothe total package. Free gifts or
promotional items, by definition, are not ``offered for sale''
and are therefore exempt under 101.9(a).

B. Compliance (101.9(g))

1. Compliance Procedures
    In discussing the agency's rationale for requiring a single
nutrient value on the label in lieu of permitting ranges of
values, FDA tentatively concluded that its current compliance
policy with respect to nutrient variability satisfied the
requirements of the 1990 amendments (56 FR 60366 at 60373).
    The compliance policy in current 101.9(e) (proposed
101.9(g)) requires that the nutrient content of the composite of
12 subsamples be at least equal to the labeled value for Class I
nutrients (i.e., added nutrients in fortified and fabricated
foods) and at least 80 percent of the labeled value for Class II
nutrients (naturally occurring or indigenous nutrients). Proposed
101.9(g)(4) specified that these requirements are applicable for
vitamins, minerals, protein, total carbohydrate, complex
carbohydrate, dietary fiber, unsaturated fat, and potassium
content. Likewise, in proposed 101.9(g)(5), the nutrient content
of the composite is required to be no more than 20 percent above
the labeled value for calories, sugars, total fat, saturated fat,
cholesterol, and sodium.
    238. The agency received a number of comments regarding its
compliance policy as stated in proposed in 101.9(g) (56 FR 60366
at 60391). A few comments agreed with allowing an 80 to 120
percent leeway in the compliance of foods. One comment notedthat
while the nutrient values may not be absolute, they are more
consistent for the consumer. Also, the present system makes it
easier for manufacturers to obtain compliance. However, the
majority of comments disagreed with FDA's compliance policy,
requesting that either a tighter or looser standard be used.
    Consumers were strongly opposed to the so-called ``80-120
rule.'' They felt the range was too lenient and stated that they
would like to see a tighter standard adopted, especially for
calories. Fat, carbohydrate, and cholesterol were also
identifiedas nutrients that should have very accurate or exact
label declarations. Some suggested other limits of acceptance,
such as a plus or minus 5 to 10 percent range. Several comments
supported a more accurate declaration of nutrients in
consideration of the needs of persons with medical conditions
requiring adherence to specialized or restricted diets. Other
consumers considered the 20 percent margin of error as being
inaccurate and misleading. Some comments considered that with
today's available technology, food manufacturers could and should
more accurately declare nutrients, notably calories, on the
labels and meet more stringent standards.
    Several comments included suggestions as how to better
declare nutrient content on the label. Suggestions included the
declaration of a tolerance standard on all product labels and an
example of what the tolerance could mean. For example, the label
of a product having a 10 percent tolerance for calories would
state the declaration as ``100 calories--could be 90 or 110
calories.'' One comment suggested that a statement be required
adjacent to the calorie value declaring that it is ``only an
approximate figure.'' Another suggested that all food labels
carry a warning of the 20 percent margin of error permitted for
calories.
    Comments from industry and trade associations considered the
80/120 percent range unduly restrictive. They supported more
flexible compliance standards that would provide ``representative
values'' of a product's nutrient content. Representative datain
one comment was defined as the mean or the mean plus or minus one
standard deviation. Their contention was that, because of the
natural variation of foods, application of FDA's compliance
procedures result in gross under-representation of some
nutrients, such as vitamin A in carrots, and gross
over-representation of other nutrients, such as sodium in soft
drinks (because of variability in water sources). The comments
took exception to FDA's assertion in the discussion on fresh
produce and seafood, in the mandatory nutrition labeling proposal
(55 FR 29487 at 29506) that nutrient content can largely be
controlled in most manufactured foods. In fact, they stated there
is greater variability in processed foods because of the
complexity of prepared foods, the further processing that is
required, the need to total the variability for each ingredient
for prepared foods, the flexibility needed for obtaining
ingredients from various sources or suppliers, and the analytical
variability for required nutrients. One comment recommended that
an 80 to 120 percent compliance range be used for nutrients with
a low degree of natural or analytical variability. For nutrients
with a high degree of natural or analytical variability, a less
stringent 65 to 135 percent compliance range was suggested.
    Another comment endorsed a more flexible compliance standard
whereby micronutrient levels need only be present at a minimum
level of 80 percent of declared levels. They recommended that no
maximum compliance level be set. This comment was particularly in
reference to the difficulty of achieving compliance for a product
that has a standard of identity, such as pasta, where maximum and
minimum levels of enrichment are specified by the standard. The
comment stated that levels of added nutrients may vary depending
upon the method of enrichment, indigenous nutrient levels in the
wheat, analytical error, rounding of values declared on the
label, and loss of nutrients during the drying process.
    The agency disagrees with establishing more stringent
requirements for label values. FDA shares concerns about
individuals with very specific health problems where diets must
be closely monitored and controlled. However, no data have been
presented,and FDA is not aware of any such data, to suggest that
health problems have been created because of the allowable
variances. Therefore, the agency considers health management
under professional guidance satisfactory using the nutrient
values on the labels based on current regulations. In addition,
it should be noted that the natural variability of foods may lead
to both under- and over-reporting within the allowable variances
for individual foods. These variances will tend to balance out
over the entire day's diet.
    While it is highly desirable to have a precise nutrient value
on the label, it is impractical. The natural variability of a
food is dependent upon a number of factors. Among them are the
season of the year, soil type, variety (cultivar), and
weatherconditions. The processing that a food undergoes also
alters its nutrient content. In addition to these variables, the
agency places restrictions on the label declarations in regard to
the rounding of nutrient values. These rounding rules are to
avoid the and understand. To declare nutrient values more
accurately or precisely than is presently required would place an
onerous burden on the manufacturer. The costs associated with the
excessive controls to provide more exact label declarations are
unreasonable and would not be commensurate with any possible
additional health benefit.
    The agency rejects the suggestions that declared values be
qualified by statements that they reflect tolerance levels or
margins of error. Such statements on the label informing
consumers of the possible variation between labeled and
analytical values would cause great confusion with no real
benefit.
    Similarly, FDA disagrees with the comments that suggested
establishing less stringent requirements for determining
compliance with declared label values. As seen in comments,
consumers rely on the declared values, and the accuracy of those
values isimportant. FDA does not believe larger suggested ranges,
such as 65 to 135, would give consumers the information that they
need to adequately evaluate their nutrition intake. Therefore,
the agency is not making requested changes in 101.9(g).
    FDA advises that it has not set maximum compliance levels in
101.9(g) for Class I and Class II nutrients, nor has it set
minimum compliance levels for nutrients specified in101.9(g)(5)
(i.e., calories, sugars, total fat, saturated fat, cholesterol,
and sodium). The 20 percent variability permitted is not a range
but rather a lower or upper limit, depending on the nutrient. The
only regulatory limit on overages of ClassI and II nutrients is
given in 101.9(g)(6), which states that ``reasonable excesses''
are acceptable within current good manufacturing practice.
Likewise, 101.9(g)(6) also states that ``reasonable
deficiencies'' of calories, sugars, total fat, saturated fat,
cholesterol, and sodium under labeled amounts are acceptable
within current good manufacturing practice. FDA anticipates that
manufacturers will be diligent in their own behalf in not
underdeclaring Class I and II nutrients, such as vitamins and
minerals, and in not overdeclaring nutrients such as calories and
fat.
    Regarding maximum levels of micronutrients in standardized
enriched pasta products, regulatory relief cannot be achieved
through modifications of 101.9 but require changes in the
standards of identity of such products.
    239. Several comments suggested that the 80 and 120 percent
criteria should only be a guideline or screening tool. A few
comments expressed the position that FDA should not declare a
product misbranded until the manufacturer has had an opportunity
to establish that the variations are reasonable under the
circumstances.
    Other comments suggested that the 80 and 120 percent criteria
be waived when there are small quantities. (The quantity limits
suggested were 10 and 20 or fewer ``units.'' ``Units'' were
interpreted to be units of measurements, such as 10 or 20
calories or 10 or 20 mg of sodium.) The comments noted that small
numbers combined with rounding rules and analytical variability
result in inequities for label compliance (i.e., the analytical
variance for some low levels of nutrients is greater than the
allowed regulatory variance). For example, the comment stated
that if a mean value of 1.3 units was rounded for label
declaration to the nearest whole unit (i.e., 1 unit), then the
acceptable range would be 0.8 to 1.2 units when applying the 80
and 120 percent criteria. The range would be below the true mean
value which could result in many products being found out of
compliance. Furthermore, these small differences of 0.2 units may
not be within the accuracy of many methods, so that the
analytical variance could be greater than the allowed regulatory
variance. For these small quantities of 20 or fewer units, the
comments recommended that a 50 to 150 percent rule be applied.
One comment recommended that FDA clarify in the final rule that
the rounding of nutng nutrient content claims to meet compliance
criteria as well as standards of identity.
    An alternative suggestion in another comment to avoid an
extreme over- or under-declaration when the value is small is to
declare the nutrient content to the nearest whole unit with
compliance based on a fixed percentage (e.g., within 80 percent)
ora fixed unit amount (e.g., one unit or 2 percent U.S. RDA, the
basic increment of rounding). The regulation would then require
that declared amounts be within 80 percent or one unit (such as a
g) for Class II nutrients or within 120 percent or one unit for
nutrients such as calories, fat, or sodium.
    The agency is not persuaded that the current or proposed
acceptance criteria for compliance evaluation should be changed.
The compliance criteria permit reasonable excesses over labeled
amounts or deficiencies under labeled amounts, dependent upon the
nutrient being evaluated, (current 101.9(e)(6), redesignated as
101.9(g)(6)) within current good manufacturing practices. As
discussed in section IV.A. of this document, the level of
``reasonable'' is not specified.
    It is the manufacturer's responsibility to target labeled
values to correspond to actual nutrient levels so that products
will meet compliance requirements. This responsibility includes
taking into consideration the effects of rounding. Any effect
caused by the rounding of labeled values to meet the agency's
requirements in 101.9(c) should be accounted for by the
manufacturer in developing a label value and would be included in
the evaluation of a ``reasonable'' level by the agency.
Analytical variance is also one of the factors in determining
compliance acceptance. This fact is stated in 101.9(g) (4) and
(5) in this final rule.
    Manufacturers should perform shelf-life stability studies to
substantiate the declared nutrient levels of the product and to
demonstrate that a product can meet label claims over the shelf
life of the product. FDA does not believe that incorporationinto
the regulations of any additional explicit provision or
compliance position for low level nutrients or small labeling
increments would provide added protection for manufacturers.
    240. One comment strongly recommended that FDA address
sampling issues. It suggested that the current procedure
in101.9(e)(2) (and in proposed 101.9(g)(2)) of preparing a
composite of 12 subsamples taken from a single lot be changed.
Instead, it was suggested that a sample composite for analysis
represent 12 different lots.
    The agency disagrees with the suggested change in sampling
procedures. The comment's suggestion reflects a sampling
objective that appears to focus on estimating the nutrient
content of product for a specified quantity (e.g., a company's
production). FDA's sampling objective is to determine whether the
average, within a given lot (a quantity that is defined in
current 101.9(e)(1)), meets label claims. From a compliance
evaluation standpoint, the suggested sampling scheme is not a
feasible alternative because the results obtained would not be
traceable to a specific lot should an overage or deficiency be
encountered. Instead of a compliance action against a smaller
quantity (a single lot), it might be necessary to take a
compliance action against a larger quantity (e.g., a company's
production for a larger specified point in time). Therefore, FDA
is making no change in 101.9(g)(2) in response to this comment.
    241. Several comments that disagreed with the agency's
compliance policy provided suggestions to clarify the codified
language. One comment recommended the elimination of total
carbohydrate, complex carbohydrate, and unsaturated fat from the
Class Icategory of nutrients at 101.9(g)(4)(i). It maintained
that these three nutrients are unlikely to be ``added'' but are
the result of having used ingredients that inherently have these
nutrients.
    FDA agrees with the recommendation. Therefore, the agency is
amending 101.9(g)(4)(i) to delete total carbohydrate, complex
carbohydrate, and unsaturated fat from the Class I category. This
deletion should allay the concerns ofhaving the cited nutrients
meet Class I nutrition labeling requirements. These nutrients
remain in the Class II category (101.9(g)(4)(ii)), although in
accordance with the changes made in section III. of this
document, complex carbohydrate is changed to other carbohydrate
and unsaturated fat to poly- and monounsaturated fat.
    To clarify the compliance policy concerning variability
because of analytical methodology for Class I and Class II
nutrients, FDA is modifying 101.9(g)(4) by making a new
paragraph out of the last sentence which begins with the word
``Provided.'' This change should make clear that the proviso
information regarding consideration of regulatory action is
applicable to both Class I and Class II nutrients. This
qualifying information was inadvertently moved under the
paragraph on Class II nutrients in the July 19, 1990 mandatory
nutrition labeling proposal, and the error was carried forward in
the supplementary proposal.
    242. One comment stated that manufacturers should be able to
use mean values in all cases, except that statistical outliers
should be ignored. The comment also urged the agency to codify
its compliance policy to the extent that if a nutrient is found
out of the 80 to 120 range of the labeled amount, the product
would not be deemed out of compliance as long as the manufacturer
can demonstrate that the label declarations represent mean values
based on reasonable and adequate sampling and analyses.
    FDA disagrees with the comment. The agency's position on the
use of mean values is summarized in the preamble of the
supplementary proposal (56 FR 60366 at 60373). This position is
discussed in more detail in the ``FDA Nutrition Labeling Manual:
A Guide for Developing and Using Databases'' and in section
VII.B. of this document. In short, FDA will allow the use of mean
values derived from satisfactory data bases if the coefficient of
variation is equal to or less than the maximum coefficient of
variation specified in the above manual. The coefficient of
variation is the standard of deviation (a measure of variability)
expressed as a percentage of the mean.
    243. A recommendation was made in one comment to
amend101.9(g)(5) by adding: ``Provided, that no regulatory
action will be based on a determination of a nutrient value which
falls above this level by a factor less than the variability
generally recognized for the analytical method used in that
product at the level involved.'' The comment noted this addition
would extend the allowance for analytical variability permitted
for vitamins, minerals, protein, total carbohydrate, complex
carbohydrate, dietary fiber, unsaturated fat, and potassium, as
given inproposed 101.9(g)(4) to the nutrient declarations for
calories, sugars, total fat, saturated fat, cholesterol, and
sodium.
    The agency agrees that analytical variance is a valid
consideration when contemplating regulatory action for all
nutrients. Accordingly, the proviso stated in 101.9(g)(4), which
is applicable for Class I and Class II nutrients, is added to
101.9(g)(5).
    244. One comment recommended a two-stage enforcement
procedure. The first stage would involve analysis of a single
12-sample composite to determine whether the product passes the
compliance standard of 80 to 120. If it passed, the agency would
have no enforcement issue. If it did not pass, the agency would
collect and measure the nutrient content in three other lots. The
average of all four lots tested would be evaluated for compliance
purposes.
    The agency is not making the change in its procedures that
was suggested by this comment. As recognized in several comments
from manufacturers, it is the manufacturer's responsibility to
accurately declare the nutrient content of a product. As
discussslated to other lots. Factors that could have altered the
nutrient content of one lot may not be present for subsequent
lots. Through quality control programs and careful consideration
of declared nutrient amounts according to the guidelines in ``FDA
Nutrition Labeling Manual: A Guide for Developing and Using
Databases,'' manufacturers can help to ensure that each lot meets
compliance standards. The suggestion made in this comment or in
comment 240 of this document in regard to evaluating results
representing the analysis of a composite from 12 different lots
or an average of results from four composites could be
implemented in a manufacturer's quality control procedures to
assure compliance with 101.9(g).
    It should be noted that the analyzed nutrient content is not
the sole factor in determining whether the agency will bring a
particular enforcement action. Other factors that it considers
include the effect of matrix upon the analyte, the level of the
analyte in the food, information obtained during an establishment
inspection of a firm, consumer complaints, past compliance
history of the firm, and the firm's demonstrated ability to
adequately perform the analysis for a nutrient.
    245. A comment recommended that a new section be added to the
codified language to the effect that ``The metric declaration of
the serving size shall be used to determine compliance under this
section.'' The comment said that this change would eliminate any
confusion about which of the dual declarations required for
serving size would be the determining factor for nutrient
declaration.
    FDA agrees with the comment. In its serving size proposal (56
FR 60394 at 60410), the agency stated that in addition to the
more approximate household measure, it needed a precise weight
statement for serving size for compliance purposes. Accordingly,
the agency proposed in 101.9(b)(7) that the serving size in
common household measures must be followed by the equivalent
metric quantity. However, FDA did not specifically state that
this metric measure would be used for compliance purposes.
Therefore, for purposes of clarity, the agency is adding a new
section, 101.9(g)(7), to correct this oversight. Consequently,
the remaining paragraphs in 101.9(g) are redesignated.
    246. One comment expressed concern that net weight
regulations must be considered when evaluating a product against
the 80 to 120 rule for compliance. The comment stated that
manufacturers are required to sell products at levels above the
declared label weight. The comment concluded that this resulted
in a discrepancy between labeled nutrition information and actual
nutrition values.
    The agency does not consider this issue to be a valid
concern. Because of the economic considerations of manufacturing,
most products are close to label claims for net weight.
Additionally, while an overage or underage of the net weight may
slightlyalter the nutrient content of the container (and
particularly if the container is a single serving size), the
serving size is the factor by which the nutrients are evaluated.
As discussed in the preceding comment, FDA will composite samples
and then usethe metric weight declared as the label serving size
to evaluate the accuracy of declared nutrient values.
2. Data Bases
    247. FDA received a large number of comments regarding the
use of data bases as sources of nutrient information for
nutrition labeling. Most comments supported the use of data
bases, giving as reasons that the use of data bases would reduce
costs toindustry (especially to small businesses), moderate food
cost increases to be passed on to consumers, promote fair
competition, save time, reduce the use of laboratory chemicals,
provide sufficient accuracy, and ease compliance verification
procedures. Co
    FDA appreciates the important role data bases can play in
nutrition labeling. Industry-wide data bases were first suggested
in 1979 as a possible means of reducing the cost of developing
nutrition labeling for individual companies. FDA, USDA, and the
Federal Trade Commission encouraged this concept in a notice of
proposed rulemaking published in the Federal Register of December
21, 1979 (44 FR 75990) describing the agencies' policies and
intentions with respect to numerous food labeling issues. In that
notice, FDA, while not agreeing to approve data bases, stated
that it would work with industry to resolve any compliance
problems that might arise for food labeled on the basis of a data
base that the agency had accepted.
    FDA is concerned about the reliability of data bases to meet
compliance requirements for nutrition labeling. Nutrient data may
be valid for some purposes and not for others. For example, data
bases that were developed largely for determining averagedaily
dietary intakes generally serve that purpose well. However, such
data bases are usually not adequate to determine natural
variability of a particular food or to develop labeling values
that are in compliance with FDA nutrition labeling regulations.
    Despite these concerns, FDA continues to acknowledge the
potential usefulness of data bases to reduce costs associated
with nutrition labeling. The agency set out its general policy on
the use of data bases most recently in the proposed and final
rules on the voluntary nutrition labeling program for raw produce
and fish (56 FR 30468 at 30474, July 2, 1991 and 56 FR 60880 at
60884, November 27, 1991, respectively) and the supplementary
nutrition labeling proposal (56 FR 60366 at 60373). In
addition,the agency announced in the Federal Register of July 23,
1992 (57 FR 32796) the availability of a draft manual entitled
``FDA Nutrition Labeling Manual: A Guide for Developing and Using
Databases.'' The manual, which replaces the former guide
``Compliance Procedures for Nutrition Labeling,'' is intended to
aid companies and trade organizations in developing and using a
data base for nutrition labeling that would meet the regulations
proposed as a result of the 1990 amendments. It also discusses
the conditions under which the mean value derived from a
satisfactory data base may be used for nutrition labeling.
Comments were requested on the draft manual. These comments have
been considered, and the agency is hereby announcing the
availability of the final manual. The manual may be obtained from
the Division of Nutrition (HFF - 260), Office of Nutrition and
Food Sciences, Center for Food Safety and Applied Nutrition, Food
and Drug Administration, 200 C St. SW., Washington, DC 20204.
    FDA anticipates that this manual will be of assistance in
identifying a data base that is of a quality to provide an
adequate basis for nutrition labeling. The use of such a data
base to calculate the final composition of a product formulated
from several ingredients presents additional problems, however,
in that there are no allowances or determinations of the loss of
nutrients that may occur during further processing. Depending on
the type and amount of processing, significant amounts of
nutrients may be lost. The agency is willing to work with
manufacturers and trade associations to determine if successful
models can be developed showing the relationship between
ingredient composition and final product composition that account
for losses duringprocessing. While extensive analyses of
ingredients and final products would be required to develop and
validate a successful model, such action could lead to an
acceptable data base.
    If a manufacturer wishes to use a data base for nutrition
labeling, it is advantageous to follow the statistical procedures
outlined in the manual and have the data base accepted by FDA. If
the agency finds that the nutrition label of a product whicht the
product, provided that the company has followed good
manufacturing practices in producing the food. Instead, the
agency would work with the manufacturer to resolve the compliance
issue. Action would be taken only if noncompliance was the result
offailure to follow good manufacturing practices.
    It must be noted that submission of a data base to FDA for
review and acceptance is voluntary. The agency has not prescribed
how an individual company is to determine nutrient content for
labeling purposes. The choice of a data source is the prerogative
of the manufacturer. The manufacturer needs to be judicious in
this selection, however, to ensure that the product labeling is
in compliance with the regulations. The compliance policy of the
agency remains unchanged from current101.9(e). An FDA
investigator/inspector will collect random units of food (e.g.,
consumer packages, items of product) from each of 12 different
randomly chosen shipping cases of the same code/lot, and an FDA
laboratory will prepare a single composite from the 12 units of
food. Analysis of the composite will be performed using methods
of the AOAC or other reliable and appropriate methods. FDA will
then compare the values declared in the nutrition label and
labeling with the results from the laboratory analyses. Section
101.9(g)(8) provides for the use of an approved data base.
    248. Many comments were received expressing support for use
of data bases because they can be used by a company one time, and
there would be only a one time cost to determine the nutritional
values for the label without regard to future changes in the
product.
    The agency is concerned that there is a misunderstanding
regarding the use of data bases. Data bases are not static but
dynamic because of changes in products. Those data bases
submitted to the agency or used by companies are expected to
reflect thenutritional content of products being offered for
sale. Changes in variety, supplier, recipe, or manner of
processing could lead to very different nutritional values for
the product than those in the original data base. The agency, in
monitoring products for compliance, will also review the
maintenance of these data bases to ensure that the information in
the data bases reflects the nutritional content of the products
being offered for sale. Maintenance of a data base means that
laboratory analyses of the product are done on a periodic basis
to ensure that the nutritional values of the product are within
the limits of the data base values. Proper maintenance of the
data base is left to the originators of the data base. Frequency
and type of maintenance are determined by the data base holders
based on their knowledge of the changes in the products.
Satisfactory data bases could be useful for periods of up to 10
years based on the size of the data base, plan for maintenance,
and the complexity of the product. FDA reviews will be based on
the amount of supplementation of the data bases with additional
laboratory data (maintenance) during the period of use and the
changes in the products covered by the data base.
3. ``Nutrition Labeling Manual: A Guide for Developing and Using
Data Bases''
    FDA announced in the Federal Register (57 FR 32796) on July
23, 1992, the availability of a draft manual entitled ``Nutrition
Labeling Manual: A Guide for Developing and Using Data Bases''
(the manual). This manual is intended to aid companiesand trade
organizations in developing and using a data base for nutrition
labeling that meets the regulations resulting from the 1990
amendments. Comments on the manual were accepted until September
8, 1992. This manual provides generic instructions on how to
develop and use a data base in preparing nutrition labeling for a
food product. Eighteen comments were received from companies or
trade associations. The following summarizes the comments and
provides the agency response to those comments.
    249. Almost all of the comments were opposed to the limit of
80 percent of the label claim for Class II (indigenous) nutrients
and 120 percent for calories, sugars, total fat, saturated fat,
cholesterol or sodium. The comments argued that these limits were
overly restrictive and should be widened or average values should
be used with no consideration of limits.
    The use of 80 percent or 120 percent as limits for regulatory
purposes is established in 101.9(g) and has been addressed above
in section VI. of this document. The manual was developed to aid
in the calculation and construction of data or data bases to meet
the regulatory requirements of the agency. Should the agency
change its regulations, it will reflect those changes in the
manual. Until such changes are made, however, the manual must
reflect the applicable regulatory limits.
    250. All but one of the comments addressed the use of recipe
data bases to calculate final composition of mixed products. The
comments cited the savings in money to small businesses, the
constant changes in recipes that make it too costly to do
analytical testing of products, and the cost of analyzing a large
number of products for which the volume is low. The expressed
belief was that calculated values better represent the nutrient
content over time. Several comments suggested criteria for a good
recipe data base. One comment offered the following four proposed
principles of good ingredient composition data bases:
    (1) Confidence in the quality of data, supported by
documentation of data sources. Companies maintaining or using
ingredient composition data bases must be able to demonstrate the
data source used for each type of product and each nutrient for
whichingredient composition data bases are utilized.
    (2) Proper maintenance of the data base. Companies developing
or using ingredient composition data bases must have procedures
in place to ensure that the values in the ingredient composition
data bases are reviewed and updated as needed and on a regular
basis.
    (3) Specificity with respect to ingredients, product
formulations, and processes. Companies using ingredient
composition data bases must have procedures in place to ensure
that the nutrient values are used only for specific applications.
For example, a company should have a procedure to ensure that
nutrient data specific for one product formulation or process are
not used to prepare nutrient declarations for similar product
formulations or processes, without assurance that the data are
applicable to those products or processes.
    (4) Validation of the data base. Companies developing or
using ingredient composition databases must have procedures in
place to ensure that nutrient values receive reviews, audits, and
confirmation through nutrient analyses as often as necessary.
    Other comments suggested that manufacturers should be
required to substantiate any nutrient content or health claims
with analytical data.
    The agency agrees that the principles suggested by the
comment are worthwhile and necessary for construction of a proper
ingredient composition data base. This was the intent of the
statement in the manual that calculation of the final composition
of a mixed product using data bases of the nutrient composition
of ingredients might be acceptable if properly modelled. The
agency wanted to assure itself that the ingredient composition
information was adequate, and that the calculation of the final
nutrient value of the finished product reflected any possible
loss of nutrients during processing. In addition, a successful
mathematical model used for this purpose should be augmented over
time with a review of its applicability by laboratory analysis of
the nutrient content of both ingredients and final products.
Models constructed with the features described above, and applied
to a limited range of appropriate products, would receive serious
consideration from the agency. The above features of an appropr
    The agency believes that in time the calculation of the final
composition of mixed products from ingredient data bases may be
acceptable for a range of food products. At this time, however,
the agency believes that the data that make up ingredient data
bases are of mixed quality and, therefore, of limited value.
Companies that wish to use ingredient data bases must look at the
individual analytical values of each ingredient to evaluate the
data to assure themselves that the data are sufficient, meet the
requirements expressed in the manual for representativeness, are
valid from an analytical standard, and are sufficient to account
for any variation in the ingredient.
    The agency has stated that the company bears the final
responsibility for the accuracy of the label. This principle has
not changed and was repeated and supported by several of the
comments.
    251. Comments were received on changing various aspects of
the agency's regulatory policy such as larger number of lots
sampled and the average taken, composite samples consisting of
several lots, exemption from compliance procedures when data
basesare used regardless of whether the agency has accepted them
or not, exemption from compliance procedures for nutrients that
have a low concentration, and exemption from compliance
procedures for companies/associations that have submitted basic
data and a plan for data base development over time.
    The manual is intended to aid manufacturers/associations in
meeting the compliance regulations of the agency. The manual does
not set compliance policy but rather offers some explanation for
the compliance policy and provide different means of complying
with the nutrition labeling regulations. Should the compliance
policy of the agency change, the manual will also be changed to
reflect those changes.
    252. Many comments were received regarding the
confidentiality of the submitted data bases. Developers of data
bases did not want to see the information gained through analyses
of products and ingredients released through freedom of
information requests or used in unacceptable ways or for
inappropriate products. In addition, development of data bases is
a program with costs shared among the participating companies.
The comments sought assurance that the data would not be
available at no cost to companies that did not participate in its
development. Formulations that are used to produce mixed products
are also regarded as confidential company information, and the
comments sought assurance that they would not be available to
anyone who requests the information.
    The agency is aware that the development of a data base is
costly, and that it may contain information that is of a
confidential nature. The agency agrees that release of a data
base could vitiate substantial proprietary interests in valuable
documents submitted to the agency. Furthermore, it has never been
the agency's intent, nor does it have the resources, to maintain
and manage data bases that are developed by manufacturers or
associations. The agency believes that the availability of a data
base is therefore the primary responsibility of the developer.
The agency will continue with the policy of assisting the
developers of data bases, providing guidance to those who ask for
it, and accepting adequate data bases for the products submitted
for review. Only those data sufficient to support the agency's
decision to accept or not accept a data base will be retained.
Confidentiality of such data will be determined and maintained in
accord with regulations in part 20 (21 CFR part 20).
    Those data base developers who choose to do so are encouraged
to make their information available through such compilations as
the USDA Handbook No. 8 so that all may benefit from the
additional analytical information.

C. Proposed 101.9(h)

    253. A few comments objected to the requirement in proposed
101.9(h) that nutrition information provided by manufacturers or
distributors directly to professionals (e.g., physicians,
dietitians, educators) must contain or have attached to it the n
required by 101.9. The comments stated that it was inappropriate
for a Federal agency to regulate the transfer of information in
this manner.
    FDA notes that this section of the regulation has been
carried unchanged since 1973 (49 FR 6961, March 14, 1973). At
that time, the agency stated that it did not want to restrict the
flow of information from food manufacturers to professionals
(suchas more precise amounts rather than the increments used in
nutrition labeling) but rather wanted nutrition information
included or attached to it in the form it would be provided to
consumers. Inasmuch as nutrition labeling is now mandatory, so
that consumers will have the required information available to
them on food product labels, FDA has decided to delete this
requirement and is doing so in this final rule.

D. Section 101.9(k)

    254. Many comments objected to proposed 21 CFR 101.9(k) which
details types of nutrition-related claims that cause a food to be
misbranded. Most of these comments asserted that the provisions
of 101.9(k) are contrary to the intent of the 1990 amendments
and contrary to the will of Congress.
    Many comments offered specific objection to proposed101.9(k)
(3) and (4) and asserted that manufacturers should be allowed to
provide information about the effects of soil, storage,
transportation, or cooking on the nutrient content of foods. Some
comments maintained that the restriction of such information is
unconstitutional. A number of comments felt that labels should be
required to provide information as to the exact identity of the
contents (including substances of no nutritional value), the
source of the contents, the amounts of all ingredients, and the
techniques and dates of processing. One comment proposed that
manufacturers should be required to put toll-free telephone
numbers on all of their products so that consumers could call for
information about those products.
    Many comments asserted that proposed 101.9(k)(5) is
arbitrary and restrictive and expressed a belief that certain
naturally-occurring food constituents will be rendered
unavailable by this provision. A number of comments maintained
that there isno legitimate reason for prohibiting substances
found in nature from being incorporated into nutritional products
and listed on the label. Some comments suggested amending
proposed 101.9(k) to allow the use of naturally-occurring
constituents of foods and herbs, unless there is sufficient
evidence that any specific such substance is harmful to human
health.
    Some comments also objected to proposed 101.9(k)(6),
maintaining that vitamins that are naturally present are better
than added synthetic vitamins. These comments expressed a strong
desire to know whether vitamins contained in any specific product
are naturally-occurring or synthetic.
    FDA regrets that its publication of 101.9(k)(2) through (6)
in the November 27, 1991 (56 FR 60393) proposal has created
confusion. The publication of 101.9(k)(2) through (k)(6) did not
constitute a proposal of new regulations. It merely represented a
proposed redesignation and republication of existing regulations
for clarity and completeness.
    The provisions embodied in current 101.9(i)(2) through
(i)(6), redesignated in the November 27, 1991, supplementary
proposal as 101.9(k)(2) through (k)(6), to which the comments
directed their objections, were first proposed in the Federal
Register of March 30, 1972 (37 FR 6493), and were promulgated and
published in the Federal Register of January 19, 1973 (38 FR
2125), as 1.17(i)(2) through (i)(6). Following an appropriate
comment period, these regulations were modified and published as
final regulations in the Federal Register of March 14, 1973 (38
FR 6961). The regulations were subsequently applied, with certain
exemptions, to all food labeling ordered after December 31, 1973,
and all labeling used for food productspublication of section 21
of the Code of Federal Regulations that appeared in the Federal
Register of March 15, 1977 (42 FR 14308), 1.17(i) was renumbered
as 101.9(i). No changes were made to the original codified
language of the subject paragraphs during any of these
renumberings, and those regulations remain as adopted in 1973.
    The only change in 101.9(k) in the supplementary proposal
was in the document entitled ``Labeling; General Requirements for
Health Claims for Food'' (56 FR 60537, November 27, 1991). This
document proposed to amend current 101.9(i)(1), redesignated as
101.9(k)(1), by adding a second sentence that reads
``Information about the relationship of a dietary property to a
disease or health-related condition may only be provided in
conformance with the requirements of 101.14 and subpart E of
part 101.'' No comments voiced specific objection to this
proposed change.
    FDA notes that the provisions of current 101.9(i) had long
been in effect at the time Congress drafted the 1990 amendments.
While Congress did enact provisions under the 1990 amendments
that allow for health claims on foods, nothing in the act or in
the legislative history of the act suggests that Congress
intended that current 101.9(i) should be changed. The agency
therefore finds no basis for the assertion that the provisions of
current 101.9(i), redesignated as 101.9(k), are now contrary to
the intent of the 1990 amendments.
    FDA has reconsidered the requirements of 101.9(k)(5),
however, in light of the comments. The agency concludes that
there is no reason to prohibit safe substances from being
incorporated into conventional foods or dietary supplements of
vitamins and minerals as long as their presence is noted in the
ingredient list, and the product's label or labeling does not
state or imply that the food has special dietary properties
because of the presence of the substance when, in fact, its
usefulness has not been established. Section 411(b)(2) of the act
provides that vitamin and mineral products may contain substances
that are not vitamins or minerals as long as the substances are
only identified as a part of the ingredient list. Therefore,
proposed 101.9(k)(5) is amended by deleting the second and third
sentences.
    Questions have been raised as to whether the amounts of these
substances that are not vitamins or minerals can be included on
the food label. Such information can be included in the
ingredient list if, in addition to listing the ingredients in
orderof predominance by weight, quantitative information on each
of the ingredients in the food is presented. However, information
about the ingredients that are not vitamins and minerals may not
be presented in a way that suggests that the dietary usefulness
of these substances has been established.
    While the comments raised objections to the other provisions
of proposed 101.9(k) (i.e., (k)(3), (k)(4) and (k)(6)), none
provided arguments that convinced the agency that deletion or
revision of those provisions was either appropriate or necessary
in fulfilling the mandates of the 1990 amendments. The objections
that were raised, however, suggest that a clarification of the
intent of those provisions would prove helpful to those who
voiced the objections. Such a clarification was provided in the
Federal Register of March 14, 1973 (38 FR 6961). In that
document, the agency noted that1.17 (i)(3) and (i)(4)
(redesignated as 101.9(k)(3) and (k)(4)) are aimed at
prohibiting unsubstantiated generalizations about nutrient losses
because of soil, transportation, or processing and do not
preclude a producer, manufacturer, or vendorfrom indicating a
higher nutrient retention in a particular product as compared to
other similar products. Nor do they preclude an indication that
such retention results from special handling of the product,
provided that such indications are factual. Furat would result in
optimum nutrient retention. While the agency recognizes that such
information may be useful to consumers, it does not believe that
it would be appropriate to require manufacturers to provide such
information, either on the labeling orthrough other media.
    Current 101.9(i)(6), redesignated in the supplementary
proposal as 101.9(k)(6), prohibits any suggestion that a
naturally-occurring vitamin is superior to an added vitamin. The
agency finds no basis for such an assertion, and the comment
offered no data in support of such an assertion. As the agency
clarified in the repromulgation of March 14, 1973 (38 FR 6950 at
6958), this section (then 1.17(i)(6)) ``forbids any suggestion
that a natural vitamin is superior to an added vitamin, but
permits any truthful designation of any nutrient as natural in
origin.''
    FDA acknowledges its inadvertent oversight in not including a
reference to proposed 101.36, Nutrition Labeling of Dietary
Supplements of Vitamins and Minerals, in proposed 101.9(k). The
inclusion of this reference is a logical outgrowth of the
agency's stated intention that ``nutrition labeling of vitamin
and mineral supplements appear as similar as possible to the
nutrition labeling of other foods'' (56 FR 60366 at 60382).
Section 101.9(k) applies to all foods, including dietary
supplements of vitamins and minerals, and the agency did not
intend to narrow its scope. Therefore, FDA tentatively concluded
that it should correct this oversight by including an appropriate
cross reference to 101.36 in the final rule. However, the agency
will propose its position on this issue following the DS Act. For
completeness, FDA is inserting the word ``label'' in the first
paragraph of 101.9(k) to clarify that this section pertains to
food labels as well as labeling.
    255. One comment asserted that the phrase ``represents,
suggests, or implies'' in the opening sentence of 101.9(k) is
unconstitutionally vague.
    FDA disagrees with the comment's assertion that the phrase
``represents, suggests, or implies'' is unconstitutionally vague.
The agency notes that the vagueness doctrine is generally applied
to strike down prohibitions on speech that leave individuals
without clear guidance on the type of speech that is prohibited.
See, e.g., Village of Hoffman Estates v. Flipside, Hoffman
Estates, Inc., 455 U.S. 489, 498-99 (1982); Grayned v. City of
Rockford, 408 U.S. 104, 108 (1972). The provisions of 101.9(k)
are narrowly tailored and clearly delineate the types of
statements about nutrients that will render a food misbranded.
Thus, 101.9(k) provides clear and precise guidance on the type
of speech that is prohibited.

E. Conforming Amendments

    256. A trade association wrote in support of the multiunit
retail exemption in 101.9(j)(15) and requested that1.24(a)(14)
be amended to reflect the change by including a reference to
section 403(q) and (r) of the act. The comment stated ``we submit
that this amendment is fully consistent with the requirements of
the 1990 amendments and the provisions of theproposed
101.9(j)(13) in that nutrition labeling will be provided on the
outer carton together with the other information required under
the referenced sections.''
    The agency agrees that 1.24(a)(14) of the General
Enforcement Regulations should be amended to reference 403(q) of
the act, as amended by the 1990 amendments. This change merely
conforms 1.24(a)(14) to the rule that FDA is adopting in
101.9(j)(15). Accordingly, the agency is amending 1.24(a)(14)
to read as follows: ``The unit containers in a multiunit or
multicomponent retail food package shall be exempt from
regulations of section 403(e)(1), (g)(2), (i)(2), (k), and (q) of
the act wi
    However, FDA cannot grant the comment's request with respect
to section 403(r) of the act. Any container that bears a nutrient
content claim or a health claim must comply fully with the
requirements of that section of the act and of the regulations
that implement it.
    257. A food trade association requested that FDA
amend101.100(d) to include section 403(q) and (r) of the act to
provide that products shipped in bulk for further processing,
labeling, or repacking in substantial quantities at an
establishment other than where originally processed or packed,
are exemptduring the time of introduction into, and movement in,
interstate commerce and during the time of holding in such
establishment.
    FDA agrees that 101.100(d), Exemptions From Food Labeling
Requirements, should be amended to include 403(q) of the act.
Again, this modification merely reflects the rule that FDA is
adopting in 101.9(j)(9). However, for the reason explained in
response to the previous comment, FDA is not granting the request
with respect to section 403(r) of the act. Accordingly, FDA is
amending 21 CFR 101.100(d) to read as follows:

    Except as provided by paragraphs (e) and (f) of this section,
a shipment or other delivery of food which is, in accordance with
the practice of the trade, to be processed, labeled or repacked
in substantial quantity at an establishment other than that where
originally processed or packed, shall be exempt, during the time
of introduction into and movement in interstate commerce and the
time of holding in such establishment, from compliance with the
labeling requirements of section 403 (c), (e), (g),(h), (i), (k),
and (q) of the act if: * * *.''

 Nutrition Labeling of Dietary Supplements of Vitamins and
Minerals

    258. Most comments, including those from supplement
manufacturers and trade associations, supported the general
concept of nutrition labeling for dietary supplements of vitamins
and minerals. One comment, however, suggested that any decisions
on nutrition labeling of vitamin and mineral supplements be
deferred until the agency decides how it intends to regulate
dietary supplements, in general. This comment is referring to
FDA's Task Force on Dietary Supplements. The comment argued that
the proposedlabeling requirement would create a label with large
amounts of information that is of little value to the consumer,
particularly for single vitamin and mineral supplements.
    As pointed out in the supplementary proposal (56 FR 60366 at
60381), section 403(q)(5)(E) of the act states that if a food to
which section 411 of the act applies (i.e., dietary supplements
of vitamins and minerals) contains one or more of the nutrients
required to be listed in nutrition labeling, ``the label or
labeling of such food shall comply with the requirements of
paragraphs (1) and (2) [of section 403(q) of the act] in a manner
which is appropriate for such food and which is specified in
regulations of the Secretary.'' In the supplementary proposal (56
FR 60366 at 60381 through 60382), the agency also stated that
vitamin and mineral supplements are required to bear nutrition
labeling under section 403(q)(5)(C) of the act even if they do
not contain any of the nutrients required to be in nutrition
labeling. Section 403(q)(5)(C) of the act provides that nutrition
labeling is not required in such circumstances unless a claim is
made about the nutritional value of the food. The statement of
identity for supplements of vitamins or minerals, including
single vitamins or minerals, may be a claim about the nutritional
value of the supplement. However, because the DS Act imposes a
moratorium on the implementation of the 1990 amendments, FDA is
not taking any action to implement section 403(q)(5)(c) of the
act at this time.
    FDA's Task Force on Dietary Supplements is consequently
irrelevant to this issue. Nothing in the Task Force's report
would relieve the agency of the obligation to adopt regulations
to implement the explicit requirements of the law.
    Furthermore, the agency does not agree that nutrition
information for vitamin and mineral supplements is of little
value to the consumer. These products are represented and sold
for their vitamin and mineral content. Thus, nutrition
information about them will assist consumers in determining the
role that the products can play in maintaining healthy dietary
practices. Moreover, the agency notes that most vitamin and
mineral supplements already bear nutrition information on their
labels.
    For the reasons stated, the agency tentatively concluded that
it is not necessary to modify its requirement for mandatory
nutrition labeling on labels of dietary supplements of vitamins
and minerals. However, the agency will propose its position on
this issue based on the provisions of the DS Act.
    259. In testimony at one of the public meetings held by FDA,
there was a comment suggesting that small packages of dietary
supplements of vitamins and minerals be exempt from mandatory
nutrition labeling.
    Section 403(q)(5)(B) of the act provides for an exemption for
foods in small packages ``if the label of such foods does not
contain any nutrition information.'' This provision is
implemented in 101.9(j)(13)(i). Thus, the question is raised as
towhether the statement of identity for supplements of vitamins
and minerals is a claim. FDA will address this question and the
implementation of section 403(q)(5)(B) with respect to dietary
supplements in accordance with the terms of the DS Act.
    260. One comment recommended deleting proposed101.36(b)(1)
that required the number of units recommended per day to be
declared in the nutrition label on the basis that it is
repetitious of information that is on the label in other places.
    Likewise, a few comments were opposed to
proposed101.36(b)(2) that required declaration of ``Units per
container'' in the nutrition label. These comments asserted that
such a requirement is redundant and unnecessary because the
number of units per container is already listed on the principal
display panel of dietary supplements as part of the net contents
declaration.
    FDA considered these comments and agrees that, for
supplements in which the unit is a discretely defined unit (e.g.,
tablets or capsules), ``Units per day'' could be confusing. The
agency is concerned that consumers could be confused by a
statement that indicates that several units are to be taken per
day (e.g., 3 tablets per day) when the nutrient information is
given for one unit. If consumers do not look at the column legend
that states that the nutrient information is ``per unit,'' they
might assume that the nutrient information is for the amount
specified for consumption per day. To avoid the possibility for
confusion, FDA tentatively concludes that the subheading ``Each
unit contains:'' should be allowed for supplements in which the
unit is a discretely defined unit (e.g., tablets or capsules).
Directions concerning the number of units to be consumed per day
should be given outside of the nutrition label.
    The agency also agrees that, since 101.105(a) requires the
net quantity of content declaration to include a numerical count
when appropriate, there is little benefit to be derived from
information on the number of units per container appearing in two
different places on the label. However, when the supplement is in
a liquid or powdered form, FDA believes additional information
similar to that on conventional foods best informs the consumer
about the dosage unit. Therefore, FDA tentatively concludes that
for dietary supplements of vitamins and minerals in liquid or
powdered form, ``Serving size'' and ``Servings per container''
should be stated consistent with 101.9(d). The agency will
propose its position on these issues in the rulemaking required
by the DS Act.
    261. Several comments, mostly from the dietary supplement
industry, opposed the dual labeling of nutrient content ``per
unit'' and ``per day'' if more than one unit is specified for
consumption per day. Comments argued that dual declaration is
impractical and will result in overcrowding of already small
labels, creating consumer confusion and obfuscating the label's
message. Other arguments against dual declaration were that such
a requirement may discriminate against supplements that are not
in the one-per-day format, and that it would force the industry
to reformulate products so that labels can accommodate all of the
information. One comment pointed out that the proposed regulation
does not address how required information should be presented
when the recommended daily dose is a range, e.g., 1 to 3 tablets
per day.
    Among the comments opposing dual declaration, however, there
was disagreement as to which declaration is preferable, ``per
unit'' or ``per day.'' Some comments stated that it was the total
amount of nutrients that is important, and therefore, declaration
should be on a ``per day'' basis. These comments pointed out that
FDA regulations (105.77) promulgated in 1973 specified that
dietary supplements be labeled according to the quantity
specified for consumption during one day. The comment stated that
although these regulations were withdrawn in 1979, most companies
still comply with them.
    Other comments stated that consumers may deviate from the
recommended dose and should be given credit for being able to
multiply quantities of nutrients by the number of units consumed.
Therefore, these comments stated that declaration should be on a
``per unit'' basis. Comments pointed out that the U.S.
Pharmacopeia is developing quality standards for dietary
supplements in which they propose that nutrient information be
presented ``per dosage unit.''
    Other comments suggested that as an alternative to just one
form of declaration on the label, the label could reference other
labeling such as package inserts that contain all of the required
information, or could permit either ``per unit'' or ``perday''
listing as long as the label clearly states which type of
information is provided. A few comments favored dual declaration.
One comment stated that omitting either declaration might confuse
people who think that the nutrition information for one unit
applies to a day or vice versa.
    The agency is persuaded that dual declaration of nutrition
information ``per unit'' and ``per day,'' when a daily dose of
more than one unit is recommended, may create a readability
problem for consumers, given the limited label space available on
most dietary supplement products. FDA also agrees that
recommended daily consumption of other than well defined dosages
(e.g., ``consume 1 to 3 tablets per day'') would pose a problem
in terms of labeling on a ``per day'' basis.
    FDA is concerned that consumers have nutrition information
available at the point of purchase upon which to base purchase
decisions. Therefore, the agency is not considering package
inserts which could be viewed only after purchase of the product.
Additionally, rather than allowing manufacturers to label on a
``per day'' or ``per unit'' basis, the agency favors one
consistent method of labeling. A consistent method will allow
consumer education programs to explain how nutrition labeling is
to always appear and to teach consumers how to calculate their
individual consumption levels if their intake differs from the
amount specified within the nutrition label. The agency believes
labeling ``per unit'' is more useful in that the product will
always beconsumed ``per unit,'' however, consumers may not always
follow a manufacturer's recommendation to consume a certain
number of units per day and therefore may not actually consume
the amount indicated ``per day.''
    For these reasons, and to harmonize with the U.S.
Pharmacopeia, the agency tentatively concludes that nutrition
information should be declared on a ``per unit'' basis. FDA
intends to propose its position in the rulemaking that is
required under the DS
    262. In the supplementary proposal, FDA proposed that
nutrition labels for dietary supplements of vitamins and minerals
include a column of quantitative amounts by weight and a second
column of percent of RDI's, expressed as ``Percent Daily
Value.''Comments were requested on the usefulness of a list of
DRV's and the percent of the DRV for fat, saturated fat,
cholesterol, carbohydrate, dietary fiber, and sodium provided by
the supplement when they are declared (i.e., when they are
present in the supplement in more than insignificant amounts) (56
FR 60366 at 60383). In the format proposal, FDA stated that it
anticipated modifying 101.36, Nutrition Labeling of Dietary
Supplements, to be as consistent as possible with the nutrition
labeling of other foods and requested comment (57 FR 32058 at
32072).
    Several comments to both the supplementary proposal and the
format proposal addressed the format for declaring amounts of
nutrients present. About half of the comments supported FDA's
position. However, one comment argued that the unique
characteristics of dietary supplements demand a different
approach to their nutrition labeling. Characteristics identified
included: (1) The vast majority of supplements are marketed in
relatively small packages, (2) the nutrition profiles for these
products typically reflect high levels of micronutrients and
relatively insignificant amounts of macronutrients, (3) consumers
look for and expect nutrition information on supplements that is
different from that on conventional foods, and (4) consumers of
supplements will already be asked to search through an array of
nutrient names and units of measure to find the information they
look for most: The percentage of their daily nutritional
requirements that the supplement provides.
    One comment from a manufacturer stated that there was no need
to make significant changes in dietary supplement labels because
current labels that have been used for many years are widely
accepted and present the necessary data on vitamins and minerals
in a logical and readily understandable form. A comment from
another manufacturer opposed the required declaration in separate
columns of quantitative amounts by weight of nutrients and by the
percent RDI or percent DRV (expressed as ``Percent of Daily
Value''). The comment argued that only percent of daily value
should be mandatory, and that listing of quantitative amounts by
weight should be voluntary, because there is no congressional
mandate to list quantitative amounts on two bases, no agency
justification that two bases are useful to consumers, and a
potential to confuse consumers with little understood terms,
e.g., mg alpha-tocopherol. The comment also asserted that a
requirement for too much information is discriminatory against
products with larger numbers of nutrients and might discourage
use of smaller packages that are less expensive to consumers. The
comment also stated that a requirement for declaration of only
percent of daily value would be consistent with the requirement
for vitamins and minerals on conventional food labels.
    A few comments objected to the required inclusion of a list
of daily values in addition to the quantitative amounts by weight
and the percent of daily value on the label. The comments stated
that this additional information will produce an even
morecluttered appearance and further contribute to the
proliferation of numerical values on dietary supplement labels.
One comment argued that ``The goal of meeting the supplement
consumer's need for relevant, comprehensible nutrition
information should notbe sacrificed out of a blind concern for
consistency.'' The comment concluded that consumers of
supplements are already familiar and comfortable with the concept
of percent of daily value and their focus on this information
should not be diverted by additional unnecessary and potentially
confusing information.
    While FDA continues to believe it is helpful to consumers to
minimize inconsistencies in the label format between types of
foods, the agency is persuaded that the unique characteristics of
dietary supplements require a reevaluation of whether the
formould be carried over to dietary supplements. For example, the
agency believes that the declaration of quantitative amounts on
two bases (i.e., both by weight and by percent of daily value)
needs to be considered for dietary supplements in terms of its
usefulness to consumers. In that regard, the agency considers
dietary supplement consumers to have special needs for
quantitative nutrition information about the products they use by
virtue of the way such products are formulated, marketed, and
used. Dietary supplements are often formulated and marketed on
the basis of offering specific amounts of certain nutrients to
consumers. Dietary supplement product users are often trying to
maintain a certain quantitative intake of specific nutrients in
their diets and use the product to obtain this quantitative goal.
Some of the nutrients contained in dietary supplements and
declared on the nutrition label are not well known to many
consumers. The quantitative goals that are importantly relevant
to consumption of dietary supplement products may be stated in
various units including units of weight or of percent of RDI's or
DRV's. FDA intends to address this issue in the rulemaking that
it will undertake in response to the DS Act.
    In its reevaluation of format requirements for dietary
supplements, the agency also looked at the requirement in
nutrition labeling of conventional foods for a list of daily
values for all nutrients declared on the label. After careful
considerationof the comments, the agency tentatively concludes
that it is not necessary or appropriate to require the inclusion
of the DV list on dietary supplements. Because of the small size
of most supplement packages and the duplication of the more
complex nomenclature of units for vitamins (e.g., mg
alpha-tocopherol) that would be required in a DV list, the agency
believes that the added complexity and proliferation of numerical
values would interfere with consumers use of the quantitative
information by weight and by percent of daily value. FDA will
propose its position regarding the format of the nutrition label
for dietary supplements based on the provisions of the DS Act.
     263. A few comments opposed the requirement for declaration
of the quantitative amount and the percent of the DRV of fat,
saturated fat, cholesterol, carbohydrate, dietary fiber, sodium,
and potassium when these nutrients are present in a supplement in
more than insignificant amounts. One comment suggested that the
declaration either be optional or be required only when these
nutrients are present at levels greater than 10 percent of their
respective DRV's. The comment stated that: (1) Excessive and
useless information would detract from the importance of a
product's vitamin and mineral content; (2) even though the vast
majority of supplements lack these substances, all products would
have to undergo extensive and expensive testing to determine
whether listing of these components is necessary, thus burdening
small companies with diverse supplement product lines; and (3)
these requirements would hinder product development and increase
the cost of bringing innovative products to market. Another
comment stated that declaration of fat should be required only
for fatty acid supplements of 1 g or more per unit since
declaration of smaller amounts would clutter the label and be
difficult to read.
    As discussed in the preceding comment, FDA agrees that the
declaration of the amount of the DRV's (i.e., the DV list) is not
necessary on labels of dietary supplements of vitamins and
minerals. However, FDA continues to believe that the
quantitativeamount and the percent of the DRV should be declared
for total fat, saturated fat, cholesterol, total carbohydrate,
dietary fiber, sodium, and potassium when these nutrients are
present in a supplement in more than insignificant amounts.
Information abouttions, and the general public. Moreover,
supplements are formulated products, and manufacturers should
know from the ingredients that they use to make these products
whether a supplement contains more than insignificant amounts of
any of the nutrients for which DRV's have been established. When
such ingredients are not used, laboratory analyses for such
nutrients would be unnecessary. As discussed in comment 176 of
this document, the definition of insignificant has been modified
so that an ``insignificant amount'' of total carbohydrate,
dietary fiber, and protein is an amount that allows a declaration
of ``less than 1 g'' in the nutrition label. The agency will
propose its position on this issue in the rulemaking that it does
in response to the DS Act.
    264. Several comments, predominantly from manufacturers and
trade associations, disagreed with FDA's statement in the
supplementary proposal that dietary supplements of selenium and
chromium are not permitted because there are no regulations
declaring these nutrients as approved food additives, GRAS, or
prior-sanctioned ingredients. A trade association stated that
nutrition labeling regulations are not the appropriate place to
announce decisions about the GRAS status of nutrients. Comments
argued that selenium and chromium are recognized as essential
nutrients for humans. They pointed out that the National Research
Council's Food and Nutrition Board has established a Recommended
Dietary Allowance for selenium and an ESADDI for chromium (Ref.
23), and that FDA proposed RDI's for these nutrients. Moreover,
comments stated that FDA has advised for many years that these
trace minerals are ``safe and suitable'' for use in supplements.
The comments argued that, therefore, FDA should adopt a general
policy that trace minerals for which a Recommended Dietary
Allowance or ESADDI has been established are GRAS, at least at
levels not to exceed their respective Recommended Dietary
Allowance/ESADDI.
    The agency agrees that this rulemaking is not the appropriate
place to announce decisions about the GRAS status of nutrients.
Therefore, FDA reiterates that there are currently no sources of
selenium or chromium that are either affirmed as GRAS or approved
food additives for addition to human food. Any direct addition of
these trace minerals to a food is based solely on the
manufacturer's judgment that the nutrient sources are GRAS and is
not sanctioned by the agency.
    265. One comment advocated the parenthetical listing of the
source of each vitamin or mineral immediately following its
declaration on the nutrition information panel in lieu of a
separate ingredient list. The comment argued that this listing
would avoid confusion by enabling consumers to readily identify
the nutrient source and would save limited label space.
Furthermore, the comment stated that it is already common
practice in the supplement industry. The comment suggested that
information aboutthe source of the nutrient would allow the
consumer to identify whether the source is the most
physiologically desirable, e.g., beta-carotene versus vitamin A
palmitate.
    FDA advises that dietary supplements, like any food, are
required to bear a complete list of ingredients under section
403(l)(2) of the act, and such list should not be confused with
the nutrition label. Ingredient listing, moreover, is needed for
substances other than vitamins or minerals, e.g., lactose, other
fillers, artificial colors, flavors, binders, and excipients.
Consumers desiring to know the source of a nutrient can merely
look at the list of ingredients just as they would for a
conventional food product. Therefore, in accordance with
ingredient labeling regulations, the specific source of vitamin A
must be shown in the ingredient list.
    However, in response to this and another similar comment (see
comment 81 of this document), the agency is allowing for the
declaration of the percent of vitamin A present as beta-carotene
in 101.9(c)(8)(vi). The agency will propose its position ren in
nutrition labeling regulations pertaining to dietary supplements
of vitamins and minerals following provisions of the DS Act.
    266. One comment objected to the listing of the quantitative
amounts of vitamins and minerals to the nearest unit of the same
level of significance at which the RDI's are specified in
101.9(c)(8)(iv). The comment stated that it would be potentially
confusing to consumers for thiamin, for example, to be declared
to the first decimal place, e.g., 100.0 mg, and niacin to be
declared to the nearest whole number, e.g., 100 mg. The comment
suggested that decimal places be dropped, and that all nutrients
be listed to the nearest whole number when nutrient levels are
ten or more times the RDI.
    While FDA intends to deal with this issue in its rulemaking
that responds to the DS Act, the agency offers the following
comments. FDA is not persuaded that consumers would be confused
by decimals for some nutrients and not others. In addition,
requiring only whole numbers would introduce a large amount of
imprecision in the declarations of some nutrients. For example,
it would cause 1.5 mg of thiamin (i.e., 100 percent of the RDI)
to be rounded up to 2 mg--a 33 percent increase.
    However, when the decimal is followed by a zero, the agency
generally has no objection to the zero being dropped. In fact, in
the supplementary proposal, this was done in the declaration of
the amount of vitamin B12 in the hypothetical sample label for
``Daily Vitamins Plus Iron'' (56 FR 60366 at 60383). Since RDI's
in 101.9(c)(8)(iv) are established only in whole numbers or in
tenths of a unit, allowing zeros following decimals to be
dropped, in effect, allows all nutrients to be declared to
thenearest whole number when nutrient levels are ten times the
RDI.
    267. A couple of comments objected to FDA's proposal that
compliance with the requirements for labeling of dietary
supplements be determined in accordance with 101.9(g), i.e., 100
percent of label claim for Class I nutrients. Comments argued
that the 100 percent requirement is unreasonable in that it is
more stringent than United States Pharmacopeia (USP) requirements
for certain vitamin and mineral products, which generally allow
lower limits of 90 percent to 95 percent of label claim.
    FDA intends to deal with this issue in the rulemaking that
responds to the DS Act. However, the agency notes that dietary
supplements are fabricated products. Therefore, the question is
raised why they should not be held to the same Class I
nutrientstandards as conventional foods that are fortified or
enriched. Based on the agency's current compliance policy it has
informed USP that anything less than 100 percent of the label
claim for vitamin and mineral products is not acceptable to FDA,
and that the only permissible deviation from this requirement
would be the variability of the analytical method (Ref. 118).
    The agency notes that, contrary to the statement in the
comments, the General Notices of the USP state that a dosage
should be formulated to provide 100 percent of the labeled amount
(Ref. 119). The limits in the monographs allow for overages of
ingredients known to decrease with time, for analytical error,
for manufacturing and compounding variations, and for
deterioration to an extent considered insignificant under
practical conditions (Ref. 119).
    268. One comment asserted that manufacturers should be
prohibited from labeling a supplement in such a way as to confuse
the weight of a unit of supplement with its nutrient content. For
example, a calcium supplement that contains 250 mg of elemental
calcium as calcium chloride should not be labeled as
``calcium--625 mg'' anywhere on the label.
    FDA concurs that such labeling is potentially misleading to
consumers. Section 403(a) of the act provides that a food will be
deemed to be misbranded if its labeling is false or misleading in
any particular. FDA concludes that existing statutory authot is
specified on the label in a manner that is likely to mislead
consumers into thinking that that is the weight of the nutrient
contained in the product if those amounts are different.

IX. Consumer Education Program

    Section 2(c) of the 1990 amendments directs the Secretary
(and FDA, by delegation) to carry out activities that educate
consumers about nutrition information on the food label and the
importance of that information in maintaining healthy dietary
practices. To achieve this purpose, FDA and USDA have jointly
initiated a multi-year food labeling education campaign. The
major goals of this campaign are to increase consumers' knowledge
and effective use of the new food label to make accurate and
sound dietary choices; to integrate food labeling education into
existing and new nutrition and health education programs; and to
build extensive partnerships capable of developing and evaluating
labeling education targeted to the dietary needs of diverse
populations, such as low literacy consumers, minorities, older
Americans, children, and people with dietary restrictions.
    As part of this effort, the agencies have established the
National Exchange on Food Labeling Education which includes an
information center housed in the Food and Nutrition Information
Center at the National Agricultural Library. The National
Exchange on Food Labeling Education provides the general public
and professionals with access to information about food labeling
research and educational activities (projects, programs, and
materials) from both the public and private sector.
    FDA and USDA have also worked to establish cooperative
projects with diverse organizations and to facilitate the
communication of information that targets various subpopulations
as well as the general public. The agencies have thus developed
extensive food label education networks that include consumers,
health professionals and organizations, educators, trade
associations, Federal and local government, and many others to
assist in the dissemination and development of information and
activities.
    To ensure that consumers have accurate and adequate resource
materials and information, the agencies have begun, and will
continue, to conduct and report on existing and planned food
labeling research; to develop education initiatives at the
national and local level; to hold regularly-scheduled meetings to
build labeling education exchanges; to produce videos; and to
produce an array of public education materials, including a
special edition of FDA Consumer magazine that summarizes the
final food labeling regulations, and brochures (in English and
other languages) on the new label and how to use it to meet the
Dietary Guidelines for Americans (Ref. 4). These materials will
be targeted to the general public, nutritionists, such special
groups as ethnic minorities, and others. Organizations will also
be able to use these resource materials to develop educational
materials of their own.

X. Economic Impact

    In its food labeling proposals of November 27, 1991 (56 FR
60366 et seq.), FDA stated that the food labeling reform
initiative, taken as a whole, would have associated costs in
excess of the $100 million threshold that defines a major rule.
Thus, inaccordance with Executive Order 12291 and the Regulatory
Flexibility Act (Pub. L. 96-354), FDA developed one comprehensive
regulatory impact analysis (RIA) that presented the costs and
benefits of all of the food labeling provisions taken together.
ThatRIA was published in the Federal Register of November 27,
1991 (56 FR 60856), and along with the food labeling proposals,
the agency requested comments on the RIA.
    FDA has evaluated more than 300 comments that it received in
response to the November 1991 RIA. FDA's discussion of these
comments is contained in the agency's final RIA published
elsewhere in this issue of the Federal Register. In addition,
FDAng final rules. The final RFA will be placed on file with the
Dockets Management Branch (HFA - 305), Food and Drug
Administration, rm. 1 - 23, 12420 Parklawn Dr., Rockville, MD
20857, and a notice will be published in the Federal Register
announcing its availability.
    In the final RIA, FDA has concluded, based on its review of
available data and comments, that the overall food labeling
reform initiative constitutes a major rule as defined by
Executive Order 12291. Further, the agency has concluded that
although the costs of complying with the new food labeling
requirements are substantial, such costs are outweighed by the
public health benefits that will be realized through the use of
improved nutrition information provided by food labeling.

XI. Environmental Impact

    The agency previously considered the environmental effects of
the action being taken in this final rule. As announced in its
nutrition labeling proposed rules published in the Federal
Register of November 27, 1991 (56 FR 60366 et al.), the agency
determined that under 21 CFR 25.24(a)(11), these actions are of a
type that do not individually or cumulatively have a significant
impact on the human environment. Therefore, neither an
environmental assessment nor an environmental impact statement
was required.
    In its November 1991 nutrition labeling proposed rules, the
agency proposed that the final rules for these actions would
become effective 6 months following their publication in the
Federal Register. Several comments on the nutrition labeling
proposed rules suggested that there would be significant adverse
environmental effects from the final rules unless the agency
allowed more time between the publication of the final rules and
their effective dates. The concern in these comments was that, if
the agency did not allow firms more time between the publication
of the final rules and their effective dates to use up existing
label inventories, large stocks of labels and labeled packaging
would have to be discarded. These comments questioned whether the
agency had sufficiently examined the impact of disposing of
obsolete labels and labeled packaging on this country's solid
waste disposal capabilities. Two comments estimated the amounts
of labeling from their respective industries, i.e., dairy and
confectionery, that would need to be discarded following
publication of FDA's final rules on several food labeling
actions, including this action. However, these comments did not:
(1) Provide details on how these estimates were derived, (2)
identify what portion of the estimated amounts are attributable
to these two actions, or (3) describe what impact the discarded
labels and packaging would have on the disposal of solid waste.
    Based on its review of available data and comments received,
the agency has decided to allow additional time for companies to
use up their old labels. Thus, the nutrition labeling final rules
will not be effective until May 8, 1994. FDA believes there will
thus be ample time for food companies to use up most of the
existing labeling and packaging stocks and to incorporate
labeling language that complies with FDA's regulations into their
food labels. Consequently, the comments on the potential for
adverse environmental effects do not affect the agency's previous
determination that no significant impact on the human environment
is expected and that an environmental impact statement is not
required.

XII. References

    The following references have been placed on display in the
Dockets Management Branch (address above) and may be seen by
interested persons between 9 a.m. and 4 p.m., Monday through
Friday.

    1. Committee on the Nutrition Components of Food Labeling,
Food and Nutrition Board, IOM, National Academy of Sciences,
``Nutrition Labeling, Issues and Directions for the 1990s,''
Washington, DC, National Academy Press, 1990.
    2. U.S. Department of Health and Human Services (DHHS),
Public Health Service (PHS), ``The Surgeon General's Report on
Nutrition and Health,'' Washington, DC, DHHS (PHS) Publication
No. 88-50210, U.S. Government Printing Office, 1988.
    3. Committee on Diet and Health, Food and Nutrition Board,
Commission on Life Sciences, National Research Council, ``Diet
and Health: Implications for Reducing Chronic Disease Risk,''
Washington, DC, National Academy Press, 1989.
    3a. Letter to Senator Frank Lautenberg from Hugh C. Cannon,
dated March 4, 1991.
    4. USDA and DHHS, ``Nutrition and Your Health, Dietary
Guidelines for Americans,'' Washington, DC, Home and Garden
Bulletin No. 232, 3d ed., U.S. Government Printing Office, 1990.
    5. Minister of Supply and Services, Canada, ``Departmental
Consolidation of the Food and Drugs Act and of the Food and Drug
Regulations with Amendments to May 3, 1990, Department of
National Health and Welfare, Canada, 1991.
    6. Joint FAO/WHO Food Standards Program, ``Codex Standards
and Guidelines for the Labeling of Foods and Food Additives,''
Codex Alimentarius Commission, vol. VI, 2d ed., Rome, 1987.
    7. Saccomandi, V., ``Council Directive of 24 September 1990
on Nutrition Labeling for Foodstuffs,'' Council of the European
Communities, Brussels, September 24, 1990.
    8. FAO of the United Nations, ``Protein Quality Evaluation,
Report of a Joint FAO/WHO Expert Consultation,'' FAO of the
United Nations, Rome, and WHO, Geneva, 1990.
    9. Young, V. R. and P. L. Pellett, ``Protein Evaluation,
Amino Acid Scoring and the Food and Drug Administration Proposed
Food Labeling Regulations,'' Journal of Nutrition, 121:145 - 150,
1991.
    9a. Codex Alimentarius Commission, Report of the 18th
Session, Geneva, 3 - 12 July 1989, Alinorm 89/40, FAO/WHO, Rome,
Italy, 1989.
    10. Joint FAO/WHO/United Nations University Expert
Consultation, ``Energy and Protein Requirements,'' Tech. Report
Ser. No. 724, WHO, Geneva, Switzerland, 1985.
    10a. Codex Alimentarius Commission, Report of the 17th
Session of the Codex Committee on Nutrition and Foods for Special
Dietary Uses, Bonn-Bad Godesberg, Germany, 18 - 22 February 1991,
Alinorm 91/26, FAO/WHO, Rome, Italy, 1991.
    11. Market Facts, Inc., ``Summary Report, FDA Nutrition
Labeling Focus Groups,'' prepared under FDA Contract No. 223 - 89
- 2170, Washington, DC, FDA, DHHS, 1990.
    12. Mensink, R. P. and M. B. Katan, ``Effect of Dietary Trans
Fatty Acids on High-Density and Low-Density Lipoprotein
Cholesterol Levels in Healthy Subjects,'' The New England Journal
of Medicine, 323:439 - 445, 1990.
    13. Grundy, S. M., ``Trans Monounsaturated Fatty Acids and
Serum Cholesterol Levels,'' The New England Journal of Medicine,
323:480 - 1, 1990.
    14. Levy, A. S., S. B. Fein, and R. E. Schucker, Division of
Consumer Studies (HFF 240), Center for Food Safety and Applied
Nutrition, FDA, ``A Study of Nutrition Label Formats: Performance
and Preference,'' 200 C St. SW., Washington, DC 20204, March
1991.
    15. NFPA, ``Summary of Findings, Food Labeling and Nutrition
* * * What Americans Want,'' Washington, DC, 1990.
    16. Committee on Energy and Commerce, Report 101 - 538,
Nutrition Labeling and Education Act of 1990, U.S. House of
Representatives, June 13, 1990.
    17. Stokes, R. C. and R. Haddock, ``Interim Report of the
First Two Phases of the CRI/FDA Nutritional Labeling Research
Program,'' August 1972, Consumer Research Institute, Inc.,
Washington, DC.
    18. Park, Youngmee K., Division of Nutrition (HFF - 265),
Center for Food Safety and Applied Nutrition, FDA, memo to file,
March 15, 1991.
    19. Bender, M., Division of Consumer Studies (HFF - 240),
Center for Food Safety and Applied Nutrition, memo to file,
January 18, 1991.
    20. FDA, ``Compliance Program Guidance Manual,'' Chapter 21,
Program No. 7321.002, FDA, 1990.
    21. American Society for Parenteral and Enteral Nutrition
(A.S.P.E.N.), ``Definitions and Terms Used in A.S.P.E.N.
Guidelines and Standards,'' A.S.P.E.N., Silver Spring, MD, 1988.
    22. Talbot, J. M., ``Guidelines for the Scientific Review of
Enteral Food Products for Special Medical Uses,'' prepared for
FDA under Contract No. FDA 223 - 88 - 2124 by the Life Sciences
Research Office, Bethesda, MD, 1990.
    23. Subcommittee on the 10th Edition of the Recommended
Dietary Allowances, Food and Nutrition Board, Commission on Life
Sciences, National Research Council, ``Recommended Dietary
Allowances, 10th Edition,'' Washington, DC, National Academy
Press, 1989.
    24. Food and Nutrition Board, Division of Biology and
Agriculture, National Research Council, ``Recommended Dietary
Allowances, 7th ed., 1968,'' Publication 1694, Printing and
Publishing Office, National Academy of Sciences, Washington, DC,
1968.
    25. Congressional Record, July 30, 1990, H5840 - 41.
    26. USDA, ``USDA's Food Guide Pyramid,'' Human Nutrition
Information Service, Hyattsville, MD, Home and Garden Bulletin
Number 249, 1992.
    27. USDA, ``Nutrition and Your Health, Dietary Guidelines for
Americans, Avoid Too Much Fat, Saturated Fat, and Cholesterol,''
Human Nutrition Information Service, Hyattsville, MD, Home and
Garden Bulletin, No. 232 - 3, 1986.
    28. ``Handy Reference for Grams of Fat Allowed at Thirty
Percent Fat Level,'' Environmental Nutrition, vol. 11, No. 11, p.
3, November 1988.
    29. McNutt, K., ``Nutritious Science or Does Language Limit
Our Nutrition Communications?'', Nutrition Today, vol. 23:4, p.
19, August 1988.
    30. National Heart, Lung, and Blood Institute, ``Report of
the Expert Panel on Population Strategies for Blood Cholesterol
Reduction,'' NCEP, PHS, National Institutes of Health, DHHS,
Washington, DC; NIH Publication No. 90 - 3046, 1990.
    31. Grundy, S. M., ``Evaluation of Publicly Available
Scientific Evidence Regarding Certain Nutrient-Disease
Relationships, a. Lipids and Cardiovascular Disease,'' Prepared
for The Center for Food Safety and Applied Nutrition, FDA, FDA
contract no. 223 - 882123, Task Order No. 7, Life Sciences
Research Office, Federation of American Societies for
Experimental Biology, 1991.
    32. Carpenter, D. E. and S. Lee, ``AOAC Methods and
Determination of Fat,'' The Referee, AOAC International, vol. 16,
No. 9, p. 1, September 1992.
    33. Peters, J. C., B. N. Holcombe, L. K. Hiller, and D. R.
Webb, ``Caprenin 3. Absorption and Caloric Value in Adult
Humans,'' Journal of the American College of Toxicology, 10:357 -
367, 1991.
    34. Hoak J. C. Ed., ``Workshop on Dietary Fatty Acids and
Thrombosis,'' Sponsored by the American Cocoa Research Institute;
Institute of Shortening and Edible Oils; International Life
Science Institute--Nutrition Foundation; National Heart, Lung and
Blood Institute; National Live Stock and Meat Board; USDA; and
FDA, 1991.
    35. Khosla P. and K. C. Hayes, ``Comparison Between the
Effects of Dietary Saturated (16:0), Monounsaturated (18:1), and
Polyunsaturated (18:2) Fatty Acids on Plasma Lipoprotein
Metabolism in Cebus and Rhesus Monkeys Fed Cholesterol-Free
Diets,'' American Journal of Clinical Nutrition, 55:51 - 61, 1992
    36. Hayes K. C. and P. Khosla, ``Dietary Fatty Acid
Thresholds and Cholesterolemia,'' Journal of the Federation of
American Societies for Experimental Biology, 6:2600 - 2607, 1992.
    37. National Heart, Lung, and Blood Institute, ``Report of
the National Cholesterol Education Program Expert Panel on
Detection, Evaluation, and Treatment of High Blood Cholesterol in
Adults,'' Archives of Internal Medicine, 148:36 - 69, 1988.
    38. Hegsted, D. M., R. B. McGandy, M. L. Myers, and F. J.
Stare, ``Quantitative Effects of Dietary Fat on Serum Cholesterol
in Man,'' American Journal of Clinical Nutrition, 17:281 - 95,
1965.
    39. Keys, A., J. T. Anderson, and F. Grande, ``Serum
Cholesterol Response to Changes in the Diet IV. Particular
Saturated Fatty Acids in the Diet,'' Metabolism, 14:776 - 87,
1965.
    40. Grundy S. M. and M. A. Denke, ``Dietary Influences on
Serum Lipids and Lipoproteins,'' review, Journal of Lipid
Research, 31:1149 - 72, 1990.
    41. WHO Technical Report, ``Diet, Nutrition, and the
Prevention of Chronic Disease,'' WHO, Geneva, 1990.
    42. Life Sciences Research Office, Federation of American
Societies for Experimental Biology, ``Health Aspects of Dietary
Trans Fatty Acids,'' Federation of American Societies for
Experimental Biology, Contract Number FDA 233 - 83 - 2020,
Bethesda, MD, 1985.
    43. Zock, P. L. and M. B. Katan, ``Hydrogenation
Alternatives: Effects of trans Fatty Acids and Stearic Acid
Versus Linoleic Acid on Serum Lipids and Lipoprotein in Human,''
Journal of Lipid Research, 33:399 - 410, 1992.
    44. Nestel, P. J., M. Noakes, G. B. Belling, R. McArthur, P.
M. Clifton, and M. Abbey, ``Plasma Cholesterol-Lowering Potential
of Edible-Oil Blends Suitable for Commercial Use,'' American
Journal of Clinical Nutrition, 55:46 - 50, 1992.
    45. USDA, ``Summary of Trans Fatty Acid Study Results,''
Agriculture Research Service, Beltsville, MD, July 27, 1992.
    46. Raper, N. R., F. J. Cronin, and J. Exler, ``Omega-3 Fatty
Acid Content of the U.S. Food Supply,'' Journal of the American
College of Nutrition, 11:304 - 8, 1992.
    47. DHHS, PHS, ``Healthy People 2000: National Health
Promotion and Disease Prevention Objectives,'' DHHS (PHS)
Publication No. 91 - 50213, DHHS, Washington, DC, 1991.
    48. American Heart Association, Committee on Nutrition,
``Dietary Guidelines for Healthy American Adults,'' Circulation,
74:1465a, 1986.
    49. Life Sciences Research Office, Federation of American
Societies for Experimental Biology, ``Nutrition Monitoring in the
United States. An Update Report on Nutrition Monitoring,''
Prepared for DHHS, Government Printing Office, Washington, DC,
1989.
    50. Lee, S. C., L. Prosky, J. W. DeVries, ``Determination of
Total, Soluble, and Insoluble Dietary Fiber in
Foods--Enzymatic-Gravimetric Method, MES - TRIS Buffer:
Collaborative Study,'' Journal of AOAC International, 75:395,
1992.
    51. Glinsman, W. H., H. Irausquin, and Y. K. Park, ``Report
from FDA's Sugars Task Force; Evaluation of Health Aspects of
Sugars Contained in Carbohydrate Sweeteners,'' Journal of
Nutrition, vol. 116(11S), pp.S1 - S16, 1986, 1986.
    52. Cotter, R., Gerber Products Co., Letter to Dockets
Management Branch, FDA, February 24, 1992.
    53. Bush, G., ``Memorandum for Certain Department and Agency
Heads, Subject: Reducing the Burden of Government Regulation,''
The White House, Washington, DC, January 28, 1992.
    54. Young, V. R., Massachusetts Institute of Technology,
Letter to Dockets Management Branch, FDA, February 24, 1992.
    55. Marcus, M., Protein Technologies International, Letter to
Dockets Management Branch, FDA, February 13, 1992.
    56. Sawar, G., H. G. Botting, and R. W. Peace, ``Amino Acid
Rating Method for Evaluating Protein Adequacy of Infant
Formulas,'' Journal of the Association of Official Analytical
Chemists, 72:622 - 6, 1989.
    57. Miller, S. A., ``A Statement by Sanford A. Miller,
Director, Bureau of Foods, Food and Drug Administration Before
the Subcommittee on Natural Resources, Agriculture Research and
Environment, Committee on Science and Technology,'' U.S. House of
Representatives, Washington, DC, April 29, 1982.
    58. AOAC International, ``Nutrient Labeling Task Force
Report,'' The Referee, AOAC International, vol. 16 No. 7, p. 1,
July 1992.
    59. Sheppard, A. J., ``Lipid Manual, Methodology Appropriate
for Fatty Acid-Cholesterol Analysis,'' Wm. C. Brown Publishers,
Dubuque, Iowa, 1992.
    60. FDA, ``Food Additives Analytical Manual,'' Published by
AOAC, Arlington, VA, vol. I, 1983 and vol. II, 1987.
    61. Life Sciences Research Office, ``A Perspective on the
Application of the Atwater System of Food Energy Assessment,''
Federation of American Societies for Experimental Biology,
Bethesda, MD, 1983.
    62. USDA, ``Composition of Foods: Finfish and Shellfish
Products,'' Agriculture Handbook 8 - 15, Human Nutrition
Information Service, USDA, p. 91, September 1987.
    63. Merrill, A. L. and B. K. Watt, ``Energy Value of Foods *
* * Basis and Derivation,'' USDA, Agriculture Handbook No. 74,
Slightly Revised 1973.
    64. Horowitz, W., A. Albert, M. J. Deutsch, ``Precision
Parameters of Methods of Analysis Required for Nutrition
Labeling. Part 1. Major Nutrients,'' Journal of Association of
Official Analytical Chemists, 73:661, 1990.
    65. ``Position of the American Dietetic Association:
Nutrition and Health Information of Food Labels,'' Journal of the
American Dietetic Association, 90:583 - 585, 1990.
    66. American Institute of Nutrition and the American Society
of Clinical Nutrition, ``Position Statement on Food Labeling,''
Journal of Nutrition, 120:534 - 536, 1990.
    67. 136 Congressional Record H5836 - H5845 (July 30, 1990)
(Statement by Congressman Waxman).
    68. 136 Congressional Record S16607 - S16612 (October 24,
1990) (Statement by Senator Hatch).
    69. Levy, A., S. B. Fein, and R. E. Schucker, ``More
Effective Nutrition Label Formats are not Necessarily More
Preferred,'' Journal of the American Dietetic Association,
92:1230 - 1234, 1992.
    70. Levy, A. S., S. B. Fein, and R. E. Schucker, ``Evaluation
of Nutrition Label Formats: FDA Study 2,'' Center for Food Safety
and Applied Nutrition, FDA, Washington, DC, March, 1992.
    71. Gardner, T. S. and E. T. Guarino, Grocery Manufacturers
of America, Inc., Letter to Dockets Management Branch, June 3,
1992.
    72. Geiger, C. J., University of Utah, Letter to Dockets
Management Branch, August 18, 1992.
    73. Byrd-Bredbenner, C., Montclair State College, Letter to
Dockets Management Branch, August 17, 1992.
    74. Frito-Lay, Inc., ``Nutritional Labeling Comprehension
Test,'' 1991 and 1992.
    75. Burton, S., Louisiana State University, Letter to Dockets
Management Branch, July 14, 1992.
    76. Jacoby, J., R. W. Chestnut, and W. Silberman, ``Consumer
Use and Comprehension of Nutrition Information,'' Journal of
Consumer Research, 4:119 - 128, 1977.
    77. Sproles, G., L. V. Geistfeld, and S. B. Badenhop,
``Informational Inputs as Influences on Efficient Consumer
Decision-Making,'' Journal of Consumer Affairs, 12:88 - 103,
1978.
    78. Sproles, G., L. V. Geistfeld, and S. B. Badenhop, ``Types
and amounts of Information Used by Efficient Consumers,'' Journal
of Consumer Affairs, 14:37 - 48, 1980.
    79. Scammon, D. L., ``Information Load and Consumers,''
Journal of Consumer Research, 4:148 - 155, 1978.
    80. Patton, III, W. E., ``Quantity of Information and
Information Display Type as Predictors of Consumer Choice of
Product Brands,'' Journal of Consumer Affairs, 15:92 - 105, 1981.
    81. Asam, E. H. and L. P. Bucklin, ``Nutrition Labeling for
Canned Goods: A Study of Consumer Response,'' Journal of
Marketing, 37:32 - 37, 1973.
    82. ORC (Opinion Research Corporation), ``Food Labeling and
Nutrition: What Americans Want,'' report prepared for the NFPA,
Washington, DC, 1990.
    83. Geiger, C. J., B. W. Wyse, C. R. M. Parent, and R. G.
Hansen, ``Nutrition Labels in Bar Graph Format Deemed Most Useful
for Consumer Purchase Decisions Using Adaptive Conjoint
Analysis,'' Journal of the American Dietetic Association, 91:800
- 807, 1991.
    84. Keller, K. L. and R. Staelin, ``Effects of Quality and
Quantity of Information on Decision Effectiveness,'' Journal of
Consumer Research, 14:200 - 213, 1987.
    85. Lewis, C. J. and E. A. Yetley, ``Focus Group Sessions on
Formats of Nutrition Labels,'' Journal of The American Dietetic
Association, 92:62 - 66, 1992.
    86. Market Facts, ``Summary Report: FDA Nutrition Labeling
Focus Groups,'' Report prepared for FDA, Washington, DC 1992.
    87. Ballin, S. D., American Heart Association, Letter to
Dockets Management Branch, August 8, 1992.
    88. Costco Connection, September 1991.
    89. Washington Post, July 29, 1992.
    90. Nutrition Action Newsletter, July/August 1992.
    91. USA Today, July 17, 1992.
    92. Daily Herald, October 24, 1991.
    93. Atlanta Constitution, July 23, 1992.
    94. New York Times, July 22, 1992.
    95. Liebman, B. and M. Jacobson, Center for Science in the
Public Interest, Letter to Dockets Management Branch, August 21,
1992.
    96. Roper Organization, Inc., for American Meat Institute,
``Nutrition Labels and Meat: Consumer Knowledge, Usage, and
Attitudes,'' 1992.
    97. Schramm, M. J., Kellogg Co., Letter to Dockets Management
Branch, August 19, 1992.
    98. Golodner, L. F., National Consumers League, Letter to
Dockets Management Branch, August 19, 1992.
    99. Food Marketing Institute, ``Trends: Consumer Attitudes
and the Supermarket, 1992,'' Washington, DC, 1992.
    100. FDA, Health and Diet Survey, 1990.
    101. Bettman, J. R. and P. Kakkar, ``Effects of Information
Presentation Format on Consumer Information Acquisition
Strategies,'' Journal of Consumer Research, 3:233 - 240, 1976.
    102. Levy, A. S., memorandum to Ed Scarbrough and Ray
Schucker, Percent Declaration Formats' Advantage on Dietary
Management Tasks, July 27, 1992.
    103. Life Sciences Research Office, Federation of American
Societies for Experimental Biology, ``Physiological Effects and
Health Consequences of Dietary Fiber,'' Bethesda, MD, 1987.
    104. Cole, C. A. and G. J. Gaeth, ``Cognitive and Age-Related
Differences in the Ability to Use Nutritional Information in a
Complex Environment,'' Journal of Marketing Research, 27:175 -
184, 1990.
    105. Bell Associates, Incorporated, ``An Evaluation of
Dietary Guidance Graphic Alternatives: Final Report,'' prepared
for the USDA, Contract No. 53 - 3142 - 1 - 1070, March, 1992.
    106. Tredici, T. J. and R. A. Goodson, ``Print Size for Food
Labels,'' Department of the Air Force, 1976.
    107. Ogilvy, D., Confessions of an Advertising Man, Chapter
7, New York, Athemeum, 1988.
    108. Adler, E. W., Print That Works, pp. 204 - 205, Bull
Publishing Co., Palo Alto, CA, 1991.
    109. Buzzard, I. M., letter to Virginia Wilkening, February
12, 1991.
    110. 136 Congressional Record S16608 (October 24, 1980).
    111. USDA, ``Composition of Foods: Spices and Herbs,''
Agriculture Handbook 8 - 2, Human Nutrition Information Service,
USDA, September 1977.
    112. USDA, ``Composition of Foods: Vegetables and Vegetable
Products,'' Agriculture Handbook 8 - 11, Human Nutrition
Information Service, USDA, pp. 217, 231, 250, 258, 288, 327, 329,
360, and 379, September 1984.
    113. Committee on State Food Labeling, ``Food Labeling,
Toward National Uniformity,'' Food and Nutrition Board, IOM,
National Academy Press, Washington, DC, 1992.
    114. NCEP, ``Report of the Expert Panel on Blood Cholesterol
Levels in Children and Adolescents,'' NIH Publication No. 91 -
2732, PHS, National Institutes of Health, DHHS, Bethesda, MD,
September 1991.
    115. American Academy of Pediatrics Committee on Nutrition,
Pediatric Nutrition Handbook, Second edition, American Academy of
Pediatrics, Elk Grove, Illinois, pp.99 - 104, 1985.
    116. Pennington, J. A. T., ``Bowes and Church's Food Values
of Portions Commonly Used,'' J.B. Lippincott Co., Philadelphia,
PA, 15th Ed., pp. 121 - 133, 1989.
    117. ``USDA-Industry Collaboration in the Generation of Food
Composition Data,'' G. Beecher, Proceedings of the 14th National
Nutrient Databank Conference, Iowa City, Iowa, p. 27 - 28, 1989.
    117a. FDA, ``Mean and Distribution of Total Label Area for a
Sample of Packaged Foods,'' FDA Food Label and Package Survey,
unpublished data, Washington, DC, 1982.
    118. Tanner, J. T., Letter to Dr. V. Srinivasan, U.S.
Pharmacopeial Convention, Inc., May 7, 1991.
    119. U.S. Pharmacopeial Convention, Inc., ``USP XXII, NF
XVII, The United States Pharmacopeia, The National Formulary,
General Notices and Requirements Applying to Standards, Tests,
Assays, and Other Specifications of the United States
Pharmacopeia,''990.
    120. Ehrenberg, A. S. C., Data Reduction: Analyzing and
Interpreting Statistical Data, Chapter 4, New York, Wiley, 1975.

List of Subjects

21 CFR Part 1

    Cosmetics, Drugs, Exports, Food labeling, Imports, Labeling,
Reporting and recordkeeping requirements.

21 CFR Part 101

    Food labeling, Reporting and recordkeeping requirements,
Incorporation by reference.

    Therefore, under the Federal Food, Drug, and Cosmetic Act and
under authority delegated to the Commissioner of Food and Drugs,
21 CFR parts 1 and 101 are amended as follows:


PART 1--GENERAL ENFORCEMENT REGULATIONS

    1. The authority citation for 21 CFR part 1 continues to read
as follows:

    Authority: Secs. 4, 5, 6 of the Fair Packaging and Labeling
Act (15 U.S.C. 1453, 1454, 1455); secs. 201, 403, 502, 512, 602,
701 of the Federal Food, Drug, and Cosmetic Act (21 U.S.C. 321,
343, 352, 355, 360, 362, 371); sec. 215 of the Public Health
Service Act (42 U.S.C. 216).

    2. Section 1.24 is amended by revising paragraph (a)(14) to
read as follows:


1.24  Exemptions from required label statements.

* * * * *
    (a) * * *
    (14) The unit containers in a multiunit or multicomponent
retail food package shall be exempt from regulations of section
403 (e)(1), (g)(2), (i)(2), (k), and (q) of the act with respect
to the requirements for label declaration of the name and place
of business of the manufacturer, packer, or distributor; label
declaration of ingredients; and nutrition information when:
    (i) The multiunit or multicomponent retail food package
labeling meets all the requirements of this part;
    (ii) The unit containers are securely enclosed within and not
intended to be separated from the retail package under conditions
of retail sale; and
    (iii) Each unit container is labeled with the statement
``This Unit Not Labeled For Retail Sale'' in type size not less
than one-sixteenth of an inch in height. The word ``Individual''
may be used in lieu of or immediately preceding the word
``Retail'' in the statement.
* * * * *


PART 101--FOOD LABELING

    3. The authority citation for 21 CFR part 101 continues to
read as follows:

    Authority: Secs. 4, 5, 6 of the Fair Packaging and Labeling
Act (15 U.S.C. 1453, 1454, 1455); secs. 201, 301, 402, 403, 409,
701 of the Federal Food, Drug, and Cosmetic Act (21 U.S.C. 321,
331, 342, 343, 348, 371).

    4. Section 101.9 is revised to read as follows:


101.9  Nutrition labeling of food.

    (a) Nutrition information relating to food shall be provided
for all products intended for human consumption and offered for
sale unless an exemption is provided for the product in paragraph
(j) of this section. A nutrition claim or any other nutrition
information on the label or in labeling or advertising in any
context, and in any form of expression, implicit, as well as
explicit, shall negate any exemption and subject a food to the
provisions of this section.
    (1) When food is in package form, the required nutrition
labeling information shall appear on the label in the format
specified in this section.
    (2) When food is not in package form, the required nutrition
labeling information shall be displayed clearly at the point of
purchase (e.g., on a counter card, sign, tag affixed to the
product, or some other appropriate device). Alternatively, the
required information may be placed in a booklet, looseleaf
binder, or other appropriate format that is available at the
point of purchase.
    (3) Solicitation of requests for nutrition information by a
statement ``For nutrition information write to
&llmdash;&llmdash;&llmdash;&llmdash;&llmdash;&llmdash;&llmdash;&l
lmdash;'
' on the label or in the labeling or advertising for a food, or
providing such information in a direct written reply to a solic
the requirements of this section.
    (4) If any vitamin or mineral is added to a food so that a
single serving provides 50 percent or more of the Reference Daily
Intake (RDI) for the age group for which the product is intended,
as specified in paragraph (c)(8)(iv) of this section, of any one
of the added vitamins or minerals, unless such addition is
permitted or required in other regulations, e.g., a standard of
identity or nutritional quality guideline, or is otherwise
exempted by the Commissioner, the food shall be considered a
foodfor special dietary use within the meaning of
105.3(a)(1)(iii) of this chapter.
    (b) [Reserved]
    (c) The declaration of nutrition information on the label and
in labeling of a food shall contain information about the level
of the following nutrients, except for those nutrients whose
inclusion, and the declaration of amounts, is voluntary as
setforth in this paragraph. No nutrients or food components other
than those listed in this paragraph as either mandatory or
voluntary may be included within the nutrition label. Except as
provided for in paragraphs (f) or (j) of this section, nutrient
information shall be presented using the nutrient names specified
and in the following order in the formats specified in paragraphs
(d) or (e) of this section.
    (1) ``Calories, total,'' ``Total calories,'' or ``Calories'':
A statement of the caloric content per serving, expressed to the
nearest 5-calorie increment up to and including 50 calories, and
10-calorie increment above 50 calories, except that amounts less
than 5 calories may be expressed as zero. Energy content per
serving may also be expressed in kilojoule units, added in
parentheses immediately following the statement of the caloric
content.
    (i) Caloric content may be calculated by:
    (A) Using specific Atwater factors (i.e., the Atwater method)
given in Table 13, ``Energy Value of Foods--Basis and
Derivation,'' by A. L. Merrill and B. K. Watt, United States
Department of Agriculture (USDA) Handbook No. 74 (slightly
revised, 1973), which is incorporated by reference in accordance
with 5 U.S.C. 552(a) and 1 CFR part 51 and is available from the
Division of Nutrition, Center for Food Safety and Applied
Nutrition (HFF - 260), Food and Drug Administration, 200 C St.
SW., Washington, DC 20204, or may be inspected at the Office of
the Federal Register, 800 North Capitol St. NW., suite 700,
Washington, DC.;
    (B) Using the general factors of 4, 4, and 9 calories per
gram for protein, total carbohydrate, and total fat,
respectively, as described in USDA Handbook No. 74 (slightly
revised 1973) pp. 9 - 11, which is incorporated by reference in
accordance with 5 U.S.C. 552(a) and 1 CFR part 51 (the
availability of this incorporation by reference is given in
paragraph (c)(1)(i)(A) of this section);
    (C) Using the general factors of 4, 4, and 9 calories per
gram for protein, total carbohydrate less the amount of insoluble
dietary fiber, and total fat, respectively, as described in USDA
Handbook No. 74 (slightly revised 1973) pp. 9 - 11, which
isincorporated by reference in accordance with 5 U.S.C. 552(a)
and 1 CFR part 51 (the availability of this incorporation by
reference is given in paragraph (c)(1)(i)(A) of this section);
    (D) Using data for specific food factors for particular foods
or ingredients approved by the Food and Drug Administration (FDA)
and provided in parts 172 or 184 of this chapter, or by other
means, as appropriate; or
    (E) Using bomb calorimetry data and subtracting 1.25 calories
per gram protein to correct for incomplete digestibility, as
described in USDA Handbook No. 74 (slightly revised 1973) p. 10,
which is incorporated by reference in accordance with 5 U.S.C.
552(a) and 1 CFR part 51 (the availability of this incorporation
by reference is given in paragraph (c)(1)(i)(A) of this section).
    (ii) ``Calories from fat'': A statement of the caloric
content derived from total fat as defined in paragraph (c)(2) of
this section in a serving, expressed to the nearest 5-calorie
increment, up to and including 50 calories, and the nearest
10-calorie increment above 50 calories, except that label
declaration of ``calories from fat'' is not required on products
that contain less than 0.5 gram of fat in a serving and amounts
less than 5 calories may be expressed as zero. This statement
shall be declared as provided in paragraph (d)(5) of this
section. Except as provided for in paragraph (f) of this section,
if ``Calories from fat'' is not required and, as a result, not
declared, the statement ``Not a significant source of calories
from fat'' shallbe placed at the bottom of the table of nutrient
values in the same type size.
    (iii) ``Calories from saturated fat'' or ``Calories from
saturated'' (VOLUNTARY): A statement of the caloric content
derived from saturated fat as defined in paragraph (c)(2)(i) of
this section in a serving may be declared voluntarily, expressed
to the nearest 5-calorie increment, up to and including 50
calories, and the nearest 10-calorie increment above 50 calories,
except that amounts less than 5 calories may be expressed as
zero. This statement shall be indented under the statement of
calories from fat as provided in paragraph (d)(5) of this
section.
    (2) ``Fat, total'' or ``Total fat'': A statement of the
number of grams of total fat in a serving defined as total lipid
fatty acids and expressed as triglycerides. Amounts shall be
expressed to the nearest 0.5 (\1/2\) gram increment below 3 grams
and to the nearest gram increment above 3 grams. If the serving
contains less than 0.5 gram, the content shall be expressed as
zero.
    (i) ``Saturated fat,'' or ``Saturated'': A statement of the
number of grams of saturated fat in a serving defined as the sum
of all fatty acids containing no double bonds, except that label
declaration of saturated fat content information is not required
for products that contain less than 0.5 gram of total fat in a
serving if no claims are made about fat or cholesterol content,
and if ``calories from saturated fat'' is not declared. Except as
provided for in paragraph (f) of this section, if a statement of
the saturated fat content is not required and, as a result, not
declared, the statement ``Not a significant source of saturated
fat'' shall be placed at the bottom of the table of nutrient
values in the same type size. Saturated fat content shall be
indented and expressed as grams per serving to the nearest 0.5
(\1/2\) gram increment below 3 grams and to the nearest gram
increment above 3 grams. If the serving contains less than 0.5
gram, the content shall be expressed as zero.
    (ii) ``Polyunsaturated fat'' or ``Polyunsaturated''
(VOLUNTARY): A statement of the number of grams of
polyunsaturated fat in a serving defined as
cis,cis-methylene-interrupted polyunsaturated fatty acids may be
declared voluntarily, except that when monounsaturated fat is
declared or when a claim is made on the label or in labeling
about fatty acids or cholesterol, label declaration of
polyunsaturated fat is required. Polyunsaturated fat content
shall be indented and expressed as grams per serving to the
nearest 0.5 (\1/2\) gram increment below 3 grams and to the
nearest gram increment above 3 grams. If the serving contains
less than 0.5 gram, the content shall be expressed as zero.
    (iii) ``Monounsaturated fat'' or ``Monounsaturated''
(VOLUNTARY): A statement of the number of grams of
monounsaturated fat in a serving defined as cis-monounsaturated
fatty acids may be declared voluntarily except that when
polyunsaturated fat is declared or when a claim is made on the
label or in labeling about fatty acids or cholesterol, label
declaration of monounsaturated fat is required. Monounsaturated
fat content shall be indented and expressed as grams per serving
to the nearest 0.5 (\
    (3) ``Cholesterol'': A statement of the cholesterol content
in a serving expressed in milligrams to the nearest 5-milligram
increment, except that label declaration of cholesterol
information is not required for products that contain less than 2
milligrams cholesterol in a serving and make no claim about fat,
fatty acids, or cholesterol content, or such products may state
the cholesterol content as zero. Except as provided for in
paragraph (f) of this section, if cholesterol content is not
requiredand, as a result, not declared, the statement ``Not a
significant source of cholesterol'' shall be placed at the bottom
of the table of nutrient values in the same type size. If the
food contains 2 to 5 milligrams of cholesterol per serving, the
contentmay be stated as ``less than 5 milligrams.''
    (4) ``Sodium'': A statement of the number of milligrams of
sodium in a specified serving of food expressed as zero when the
serving contains less than 5 milligrams of sodium, to the nearest
5-milligram increment when the serving contains 5 to 140
milligrams of sodium, and to the nearest 10-milligram increment
when the serving contains greater than 140 milligrams.
    (5) ``Potassium'' (VOLUNTARY): A statement of the number of
milligrams of potassium in a specified serving of food may be
declared voluntarily, except that when a claim is made about
potassium content, label declaration shall be required. Potassium
content shall be expressed as zero when the serving contains less
than 5 milligrams of potassium, to the nearest 5-milligram
increment when the serving contains less than or equal to 140
milligrams of potassium, and to the nearest 10-milligram
increment when the serving contains more than 140 milligrams.
    (6) ``Carbohydrate, total'' or ``Total carbohydrate'': A
statement of the number of grams of total carbohydrate in a
serving expressed to the nearest gram, except that if a serving
contains less than 1 gram, the statement ``Contains less than 1
gram'' or ``less than 1 gram'' may be used as an alternative, or
if the serving contains less than 0.5 gram, the content may be
expressed as zero. Total carbohydrate content shall be calculated
by subtraction of the sum of the crude protein, total fat,
moisture, and ash from the total weight of the food. This
calculation method is described in A. L. Merrill and B. K. Watt,
``Energy Value of Foods--Basis and Derivation,'' USDA Handbook 74
(slightly revised 1973) pp. 2 and 3, which is incorporated by
referencein accordance with 5 U.S.C. 552(a) and 1 CFR part 51
(the availability of this incorporation by reference is given in
paragraph (c)(1)(i)(A) of this section).
    (i) ``Dietary fiber'': A statement of the number of grams of
total dietary fiber in a serving, indented and expressed to the
nearest gram, except that if a serving contains less than 1 gram,
declaration of dietary fiber is not required or, alternatively,
the statement ``Contains less than 1 gram'' or ``less than 1
gram'' may be used, and if the serving contains less than 0.5
gram, the content may be expressed as zero. Except as provided
for in paragraph (f) of this section, if dietary fiber contentis
not required and as a result, not declared, the statement ``Not a
significant source of dietary fiber'' shall be placed at the
bottom of the table of nutrient values in the same type size.
    (A) ``Soluble fiber'' (VOLUNTARY): A statement of the number
of grams of soluble dietary fiber in a serving may be declared
voluntarily except when a claim is made on the label or in
labeling about soluble fiber, label declaration shall be
required.Soluble fiber content shall be indented under dietary
fiber and expressed to the nearest gram, except that if a serving
contains less than 1 gram, the statement ``Contains less than 1
gram'' or ``less than 1 gram'' may be used as an alternative, and
if the serving contains less than 0.5 gram, the content may be
expressed as zero.
    (B) ``Insoluble fiber'' (VOLUNTARY): A statement of the
number of grams of insoluble dietary fiber in a serving may be
declared voluntarily except that when a claim is made on the
label or in labeling about insoluble fiber, label declaration
shall be required. Insoluble fiber content shall be indented
under dietary fiber and expressed to the nearest gram except that
if a serving contains less than 1 gram, the statement ``Contains
less than 1 gram'' or ``less than 1 gram'' may be used as an
alternative, and if the serving contains less than 0.5 gram, the
content may be expressed as zero.
    (ii) ``Sugars'': A statement of the number of grams of sugars
in a serving, except that label declaration of sugars content is
not required for products that contain less than 1 gram of sugars
in a serving if no claims are made about sweeteners, sugars, or
sugar alcohol content. Except as provided for in paragraph (f) of
this section, if a statement of the sugars content is not
required and, as a result, not declared, the statement ``Not a
significant source of sugars'' shall be placed at the bottom of
the table of nutrient values in the same type size. Sugars shall
be defined as the sum of all free mono- and disaccharides (such
as glucose, fructose, lactose, and sucrose). Sugars content shall
be indented and expressed to the nearest gram, except that if a
serving contains less than 1 gram, the statement ``Contains less
then 1 gram'' or ``less than 1 gram'' may be used as an
alternative, and if the serving contains less than 0.5 gram, the
content may be expressed as zero.
    (iii) ``Sugar alcohol'' (VOLUNTARY): A statement of the
number of grams of sugar alcohols in a serving may be declared
voluntarily on the label, except that when a claim is made on the
label or in labeling about sugar alcohol or sugars when sugar
alcohols are present in the food, sugar alcohol content shall be
declared. For nutrition labeling purposes, sugar alcohols are
defined as the sum of saccharide derivatives in which a hydroxyl
group replaces a ketone or aldehyde group and whose use in the
food is listed by FDA (e.g., mannitol) or is generally recognized
as safe (e.g., xylitol, sorbitol). In lieu of the term ``sugar
alcohol,'' the name of the specific sugar alcohol (e.g.,
``xylitol'') present in the food may be used in the nutrition
label provided that only one sugar alcohol is present in the
food. Sugar alcohol content shall be indented and expressed to
the nearest gram, except that if a serving contains less than 1
gram, the statement ``Contains less then 1 gram'' or ``less than
1 gram''may be used as an alternative, and if the serving
contains less than 0.5 gram, the content may be expressed as
zero.
    (iv) ``Other carbohydrate'' (VOLUNTARY): A statement of the
number of grams of other carbohydrates may be declared
voluntarily. Other carbohydrates shall be defined as the
difference between total carbohydrate and the sum of dietary
fiber, sugars, and sugar alcohol, except that if sugar alcohol is
not declared (even if present), it shall be defined as the
difference between total carbohydrate and the sum of dietary
fiber and sugars. Other carbohydrate content shall be indented
and expressed to the nearest gram, except that if a serving
contains less than 1 gram, the statement ``Contains less than 1
gram'' or ``less than 1 gram'' may be used as an alternative, and
if the serving contains less than 0.5 gram, the content may be
expressed as zero.
    (7) ``Protein'': A statement of the number of grams of
protein in a serving, expressed to the nearest gram, except that
if a serving contains less than 1 gram, the statement ``Contains
less than 1 gram'' or ``less than 1 gram'' may be used as an
alternative, and if the serving contains less than 0.5 gram, the
content may be expressed as zero. When the protein in foods
represented or purported to be for adults and children 4 or more
years of age has a protein quality value that is a protein
digestibility-corrected amino acid score of less than 20
expressed as a percent, or when the protein in a food represented
or purported to be for children greater than 1 but less than 4
years of age has a protein quality value that is a protein
digestibility-correan 40 expressed as a percent, either of the
following shall be placed adjacent to the declaration of protein
content by weight: The statement ``not a significant source of
protein,'' or a listing aligned under the column headed ``Percent
Daily Value'' of the corrected amount of protein per serving, as
determined in paragraph (c)(7)(ii) of this section, calculated as
a percentage of the Daily Reference Value (DRV) or Reference
Daily Intake (RDI), as appropriate, for protein and expressed as
Percent of Daily Value. When the protein quality in a food as
measured by the Protein Efficiency Ratio (PER) is less than 40
percent of the reference standard (casein) for a food represented
or purported to be for infants, the statement ``not a significant
source ofprotein'' shall be placed adjacent to the declaration of
protein content. Protein content may be calculated on the basis
of the factor of 6.25 times the nitrogen content of the food as
determined by the appropriate method of analysis as given in the
``Official Methods of Analysis of the AOAC International''
(formerly the Association of Official Analytical Chemists), 15th
Ed. (1990), which is incorporated by reference in accordance with
5 U.S.C. 552(a) and 1 CFR part 51, except when the official
procedure for a specific food requires another factor. Copies may
be obtained from AOAC, 2200 Wilson Blvd., suite 400, Arlington,
VA 22201-3301, or may be inspected at the Office of the Federal
Register, 800 North Capitol St. NW., suite 700, Washington, DC.
    (i) A statement of the corrected amount of protein per
serving, as determined in paragraph (c)(7)(ii) of this section,
calculated as a percentage of the RDI or DRV for protein, as
appropriate, and expressed as Percent of Daily Value, may be
placed on the label, except that such a statement shall be given
if a protein claim is made for the product, or if the product is
represented or purported to be for use by infants or children
under 4 years of age. When such a declaration is provided, it
shall beplaced on the label adjacent to the statement of grams of
protein and aligned under the column headed ``Percent Daily
Value,'' and expressed to the nearest whole percent. However, the
percentage of the RDI for protein shall not be declared if the
food is represented or purported to be for use by infants and the
protein quality value is less than 40 percent of the reference
standard.
    (ii) The ``corrected amount of protein (gram) per serving''
for foods represented or purported for adults and children 1 or
more years of age is equal to the actual amount of protein (gram)
per serving multiplied by the amino acid score corrected for
protein digestibility. If the corrected score is above 1.00, then
it shall be set at 1.00. The protein digestibility-corrected
amino acid score shall be determined by methods given in sections
5.4.1, 7.2.1, and 8.00 in ``Protein Quality Evaluation, Report of
the Joint FAO/WHO Expert Consultation on Protein Quality
Evaluation,'' Rome, 1990, except that when official AOAC
procedures described in section (c)(7) of this paragraph require
a specific food factor other than 6.25, that specific factor
shall be used. The ``Report of the Joint FAO/WHO Expert
Consultation on Protein Quality Evaluation'' as published by the
Food and Agriculture Organization of the United Nations/World
Health Organization is incorporated by reference in accordance
with 5 U.S.C. 552(a) and 1 CFR part 51. Copies are available from
the Division of Nutrition, Center for Food Safety and Applied
Nutrition (HFF - 260), Food and Drug Administration, 200 C St.
SW., Washington, DC 20204, or may be inspected at the Office of
the Federal Register, 800 North Capitol St. NW., suite 700,
Washington, DC. For foods represented or purported for infants,
the corrected amount of protein (grams) per serving is equal to
the actual amount of protein (grams) per serving multiplied by
the relative proteiein. If the relative protein value is above
1.00, it shall be set at 1.00.
    (iii) [Reserved]
    (8) Vitamins and minerals: A statement of the amount per
serving of the vitamins and minerals as described in this
paragraph, calculated as a percent of the RDI and expressed as
percent of Daily Value.
    (i) For purposes of declaration of percent of Daily Value as
provided for in paragraphs (d), (e), and (f) of this section,
foods represented or purported to be for use by infants, children
less than 4 years of age, pregnant women, or lactating womenshall
use the RDI's in paragraph (c)(8)(iv) of this section that are
specified for the intended group. For foods represented or
purported to be for use by both infants and children under 4
years of age, the percent of Daily Value shall be presented by
separate declarations according to paragraph (e) of this section
based on the RDI values for infants from birth to 12 months of
age and for children under 4 years of age. Similarly, the percent
of Daily Value based on both the RDI values for pregnant women
and for lactating women shall be declared separately on foods
represented or purported to be for use by both pregnant and
lactating women. When such dual declaration is used on any label,
it shall be included in all labeling, and equal prominence
shallbe given to both values in all such labeling. All other
foods shall use the RDI for adults and children 4 or more years
of age.
    (ii) The declaration of vitamins and minerals as a percent of
the RDI shall include vitamin A, vitamin C, calcium, and iron, in
that order, and shall include any of the other vitamins and
minerals listed in paragraph (c)(8)(iv) of this section when they
are added as a nutrient supplement, or when a claim is made about
them. Other vitamins and minerals that are:
    (A) Required or permitted in a standardized food (e.g.,
thiamin, riboflavin, and niacin in enriched flour) and that
standardized food is included as an ingredient (i.e., component)
in another food; or
    (B) Included in a food solely for technological purposes and
declared only in the ingredient statement need not be declared if
neither the nutrient nor the component is otherwise referred to
on the label or in labeling or advertising. The declaration may
also include any of the other vitamins and minerals listed in
paragraph (c)(8)(iv) of this section when they are naturally
occurring in the food. The additional vitamins and minerals shall
be listed in the order established in paragraph (c)(8)(iv) of
this section.
    (iii) The percentages for vitamins and minerals shall be
expressed to the nearest 2-percent increment up to and including
the 10-percent level, the nearest 5-percent increment above 10
percent and up to and including the 50-percent level, and the
nearest 10-percent increment above the 50-percent level. Amounts
of vitamins and minerals present at less than 2 percent of the
RDI are not required to be declared in nutrition labeling but may
be declared by a zero or by the use of an asterisk (or other
symbol) that refers to another asterisk (or symbol) that is
placed at the bottom of the table and that is followed by the
statement ``Contains less than 2 percent of the Daily Value of
this (these) nutrient (nutrients).'' Alternatively, except as
providedfor in paragraph (f) of this section, if vitamin A,
vitamin C, calcium, or iron is present in amounts less than 2
percent of the RDI, label declaration of the nutrient(s) is not
required if the statement ``Not a significant source of
&llmdash;&llmdash;&llmdash;&llmdash;&llmdash; (listing the
vitamins or minerals omitted)'' is placed at the bottom of the
table of nutrient values. Either statement shall be in the same
type size as nutrients that are indented.
    (iv) [Reserved]
    (v) The following synonyms may be added in parentheses
immediately following the name of the nutrient or dietary
component:

Vitamin C--Ascorbic acid
Thiamin--Vitamin B1
Riboflavin--Vitamin B2
Folate--Folacin
Calories--Energy

    (vi) The percent of vitamin A that is present as
beta-carotene may be declared to the nearest 10-percent increment
immediately adjacent to or beneath the nutrient name (e.g.,
``Vitamin A (90 percent as beta-carotene)'').
    (9) [Reserved]
    (d)(1) Nutrient information specified in paragraph (c) of
this section shall be presented on foods in the following format,
as shown in paragraph (d)(12) of this section, except on foods on
which dual columns of nutrition information are declared
asprovided for in paragraph (e) of this section, on those food
products on which the simplified format is required to be used as
provided for in paragraph (f) of this section, on foods for
infants and children less than 4 years of age as provided for in
paragraph (j)(5) of this section, and on foods in small or
intermediate-sized packages as provided for in paragraph (j)(13)
of this section. In the interest of uniformity of presentation,
FDA urges that the nutrition information be presented using the
graphic specifications set forth in Appendix B to Part 101.
    (i) The nutrition information shall be set off in a box by
use of hairlines and shall be all black or one color type,
printed on a white or other neutral contrasting background
whenever practical.
    (ii) All information within the nutrition label shall
utilize:
    (A) A single easy-to-read type style,
    (B) Upper and lower case letters,
    (C) At least one point leading (i.e., space between two lines
of text) except that at least four points leading shall be
utilized for the information required by paragraphs (d)(7) and
(d)(8) of this section as shown in paragraph (d)(12), and
    (D) Type that is kerned (i.e., has proximity of placement) no
tighter than {dieresis} setting.
    (iii) All information except for the information required in
paragraphs (d)(4), (d)(6), (d)(9), and (d)(10) of this section
shall be in type size no smaller than 8 point. The information
required in paragraphs (d)(4), (d)(6), (d)(9), and (d)(10) of
this section shall be in type size no smaller than 6 point.
    (iv) The headings required by paragraphs (d)(2), (d)(4), and
(d)(6) of this section (i.e., ``Nutrition Facts,'' ``Amount per
Serving,'' and ``% Daily Value*''), the names of all nutrients
that are not indented according to requirements of paragraph (c)
of this section (i.e., ``Calories,'' ``Total Fat,''
``Cholesterol,'' ``Sodium,'' ``Total Carbohydrate,'' and
``Protein''), and the percentage amounts required by paragraph
(d)(7)(ii) of this section shall be highlighted by bold or extra
bold type or other highlighting (reverse printing is not
permitted as a form of highlighting) that prominently
distinguishes it from other information. No other information
shall be highlighted.
    (v) A hairline rule that is centered between the lines of
text shall separate ``Amount Per Serving'' from the calorie
statements required in paragraph (d)(5) of this section and shall
separate each nutrient and its corresponding percent Daily Value
required in paragraphs (d)(7)(i) and (d)(7)(ii) of this section
from the nutrient and percent Daily Value above and below it, as
shown in paragraph (d)(12) of this section.
    (2) The information shall be presented under the identifying
heading of ``Nutrition Facts'' which shall be set in a type size
larger than all other print size in the nutrition label and,
except for labels presented according to the format provided for
in paragraph (d)(11) of this section, unless impractical, shall
be set the full width of the information provided under paragraph
(d)(7) of this section, as shown in paragraph (d)(12) of this
section.
    (3) Information on serving size shall immediately follow the
heading as shown in paragraph (d)(12) of this section. Such
information shall include:
    (i) ``Serving Size'': A statement of the serving size as
specified in paragraph (b)(7) of this section.
    (ii) ``Servings Per Container'': The number of servings per
container, except that this statement is not required on single
serving containers as defined in paragraph (b)(6) of this
section.
    (4) A subheading ``Amount Per Serving'' shall be separated
from serving size information by a bar as shown in paragraph
(d)(12) of this section.
    (5) Information on calories shall immediately follow the
heading ``Amount Per Serving'' and shall be declared in one line,
leaving sufficient space between the declaration of ``Calories''
and ``Calories from fat'' to allow clear differentiation, or,if
``Calories from saturated fat'' is declared, in a column with
total ``Calories'' at the top, followed by ``Calories from fat''
(indented), and ``Calories from saturated fat'' (indented).
    (6) The column heading ``% Daily Value,'' followed by an
asterisk (e.g., ``% Daily Value*''), shall be separated from
information on calories by a bar as shown in paragraph (d)(12) of
this section. The position of this column heading shall allow
fora list of nutrient names and amounts as described in paragraph
(d)(7) of this section to be to the left of, and below, this
column heading. The column headings ``Percent Daily Value,''
``Percent DV,'' or ``% DV'' may be substituted for ``% Daily
Value.''
    (7) Except as provided for in paragraph (j)(13) of this
section, nutrient information for all nutrients required by
paragraph (c) of this section, except vitamins and minerals,
shall be declared as follows:
    (i) The name of each nutrient specified in paragraph (c) of
this section shall be given in a column and followed immediately
by the quantitative amount by weight for that nutrient appended
with a ``g'' for grams or ``mg'' for milligrams as shown in
paragraph (d)(12) of this section.
    (ii) A listing of the percent of the DRV as established in
paragraphs (c)(7)(iii) and (c)(9) of this section shall be given
in a column aligned under the heading ``% Daily Value''
established in paragraph (d)(6) of this section with the percent
expressed to the nearest whole percent for each nutrient declared
in the column described in paragraph (d)(7)(i) of this section
for which a DRV has been established, except that the percent for
protein may be omitted as provided in paragraph (c)(7) of this
section. The percent shall be calculated by dividing the actual
amount (i.e., before rounding) for each nutrient by the DRV for
the nutrient, except that the percent for protein shall be
calculated as specified in paragraph (c)(7)(ii) of this section.
The numerical value shall be followed by the symbol for percent
(i.e., %).
    (8) Nutrient information for vitamins and minerals shall be
separated from information on other nutrients by a bar and shall
be arrayed horizontally (e.g., Vitamin A 4%, Vitamin C 2%,
Calcium 15%, Iron 4%) as shown in paragraph (d)(12) of this
section, except that when more than four vitamins and minerals
are declared, they may be declared vertically with percentages
listed under the column headed ``% Daily Value.''
    (9) A footnote, preceded by an asterisk, shall be placed
beneath the list of vitamins and minerals and shall be separated
from that list by a hairline.
    (i) The footnote shall state:
    Percent Daily Values are based on a 2,000 calorie diet.
    Your daily value may be higher or lower depending on your
calorie needs.

ÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄ
ÄÄÄÄÄÄÄÄÄÄÄÄÄ
                                           Calories:       2,000  
    2,500  
ÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄ
ÄÄÄÄÄÄÄÄÄÄÄÄÄ
Total fat..............................  Less than....  65 g      
 80 g      
Saturated fat..........................  Less than....  20 g      
 25 g      
Cholesterol............................  Less than....  300 mg    
 300 mg    
Sodium.................................  Less than....  2,400 mg  
 2,400 mg  
Total carbohydrate.....................  .............  300 g     
 375 g     
Dietary fiber..........................  .............  25 g      
 30 g      
ÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄ
ÄÄÄÄÄÄÄÄÄÄÄÄÄ




    (ii) If the percent of Daily Value is given for protein in
the Percent of Daily Value column as provided in paragraph
(d)(5)(ii) of this section, protein shall be listed under dietary
fiber, and the DRV established in paragraph (c)(7)(iii) of this
sec
    (iii) If potassium is declared in the column described in
paragraph (d)(5)(i) of this section, potassium shall be listed
under sodium and the DRV established in paragraph (c)(9) of this
section shall be inserted on the same line in the numeric
columns.
    (iv) The abbreviations established in paragraph
(j)(13)(ii)(B) of this section may be used within the footnote.
    (10) Caloric conversion information on a per gram basis for
fat, carbohydrate, and protein shall be presented beneath the
information required in paragraph (d)(9) and shall be separated
from that information by a hairline. This information may be
presented horizontally as shown in paragraph (d)(12) of this
section (i.e., ``Calories per gram: fat 9, carbohydrate 4,
protein 4'') or vertically in columns.
    (11) If the space beneath the information on vitamins and
minerals is not adequate to accommodate the information required
in paragraphs (d)(9) and (d)(10) of this section, the information
required in paragraph (d)(9) may be moved to the right of the
column required in paragraph (d)(7)(ii) and set off by a line
that distinguishes it and sets it apart from the percent daily
value information. The caloric conversion information required in
paragraph (d)(10) of this section may be presented beneath either
side or along the full length of the nutrition label.
    (12) The following sample label illustrates the provisions of
paragraph (d) of this section.

BILLING CODE 4160 - 01 - m
BILLING CODE 4160 - 01 - C

INSERT SAMPLE LABEL 541.TIF

    (e) Nutrition information may be presented for two or more
forms of the same food (e.g., both ``as purchased'' and ``as
prepared'') or for common combinations of food as provided for in
paragraph (h)(4) of this section, for different units (e.g.,
slices of bread or per 100 grams) as provided for in paragraph
(b) of this section, or for two or more groups for which RDI's
are established in paragraph (c)(8)(iv) of this section (e.g.,
both infants and children less than 4 years of age) as shown in
paragraph (e)(5) of this section. When such dual labeling is
provided, equal prominence shall be given to both sets of values.
Information shall be presented in a format consistent with
paragraph (d) of this section, except that:
    (1) Following the subheading of ``Amount Per Serving,'' there
shall be two or more column headings accurately describing the
forms of the same food (e.g., ``Mix'' and ``Baked''), the
combinations of food, the units, or the RDI groups that are
beingdeclared. The column representing the product as packaged
and according to the label serving size based on the reference
amount in 101.12(b) shall be to the left of the numeric columns.
    (2) When the dual labeling is presented for two or more forms
of the same food, for combinations of food, or for different
units, total calories and calories from fat (and calories from
saturated fat, when declared) shall be listed in a column and
indented as specified in paragraph (d)(5) of this section with
quantitative amounts declared in columns aligned under the column
headings set forth in paragraph (e)(1) of this section.
    (3) Quantitative information by weight required in paragraph
(d)(7)(i) of this section shall be specified for the form of the
product as packaged and according to the label serving size based
on the reference amount in 101.12(b).
    (i) Quantitative information by weight may be included for
other forms of the product represented by the additional
column(s) either immediately adjacent to the required
quantitative information by weight for the product as packaged
and according to the label serving size based on the reference
amount in 101.12(b) or as a footnote.
    (A) If such additional quantitative information is given
immediately adjacent to the required quantitative information, it
shall be declared for all nutrients listed and placed immediately
following and differentiated from the required quantitative i
    (B) If such additional quantitative information is given in a
footnote, it shall be declared in the same order as the nutrients
are listed in the nutrition label. The additional quantitative
information may state the total nutrient content of the product
identified in the second column or the nutrient amounts added to
the product as packaged for only those nutrients that are present
in different amounts than the amounts declared in the required
quantitative information. The footnote shall clearly identify
which amounts are declared. Any subcomponents declared shall be
listed parenthetically after principal components (e.g., 1/2 cup
skim milk contributes an additional 40 calories, 65 mg sodium, 6
g total carbohydrate (6 g sugars), and 4 g protein).
    (ii) Total fat and its quantitative amount by weight shall be
followed by an asterisk (or other symbol) (e.g., ``Total fat (2
g)*'') referring to another asterisk (or symbol) at the bottom of
the nutrition label identifying the form(s) of the product for
which quantitative information is presented.
    (4) Information required in paragraphs (d)(7)(ii) and (d)(8)
of this section shall be presented under the subheading``% DAILY
VALUE'' and in columns directly under the column headings set
forth in paragraph (e)(1) of this section.
    (5) The following sample label illustrates the provisions of
paragraph (e) of this section:

BILLING CODE 4160 - 01 - M
BILLING CODE 4160 - 01 - C

INSERT 545.TIF

    (f) The declaration of nutrition information may be presented
in the simplified format set forth herein when a food product
contains insignificant amounts of seven or more of the following:
Calories, total fat, saturated fat, cholesterol, sodium, total
carbohydrate, dietary fiber, sugars, protein, vitamin A, vitamin
C, calcium, and iron; except that for foods intended for children
less than 2 years of age to which101.9(j)(5)(i) applies,
nutrition information may be presented in the simplified format
when a food product contains insignificant amounts of six or more
of the following: Calories, total fat, sodium, total
carbohydrate, dietary fiber, sugars, protein, vitamin A, vitamin
C, calcium, and iron.
    (1) An ``insignificant amount'' shall be defined as that
amount that allows a declaration of zero in nutrition labeling,
except that for total carbohydrate, dietary fiber, and protein,
it shall be an amount that allows a declaration of ``less than 1
gram.''
    (2) The simplified format shall include information on the
following nutrients:
    (i) Total calories, total fat, total carbohydrate, protein,
and sodium;
    (ii) Calories from fat and any other nutrients identified in
paragraph (f)(1) of this section that are present in the food in
more than insignificant amounts; and
    (iii) Any vitamins and minerals listed in paragraph
(c)(8)(iv) of this section when they are required to be added as
a nutrient supplement to foods for which a standard of identity
exists.
    (iv) Any vitamins or minerals that are voluntarily added to
the food as nutrient supplements.
    (3) Other nutrients that are naturally present in the food in
more than insignificant amounts may be voluntarily declared as
part of the simplified format.
    (4) If any nutrients are declared as provided in paragraphs
(f)(2)(iii), (f)(2)(iv), or (f)(3) of this section as part of the
simplified format, the statement ``Not a significant source of
&llmdash;&llmdash;&llmdash;&llmdash;&llmdash;'' (with the blank
filled in with the name(s) of any nutrient(s) identified in
101.9(f) and calories from fat that are present in insignificant
amounts) shall be included at the bottom of the nutrition label.
    (5) Except as provided for in paragraphs (j)(5) and (j)(13)
of this section, nutrient information declared in the simplified
format shall be presented in the same manner as specified in
paragraphs (d) or (e) of this section, except that the
footnoteuired. When the footnote and caloric conversion
information are omitted, an asterisk shall be placed at the
bottom of the label followed by the statement ``Percent Daily
Values are based on a 2,000 calorie diet'' and, if the term
``Daily Value'' is notspelled out in the heading, a statement
that ``DV'' represents ``Daily Value.''
    (g) Compliance with this section shall be determined as
follows:
    (1) A collection of primary containers or units of the same
size, type, and style produced under conditions as nearly uniform
as possible, designated by a common container code or marking, or
in the absence of any common container code or marking, a day's
production, constitutes a ``lot.''
    (2) The sample for nutrient analysis shall consist of a
composite of 12 subsamples (consumer units), taken 1 from each of
12 different randomly chosen shipping cases, to be representative
of a lot. Unless a particular method of analysis is specified in
paragraph (c) of this section, composites shall be analyzed by
appropriate methods as given in the ``Official Methods of
Analysis of the AOAC International,'' 15th Ed. (1990), which is
incorporated by reference in accordance with 5 U.S.C. 552(a) or1
CFR part 51 or, if no AOAC method is available or appropriate, by
other reliable and appropriate analytical procedures. The
availability of this incorporation by reference is given in
paragraph (c)(7) of this section.
    (3) Two classes of nutrients are defined for purposes of
compliance:
    (i) Class I. Added nutrients in fortified or fabricated
foods; and
    (ii) Class II. Naturally occurring (indigenous) nutrients. If
any ingredient which contains a naturally occurring (indigenous)
nutrient is added to a food, the total amount of such nutrient in
the final food product is subject to class II requirements unless
the same nutrient is also added.
    (4) A food with a label declaration of a vitamin, mineral,
protein, total carbohydrate, dietary fiber, other carbohydrate,
polyunsaturated or monounsaturated fat, or potassium shall be
deemed to be misbranded under section 403(a) of the Federal Food,
Drug, and Cosmetic Act (the act) unless it meets the following
requirements:
    (i) Class I vitamin, mineral, protein, dietary fiber, or
potassium. The nutrient content of the composite is at least
equal to the value for that nutrient declared on the label.
    (ii) Class II vitamin, mineral, protein, total carbohydrate,
dietary fiber, other carbohydrate, polyunsaturated or
monounsaturated fat, or potassium. The nutrient content of the
composite is at least equal to 80 percent of the value for
thatnutrient declared on the label.
    Provided, That no regulatory action will be based on a
determination of a nutrient value that falls below this level by
a factor less than the variability generally recognized for the
analytical method used in that food at the level involved.
    (5) A food with a label declaration of calories, sugars,
total fat, saturated fat, cholesterol, or sodium shall be deemed
to be misbranded under section 403(a) of the act if the nutrient
content of the composite is greater than 20 percent in excessof
the value for that nutrient declared on the label. Provided, That
no regulatory action will be based on a determination of a
nutrient value that falls above this level by a factor less than
the variability generally recognized for the analytical method
used in that food at the level involved.
    (6) Reasonable excesses of a vitamin, mineral, protein, total
carbohydrate, dietary fiber, other carbohydrate, polyunsaturated
or monounsaturated fat, or potassium over labeled amounts are
acceptable within current good manufacturing practice. Reasonable
deficiencies of calories, sugars, total fat, saturated fat,
cholesterol, or sodium under labeled amounts are acceptable
within current good manufacturing practice.
    (7) Compliance will be based on the metric measure specified
in the label statement of serving size.
    (8) Compliance with the provisions set forth in paragraphs
(g)(1) through (g)(6) of this section may be provided by use of
an FDA approved data base that has been computed following FDA
guideline procedures and where food samples have been handled in
accordance with current good manufacturing practice to prevent
nutrition loss. FDA approval of a data base shall not be
considered granted until the Center for Food Safety and Applied
Nutrition has agreed to all aspects of the data base in writing.
The approval will be granted where a clear need is presented
(e.g., raw produce and seafood). Approvals will be in effect for
a limited time, e.g., 10 years, and will be eligible for renewal
in the absence of significant changes in agricultural or industry
practices. Approval requests shall be submitted in accordance
with the provisions of 10.30 of this chapter. Guidance in the
use of data bases may be found in the ``FDA Nutrition Labeling
Manual--A Guide for Developing and Using Data Bases,'' available
from the Division of Nutrition, Center for Food Safety and
Applied Nutrition (HFF - 260), Food and Drug Administration, 200
C St. SW., Washington, DC 20204.
    (9) When it is not technologically feasible, or some other
circumstance makes it impracticable, for firms to comply with the
requirements of this section (e.g., to develop adequate nutrient
profiles to comply with the requirements of paragraph (c) of this
section), FDA may permit alternative means of compliance or
additional exemptions to deal with the situation. Firms in need
of such special allowances shall make their request in writing to
the Office of Nutrition and Food Sciences (HFF - 200), Food and
Drug Administration, 200 C St. SW., Washington, DC 20204.
    (h) Products with separately packaged ingredients or foods,
with assortments of food, or to which other ingredients are added
by the user may be labeled as follows:
    (1) If a product consists of two or more separately packaged
ingredients enclosed in an outer container or of assortments of
the same type of food (e.g., assorted nuts or candy mixtures) in
the same retail package, nutrition labeling shall be located on
the outer container or retail package (as the case may be) to
provide information for the consumer at the point of purchase.
However, when two or more food products are simply combined
together in such a manner that no outer container is used, orno
outer label is available, each product shall have its own
nutrition information, e.g., two boxes taped together or two cans
combined in a clear plastic overwrap. When separately packaged
ingredients or assortments of the same type of food are intended
to be eaten at the same time, the nutrition information may be
specified per serving for each component or as a composite value.
    (2) If a product consists of two or more separately packaged
foods that are intended to be eaten individually and that are
enclosed in an outer container (e.g., variety packs of cereals or
snack foods), the nutrition information shall be specified per
serving for each food in a location that is clearly visible to
the consumer at the point of purchase.
    (3) If a package contains a variety of foods, or an
assortment of foods, and is in a form intended to be used as a
gift, the nutrition labeling shall be in the form required by
paragraphs (a) through (f) of this section, but it may be
modified as follows:
    (i) Nutrition information may be presented on the label of
the outer package or in labeling within or attached to the outer
package.
    (ii) In the absence of a reference amount customarily
consumed in 101.12(b) that is appropriate for the variety or
assortment of foods in a gift package, 1 ounce for solid foods,2
fluid ounces for nonbeverage liquids (e.g., syrups), and 8 fluid
ounces for beverages may be used as the standard serving size for
purposes of nutrition labeling of foods subject to this
paragraph. However, the reference amounts customarily consumed
inl be used for purposes of evaluating whether individual foods
in a gift package qualify for nutrient content claims or health
claims.
    (iii) The number of servings per container may be stated as
``varied.''
    (iv) Nutrition information may be provided per serving for
individual foods in the package, or, alternatively, as a
composite per serving for reasonable categories of foods in the
package having similar dietary uses and similar significant
nutritional characteristics. Reasonable categories of foods may
be used only if accepted by FDA. In determining whether a
proposed category is reasonable, FDA will consider whether the
values of the characterizing nutrients in the foods proposed to
be in the category meet the compliance criteria set forth in
paragraphs (g)(3) through (g)(6) of this section. Proposals for
such categories may be submitted in writing to the Office of
Nutrition and Food Sciences (HFF - 200), Center for Food Safety
and Applied Nutrition, Food and Drug Administration, 200 C St.
SW., Washington, DC 20204.
    (v) If a food subject to paragraph (j)(13) of this section
because of its small size is contained in a gift package, the
food need not be included in the determination of nutrition
information under paragraph (h) of this section if it is not
specifically listed in a promotional catalogue as being present
in the gift package, and:
    (A) It is used in small quantities primarily to enhance the
appearance of the gift package; or
    (B) It is included in the gift package as a free gift or
promotional item.
    (4) If a food is commonly combined with other ingredients or
is cooked or otherwise prepared before eating, and directions for
such combination or preparations are provided, another column of
figures may be used to declare nutrition information on the basis
of the food as consumed in the format required in paragraph (e)
of this section (e.g., a dry ready-to-eat cereal may be described
with one set of Percent Daily Values for the cereal as sold
(e.g., per ounce), and another set for the cereal and milk as
suggested in the label (e.g., per ounce of cereal and 1/2 cup of
vitamin D fortified skim milk); and a cake mix may be labeled
with one set of Percent Daily Values for the dry mix (per
serving) and another set for the serving of the final cake when
prepared): Provided, That, the type and quantity of the other
ingredients to be added to the product by the user and the
specific method of cooking and other preparation shall be
specified prominently on the label.
    (i) Except as provided in paragraph (j)(13) of this section,
the location of nutrition information on a label shall be in
compliance with 101.2.
    (j) The following foods are exempt from this section or are
subject to special labeling requirements:
    (1)(i) Food offered for sale by a manufacturer, packer, or
distributor who has annual gross sales made or business done in
sales to consumers that is not more than $500,000 or has annual
gross sales made or business done in sales of food to consumers
of not more than $50,000, Provided, That the food bears no
nutrition claims or information on a label or labeling or in
advertising.
    (ii) For purposes of this paragraph, calculation of the
amount of sales shall be based on the most recent 2-year average
of business activity. Where firms have been in business less than
2 years, reasonable estimates must indicate that annual sales
will not exceed the amounts specified. For foreign firms that
ship foods into the United States, the business activities to be
included shall be the total amount of food sales, as well as
other sales to consumers, by the firm in the United States.
    (2) Food products which are:
    (i) Served in restaurants;
    (ii) Served in other establishments in which food is served
for immediate human consumption (e.g., institutional food service
establishments, such as schools, hospitals, and cafeterias;
transportation carriers, such as trains and airplanes;
bakeries,tessens, and retail confectionery stores where there are
facilities for immediate consumption on the premises; food
service vendors, such as lunch wagons, ice cream shops, mall
cookie counters, vending machines, and side-walk carts where
foods are generally consumed immediately where purchased or while
the consumer is walking away, including similar foods sold from
convenience stores; and food delivery systems or establishments
where ready-to-eat foods are delivered to homes or offices);
    (iii) Sold for sale or use only in such facilities; or
    (iv) Sold by a distributor who principally sells food to such
facilities: Provided, That:
    (A) This exemption shall not be available for those foods
that are manufactured, processed, or repackaged by that
distributor for sale to any persons other than restaurants or
other establishments that serve food for immediate human
consumption, and
    (B) The manufacturer of such products is responsible for
providing the nutrition information on the products if there is a
reasonable possibility that the product will be purchased
directly by consumers.
    (3) Food products that are:
    (i) Of the type of food described in paragraphs (j)(2)(i) and
(j)(2)(ii) of this section,
    (ii) Ready for human consumption,
    (iii) Offered for sale to consumers but not for immediate
human consumption,
    (iv) Processed and prepared primarily in a retail
establishment, and
    (v) Not offered for sale outside of that establishment (e.g.,
ready-to-eat foods that are portioned and packaged on-site and
sold by independent delicatessens, bakeries, and retail
confectionery stores where there are no facilities for immediate
human consumption, by in-store delicatessen, bakery, or candy
departments, or at self-service food bars such as salad bars).
    (4) Foods that contain insignificant amounts of all of the
nutrients and food components required to be included in the
declaration of nutrition information under paragraph (c) of this
section. An insignificant amount of a nutrient or food component
shall be that amount that allows a declaration of zero in
nutrition labeling, except that for total carbohydrate, dietary
fiber, and protein, it shall be an amount that allows a
declaration of ``less than 1 gram.'' Foods that are exempt under
this paragraph include coffee beans (whole or ground), tea
leaves, plain unsweetened instant coffee and tea, condiment-type
dehydrated vegetables, flavor extracts, and food colors.
    (5)(i) Foods, other than infant formula, represented or
purported to be specifically for infants and children less than 2
years of age shall bear nutrition labeling, except as provided in
paragraph (j)(5)(ii) and except thatsuch labeling shall not
include calories from fat (paragraph (c)(1)(ii) of this section),
calories from saturated fat ((c)(1)(iii)), saturated fat
((c)(2)(i)), polyunsaturated fat ((c)(2)(ii)), monounsaturated
fat ((c)(2)(iii)), and cholesterol ((c)(3)).
    (ii) Foods, other than infant formula, represented or
purported to be specifically for infants and children less than 4
years or age shall bear nutrition labeling, except that such
labeling shall not include listings of percent of Daily Value and
the footnote required in paragraphs (d)(7), (d)(9), and (d)(10)
of this section. Nutrient names and quantitative amounts by
weight shall be presented in two separate columns.
    (6) Dietary supplements of vitamins and minerals except that
the labeling of a dietary supplement of vitamins and minerals in
conventional food form, e.g., a breakfast cereal, shall conform
to the labeling established in this section.
    (7) Infant formula subject to section 412 of the act, as
amended, except that such foods shall be labeled in compliance
with part 107 of this chapter.
    (8) Medical foods as defined in section 5(b) of the Orphan
Drug Act (21 U.S.C. 360ee(b)(3)). A medical food is a food which
is formulated to be consumed or administered enterally under the
supervision of a physician and which is intended for the specific
dietary management of a disease or condition for which
distinctive nutritional requirements, based on recognized
scientific principles, are established by medical evaluation. A
food is subject to this exemption only if:
    (i) It is a specially formulated and processed product (as
opposed to a naturally occurring foodstuff used in its natural
state) for the partial or exclusive feeding of a patient by means
of oral intake or enteral feeding by tube;
    (ii) It is intended for the dietary management of a patient
who, because of therapeutic or chronic medical needs, has limited
or impaired capacity to ingest, digest, absorb, or metabolize
ordinary foodstuffs or certain nutrients, or who has other
special medically determined nutrient requirements, the dietary
management of which cannot be achieved by the modification of the
normal diet alone;
    (iii) It provides nutritional support specifically modified
for the management of the unique nutrient needs that result from
the specific disease or condition, as determined by medical
evaluation;
    (iv) It is intended to be used under medical supervision; and
    (v) It is intended only for a patient receiving active and
ongoing medical supervision wherein the patient requires medical
care on a recurring basis for, among other things, instructions
on the use of the medical food.
    (9) Food products shipped in bulk form that are not for
distribution to consumers in such form and that are for use
solely in the manufacture of other foods or that are to be
processed, labeled, or repacked at a site other than where
originally processed or packed.
    (10) Raw fruits, vegetables, and fish subject to section
403(q)(4) of the act, except that the labeling of such foods
should adhere to guidelines in 101.45. The term ``fish''
includes freshwater or marine fin fish, crustaceans, and
mollusks, including shellfish, amphibians, and other forms of
aquatic animal life.
    (11) Packaged single-ingredient products that consist of fish
or game meat (i.e., animal products not covered under the Federal
Meat Inspection Act or the Poultry Products Inspection Act, such
as flesh products from deer, bison, rabbit, quail, wildturkey, or
ostrich) subject to this section may provide required nutrition
information for a 3-ounce cooked edible portion (i.e., on an ``as
prepared'' basis), except that:
    (i) Such products that make claims that are based on values
as packaged must provide nutrition information on an as packaged
basis, and
    (ii) Nutrition information is not required for custom
processed fish or game meats.
    (12) Game meats (i.e., animal products not covered under the
Federal Meat Inspection Act or the Poultry Products Inspection
Act, such as flesh products from deer, bison, rabbit, quail, wild
turkey, or ostrich) may provide required nutrition information on
labeling in accordance with the provisions of paragraph (a)(2) of
this section.
    (13)(i) Foods in small packages that have a total surface
area available to bear labeling of less than 12 square inches,
Provided, That the labels for these foods bear no nutrition
claims or other nutrition information. The manufacturer, packer,
or distributor shall provide on the label of packages that
qualify for and use this exemption an address or telephone number
that a consumer can use to obtain the required nutrition
information (e.g., ``For nutrition information, call 1 - 800 -
12394567'').
    (ii) Foods in packages that have a total surface area
available to bear labeling of 40 or less square inches may modify
the requirements of paragraphs (c) through (f) and (i) of this
section by one or more of the following means:
    (A) Presenting the required nutrition information in a
tabular or, as provided below, linear (i.e., string) fashion
rather than in vertical columns if the product has a total
surface area available to bear labeling of less than 12 square
inches or ifannot accommodate a column display on any label
panel. Nutrition information may be given in a linear fashion,
only if the label will not accommodate a tabular display, and, in
that case, any subcomponents declared shall be listed
parenthetically after principal components (e.g., saturated fat
shall be declared in parentheses after total fat). The following
sample label illustrates a tabular display.

BILLING CODE 4160 - 01 - M
BILLING CODE 4160 - 01 - C

INSERT 563.TIF

    (B) Using any of the following abbreviations:
Serving size--Serv. size
Servings per container--Servings
Calories from fat--Fat cal
Saturated fat--Sat fat
Cholesterol--Cholest
Total carbohydrate--Total carb
Dietary fiber--Fiber

    (C) Omitting the footnote and caloric conversion information
required in paragraphs (d)(9) and (d)(10) of this section and
placing another asterisk at the bottom of the label followed by
the statement ``Percent Daily Values are based on a 2,000 calorie
diet'' and, if the term ``Daily Value'' is not spelled out in the
heading, a statement that ``DV'' represents ``Daily Value.''
    (D) Presenting the required nutrition information on any
label panel.
    (14) Shell eggs packaged in a carton that has a top lid
designed to conform to the shape of the eggs are exempt from
outer carton label requirements where the required nutrition
information is clearly presented in no less than 1/16-inch type
size immediately beneath the carton lid or in an insert that can
be clearly seen when the carton is opened.
    (15) The unit containers in a multiunit retail food package
where:
    (i) The multiunit retail food package labeling contains all
nutrition information in accordance with the requirements of this
section;
    (ii) The unit containers are securely enclosed within and not
intended to be separated from the retail package under conditions
of retail sale; and
    (iii) Each unit container is labeled with the statement
``This Unit Not Labeled For Retail Sale'' in type size not less
than 1/16-inch in height. The word ``individual'' may be used in
lieu of or immediately preceding the word ``Retail'' in the
statement.
    (16) Food products sold from bulk containers: Provided, That
nutrition information required by this section be displayed to
consumers either on the labeling of the bulk container plainly in
view or in accordance with the provisions of paragraph (a)(2) of
this section.
    (k) A food labeled under the provisions of this section shall
be deemed to be misbranded under sections 201(n) and 403(a) of
the act if its label or labeling represents, suggests, or
implies:
    (1) [Reserved]
    (2) That a balanced diet of ordinary foods cannot supply
adequate amounts of nutrients.
    (3) That the lack of optimum nutritive quality of a food, by
reason of the soil on which that food was grown, is or may be
responsible for an inadequacy or deficiency in the quality of the
daily diet.
    (4) That the storage, transportation, processing, or cooking
of a food is or may be responsible for an inadequacy or
deficiency in the quality of the daily diet.
    (5) That the food has dietary properties when such properties
are of no significant value or need in human nutrition.
    (6) That a natural vitamin in a food is superior to an added
or synthetic vitamin or to differentiate in any way between
vitamins naturally present from those added.
    5. Section 101.100 is amended by revising the introductory
text of paragraph (d) to read as follows:


101.100  Food; exemptions from labeling.

* * * * *
    (d) Except as provided by paragraphs (e) and (f) of this
section, a shipment or other delivery of a food which is, in
accordance with the practice of the trade, to be processed,
labeled, or repacked in substantial quantity at an establishment
other ty processed or packed, shall be exempt, during the time of
introduction into and movement in interstate commerce and the
time of holding in such establishment, from compliance with the
labeling requirements of section 403 (c), (e), (g), (h), (i),
(k), and (q) of the act if:
* * * * *
    6. Appendix B to Part 101 is added to read as follows:

BILLING CODE 4160 - 01 - M
BILLING CODE 4160 - 01 - C

INSERT 569.TIF
INSERT 570.TIF

    Dated: December 17, 1992.
David A. Kessler,
Commissioner of Food and Drugs.
Louis W. Sullivan,
Secretary of Health and Human Services.
    Editorial Note: The following appendixes will not appear in
the annual Code of Federal Regulations.

BILLING CODE 4160 - 01 - M
BILLING CODE 4160 - 01 - C

INSERT 573.TIF


                           Appendix B.--True Protein
Digestibility Value of Common Foods                          
ÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄ
ÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄ
                                                                  
                              True                 
   Major product group                Subgroup                    
  Product            digestibility   References
ÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄ
ÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄ
Cereals and grains...........  Barley....................... 
Barley......................            90             1
                               Corn and Corn Products....... 
Corn, Extruded Cereal.......            62             2
                                                             
Corn, Corn Flake............            70             2
                                                             
Corn, Degerminated Opaque 2.            94             2
                                                             
Corn, Puffed Cereal.........            76             2
                                                             
Corn, Whole.................            89             2
                                                             
Corn, Whole Opaque 2........            92             2
                                                             
Corn meal...................            84             2
                               Millet....................... 
Millet......................            79             2
                               Oats and Oat products........ 
Oat, Sugared Flakes.........            67             2
                                                              Oat
Flakes..................            70             2
                                                             
Oat, Extruded Oat/Wheat.....            76             2
                                                             
Cereal......................            89             2
                                                             
Oat, Quick Oatmeal..........            82             2
                                                             
Oat, Oatmeal................            90             3
                                                             
Oats, Rolled................  .............  ...........
                               Rice and Rice Products....... 
Rice........................            91             2
                                                             
Rice, Germ..................            87             4
                                                             
Rice, Brown, Cooked.........            72             5
                                                             
Rice, High Protein..........            85             2
                                                             
Rice, Milled, Cooked........            86             6
                                                             
Rice, Polished..............            87          2, 5
                                                             
Rice, Crisped Rice Cereal...            77             2
                                                             
Rice, Flakes................            63             2
                               Sorghum...................... 
Sorghum, Cooked.............            73             6
                               Triticale.................... 
Triticale...................            90             2
                               Wheat and Wheat Products..... 
Bread.......................            96             2
                                                             
Bread, Coarse, Brown........            91             2
                                                             
Bread, White................            99             2
                                                             
Bread, Whole Wheat..........            92             2
                                                             
Wheat, Bran.................            75            15
                                                             
Wheat, Brown, Cooked........            93             6
                                                             
Wheat Endosperm (Farina)....            98             2
                                                             
Wheat, Flour, 90% extracted.            89             2
                                                             
White, Flour, 80% extracted.            92             2
                                                             
Wheat Germ..................            81             2
                               Wheat and Wheat Products..... 
Wheat Gluten................            98             2
                                                             
Wheat, Hard Spring..........            86             2
                                                             
Wheat, Meat Analogue........            95             2
                                                             
Wheat, Puffed Wheat.........            84             2
                                                             
Wheat, Shredded.............            73             2
                                                             
Wheat, Wheaties.............            80             2
                                                             
Wheat, White Flour..........            97             2
                                                             
Wheat, Whole................            87             2
                                                             
Wheat, Whole, Hot Cereal....            85             2
                                                             
Wheat, 40% Bran Flakes......            69             2
Dairy Products...............  Casein....................... 
Acid-Casein.................            95             7
                                                             
Casein......................            96          8, 4
                                                             
Caseinate...................            95             7
                                                             
Rennet-Casein...............            94             7
                               Cheese....................... 
Cheese, Cheddar.............            99             4
                                                             
Cottage, Cheese.............            99             2
                               Lactalbumin.................. 
Lactalbumin.................            94             2
                               Milk......................... 
Milk Retentate..............            97             7
                                                             
Milk, Skim..................            94             8
                                                             
Milk, Whole.................            94             2
                                                             
Milk, Whole, Powdered.......            95             2
                               Whey......................... 
Whey Protein................            95             4
Egg and Egg Products.........  .............................  Egg
albumen.................            97             2
                                                             
Egg, Flakes.................            92             2
                                                             
Egg, Powdered, Dried........            93             4
                                                             
Egg, Dried..................            98             2
                                                             
Egg, Powdered Defatted......           100             2
                                                             
Egg, Scrambled..............            96             2
                                                             
Egg, Spray Dried............            92             2
                                                             
Egg, Whole, Unprocessed.....            97          2, 9
Legumes and Oilseed Products.  Beans (Mucunoa Spp.)......... 
Beans, Velvet...............            68             4
                               Beans (Phaseolus lunatus).... 
Beans, Butter...............            57             4
                                                             
Beans, Lima.................            78             4
                               Beans (Phaseolus vulgaris)... 
Beans, Black................            69         4, 10
                                                             
Beans, Brown, Cooked........            79            11
                                                             
Beans, Common...............            82             4
                                                             
Beans, Haricot..............            71             4
                                                             
Beans, Kidney...............            81             8
                                                             
Beans, Natal round Yellow...            80             4
                                                             
Beans, Pinto, Canned........            73             8
                                                             
Beans, Red..................            78             4
                                                             
Beans, Seafarer.............            84             4
                                                             
Beans, Snap, Frozen.........            82            11
                                                             
Beans, Spotted Sugar........            81             4
                                                             
Beans, Sugar................            69             4
                                                             
Beans, Sugar, Speckled......            78             4
                                                             
Beans, White Kidney.........            78             4
                               Beans (Vicia Faba)........... 
Beans, Broad................            87             4
                                                             
Beans, Faba.................            86            10
                               Cottonseed................... 
Cottonseed..................            78             2
                                                             
Cottonseed Meal.............            80             4
                                                             
Cottonseed, Glandless flour.            98             2
                               Flaxseed..................... 
Flaxseed....................            85             4
                               Lentils (Len culinaris)...... 
Lentils.....................            85          4, 8
                               Lupins (Lupinus albus)....... 
Lupine......................            76             4
                               Peanut Products.............. 
Peanut Butter...............            95             2
                                                             
Peanut Flour................            93             2
                                                             
Peanuts.....................            87             4
                                                             
Peanut Meal.................            91            12
                               Peas (Cajanus cajan)......... 
Pigeon Peas.................            76            13
                                                             
Pigeon Peas, Raw............            41            13
                               Peas (Cicer Arietinum)....... 
Chick peas, Canned..........            88             8
                               Peas (Pisum sativum).........  Pea
Concentrate.............            94             3
                                                             
Peas........................            88            14
                                                             
Peas, Century, Autoclaved...            83            10
                                                             
Peas, Green, Frozen.........            94            11
                                                             
Peas, Trapper, Autoclaved...            84            10
                                                             
Peas, Yellow, Cooked........            86            11
                                                              Pea
Flour...................            88            10
                                                             
Peas, Alaskan Field.........            88             2
                               Peas (Vigna unguiculata)..... 
Cowpeas.....................            79             4
                               Sesame....................... 
Sesame Seed, Dehulled.......            82             4
                               Soy Products................. 
Soybean.....................            91             4
                                                              Soy
Concentrate.............            95             4
                                                              Soy
Flour...................            84             2
                                                              Soy
Flour, Defatted.........            87             2
                                                              Soy
Isolate.................            96             2
                                                              Soy
protein, spun...........           100             2
                               Sunflower.................... 
Sunflower Seed..............            82             2
                                                             
Sunflower Seed-Flour........            90             2
Meats and Meat Products......  Beef......................... 
Beef........................            95             2
                                                             
Beef, Low Fat, Ground.......            91             2
                                                             
Beef, Powdered, Defatted....            97             2
                                                             
Beef, Salami................            98             8
                                                             
Beef, Stew..................            89             8
                                                             
Beefsteak...................            97             2
                                                             
Beef Tenderloin, Roasted....            91             2
                               Fish and Seafood............. 
Haddock.....................           100             4
                                                             
Sardine.....................            95             4
                                                             
Shark.......................            72             4
                                                             
Tuna, Canned................            90             2
                               Luncheon Meats............... 
Canned Frankfurters.........            97             4
                                                             
Chicken, Frankfurters.......            97             8
                                                             
Sausage.....................            94             8
                               Pork......................... 
Pork Loin and Tenderloin....            98             2
                               Poultry...................... 
Chicken.....................           100             4
                                                             
Chicken, Dark Meat..........            92             4
                                                             
Chicken, Light Meat.........            93             4
                                                             
Turkey Breast, Roasted......            91             2
Miscellaneous Foods..........                                
Macaroni/Cheese, Canned.....            94             3
Nuts and Nut Products........                                
Cashew......................            85             4
                                                             
Coconut, meal (defatted)....            80             4
                                                             
Pecan.......................            71             4
Starchy Roots, Tubes.........                                
Potato......................            89             4
Vegetables...................                                
Cabbage.....................            88             4
                                                             
Kale........................            85             4
                                                             
Rape........................            85             4
                                                             
Mustard.....................            82             4
                                                             
Turnip Leaves...............            86             4
                                                             
Mushrooms...................            90             4
ÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄ
ÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄÄ





Notes

    True digestibility values obtained using adult subjects were
considered first followed by data using the rat as the animal
model. When more than one value was considered the values were
averaged. Data sources:
    1. Eggum, B. O., B. Pederson, and I. Jacobson, ``The
Influence of Dietary Tea, Coffee and Cocoa on Protein and Energy
Utilization of Soybean Meal and Barley in Rats,'' British Journal
of Nutrition, 50:197 - 205, 1983.
    2. Hopkins, Daniel T., ``Effects of Variation in Protein
Digestibility,'' in Protein Quality in Humans: Assessment and In
Vitro Estimation, C. E. Bodwell, J. S. Adkins, and D. T. Hopkins
(eds.), AVI Publishing Company INC, Westport, CN, pp. 169 - 193,
1981.
    3. McDonough, F. E., F. H. Steinke, G. Sarwar., B. O. Eggum,
R. Bressani, P. J. Huth, W. E. Barbeau, G. V. Mitchell, and J. G.
Phillips, ``In Vitro Rat Assay for True Protein Digestibility:
Collaborative Study,'' Journal of the Association of Off
    4. FAO, ``Amino Acid Content of Foods and Biological Data on
Proteins'' Food Policy and Food Science Service, Nutrition
Division, FAO of the United Nations, Rome, 1970.
    5. Miyoshi, H., T. Okuda, K. Okuda, and H. Koishi, ``Effects
of Brown Rice on Apparent Digestibility and Balance of Nutrients
in Young Men on Low Protein Diets,'' Journal of Nutritional
Science and Vitaminology, 33:207 - 218, 1987.
    6. Eggum, B. O., B. O. Juliano, M. G. B. Ibabao, C. M. Perez,
and V. R. Carangal, ``Protein and Energy Utilization of Boiled
Rice-Legume Diets and Boiled Cereals in Growing Rats,'' Plant
Foods for Human Nutrition, 37:237 - 245, 1987.
    7. Frangne, R. and J. Adrian, ``Protein Value of Some
Industrial Dairy Products,'' International Journal of Vitamins
and Nutrition Research, 56:391 - 393, 1986.
    8. Sarwar, G., ``Digestibility of Protein and Bioavailability
of Amino Acids in Foods. Effects on Protein Quality Assessment,''
World Review of Nutrition and Dietetics, 54:26 - 70, Karger,
Basel, 1987.
    9. Kaneko K. and G. Koike, ``Utilization and Requirement of
Egg Protein in Japanese Women,'' Journal of Nutritional Sciences
and Vitaminology, 31:43 - 52, 1985.
    10. Sarwar, G. and F. E. McDonough, ``Evaluation of Protein
Digestibility-Corrected Amino Acid Score Method for Assessing
Protein Quality of Foods,'' Journal of the Association of
Official Analytical Chemist, 73(3):347 - 355, 1990.
    11. Lowgren M. and L. Hambraeus, ``Protein Quality in
Legumes-Effect of Soaking and Heat Treatment and Amino Acid
Supplementation on Nitrogen Retention in rats'. Nutrition Reports
International, 38(4):745 - 756, 1988.
    12. FAO of the United Nations, Protein Quality Evaluation,
Report of Joint FAO/WHO Expert Consultation, Rome, 1991.
    13. Griffiths, A. G. and G. P. Savage, ``The Composition and
Nutritive Value of Pigeon Pea (Cajanus Cajan Millsp.),''
Nutrition Abstracts and Reviews (Series A), 61(4) 261 - 283,
1991.
    14. Mongeau, R., G. Sarwar, R. W. Peace, and R. Brassard,
``Relationship Between Dietary Fiber Levels and Protein
Digestibility in Selected Foods as Determined in Rats, Plant
Foods for Human Nutrition, 39:45 - 51, 1989.
    15. Hansen, I. T. Larsen, K. E. Bach Knudsen, and B. O.
Eggum, ``Nutrient Digestibilities in Ingredients Fed Alone or in
Combination,'' British Journal of Nutrition, 66:27 - 35, 1991.

BILLING CODE 4160 - 01 - M
[FR Doc. 92 - 31501 Filed 12 - 28 - 92; 8:45 am]
BILLING CODE 4160 - 01 - C

INSERT 591.TIF
INSERT 593.TIF
INSERT 594.TIF
INSERT 595.TIF
INSERT 596.TIF
INSERT 597.TIF
INSERT 598.TIF
INSERT 599.TIF
INSERT 600.TIF