BEGENT ADVANCES IN BAGTEBIAL GENETICS

S. E. LURIA

Indiana Uniwaity, Bloomington

coNmmTa

I. Axmlysis of spontqous mutsbility. . . . . . . . . . . . . . . . . . . . . . . . . . . .
II. Incresse in mutation frequency produced by non-specific agents..
III. B8cteris1 mutstion 8nd the genetic determirmnta of b8cteri8. . . . .
IV. Specific induction of mutations ,.................................
V. Fusion and eexuslity mech8nieme ,...............................
VI. &k&ion phenomena and evolutionary considerationa. . . . . . . . . . . .

. . . . .
. . . . .
. . . . .
. . . . .
. . . . .

. .   3
. . 16
. .  18
. .  22
. .  !24
. .  26

Bacterial genetics is today at a singular point of development.  Scant knowl-
edge and lack of agreement have until recently prevailed even on the most
elementary facta of reproduction and character transmission in bacteria.  The
occurrence of sexual reproduction, although denied by most workers, was ac-
cepted by several others, mainly on the basis of suggestive but inconclusive
cytological evidence.  Variation in bacteria was interpreted by some as develop-
mental, by others as genetic; and further complications resulted from attempts to
explain by atricty physicochemical theories the supposed specific induction of
bacterial variation (76,58, Xl), making bacteriology one of the last strongholds
of Lamarckism.

Most of the sweeping generalimtiona that have taken the place of a genetics
of bacteria are based upon interpretation of qualitative observations, often ig-
noring delicate population problema involved in di&ngui&ing between cell
character and culture character.  In recent years, however, incressing attempts
have been made to approach bacterial genetics in the same way that haa been
so fruitful in the study of the genetics of higher organisms, that is, by a "quan-
tized" study of pairs or series of discrete, mutually exclusive characters.  Of
necessity, this approach has been limited until very recently to the analysis of
variation, setting a&de-though by no meana ignoring-the problem of the
mechanism of homologous transmission of characters, that is, the problem of
genetic stability of which variation is the inseparable negation. Work has
o@.ered on the mode and mechanism of origin of new characters.  Evidence
for the existence of discrete unit characters and for interaction between character
determmanta has been collected.  Attempts to compare the nature and action
of these dete rminants with those pre8ent in higher orgamam8 have been fruitful,
thus opening the path toward an integration of the genetic systems of bacteria
into a comprehensive compamtive genetics. In turn, bacterial genetic.6 has
offered to the geneticist the unique proof of the possibility of specific induction
of hereditary changes, in the c88e of induced serological type transformation
(10, 22). F'inally, the first genetically convincing evidence for some kind of
mxual process in common bacteria haa been offered within the pa& few months
(91). Although we may fmd oumelves on the threshold of a deep change in
our ideas of bacterial heredity, and possibly because of this, it (leema useful to

1


2                          8. E. LlJlUA            [VOL. 11

brins tOgether those regulta of bacteriological experimentation which satisfy
the quantitative requirements of modem geneticircts.  `II& also in view of the
fact that them have recently come to the fore of the genetic scene certain phenom-
ena-particularly cytoplasmic inheritance-whose interpretation may have
some bearing on the problems of bacterial genetics. .

This review attempts to present recent work on bacterial genetics in a co-
ordinated form, without claim either to complete coverage of the recent literature
or to a detailed re-interpret&ion of the bewildering maas of ohrvatione on
bacterial varisbility.  These have repeatedly been collected in extensive reviews
(102, 73, 23) and have recently been valuably discussed by Dubos (57).  In
the course of this review, however, we shall occasionally attempt to show how
simple genetic principles may offer the lead for a correct interpretation of some
of the most controversial aspecta of bacterial variation.

That he may not be misuhdemtood, because of his choice of material, aa
arbitrarily excluding from the field of bacterial genetics certain observations
that have previously caused much genetic speculation, this reviewer feels it
desirable to put forward at this point a few statementa of opinion:

1. The evidence for reproductive proceases other than.binary f&ion in bac-
teria is today rather suggestive; sexual, or at least fusion processee seem to take
place (91, 93). Their occurrence, however, cannot at present be general&d;
and most of the cytological evidence (102, 118, 119, 123) on the basis of which
such processes have previously been poetulated cannot be profitably discussed,
since cytological obeervationa have not been correlated with the study of trane-
mission of well defined, mutually exclusive charactera.

2. Cytological evidence for the existence of diecrete chromatinic structures
in bacteria, which may be described aa homologous to nuclei, is very convincing
(11,137,98,146,81).  This agreea with the logical expectation for the existence
of home structural device for equipartition of the genetic material at cell fission,
in order to account for such high degree of genetic stability as is encountered
in bacteria. Most of the available cytological evidence is again, however, of
little help to the geneticist, since it does not yet supply either sure proof or
analyaift of such processes aa mitoaie, meio8i8, or chromosomal grouping of genetic
determinanta.

3. It ia impossible today to decide whether 8ome caeea of bacterial variation
represent developmental (life cyclea) rather than genotypic changes. On the
one hand, proof of genotypic identity of two differing bacterial phenotypee
is difhult to obtain; and CBIKB of appamntly cyclic comae of variation can
bply be explained on the basis of mutations, reverse mutations, and d&ion
pbnoma. On the other hand, it ia on the teat of accounting for the oriein
of permanent genotypic diEerencea that the developmental theories of variation
8ppt3W to fail. Nor does it eeem profitable or juati6ed to explain most ca8e8
of bacterial variation, aa has been attempted (165), by eegreg;rrfion of &m&em
from heterosygotea in some form of autogamy rather than by mutatin.  In
fact, this only displacea the problem of the origin of the genetia differences found
in the aup& heterosygote. Mutation, chased off the front porch, is read-


19471        RECENT ADVANCES IN BACX'EBIAL GENETICS          3

mitted by the kitchen door, with the added difficulty that it must now be fitted
into the same household with a highly hypothetical guest.

In view of these considerations, the best approach to a discussion of bacterial
genetics today seems to be an analysis of bacterial mutability in its origin, mani-
festations, causes, and effects on bacterial populations.  We may then attempt
an interpretation on the basis of the available evidence. By "mutation"
we &all mean a permanent change affecting one or more properties of a bacterial
cell and of its offspring. The use of this term does not imply a priorr' identifica-
tion with the process of gene mutation or with any other type of hereditary
ahange in higher orgamsms.  Whatever similarities or differences exist should
be discovered through the study of specific cases.

I. ANALYSIS OF SPONTANBOUS MUTABILITY

1. Detect&m and fmqtmq of mutunta

In the study of bacterial mutability we are faced with the problem of deter-
mining the differential properties of individuals within a population-which
should be a pure line (36)-from the characters of the clones to which each
individual gives origin (colonies, one cell cultures). Upon plating a uniparental
population to obtsin isolated colonies, mutants can be detected directly by colony
observation, if they aEect visible properties, or by testing bacteria from individual
colonies for any desired property.  The number of colonies that can be tested
being necessarily limited, only frequent mutations can be detected by this
method. The frequent mutants present special problems (41). Assuming a
constant mutation rate (see below), the number of mutants increases during
the growth of a culture by multiplication of previous mutants and by new
mutations. In order for the parent type not to be displaced rapidly by a frequent
mutant, the increase of the latter must be kept in check either by reverse muta-
tion or by adverse selection.  If reverse mutations occur, practically every sizable
clone (visible colony) will contain a mixed population, often in equilibrium.
The condition of equilibrium is given by the expression M/N = a/b, where M
and N are the proportions of mutant and normal cells, and a and b the rates
of forward and back mutation.

If the mutant type is handicapped by adverse selection, equilibrium will be
reached at a condition defined by the equation M/N = a/(&u), where 8 is the
"selection coefficient" defined as the difference between the growth rates of
normal and mutant types (50,41).

In most actual cases, the existence of both reverse mutation and growth rate
differences makes the situation very difficult to analyze.  One of the few cases
in which only forward and back mutations are at play has been analyzed in the
beautiful studies of Bunting (30, 31). Studying color variants of Serr&a
marcem, Bunting found that in cultures maintained in the logarithmic phase
of growth various types of mutant cells arose, each giving origin to new color
types at constant ratee. From the rate at which equilibrium was approached,
forward and back mutation rates could be determined ; for the mutation "dark


4                           8. E. LDRIA             [VOL. 11

red-bright pink", these mutation rates were respectively 10-r and 3 x 104
per bacterium per generation.

The difficulty of reaognising from the character of a colony the type of cell
from which it had arisen could be overcome in Bunting's work because of the
high degree of reproducibility of the pattern of variation within each clone,
so that although each colony contained a mixture of types, the proportions
in which these were present must have been characteristic for the type of cell
from which the colony stemmed. When frequent mutations and reversions
are present, the different types of cells can be better defined in terms of such
mutational patterns and equilibria than of any one character of the clone stem-
ming from each cell.

In Bunting's work, selection phenomena were encountered as soon as the
bacterial cultures where studied in the ageing phase (32,33).  A case of frequent
mutation apparently balanced only by adverse selection was that of the mutation
R.-S in Salmu a&q&e (50,154) which was found to occur at a fixed rate
independently of the medium, whose only action was to alter the selection co-
efficient.

It is clear that, for most quantitative studies on the mutation process, fre-
quent mutations are unsuitable, because of the difficulties encountered in deter-
mining the characters of the individual cell from those of the offspring.  Rare
mutations present different problems (106,41,105). A strongly selective environ-
ment is required to detect the presence of the mutants, and only mutants capable
of growth in environments very unfavorable to the normal type can thus be
detected (mutants capable of growth in media insufficient for the parent type,
or resistant to inhibit& age$+yr producing detectable ferment&ions).  The
problem here arises of proving the' spontaneous origin of the mutations.  Since
the mutants can be detected' only after exposure to the special environment,
the hypothesis that the new hereditary cl&a&r has been induced by this
environment cannot easily be ruled out.  The typicalexample is that of Escher&
cl& co&m&bile (Massini, 117): the parent strain does not ferment lactose
but gives a stable lactose-positive variant. The opinion that the change is
induced by exposure to lactose has been held by a number of authors. The
demonstration by Lewis (96) that s 6xed proportion (about 2 X 1W`) of the
cells in a negative culture grown without lactose will give positive colonies when
tested in lactose has been considered by many authors as proof of the spontaneous
origin of the variants.  This evidence, although strong, is, however, not really
conclusive: the same result would be obtained if exposure to lactose that is
required for the final test produced the change in a constant, low proportion
of the cells. The same can be said, for ewrmple, of acquired resistance to
salts (153).

The problem of proving the spontaneous origin of'rare mutations on the
basis of the frequency distribution of mutants was analysed by Luria and
DelbrQck (106) in relation to acquired &stance to bacteriophage. The dis-
tinguishing features of the distribution of the numbers of spontaneous mutants
are those mffecting the clonal grouping of mutants in the cultures where they


19471          BBCENT ADVANCES IN BACX'EBIAL GENETIC.6        5.

originate, each clone stemming from one independent mutation. The first
eharmteristia feature is an increase in the proportion of mutants during growth.
This increase is difficult to establish for rare mutants, since the occurrence of
rare mutations, obeying the laws of chance, is subject to large fluctuations, and
successive samples taken from the same culture or from similar cultures give
very erratic resulta.  Them fluctuations, however, are themselves a distinguish-
ing feature of spontaneous mutations.  If a change were induced by the test
environment in a certam' proportion of cells, this proportion should not differ
from sample to sample, whether the samples come from the same cultureor from
different ones.  If, however, the variants originate by mutation prior to the
test, the chance occurrence of rare mutations will be reflected in large variation
in their time of occurrence, and, therefore, in the number of individuals present
.in each mutant clone. This, in turn, will result in large fluctuations in the
proportion of mutants in different wild-type cultures.  Presence of such fluc-
tuations in the number of mutants between cultures that have grown from one
or few wild-type cells is strong evidence of clonal grouping, and hence, of muta-
tional origin of the variants. This "fluctuation" test for spontaneous mutation
was applied 6rst (106) to proving the mutational origin of phage resistance,
evidence for which had previously been offered (35)' and which had been assumed
by' several authors (see 29).  This experimental material is a most favorable
one, since every single resistant cell can be isolated after quick lysis of sensitive
populations of enormous sises.  Using Eschetichia coli strain B and phage Tl,
enormous fluctuations were found in the proportions of resistant mutants present
in series of similar cultures started from few sensitive cells.  A wide distribution
,of the numbem of mutants was thus proved.  In the absence of selection, the
actual distribution of the mutants should only be a function of the mutation
rate. The theoretical distribution to be expected from the hypothesis of a
constant mutation rate (probability of mutation per cell per unit time) was not
calculated because of mathematical difhculties, but an approximation to it
(106) closely approximated the experimental distribution.

The intereat of this type of analysis, besides the proof of spontaneous origin
of mutation to phage resistance, is the possibility of defining and calculating
mutation rates as intrinsia properties of the strains.  Two methods were given
by Luria and DelbrGck for the determination of mutation rates, one of them based
on the proportion of cultures without mutants, the other on the average number
of mutants per culture.  The second method can be used only when selection
for or against the mutant does not occur.  Both of these methods yield only
rough estimates of the mutation rates; their limitations have been discussed
elsewhere (106, 105).

,

Mutation rates were measured in probability of mutation per bacterium per
generation (106).  The choice of a physiological time unit was justified, at least
foF the case in question, by the 6nding that the mutation rate thus defined
was the same in cultures of the same strain in different media, in spite of differ-
ences in growth rate, and by the demonstration that no new mutation occurred
after multiplication in the cultures stopped.  The same type of analysis was


6                         8. E. LlJRIA            [VOL. 11

applied by Demerec and Fano (47) to the study of the origin of resistance to
other strains of phage.  Mutation rates varying from 10-D to lO-' were found.
The "fluctuation" test has also been used in the study of resistance to penicillin
(4445) ad tdfc~ddea (199) in Staphyk~ccus, resistance to radiation in
Es&e&i& ccli (178), and in the cases of the mutation from hi&line dependence
to histidine independence in Es&ri&u c& (148), and to uracil independence
in CbHium 8e@iCM (150). Some of these studies will be discussed later
in relation to other aspects of bacterial mutability.

The demonstration that permanently acquired resistance to a number of
antibacterial agents is acquired by spontaneous mutation is likely to be of
general applicability to most types of resistance. It appears to contradict
those theories acaording to which acquired resistance is explained in terms
of a direct action of the antibacterial agent on the ensyme systems of the bac-
terial cell (76, 58)' although such Lamar&an theories are often revived with
rather surprising unconcern for the general outlook of modern biological thought
(151).
Experiments on acquired n&stance to antibiotics in Staphylococcus have
recently led Abraham et uZ. (la) to conclude that adaptation rather than muta-
tion is the me&a&m involved.  Their data do not offer evidence for this con-
clusion, which appears to be based on misconceptions on the occurrence of muta-
tions in pure cultures and on neglect of the population problems outlined above.

Presenae of fluctuations in the number of mutants in similar cultures can be
detected by inspection of the data of Lewis (96) for EtuAvichti w&mWk,
and of Kr&nsen (87) for several fermentative mutations in Salmonella.  The
spontaneous origin of these ferment&ive mutations seems altogether well
established.

One more result from the work on phage resistance should be noted (106) : cul-
turea ime found that contained only one mu&& c.cU.  This was taken to indicate
that a mutation can become phenotypical in the first cell in which it appears.
Should the wild-type character persist in the phenotype of the mutated cell
for one or mom cell generations, the change might be expected to affect two or
more cells at its first appearance (however, see 160).

Mutation rates for the diveme mutations listed above were found to range
from lO-1e to lO+ (105).  Their values can be considered as rough estimates
at best, particularly because in some cases them was evidence of selection for
or against the mutant. An approximate method for estimation of mutation
rata was used by Lincoln (97) in the study of colony type variants in Phylommurs
utuwartii: the ratio of the total number of mutants to the total number of cells
wast&masamaximumvalue(assmm *rig each mutant to arise from a separate
mutation); the ratio of the number of cultures in which mutants were not found
to the number of cells examined gave a minimum estimate (assuming not more
than one mutation per culture).  The mutation rates thus estimated were be-
tween 1 x 10-O and 5 x lo-`.

In au thoroughly analyzed cams, we see that bacterial variation, including
apparent hereditary adaptation, is the result of sudden s~ntaneous mutations.


19471          BECENT ADVANCES IN BACTEBIAL QENEZICS          7

There are some cases, not yet investigated quantitatively, of apparently very
slow adaptation, which might be difficult to interpret in terms of selection of
one-step mutants (138). These cases, however, may be interpretable in terms
of successive discrete mutational steps, as in the cases of quantitative resistance
to be discussed in section I, 3, a.

9. An&& oj mutant char&u

There are two approaches to an understanding of the mechanism of mutation.
On the one hand, one can focus the attention on the mutational step itself,
its statistical regularities, its independence of or interdependence on other
mutations or on environmental conditions.  On the other hand, one can analyze
the effects of mutations in terms of specific physiological and biochemical
changes, and attempt to retrace the primary mutational change from ite end-
effects.  This corresponds to the study of physiological genetics in higher organ-
isms, as examplified by work on hereditary anomalies of metabolism in man
(65a), on pigment inheritance in mammals (183) and in insects (62)' on flower
color in plants (15Oa), and on biochemical syntheses in Neurospora (16, 13).
The results of work on biochemical genetics have been summarimd recently
by Beadle (14).

The concept has become. widely accepted that a gene affects a character by
determining the presence and specificity of one of the ensymes whose action is
necessary for the appearance of the character.  Beadle and his collaborators
have put forward and experimentally supported the generalisation that each
gene acts by controlling one specific ensyme.  If the products of the reaction
catalyzed by the enzyme are utilised in several chains of reactions, a gene change
may affect more than one character.  This "one gene, one ensyme" theory
(170) has proved fruitful as a stimulw to studies in biochemical genetics, that
in turn have provided a powerful tool for the analysis of biochemical syntheses.
Gene mutations can produce specific metabolic blocks by suppressing the activity
of specific ensymes. By studying the influence of individual gene mutations
on production of nutritional re quimments and on accumulation of intermediary
metabolites, numerous chains of reactions have been traced in some detail,
even though the corresponding exmymea have not been isolated. As to the
mechanism of en- regulation by the gene, no conclusive evidence has ap-
peared.  Emerson (61) dimusmd the hypothesis that geneenzyme relation may
. depend on a kind of complementary surface action-similar to that suggested
by l%uling for antibody formation (135)-m which the gene would act by pro-
viding, directly or indirectly, a specifia template for the synthesis of the ensyme
molecule.  The supposed analogy between antibody and ensyme formations
has led to the suggestion that antigens may actually be primary gene products
(74,167)' and that antibodies may affect the genes themselves with production
of mutations, suggestion for which experiments on Neurospora, still awaiting
confinzurtion, have offered some support (60).

.

It must be said that the hypothesis that each gene operates by regulating the
presence and activity of one specific ensyme cannot be considered as more than


8                         8. 1. LumA            [VOL. 11

a fruitful working hypothesis.  Most of the metabolic studies on Neurospora
that support this hypothesis dealt with mutations causing nutritional deficiencies,
bound to be due to suppression of specific ensyme systems.  It is not apparent
today how crucial proof for or against the hypothesis may be obtained (41a).

In bacteria, most studies on variation have dealt with characters whose inter-
pretation in biochemical terms is still difficult. The vast field of antigenic varia-
tion and of variation in virulence (see 57) belongs in this category.  An inter-
pretation of pigment variation in terms of changes in the undoubtedly
complicated synthetic reactions involved has not yet been attempted. Mutations
involving changes in specific enzymatic reactions, however, have been described
in recent years in increasing number.

Mutants with increased growth requirements (loss of ability to synthesize
an essential cell constituent) can be detected directly by picking from individual
colonies into a complete medium, and, after growth, testing the individual
cultures for ability to grow in a minimal medium sufficient for the parent strain
(147, 69).  Mutants with nutritional deficiencies will not grow; their require-
ments can then be identified by determining the additions necessary to permit
growth.  The limited number of colonies that can be tested restricts this method
to the detection of rather frequent mutants, although improved "screening"
techniques can increase its efficiency (92).

Roepke, et al. (147) isolated by this method a number of deficient mutants
of Eezherkhh wli, requiring nicotinami&, thiamine, methionine, cystine, lysine,
arginine, threonine, and tryptophane, respectively. One additional mutant
required either glycine or serine.  Some of these mutants were obtained from
x-ray treated cultures, others from non-irradiated controls.  Biotin and threonine
de6cient mutants of EacheridrM wli, and four mutants of Acetobactar mdano-
genum requiring glycine, serine, leucine, and adenine or adenosine, respectively,
were isolated by Gray and Tatum after x-ray treatment (69). Additional
deficient mutants of Esehetidtio wli were isolated later by Tatum and his collab-
orators (166, 169)' and mutations of the same type were also discovered in
Badha subtilis (see 169) and Cbstniiim 88pticum (150). The amounts of
growth factors required in each case to produce maximum growth varied in
individual cases between 104 and 5 micrograms/ml, being generally lower for
vitamins than for amino acids, purines, or pyrimidines.  A discussion of the
individual Wings in their relation to the chemistry of bacterial syntheses would
be beyond the scope of this review; many of the pertinent data have been dis-
cussed by Tatum (169). All evidence indicates a basic similarity between
mutation-produced deficiencies in bacteria and in Neurospora, suggesting
similarity of the genetic mechanisms involved.

Synthetic deficiencies have also been detected in mutants isolated because
of some other effecte of the mutations.  Anderson (4, 5) found that a certain
mutation to phage resistance in Ea&ert&u wli strain B, produced inability
to grow in synthetic media, and identified the required nutrilite as tryptophane.
The mutants are, moreover, unable to utilize ammonia nitrogen unless supplied
with a relatively large amount of any of a number of amino acids.  Deficiencies


1947]         RECENT ADVANCES IN BAcrERLuJ GENETICS          9

for proline and for some other unidentified nutrilites have also been found coupled
with phage resistance (181i b). These amociations were proved to be due to
multiple e!Iects of the same mutations and not to fortuitous coincidences of
several mutations in the same cell.

The association of metabolic alterations with phage resistance is particularly
interesting for a number of reasons. Fit, it offers the possibiity of using
phsge n&stance as the selective agent in isolating the mutants, so that all
cells with the double mutated character present in a population can be isolated
and counted. Second, it indicates the possibility of interpreting phsge re-
sistance in terms of specific metabolic changes, and hence, of deriving information
on the biochemistry of phage growth.

An interesting metabolic mutation is the loss of ability to synthesise methionine
in Eschericlria wli while acquiring sdf~namide resistance upon transfers in
broth containing sulfonamide (&i). This case illustrates the role of selection
in the establishment of a mutant type. Sulfonamide is known to interfere
with the synthesis of methionine (156).  Growth in the presence of methionine
and sulfonamide must have selectively favored a mutant in which the sul-
fonamide-sensitive reaction leading to methionine synthesis was blocked. In
the absence of methionine the mutation would have been lethal.

In contrast to the mutations just discussed, another group of mutations has
been described which produces increase in synthetic powem (decreased growth
requirements). These mutsnts are more easily detected, since they grow selec-
tively in a deficient medium on which the parent strain cannot multiply.  Even
few mutants in a large population can be detected.  Some of these mutations
occur in strains which were deficient upon first isolation; others appear as re-
versions of mutations producing synthetic deficiencies. To this category
probably belong the cases of so-called "training" to dispense with essential
metabolites, for example, tryptophane adaptation in EberUleZZu @hposcr (64),
nicotinamide adaptation in ShigeUa ptarad~s~ (86) and in Proteus tm&wi~
(140), and thiamine independence in F'ropionibacterium (157). UxaciJ inde-
pendence in CZ&ridi~~ septicunr was shown (159) to be produced by mutation.
Histidine independence from a histidine deficient mutant of Escherichia wli
(148) occurred at a rate of RF* per cell per generation.

Such casw explain the features of training experiments.  Traiing is usually
obtained either by transfer of heavy inocula to deficient media (chance of
transferring at least one mutant), or by successive transfers in decreasing amounts
of the nutrilite (selection for mutants that may arise during the initial growth
of the normal strain in the partially deficient medium) (see also 65).

To designate a mutant which regains the ability to grow in media not contain-
ing any growth factor, Ryan and Lederberg (149) introduced the term "proto-
troph". When prototrophic mutants appear 88 reversions after mutations to
deficiencies, it is important to establish if one is dealing with true reverse muta-
tion or with a different mutation supplying an alternate pathway for the syir-
thesis previously blocked. One possible way of deciding the question might
be based on the expectation that a reverse mutation should reestablish the


10                        8. a. LumA             [VOL. 11

ata& quo, restoring not only the primitive character, but also its mutability
to the deficient form at the same origind rcrie. This expectation may, however,
be misleading, because of possible interactions with other mutations which may
have accumulated in the meantime, and may act as modiflem of character ex-
pression or of mutability (see section VI, b).  In cases of metabolic deficiencies
associated with changes in other characters, it has occasionally been possible
to prove that the return to prototrophic condition was not due to reverse muta-
tion. For example, Wolhnan (181) found that proline deficiency and phage
resistance could be produced by one mutation in Emherichk wli, but the apparent
reversion to proline independence gave mutants which were still phage resistant.
The apparent reversion must have been due to an independent mutation.

Other mutations producing what appear to be increased synthetic powem are
those causing increased formation of some metabolitc (pantothenate in pan-
toyltaurine resistant Coryebcretetium dips (116), paminobensoic acid
in sulphonamide msistant staphylococci and other organisms (38, 89). In-
creased synthesis in these cases may be more apparent than real, since it may
result from accumulation by the mutant of an intermediate metabolite more
completely utilised by the parent type.

Appearance of new ensymatic functions by mutation is the feature of a whole
group of phenomena involving utilization and fermentation of simple carbon
sources. These phenomena include the classic examples of fermentative varia-
tion in Escherichicr w&mu&bile and other gram negative bacteria, and the varia-
tion in utilisation of various organic compounds as sole course of carbon.  A
typical example is that of MotczzcUu Zwofi (110, 8, 111).  This bacterium can
utilise alcohols but not sugars nor several fatty acids as sources of carbon.
With heavy inocula of normal (N) cells in a medium containing succinate only,
late growth appears, due to the presence of mutants (S+) in a proportion of
about one in 10' normal cells. Succinate can be replaced by either fumarate
or malate. The mutation involves a stabilisation of the ability to decarboxylate
oxaloacetic acid with the production of pyruvic acid: this ability disappears
in one hour in strain N, while it persista in S+.  A study of the effects of suc-
&ate, fumarate and malate on the splitting of oxaloacetic acid by S+ led Lwoff
(109) to suggest that the mutation involves a change in an ensyme (or in a
common precursor of a group of ensymes) catalyriing the first reaction in the
attack of 4-carbon acids (phosphorylation?).  It was not decided whether the
mutation produced the transformation of the enzyme into a more stable form
or the removal of an en&me inhibitor. Two other mutations in MoroxeUa
Zwa$i, occurring independently of S+, produce ability to grow with glutamate
or glutarate respectively as carbon source.

Mutational acquisition of ability to utilise citrate by E8cherkhk wli was
described by Parr and his collaborators (132, 133, 134). It seems that the
normal cells can undergo some degree of multiplication on citrate ; growth stops
early, however, whereas growth of the mutant is much more vigorous. In
Parr's experiments, the-citrate positive mutants appeared long after growth of
the normal cells had stopped.  The mutants can, however, be detected imme-


19471        RECENT ADVANCES IN BAcrE* cmNmIcs         11

diately in a negative culture provided large enough inocula are used (184).
In complete media, the mutants are spec&ally inhibited by some unknown prod-
uct of the metabolism of the normal cells (184).

Parr and Simpson (134) described the occasional finding of some stable citrate
negative variants not giving any positive mutant.  The data do not encourage
speculation as to the possible origin of stable citrate positive and citrate negative
forms as homosygotes by segregation in a heterosygotic mutable type.  Still
less just&d appear such speculations when applied (165,166) to cases in which
,a mutable non-fermenting bacterial strain constantly throws off some apparently
stable fermenting variants, but no stable non-fermenters.  The classic case is
that of Escherichio wZ&nWtabi~ (117,96) although many similar cases have been
described in the literature (see 67). &che&hia wli-nutabile, not fermenting
lactose, gives a Gred proportion of lactose-fermenting mutants, which also can
grow with lactose as sole source of carbon; the mutants appear to be stable,
which only means, however, that they do not give lactose negative mutants
in detectable numbers.  It has been suggested (39, 46) that the difference be-
tween lactose negative and lactose positive cells consists in different permeability
to lactose.

More light on the mechanism of these mutations came from Monad's studies
on E8cherkhti wZi-mut&ik strain M.L. (129). Ability to ferment lactose is
dependent on an adaptive enzyme system: the mutation L- to L+ produces
adaptability.  The same strain was found to give also an interesting mutation
affecting galactose utilisation.  The normal strain G- grows slowly in gala&se,
while the mutant G+ grows very fast. Experiments proved that the galac-
tosymase activity of G- is inhibited by 8ome product of galactose utilization,
whereas G+ produces less of this inhibitor and is not inhibited by it.  This
type of mutation in which an apparent increase in ensymatic activity is actually
brought about by the overcoming of an ensyme inhibition, has been called
,`anaphragmic" (from ana = over, and phragmos = barrier) by Lwoff (199),
who suggests that many types of apparently positive mutations, for example,
the citrate positive mutation, may fall into this category.  Interactions between
the mutations L+ to L- and G+ to G- in the same strain has led Monod to
suggest that both mutations affect a common ensyme precursor.

Altogether, biochemical studies of bacterial mutants show that many muta-
tions affect specific enzyme systems, and it is often possible to attribute the
effects of a mutation to a change in one speciik enzyme. Bacteria may actually
offer a most favorable material for the study of mutational enzyme changes.
The relation between biochemical variation and evolutionary trends will be
discussed later.  We should point out here, however, that bacterial mutations
suppressing synthetic ability will often act as lethals.  A block of an essential
ensymatic synthesis will suppress growth, and therefore be lethal, unless the
organism happens to find in ita environment, and absorb from it, the product or
products of the missing reaction.  In some cases, the missing essential metabo-
lites may be replaceable by the products of some other ensyme reaction. A
mutation lethal under the conditions of a given experiment will generally fail


12                       8. m. LUEUA            [VOL. 11

to be detected unless occurring at extremely high rate, in which case it may
reveal itself by reduced viability of the population as a whole.
It should be kept in mind that mutations might affect bacterial characters
by mechanisms other than ensymatic changes.  This may be the ~888, for exsm-
ple, for antigenic variation, including the well known cases of "phase" trans-
formation, which, although seldom analysed from the point of view of the mecha-
nism of their origin, are probably caused by spontaneous mutations.  The role
of specific antisera in bringing forward antigenic variants has not been analysed
sufficiently, but it seems probable that in most cases the antiserum acts by
inhibiting growth of the cell possessing the antigens with which they combine
and allowing the variant cells to grow undisturbed or with less inhibition (57).
It is difficult to decide by which mechanism a mutation alters the antigenic
pattern; the primary change may be supposed either to affect the mold or
template on which the antigen is shaped (61), thus directly affecting the antigenic
structure of the cell, or to alter some enzyme system involved in antigen syn-
thesis. The frequent association of antigenic variation with metabolic changcs
might favor the second hypothesis. An int.eresting observation is that of
P. Rordet (25) that growth at room temperature (20') causes a completely re-
versible transformation of a smooth strain of Ewherichiu wli & into a phenocopy
#A320 of the stable rough mutant #R, which can originate by mutation from $S.
Upon growth of fi at 26"C, the glucolipidic antigen is not formed, although the
potentiality to produce it is present as can be shown by returning #XUl to 37".
It appears that in this case the mutation S+R causes permanent suppression
of a synthetic reaction which in the S strain does not take place at 20".
It is to be expected that any mutation altering the chemical structure of some
bacterial protein or of some compound with haptenic properties may result in a
change in the antigenic properties of the cell if the compound afIected is located
on the cell surface.

S. R&tiolls betwtm mutant t.hmet.ers

a. Independent mutations. Most bacterial strains can undergo changes in
a variety of characters. This great variability of bacteria, often interpreted'
as a biological peculiarity of these organisms, can simply be explained by the
relatively enormous sise of bacterial populations, which offers an opportunity
for occurrence and detection of even rare mutations.  That various characters
of the same strain can vary independently has repeatedly been observed, and
interpreted as not supporting the "life-cycle" theories of variation (142, 82,
143, 144). The independent variability of different characters has been also
interpreted (141) as suggesting mutations of different genes.  Quantitative stud-
ies, however, on the independence and interdependence of mutations have only
recently appeared. Before analysing their results we must briefly deal again
with the problem of multiple effects of mutations.

The already mentioned cases of association of synthetic deficiencies with phage
resistance as a result of the same mutational step are good illustrations of such
multiple effects (4, 5, 181). Other changes frequently associated with phage


19471         RECENT ADVANCES IN BACTERIAL C+ENETICS          13

resistance are variations in colony type and antigenic properties (35, 47, 175).
An interesting type of variation is that of changes in growth rate: a large number
of phage resistant strains, for example, has been found to grow at a slower rate
than the parent type (105) in the regular media.  Differences in growth rate and
death rate have also been found associated with S+R variation (50, 26).  In
many such cam, it could be proved that the various changes resulted from the
same mutational step. The simplest explanation of these cases, which we might
call "pleiotropic" mutations (pleiotropic = producing more changes), in analogy
to the expression "pleiotropic genes" (54)' is that the mutation affects an en-
symatic reaction involved in more than one chain of reactions.  The phenotypic
result of the mutation will be a change in all characters controlled by the affected
reaction chains. Phage resistance may be associated with inability to synthesize
a certain amino acid because of a block in a reaction responsible for the synthesis
of a precursor of both the amino acid and the surface receptor for the phage (4).
Need for two or more growth factors may arise from a mutation affecting the
synthesis of a common precursor of the amino acids or of the enzymes involved
in their synthesis (169).

The above interpretation is in line with the "one gene, one ensyme" theory
of gene action. It is, however, possible that multiple effects of mutations may
actually result from changes in multiple primary functions of the same deter-
minant center (gene). The problem of pleiotropic gene effects, controversial
in the case of higher organisms, is even less susceptible of fruitful discussion at
the present stage of bacterial genetics.

Independently of their significance for the mechanism of the mutational
p&xss, multiple effects of mutations are important because, in the absence of
crossing test, they facilitate the identification of a given mutational step when
it occurs in strains already differing from one another by one or several mutations.
Thus, it becomea possible to study the influence of various genotypes on the
frequency and effects of one mutation.

In the clearer cases, the same mutation can be proved to occur in strains al-
ready differing in one or more mutant characters.  This was particularly well
demonstrated for mutations to phage resistance in Escherichiu wli strain B (104,
47). Demerec and Fano, in particular, showed that a mutation to resistance to
a given phage generally occurs at the same rate in the wild type and in a series of
mutants (47). This was considered as suggesting `changes at different genetic
loci &her than a series of allelic changes at the same locus.  Similar results
were obtained (8) for three independent mutations causing ability to utilize
tierent dicarboxylic acids in Moraxellu lwofi.  It is lmown that mutations caus-
ing synthetic deficiences in bacteria also occur independently, and it has actually
become common practice in their study to utilize, as a source of new mutations,
strains "marked" by one or more mutant characters (genetic markers) in order
to insure against misinterpretation of accidental contaminanta as mutants (163).
The new mutants must show the original mutant character in addition to the
new ones.

The proof of independent mutability at different loci has made the study of


14                         8. ID. LlnuA            [VOL. 11

"mutational patti" (41, 1OS) the most suitable method for an analysis of

bacterial genotypes, because of the possiblity of tracing the presence of certain
genetic loci in different bacterial strains.

An in-g wee is that of several mutations &ecting the same hereditary
trait.  This is well exemplified in the study of quantitative characters, for exam-
ple, of reeistance to various concentrations of drugs or antibiotics.  Demerec's
work (44,45) on resistance of staphylococci to penicillin showed that resistance
to increasing concentrations of the antibiotic is acquired by a series of successive
mutations, each producing further resistance.  Since the various mutations occur
at comparable low rates, a sensitive strain will not directly give highly resistant
mutants: these will appear only after the low grade resistants have been selected
in presence of low concentrations which allow them to grow.  Are we dealing
with a series of mutations tiecting, in different degrees, the same function or
with independent mutations aflecting different metabolic functions involved in
penicillin sensitivity7

The question could be partially answered by the study of the genetics of sul-
fonamide resistance in a strain of Staphylococcw aureus (129).  Here too, there
are a number of mutations-at least five-that cause small increases in resist-
ance; some of them can be distmgui&d because they produce different levels
of resistance when occuring in the same strain.  The study of associated efIects
of these mutations made it possible to single out one or two of them as causing
constant increase in extracellular production of paminobensoic acid (see also 88,
89, 152).  These mutations could thus be recognixed when occurring in strains
already having dXerent sulfonamide tolerance.  It is apparent that resistance
can be produced by alteration of a series of different sulfonamide-sensitive cell
functions, each of which can be &ect..ed by one or more non-lethal mutations
(see I!%).  Apparent increase in p-aminobensoic acid production may actually
be due to increased excretion because of a non-lethal block of its utilisation,
which, naturally enough, Would result in increased sulfonamide tolerance.

Permanent, hereditary resistance to various salts, as distinguished from tem-
porary .adaptation (55) is also acquired by a stepwise mutational process (153).
A similar situation, however, does not hold for all cases of quantitative resistance.
Ultraviolet sensitivity in Etxkrichia wZi atrain B seems to be affected by one
mutation only (178, 179), which produces a moderate degree of resistance, ap-
parently by suppression of the ability of the wild type to react to small doses
of ultraviolet radiation with an inordinate, semi-lethal increase in synthesis of
protoplasm (elongation not followed by cell division).

b. Nca-i&p,&ent mubtitma.  Up to now, we have discussed cases of differ-
ent mutations occurring independently, possibly at difIerent genetic loci, even
when they affect the same phenotypic trait. In the study of resistance of
Eacherkhia WE strain B to phages Tl-T7, however, cases of complex interrela-
tions are found, which may require additional assumptions for their intarpreta-
tion (106).

Reeiatance to one given phage can reeult from any one of a number of di&rent
mutations recognisable by other e&cts+nmh as msistance to some other un-


1947]         RECENT ADVANCES IN BACTERXAL QENETICS         15

related phage. If we indicate msistance to a phage Tn by the symbol /n,
resistance to phage Tl, for example, can result either from the mutation /I, also
producing tryptophane requirement, or from the mutation /1,5 causing also
resistance to phage T5. It is easy to prove the independence of these two muta-
tions, which occur one after another at similar rates in the same clone, with
resulting superposition of the corresponding phenotypes (104).
Another group of mutations, however, occurring more rarely, produce re-
sistance to a number of phages, and the pattern of resistance resulting from
each mutation is the exact superposition of that which can be produced by two
other mutations also known to occur separately at independent rates (105).
For instance, a mutation /1,5,3,4,7 produces exactly the same phenotype obtain-
able by successive mutations /1,5 and /3,4,7, or by the reverse series /3,4,7/1,5.
This can be confirmed by examination of many other effects of the mutations
involved, which are all found in the complex mutant.  The complex mutation,
although occurring more rarely, is too frequent to be due to chance occurrence
of the two simpler ones together.  Occasionally, some complex mutations are
found to produce, in addition to character changes caused also by the simpler
mutations, additional phenotypical efIects, most of them indicative of deep
metabolic disturbances (very slow growth, lack of gas production from sugars).
How are these complex mutations to be interpreted? According to the one
gene, one ensyme hypothesis, one could simply assume that the two simpler
mutations atrect different ensymatic me&msms
                                      .   blocking separate reaction
chains, while the complex mutation blocks a third reaction common to both
reaction chains. This interpretation seems not only improbable-in view of
the extreme complication of the reaction chains to be postulated in order to
explain even a limited number of actual cases-but also rather pointless. In
effect, such type of explanation, if repeated ad in$nitum, might become a purely
verbal interpretation, impossible to disprove (41a) and interfering with the recog-
nition of other possible genetic me&anisms. Another view might be taken by
assuming that the simpler mutations result from allelic changes at the same ge-
netic locus, while the complex mutations represent a third allelic change.  This
interpretation seems unlikely, in view of the completely independent occurrence
of the eimple mutations, as discus& above.
A more likely mechanism appears to be one by which several mutations at
different loci can occur together, by a deeper change in some material center
carrying the hereditary determinants. This seems supported by the occasional
association of deep metabolic disturbances with the complex mutations. It
would be unjustifiable to debate now whether this center may be a complex
molecular unit endowed with several, independently mutable specificities, or
a more complex unit similar to chromosomes of higher orgsnisms.
It could &ally be suggested that complex mutations may occur because of
some special conditions enhancing mutabiity and aBecting simultaneously two
or more distinct functions of the same cell.
Another type of unusual interaction between mutations to phage resistance
has recently been found to involve an elect of one mutation on the rate of ap-


16                         8. E. LTJRIA            [VOL. 11

pearance of a diEerent one (unpublished experiments by the reviewer). In-
fluences of the genotype on the pattern of variability have often been described
in bacteria, but in most cases they are likely to represent effects of the genotype
on the rate of selection for or against the mutants, rather than effects on mutation
rates.  In Edwich~ wli strain B, however, it was found that a mutation /2
(causing resistance to phage T2) does not occur with any appreciable frequency
in the wild-type strain B but is found to occur regularly at rather high rate
(about lo-' per cell per generation) in a series of mutants B/3,4,7 distinguishable
from one another by a number of minor difIerences.  The different rate of ap-
pearance of the /2 mutation is not due to d.ifIerent selection; it is also unlikely,
for a number of reasons, that the mutations /2 and /3,4,7 represent allelic
changes. We must then consider, either an effect of the mutation /3,4,7 on
mutability at a different locus-such as have been found in higher organisms
(145)-or an interaction between the effects of two independent mutations.  The
mutation /2 might actually occur at the same rate in wild-type and B/3,4,7
mutants, but its effect may remain masked in the former because of a "suppres-
sor" effect by the wild-type genotype, which effect is eliminated by the mutation
/3,4,7.

This possibility has been mentioned because of the interest which the suggested
mechanism might have for the genetics of bacteria (see Section VI).  Moreover,
complex mutations involving changes in a number of independently variable char- .
acters might be simulated by the occurrence of a "revealing" mutation in a cell
already carrying a "suppressed" one, although this is probably not the explana-
tion of the complex mutations discussed above.

Altogether, it appears that intensive study of the mutability patterns of
some typical representative bacterial strains and of their mutants should offer
a most interesting way of gaining insight in the basis of hereditary processes in
bacteria.  More work in this direction is greatly desirable.

II. INCREASE IN MUTATION FBEQUENCY PRODUCED BY NON-SPECIFIC AGENTS

Induction of bacterial variants by a variety of environmental agents--them-
icals, antisera, high temperatures-has often been claimed, and a Lamar&an
belief in the inheritance of acquired characters has persisted longer among
bacteriologists than among any other group of biologists.  We shall see later
that most of these cases may find a better explanation by the assumption of
differential selection of spontaneous mutants.  A number of agents, however,
have been proved truly to affect mutability in bacteria, that is, to increase muta-
tion rates and to cause the appearance of new mutations which had not yet been
found to occur spontaneously.

The most useful. agent of this type is radiation.  Increases in the rate of dis-
sociation ratio of mutant to normal colonies and in the rate of appearance of
other variants have been described repeatedly after exposure to various types
of radiation (71,72).  Most data, however, do not allow a decision as to whether
the effect was due to selective killing or to increased mutability, and, in this case,
whether by immediate or delayed action.  Gowen (67) described a large increase


19471         RECENT ADVANCES IN BACTEIUAL GENETICS         17

in mutation rate in Phytommua &wartG after exposure to x-rays, and found rates
of induced mutation of the same order for individual mutations in bacteria,
viruses, and Drosophila. X-rays have been used to produce biochemical mutants
in bacteri8 (147,69, 163,169), resulting in isolation of a variety of nutritionally
,deficient mutants. The mutations encountered were of the same types as those
found to occur spontaneously, although a gre8t many new types were also ob-
tained. Because of the hit-ormiss mode of their detection, these various types
of mutants are not very suitable for 8 study of the mechanism of action of rsdia-
tion in inducing mutations. Changes whose spontaneous rates c81.1 be deter-
mined fairly accurately, and in which 8ll mutant individuals .can be detected,
are better suited for the purpose. Typical of this are mutations to bacteriophage
resistance. A very important study by Demerec (46) showed that ultraviolet
radiation, as well 8s x-rays, increases the rate of mutation to resistance to phage
Tl in Exherichia wli B, higher doses producing higher increases.  The remark-
able fact was discovered that mutations continue to occur at higher rate for 8
relatively long time after irradiation, and the mutation rate does not return to
normal until several hours later, after the bacteria have possibly undergone 8s
many as 13 generation cycles. Apparently all mutations thus produced belonged
to the same types that also occur spontaneously.  The data on delayed effect
of radiation on mutability were obtained by an ingenious technique which per-
mitted counting the number of mut8tions that occur in the population in 8 given
interval of time directly rather than calculating it indirectly from the number of
mutant cells present.

In further expansion of this work, the dosage effect w&s quantitatively studied
(48). The number of mutations produced was proportional to the dose for
x-rays, but increased more rapidly than the dose for ultraviolet rays.  This held
for both immediate and delayed mutations. The ratio `%mm?di8te/delByed"
incre8sed rapidly with the dose. Mutation frequencies 8s high as 2.8oJc were
obtained with very high doses of radiation.

A delaya effect of radiation seems to be present also in the production of
biochemical mutations (169). The rate of the mutation "succinate positive"
in MuruzeUu Zwoji (36a) could also be increased by x-ray treatment.

Before we discuss these results, we wish to point out that induction of mutation
by radiation is a completely aspecific process.  The accepted theories of the
mode of action of radiation (63, 90) indicate a direct action ,on molecules by
transfer of radiation energy in elementary acts of absorption. Whether the
molecule thus activated will undergo a certain change depends on the properties
of the molecule and on probability considerations, but not on the nature of the
radiation, provided the energy transferred in one act of absorption is greater
than 8 given threshold. Radiation is accordingly supposed to cause mutations
by raising the probability of occurren ce of a multitude of mutations, and not
by affecting specifically this or that mutable determinant.

An interesting c8se is that of a mutation to radiation resistance in EschawIia
coli, which, besides occurring spontaneously, was found to be induced by radia-
tion (179). In this c8se, there is simulation of a specific effect, but radiationis


18                         8. E. LuBL%            [VOL. 11

again likely to act only by incmasing the overall mutability, irmluding mutability
to radiation r&stance, rather than by specifically acting on the latter.

The other group of agents, which, besides radiation, have been found to pro-
duce mutations both in higher organisms (9) and bacteria are the chemical
compounds commonly known 8s nitrogen and sulfur mustards &hloroethyl
amines and sulfides, 66).  These also act in a strictly unspecific way, due to the
high reactivity of certain groups in their molecules which allows them to react
with 8 variety of substances.  A number of biochemical mutants in bacteria hsve
been obtained by treatment with nitrogen mustards (169). Treatment with
0.1% mustard for 30 minutes produced as much as one mutation per 106 treated
cells, comprising 8 variety of biochemically mutated types.  In this case too,
most mutations seem to manifest themselves after a certain delay, indicating a
mechanism baeically similar to that induced by radiation.

Ark (6) reported production of some mutants in plant pathogenic bacteria
by treatment with acenaphthene, a compound which has been found to be a power-
ful inducer of polyploidy in higher organisms.  Ark's results do not prove, how-
ever, that actual induction of mutations is involved, and rather suggest selection
for mutants, possibly spontaneous, in the presence of acenaphthene.  It is in-
teresting to note that attempts to produce bacterial mutations with colchicine
have given negative re8ult.s (176).

Interpret&ion of the data on nonspecific induction of mutations by radiation
and mustards brings us to the question of the mechanism of bacterial mutations,
which will be discussed in the following section.

III. BACl'BBUL MJTATION AND TEB GENETIC DBTEBMINANTS OB BACTERIA

The mutational processes in bacteria, as described above, present so many
+ilarities with gene mutations in higher orgamsms, where the existence of
discrete genea cBn be proved by crossing-over and chromosomal break experi-
ments (124), that a comparison of the mechanisms involved is indeed appealing.
In both types of organisme, mut8tions occur in a mndom, apparently spontaneous
way and at specific, generally low rates independent of physiological conditions.
Once a mutation h8s appeared the new character has often a stability of the same
order 8s that of the alternate character.  Mutation rates spread very much over
the same range of values, although the sixe of bacterial populations permits
recognition of mutations rarer than can be detected in most higher organisms.
The functions affected by mutations in bacteria belong to the same type as those
tieb.ed by gene mutations in higher orgar&D~ (166) : in both c8ses, it seems that
mutations often 8fIect enqyme specificity. These bacterial mutations c8n be
induced by the same agents, radiations and mustards, that produce gene muta-
tions.  Specific induction of mutations by environmentally induced adaptation
and inherit8nce of the acquired character c8n be disproved in almost every ~888.
The most remarkable exceptions-type transformations induced by specific
bacteri81 extracts (10,22), to be d&ussed in a later section-belong to 8 separate
category and, by their very nature, do not encourage a similar interpretation of
the more common types of variation.


19471          RECENT ADVANCES IN BAcl2ZU.L GENETICS         19

Mutations affecting different character occur as a rule independently of one
another in bacteria as well as in higher plants and animals: multiple &e&s of
mutations may in both cases be ascribed to multiple results of a primary change.
Apparently true reversion of mutant characters, a possible indication of the
presence of determinan ts self-reproducing in the mutated form, has been observed
even more frequently in bacteria than in other organisms (150).

Are we justifled then, on the basis of these analogies, in assumkg the existence
in bacteria of discrete mutable dete rminants comparable to genes in higher
organisms, and similarly endowed with the property of homologous reproduction
both in the original and in the mutated form?  That some mechanism for orderly
segregation of character dete rminanta at cell division exists in bacteria is a neces-
sary postulate in order to explain the stability of cell characters.  Although this
can be visualised better in terms of distribution to the daughter cells of discrete
materlal elements concentrated in some structural unit (nucleus?), we must ad-
mit that fairly orderly segregation might be obtained by equational partition of
enzyme molecules present in large numbers; mutations could then appear when,
by fluctuations in the division process, one enzyme happened either to be absent
or present in amounts lower than a given threshold in one of the daughter cells.
The constancy of spontaneous mutation rates, however, is hardly in favor of
this hypothesis. Besides, reversion might be difficult to explain if mutation
were due to the chance loss of a self-reproducing ensyme.  It must be remem-
bered that induction of specific mutations by changes of substrate was considered
as the strongest evidence for the hypothesis of mutation by induced ensyme
change (76).  we have seen above that in all well investigated cases this type of
`mduction has been disproved.

Interpretation of variation as due to segregation in heteroaygotic diploid cells
(166, 2), besides being devoid of experimental basis, would still leave open the
problem of the origin of heterozygxis.

The possibility that VW fn?quent mutations in actinomycetea 8ppWing 8s
sectma in colonies may be due to segregation of charactera upon germination of
heteroaygotic conidia has been suggested by Badian (12) on the basis of cytologi-
cal findings that require confirmation.  Thie type of explanation encountera
the obvious objection of failing to account either for the origin of heteroeygosis,
or for sector formation at stages of colonial growth when no conidia are formed.

Mechanisms of cell fusion, even if proved to be of more general occurrence
than is now known (91, 93)' cannot account for most instances of varktion,
since fusion within pure line clones should not bring about new characters.  As
for other "life-cycle" interpretations of mutation, we have already stated that
there ia no evidence in favor of them since independent and random variation
is the rule; we shall see later that apparent directional series of variation may
find their'explanation in diEerentia1 selection for certain mutant types common to
a large number of b%!tk81 Speciee.

Beaides this negative evidence, do we have any positive one for the existence
of discrete genetic dete rminanta in bacteria? The results of radiation experi-
ments, although still of preliminary type, offer some pertinent evidence. It


20                         8. E. LUIUA            [VOL. 11
has been found (43) that, at least for x-rays, the number of induced mutations to
phage resistance is proportional to the dose. According to the generally accepted
interpretations of radiation effects, this result should indicate a "one-hit" action
in the production of a mutation. The prhnary mutational change must be the
result of 8 single photochemical reaction process, involving a direct action on one
or a few molecules within a limited spatial domain.  Such a process could hardly
produce mutations by mass inactivation of an enzyme scattered over the whole
cell volume; action on a spe&lised center, photochemically reacting as a unit,
eeems indicated. The gene has been considered to be such a center (126, 171)
and the one-hit interpretation of radiogenetic experiments is considered ,one of
the main supports for the hypothesis of the gene being something like anucleo-
protein molecule (see 96). Results of radiation experiments have been consid-
ered before as supporting the hypothesis of 8 basic similarity between bacterial
mutations and gene mutations (67).
It must be said that, although direct proportionality of the number of muta-
tions produced to the dose is indication of a one-hit direct effect, it is not in
itself sufficient proof.  Only proof that the effectiveness of a given dose of radia-
tion is independent of the temperature snd of the intensity of irradiation (dose
per unit time) would be completely satisfactory (90).  It is to be hoped that such
proof will be forthcoming.
The presence of delayed effects of radiation (46) and the non-linear relation
between dose and effect in the case of mutations produced by ultraviolet light
(4.8) indicate some complexities which the simple picture does not account for.
Production of the mutational change by ultraviolet may require accumulation
of a number of primary reactions, if each of these affected one of several equiva-
lent portions of a material determinan t. One act of x-ray absorption, producing
a greater transfer of energy, may affect the whole structure producing the effect
at once whereas each ultraviolet quantum may affect only one of the several
portions. This might also explain the delayed appearance of some of the muta-
tions. The presence of duplicated genes at the time of irradiation has already
been suggested by Muller (126) and others to explain delayed genetic effects of
radiation in Drosophila. Other interpretations of the delayed effect have also
been suggested (46, 48). In spite of these complications, the presence of the
"one-hit" type of action for induction of bacterial mutations by x-rays, if con-
firmed by further studies, would appear to be the strongest evidence for a direct
action on discrete material units, comparable to ,genes, which determine the
hereditary characters of the bacterial cell. The best cytological evidence avail-
able (81,146) can be viewed as supporting this hypothesis by 8fIording proof of
the existence of discrete maws with the microchemical properties of desoxyribo-
nucleoproteins, comparable to nuclei or chromosomes, in many and possibly all
types of bacterial cells. Recent work (91'93) demonstrating fusion with genetic
recombinations in bacteri81 cultures (see Section V) may provide direct genetic
proof for the existence of discrete heredity determin8nta.
Reed, one of the strongest sdvocstm of the presence of genes in bacteria, has
proposed a theory of b8cterial variation (141) based on a mech8nism of unequal


19471        BECENT ADVANCES IN BACTEBIAL GENETICS          21

segregation of genes. By failure of a gene to divide simultaneously with the
others, or by failure of the products of division to migrate to one of the daughter
cells at the proper time, differences in genotype could arise. Transitional
unstable forms would depend on successive unequal divisions in genes present
in multiple copies. This hypothesis can be made to account for every type Of
bacterial variation, and is not clearly in opposition to the result of induction of
mutations by radiation, which might affect the orderly division and segregation
of genes rather than their structure. The following reservations should be made,
however. If all bacterial mutations depended on irregularities of gene segrega-
tion, the independent occurrence of several mutations, some of them at very higb
rates, should be explained by assuming a less precise mechanism for gene segrega-
tion than is present in the chromosomal apparatus of higher organisms.  We
incline, however, to believe that some very precise mechanism for equal segre-
gation of genetic dete rminanta is neceaeary to explain the high degree of heredi-
tary stability of bacteria. Moreover, the high reproducibility of the frequency
of rare mutations, and its independence of physiological conditions (50, 106) do
not seem to favor this interpretation. Unequal division, or loss of some segments
of the hereditary material, may be responsible for the occurrence of complex
mutations producing the .s8me effects 8s two or more other mutations (105).
In trying to assimilate bacterial mutations to genetic changes in higher organ-
isms we should not forget the existince in the latter of a group of phenomena
which have come to the fore of the genetic scene within the past few years.
These phenomena, only partially understood, involve cases of cytoplasmic in-
heritance and give evidence of the exist&xx of cytoplasmic determinants of
heredity, whose occurrence may be more common than has hitherto been recog-
n&d. Besides the semi-independent plastid inheritance, other types of cytoplas-
tic d&mninantu ("plasmagenes") have been described, particularly in
unicellular organisms. These dete rminanta may show various degrees of de-
pendence on nuclesr genes. In certain races of Paramefkm aur&a (153,169)
the presence of a given gene is required to insure continued production of e8ch
cytoplasmic dete rminant but is not sufficient to initiate its production. In
yeasts, a situation has been described (99, 100, 161) in which 8 gene is needed
to initiate production of 8 given ensyme but this production can then continue
in the presence of substrate even after the gene is removed by appropriate
crosses. The self-reproducing unit is supposed to be, not the enzyme itself, but
a nucleoprotein (plasmagene) regulating enzyme production (162).  It must be
said that these-experiments on yeast still require confirmation; Other possible
examples of the role of plasmagenes in heredity have been discussed by Darling-
ton (37).
It is thus likely that there occur various types of self-reproducing, mutable
cytoplasmic determinants of heredity in plant and animal organisms. Their
recognition is particularly important as they may offer 8 key to interpretation
of differentiation in the course of development (158): char8cter differences be-
tween da with the same genotype might arise by differential segregation, irreg-
ularity of reproduction, or mutetion of plasmagenes.


22                        8. B. LuBL4            [ML. 11

It is interesting to speculate on the possibility that bacterial mutations corre-
spond to plsnmnrr#ne changes rather than gene mutations, that is, to changes in
cytoplasmic rather than in nuclear hereditary dete&nants.  A choice, however,
would be impossible to make at the present time, sines we know even less about
phWmagene mutations than about bacterial mutations. It may be of value,
however, to suggest that plasmagene inheritance may prove less stable and more
susceptible to environmental influences than gene inheritance.  Cases of bac-
terial variation apparently caused by the environment, or regularly reversible,
have been supposed (28) to be more similar to `~Dawtm.odaj%aGmd' (78), as
described in protosoa, than to gene mutations. The mechanism of Dam
modijikdioncm is unknown, but it seems likely .that their interpretation may lie
in planmapnnic e&3&1.

It is this reviewer's opinion that an important task of bacterial genetics today
might be a critical reinvestigation with appropriate techniques of those cases of
vazi8tion which appear to involve slow progressive hereditary changes under
the influence of &anging environment.  &en if most of them should again
prove, as we consider likely, to correspond simply to the ordinary type of spon-
taneous discontinuous mutations-selection phenomena complicating the course
of variation4covery of some new type of genetic mechanism might be forth-
coming.
To conclude, we wish to suggest that a distinction between gene and plssma-
gene in bacteria might not be feasible.  Ditrerentiation between nuclear and
cytoplasmic determinants may not have arisen in or&mm3 which, as a rule,
undergo little developmental d8erentiation and do not require a nuclear ap-
pamhis 88 elaborately organized as is needed for carrying out the meiotic process
in sexual orgamsms.  In such case, we might also envision the existence in bac-
teria of a more direct type of gene action than in 0rgaGms with genetic systems
of higher complexity.

While this review was in press, there appeared an important article by
McIlwain (11&Q, suggesting that a number of enzymes may be present in one
or a few copies in each bacterial cell.  The suggestion avas based on a comparison
between the number of molecules of certain vitamins per cell and the turnover
number of several ensymes (number of molecules of substrate used up per second
per molecule of ensyme), assummg that similar valuea obtain for the enrym~
involved in vitamin synthesis. This suggestion leads to the hypothesis that
ensyme production may be directly associated with gene reproduction, and that
in bacteria some enzymes may actually be identifiable with the gene themselves,
the latter having both autocatalytic (hereditary) and heterocatalytic (ermymatic)
activities.

XV. BPZCIPIC INDUCEION OP MUTATIONS

'There is a group of phenomena in the field of bacterial genetics whose unique
&aracter makes them of paramount interest for geneticists and biologists in
general as well as for bacteriolcgists.  These are cases of true induction of heredi-
tary changea by specific treatments which seem to reach into the very core of


1247]          RECENT ADVANCES IN BACl'ERIAL G-CS         23

the genetic make-up of the bacterial cell. The singular importance of these
phenomena has not been recognized as early as desirable, first, because of con-
fusion with many indiscriminate claims to induction of bacterial variation by
practically every kind of environmental change, and second, because only
recently have rapid advances been made towards the elucidation of the phe-
nomena.

A number of cases in which bacteria appear to acquire, after growth in the
presence of products of other strains, some characters of the latter have been
reported (34,130,94).  The characters affected may be virulence, pigmentation,
thermoagglutinabiity, agglutinabiity by specific antisera. Some similar
changes were reported as resulting from growth of two organisms in "parabiosis"
in Asheshov tubes separated by collodion septa (101).  To some of these results
it may be objected that the changes might have resulted from selection of spon-
taneous mutants in the environment containing products of other bacterial types.

The phenomenon of type transformation in pneumococci is not subject to
such doubts. The subject has recently been reviewed (112)' and here we need
Oldy l'fX8~ the mOSt i!dkllt f8CtS.  A non-capsulated R form of Pneumococcus,
derived for example from an 8 culture Type II, can be transformed into cap-
sulated s forms of Type I, II, III, . . , by growth with dead pneumococci of the
respective type in I&O (70) or in do (38) or by growth in presence of cell-free
extracts of each specific type (1).  The presence of serou~ fluids is required for the
transformation to take place. Avery and his collaborators (10, 113, 114, 115)
have brilliantly developed this work to prove the following facts: 1. The specific
component in the inducing extract (TP = transforming principle) is a highly
polymer&d nucleic acid containing desoxyribose and specific for each pneumo-
coccal type.  This was confirmed by 8 number of methods, including inactiva-
tion of TP by purified, crystalline desoxyribonuclease. 2. The specifically
active TP represents only a small fraction of the total desoxyribonucleic acid
extracted from a cell, which is to be expected, since it should only consist of that.
fraction of the nucleic acid which is concerned with the particular character
under study.  Its activity must be enormous, since transformation can be pro-
duced by aa little aa 0.003 microgram of the total desoxyribonucleic acid fraction.
3. Since the transformed character persists in the absence of externally supplied
TP, the TP must be reproduced indefinitely in the transformed cells.  4. Under
optimal conditions, the R+S transformation can affect as many as 0.5 per cent.
of the cells of the R culture.  This high proportion makes it unlikely that the
transformed cells represent spontaneous mutants that only need TP for manifes-.
tation of the mutant character.  It seems practically certain that the change is
actively induced by the action of TP in what probably amounts to a random:
sample of the exposed population of R cells.  5. The role of serous fluids in the
reaction has been partially clarified by recognizing the presence in them. of a
number of fractions involved in various phases of the transformation reaction,
in particular in the sex&it&ion of the R cells to the transforming action of the
nucleic acid (115).

Recently, Boivin and his collaborators (22, 23, 24, 173) have obtained in


24                         8. E. LURLi            [VOL. 11

E8Mh wZi reaulte codming entirely thoee described above for Pneumococ-
CUB. A non-caps&ted R type der@ed from qsulated, ant&e&ally specific
type Cl (or C2) can be transformed into either of the capsulated types by
growth in presence of desorryribonucleic acid extracted from the capsulated cells
of the appropriate type. Work seems to have been facilitated in this c8se
through the circumstance that the transformation occurs in plain media without
serous fluids, and that the nucleic acid appears to be more stable and, therefore,
easier to extract in active form.

The significance of these results is manifold and far reaching.  First of all,
they prove that biological specificity of nucleoproteins can be carried not only
in the protein, but also in the nucleic acid moiety.  It is not known whether this
nucleic acid specificity results from different proportions of certain components
of individual nucleotides, or from different spatial orient&ion of common com-
ponents.  Even more important, the results show the possibility of altering the
heredity of 8 cell by supplying 8n alternative form of desoxyribonucleic acid,
a specific component of chromosomes, and possibly of the gene itself (121).  One
might speculate whether the new form of nucleic acid thus introduced is directly
incorporated into the hereditary material to yield a self-reproducing nucleopro-
tein endowed with the new specificity; or, by its presence in the cell, causes a
change in the synthesis of new nucleic acid M.er used in gene formation; or else,
if it affects the specificity of some other determinant of heredity. It would
certainly be of great interest to attempt production of other types of bacterial
-variation by specific bacterial extracts.

Substances causing type transformation in bacteria h8ve been compared with
viruses (see 180, 164) long before their nucleic acid composition was known.
Both types of agents have in common the ability to induce new synthetic proper-
ties in a sensitive cell.  How far the analogy supports the endogenous theories
of virus origin can hardly be decided at the present time.
It is intere8ting to point out that phenomena of the 8ame type, though not yet

as thoroughly investigated, have been described in viruses.  Fibroma virus can
be transformed into myxoma virus by injection into rabbits of a mixture of
active fibroma and inactivated myxoma virus (19,20,21).  &changes in proper-
ties between different bacteriophages growing in the s8me host-cell (42'75, and
experiments by this reviewer, to be published), although still incompletely under-
stood, may bear a relation to the phenomena of type transformation in bacteria.
These phenomena again point to a more accessible genetic system in bacteria
and viruses than hae been proved to exist in higher org8nisms, since in the former
the genetic dete rminante can be reached and altered by specific components of
the nuclear material supplied from the outside.

V. FUSION AND SEXUALITY bfEcEA.NIt)M8

The occurrence of fusion and sexuslity processes in bacteria has been czsimed
so often (and as often disputed) on the basis of controversial cytologic81 evidence,
that it would hardly be possible tod8y for the worker without person81 cytological
experience either to reach a decision, or even to select reliable examples.  It is,
however, important to point out that most of the older material presented in


10471          BBCENTADVANCESINBA~BIALGENBTICS          25

support of the hypothesis of sexuality in bacteria (102, 118) cannot be used as
genetic evidence because of the lack of information on the exchange or recombina-
tion of discrete hereditary characters in the course of the supposed sexual fusion.

Cases like those described by Almquist (3) of "hybrid" forms with double
serological specificity in mixed cultures of two different organisms can easily be
criticised, among other reasons, because of the possibility of spontaneous varia-
tion or of induction by soluble products.  The whole problem of formation of
"large bodies" is controversial (52,30,136) and their interpretations range from
sexual forms, to symbiotic growth of pleuropneumonia-like organisms with regu-
lar bacteria, to involution forms.  The constancy of their formation at the line of
contact between growth of different cultures of Proteus (51) might offer a suitable
material on which conclusive genetic evidence for or against their origin by fusion
could be obtained by working with genetically marked strains. ,Fusion wit.h
exchange of characters might have been involved in cases of transfer of proper-
ties between bacteria growing in mixed cultures, mentioned in the preceding
section (34, 130, 94) ; but the mechanisms involved were not analyzed.

Strong evidence in favor of recombination of discrete unit characters in mixed
cultures, although still without cytological co&mation, has recently been sup-
plied by experiments with carefully controlled genetic material. Earlier
attempts in this direction (68)' although employing the correct technique of
trying to hybridise mutants from the same strain Wering by one or more visibIe
characters, had given negative results, possibly because of the necessary in-
efficiency of the methods available for the detection of visible colonial variation.

The discovery of biochemical mutations in bacteria with production of specific
growth factor deficiencies permitted Lederberg and %tum to demonstrate by
a brilliant technique the recombination of characters in mixed cultures of different
mutants (91, 93).  These studies, still in the preliminary stage, appear to be
among'the most fundamental advances in the whole history of bacteriological
science.

Mutant strains deficient for two or more growth factors were produced by
irradiation of a strain of Eschsn'chia wli.  Two strains, each carrying a different
pair or group of biochemical deficiencies (double biochemical mutants), were then
grown together in a complete liquid medium.  After growth, large inocula were
plated on minimal medium agar on which neither of the two strains could grow.
Colonies appeared, consisting of cells that had permanently acquired the ability
to grow on the minimal medium like the original strain of Each&&a wZi (proto-
trophic cells).  These cells must therefore have the ability-to synthesise all .four
growth factors, combining the synthetic powers of the two parental strains.
The frequency of prototrophs in mixed cultures was of the order of 1 in 10' bac-
teria.                                                  
Since reversion of one biochemical deficiency was never found to occur at

rates as high as lb6, the chance occurrence of two reversions in the same line
should be much too rare to be detected.  In fact;prototrophih forms do not ap-
pear in pure cultures of each of the double biochemical mutants.  This ilhrstrates
the importance of using double mutcmts for any study of recombination..

The prototrophic forms seem, therefore, to originate from true recombination


26                        8. E. LUIUA             [VOL. 11

between cells of the two strains grown together.  This recombination appears to
involve segregation rather than formation of double cells. Experiments with
triple mutants (sometimes including phage resistance as a marker) showed, in
fact, that exchanges of only one or two out of three characters can occur with
frequency comparable to that of prototroph formation.  This also proves that
prototrophic growth does not represent a symbiosis of the two parent types.
Moreover, it has been found that segregation of characters is not random (per-
sonal communication from J. Lederberg).  This may be an indication of some
type of linkage of determinants in 8 material unit (chromosome?).

These experiments appear to prove the existence in bact4xia of fusion followed
by exchange of genetic determinants, similar to crossing-over, followed by separa-
tion of the fused cells. The possibility that the changes are produced not by
fusion but by induction through the action of ditrusible products has not been
ruled out, but seems rather remote.  F'iltrates of one mutant did not cause the
appearance of prototrophic forms from the other double mutant.

When fusion occuw, it may lead to the formation of heterocaryons, that is,
of cells containing nuclei of two types in a common cytoplasm, as shown to be
produced in a variety of fungi by hyphal fusion (15, 139).  It seems unlikely,
however, that the prototrophs obtained by fusion represent heterocaryons, be-
cause of the apparently independent segregation of characters, with the possible
exception of cases of linkage.  If the genetic determinants are concentrated in
a nucleus, nuclear fusion must be postulated to expIain these resuha.

Temporary fusion, followed by exchange of genetic determinants and separa-
tion of the fused cells, seems to be the correct interpretation; this would then
represent a true form of sexuality in a very simpIe bacterium.  The fused forms
may represent a sporophyte, while the regular type of vegetative cell represents
part of the gametophyte.  It would be interesting to know how long the cells
remain in the fused condition, whether they can divide while fused, or whether
the sporophyte lasts only one cell generation.

It must be pointed out here that, independently of the tremendous importance
of these results, the range of applicability of the conchrsions derived from them
cannot yet be evahrated.  F'usion mechanisms could not be detected, for example,
in another strain of Escherichio wli, either by using biochemical mutants (Leder-
berg, personal communication) or by using phage msistant mutants (unpublished
experiments by this reviewer). In particular, it must be emphasised that there
is as yet no evidence that fusion phenomena of this type may be responsible for
*he ordinary type of bacterial mutations. The phenomena of exchange of here&
by pp-tieg between phage particles growing inside the same host-cell, men-
ltioned in the preceding section, present some analogy with the fusion phenomena
in bacteria described above.

VI. SELECMON PHENOMENA AND ~VOLVTIONARY CON6IDEaATIONB

a. Selsclion pherumma.  Although the evidence discussed in the preceding
section6 indicates that most bacterial mutations occur spontaneously rather than
aa a responss to the environment, the latter plays an important role in determin-


19471          BECENT ADVANCES IN BACX'IZFUAL QENETICS         n

ing the course of bacterial variation.  This role may be twofold.  On the one
hand, bacterial mutations of apparently adaptive character may require the
activity of the environment to render phenotypical a change that in a dilIerent
environment would have remained masked.  This is certainly true in the csse
of mutations permitting the production of adaptive enve8, where the substrate
ie nary to reveal the new potentiality brought about by mutation.  It is
possible that similar mechanisms are present in other casea.  Mutationa to phage
resistance, for example, might conceivably become phenotypical only after the
phage haa actually attached the mutated cell (106).

On the other hand, the environment acta by selectively favoring growth of
certain phenotypes.  We have already seen that bactericidal and bacterio-
static t3ub6tancerr act as powerful selective agents permitting the detection of
rekstant mutants.  The ame is true of deficient media used in the isolation of
mutants capable of dispensing with the missing nutrient.

How normal and mutant types will compete in an environment in which they
both can grow depends on the effects of the mutation on metabolic processes
determining growth characteristics.  A bacterial mutation can bring about
change8 in a number of different characters (pleiotropic mutations).  Changes
in growth characteristica will affect the ability of the mutsnt to grow in competi-
tion with the normal type. Bacterial mutations may actually offer an
uncommonly favorable material for the study of selection phenomena.

Mutations producing apparent increasea in biochemical activities may prove
unfavorable for survival under ordinary circumatsnces, although useful in the
exceptional environment.  Typical is the case of the succinate positive mutant in
MoraxeUa Zwo$i (110).  Although capable of growth with succinate aB eole carbon
source, the mutant is rapidly overgrown by the normal type in media in'which
both of them can grow.  The same is true of the phage reeistant mutant8 of
EstAaichb coli B (105).  While some of them grow at the 883336 rate as the
normal type in broth, a gre8t many are found to grow more slowly, in some ~8~8
the growth rate being half 88 rapid. Unless in the presence of the specific phage,
these mutations appear to bc of no value to the strain, and the mutant8 will be
more or ha rapidly eliminated.  A similar situation seems to obtsin in the c8se
of salt reclistant mutant8 of salmonella (153) and for a number of other types of
bacterial variants (7, 55, 50). Also in the c8se of biochemically deficient
mutants, it is likely that in a complete medium mutant and normal typea may
not show the same growth characteristica (148).  Which of two phenotypes wilI
establish itself in a mixture (~8 the predo minant one is not always predictable from
the study of growth rates of the two typea when growing separately (105, 148)-
Among the mutant8 from EmAerichk ccli strain B, some of the phage resistant.
mutant8 appear indeed to grow in mixturea with the wild type or with one another
aa they would in separate cultures, without appreciable interactions (unpublished
experimenti).  The radiation re&tant mutant, however, was found in careful
studies (179) to behave differently.  Growing alone in nutrient broth, the mutant
haa the same generation time and the same maximum viable titer aa the normal,
and a shorter lag phase; it might accordingly be expected, not only to hold its


28                         8. E. LURLi             [VOL. 11

ground, but to be successful if grown and subcultured in mixture with the normal.
It was found, instead, that the mutant is rapidly overgrown in such mixturea,
so that the proportion of mutants in mixed cultures diminishes rapidly.  Such
interactions indicate competition for substrates or effects of diffusible products
of the metabolism of one strain on the growth of the other, and may give ad-
ditional information on the biochemical effects of the mutations.  Similar phe-
nomena have been observed also in fungi, where competition occurs between
nuclei carrying different alleles of one gene in heterocaryotic mycelia (149).

An interesting point is that differences in growth characteristics have been
found to occur between cells of smooth and rough variants of the same strain
(49,50,26).  These differences are dependent on the medium used, and certainly
play a large role in determining dissociation percentages, as has been indicated
clearly in Braun's work on Brucella (26).  The proportion of S and R cells in
cultures at various stages of growth is the result of competition between the two
types, competition that becomes very keen in the late phases of the life of the
culture, when crowding brings about strong population pressure.  Late growth
and death proceed side by side, as shown by the increase in total cell count with
constant or decreasing viable count.  This situation favors the type which can
grow better and survive longer under such crowded conditions. In Braun's
studies, the R cells were found to fulfill these requirements, which explained their
relative increase in ageing cultures.  In a further study (27) the growth of S and
R cells was investigated in presence of antisera against each phase, showing
enormous selective advantage for the heterologous cells.  An interesting meta-
bolic difference between S and R variants in Proteus VU&&S, involving in-
creased requirement for nicotinic acid, was descriid by Morel (123).

Growth rate differences of various cell types in a colony will give rise to sectors
whose significance has been discussed by Shinn (155) and,.in relation to fungi,
by Pontecorvo and Gemmell (139a).  A mathematical analysis of the relation
between growth rates and shape of the sectors has been given by Waddell (174).

It is the opinion of this reviewer that studies of this type will provide the key
to an explanation of most cases of apparently "directional" phase variation, in
which different cultures appear to undergo similar eerics of orderly changes (73).
We can imagine that in many merent strains homologous mutations occur,
producing similar colony typea and also bringing about changes in growth
characteristics which determine whether they will be favored or eliminated.  The
same mutation may be favorable when occurring before another and unfavorable
if occurring after it ; the apparent series of successive phases as in a developmental
process would thus be explained.  As pointed out in section I, 1, frequent mu-
tations producing growth advantages can be expected to be checked by some
degree of reversion, which is probably the cause of the apparently cyclic course of
most dissociative patterns, with reappearance of the Origiinal type.

Selection phenomena probably explain most cases of supposed induction of
mutations, for example, by antisera or by salts.  It has recently been shown, for
instance, that variants of Chrmbactffium violuce~m appearing in presence of
I.,iCl show different viability in presence of this salt as compared with the parent
strain (79); these differences can explain the apparent dissociative action of the


19471          RECENT ADVANCE8 IN BACTERIAL GENETICS          29

salt as due to its selective effect on various phenotypes.  One should be par-
ticularly cautious before claiing induction of mutation by environmental agents
when the' change appears to affect the whole population exposed.  It is very
likely that in such cases a type arisen by spontaneous mutation has completely
displaced the original type because of favorable selection by the special en-
vironment.

b. Bacterial mutations arid wolution. The large amount of bacterial variability
brought about by mutation provides ample material for natural selection to act
and lead to the establishment in given environments of those biotypes whose
combinations of genetic determinants represent "adaptive peaks" in the field of
the available genotypes (53).

A great number of bacterial mutations involve loss of ability to perform certain
metabolic tasks.  Some of the mutations that appear to bring about new bio-
chemical abilities are accompanied by associated changes which make their
survival and establishment unlikely.  These facts are found to be in agreement
with the hypothesis of a "regressive physiological evolution", developed partic-
ularly. by Knight (84) and Lwoff (107, 108), whose monographs should be
consulted for a detailed account of the basis and implications of the hypothesis.
According to Lwoff (lo@, one can trace through a number of evolutionary
series, in m.icro5rganisms and also in higher plants and animals, a progressive
loss of synthetic and metabolic potentialities. In bacteria, examples are
seen in the transition from coliform to typhoid to dysentery bacilli, where
there seem to occur successive losses of antigens, of fermentative capacity,
and of synthetic powers (increased growth factor requirements). According
to White (177) the antigenic evolution of the Salmonella group has taken place
by successive and independent losses of antigenic components, all present in a
hypothetical common ancestor.  Evolution in bacteria (108) is supposed to have
proceeded from autotrophic organisms, endowed with high synthetic power and
ability to utilise light or inorganic compounds as energy sources, to organisms
requiring some growth factors and deriving energy from the oxidation of organic
carbon compounds.  Further losses narrowed the range of energy sources
utilizable and increased the number of required growth factors.  Anaerobes seem
t.o have originated from aerobes by loss of enzyme systems, among them those
involving cyt&uome and hem&in.  Obligate parasites finally derive from free
living forms if after a number of mutational losses of synthetic power the cells
find only in a living host the necessary materials for their growth.  Extreme
cases of loss of functions could bring to intracellular parasitism, and possibly to
VilllB Origin.

In parallel with these losses of metabolic activities, one observes increased
specialisation, and often increased ability to perform certain specific functions.
According to Lwoff (108) this specialization can by no means be considered as a
true progress, since it is accompanied by reduced adaptability, and, therefore, by
reduced chance of survival. The more special&d the metabolism of abacterium,
the more dependent it will be on particular sets of environmental conditions.
Even slight changes in these conditions may mean extinction for the species.
According to Lwoff (108) the changes underlying physiological evolution are


30                          8. E. LUIUA             [VOL. 11

at kast in part the result of tendencies inherent in the heredity of the bacterial
cell, although the environment would play a partly active role in determining
mutational changes, as abo supposed by Knight (84).  The discu~~iUn of bac-
terial mutability which forms the major part of this review suggests, however,
that Spontaneous mutability is the mechanism that, in bacteria as well as in
higher organisms, brings about a variety of phenotypes on which the environment
exerta its selective role.  We feel, moreover, that the idea of special tendencies
to variation in given directions may be m&leading if interpreted in any other way
than a~ the identification of each genotype with a set of independently mutable
determinants.  Some of the supposed "regressive orthogenetic series" described
by Lwoff (108)-particularly in the case of losses of individual reactions in the
same reaction chain-are probably chance directions of evolutionary change
followed under the pressure of random mutability and natural selection.

One special factor is likely to operate in the apparently progre&ve loss of
BucceBsive 6tep6 in given reaction chains. Lack of ability to synthegize an
intermediary metabolite will result in a requirement for either the metabolite
itself or the end product of the reaction chain in which the metabolite is involved.
If the new habitat contains the end product-which may well be mom readily
available than the intermediary metabolite-mutations producing blocks of
other steps in the same reaction chain can then be accumulaw without adverse
selection, and the process may lead to loss of the whole series of reactions involved
in the ~ymbesi~ of the end product.  Thus, a mutant unable to perform any one
of the steps necessary for synthesis of thiamine may undergo successive losses of
the ability to perform the other reaction steps needed for thiamine ByntheBiB.

Although the evolutionary role of biochemical loss mutations is undeniable, it
is possible that this role is not nearly so unique a~ seems suggested by Lwoff.
Mutations by acquisition of new synthetic abilities have been supposed to have
played a fundamental role in the early stages of life on earth (77), a~ the supply
of organic compounds of prebiologica1 origin (130) began to run low.  Even in
the present highly complex organic environment, a number of positive bio-
chemical mutations have chances to a&& evohitionary trends. It might
actually be expected, as suggested to this reviewer by Dr. R. Y. Star&r, that
every time in the course of evolution a new compound was synthesized and set
free in nature, some microorganism must have been present that possessed, or
developed by mutation, the ability to attack the new compound.  It is likely that
the examples of positive mutants being at a disadvantage when competing with
their parent strains (110) do not havethe general significance attributed to them.
It stands to reason that most mutations occurring in a well established genotype
will be somewhat detrimental in the original environment to which the parent
type is well adapted.  They will, however, have definite survival value if a change
of environment happens to require the newly acquired biochemical property-

It has been eugge~ted (108) that syntheticahy deficient mutants, if properly
supplemented, may draw an energetic advantage from not having to perform the
missing synthesis.  A situation suggesting confirmation of this poasibihty has
been described in Neurospora (149).


19471          BECENT ADVANCES IN BACTEBIAL cIENEI!ICB         31

A word is possibly in order concerning the relative survival value of mutations
and reverse mutations, and the irreversibility of evolution.  In bacteria as in
higher oqanisms (X87), we may expect that if reverse mutation occurs after a
mutant has grown for a certain time, the resulting type may not restore the
original situation, so far as survival capacity or even gross phenotypic efIects are
concerned.  During the time in which the mutant type has grown in a certain
environment, mutations at other genetic loci may have occurred, and been
selected for, that altered the genotype in such a way as to render our mutant better
suited for life, possibly by taking over some of the functions in which the mutant
determinant was handicapped in comparison with its wild-type allele. After
this has occurred, reversion of this determinant to the wild-type allele will not
lead to the sti quo ante, but may actually give a less favored type.  We must
keep in mind that natural selection is always at work on the genotype as a whole
rather than on individual characters, with the result of making a strain better
fitted to life in the environment in which it has grown for any length of time.  Of
course, if the environment is highly special&d, the increase in adaptation may
result in lack of surviving ability in a less specialized milieu.

It is interesting to note that in bacteria (31), as well as in higher organisms
(127), the expectation is verified that very mutable characters often present rates
of reverse mutations higher than the direct mutation rates.
If bacteria are throughout all or most of their life in a haploid condition, natural
selection may be expected to work rather exactly, since all mutations are likely to
find immediate phenotypic expression.  This in turn will tend to reduce adapta-
bility, since immediate selection for or against one mutant character will reduce
the number of genotypic combinations available in a population. This is
possibly counteracted, in the case of bacteria, by the enormous size of the popula-
tions, which increases the variety of mutant types presented to the changing
whim of selective forces.

It is also possible that interactions between mutations (see section I, 3, b)
may provide mechanisms by which a larger variety of genotypes is available in
, $@@a1 populations.  If fusion and recombination mechanisms, discussed in a
>-&t& section, were proved to be of general occurrence, they would certainly play
a tremendoua role in increasing the range of genotypes, and therefore the evo-
lutionary potentialities of bacteria (53).

c. Bachmid geneties wad cl.usc#&&n~ A few remarks may be added concerning
the bearing of genetic research on the problem of bacterial classification.  As
has repeatedly been pointed out (163, 172), most of the schemes of bacterial
classification in current use are determinative keys rather than natural classi-
fication systems. A determinative key is meant for practical use by a certain
group of special&d workem, and as such can emphasise whatever category of
bacterial similarities or difTerences these workers are interested in.  For such
purposes, the best definition of a species remains that by W. Benecke (17): a
species is "what the worker who defines it includes in it according to his scientific
tact".

The geneticist has no direct interest in detc rminative keys, but only in natural


32                          8. E. LUIUA            [VOL. 11

classification. For the present, genetics can contribute little to bringing under
control the hornet's nest of bacterial taxonomy, but may suggest some useful
precautions in approaching it.

It is first of all important to realize that a bacterial species cannot be considered
the strict equivalent of the taxonomic species in organisms with recognised
sexuality, since in bacteria we lack the important criterion of partial or complete
sexual isolation (53). A species, a genus, a tribe, or a family can only be a larger
or smaller section of a clone including certain biotypes recognisable as suf-
ficiently stable, similar to one another, and distinguishable from representative
biotypes in other groups.

In assigning taxonomic positions in a hierarchical order a given differential
criterion should be considered the more fundamental, the larger the number of
differences in independently variable characters it involves. Phenotypic dis-
tinctions resulting from differences in a number of individual unit characters
are extremely unlikely to be erased or to merge into one another by any sudden
genetic change. On the contrary, differences that can be brought about by a
single mutational step, even if phenotypically striking, are of little value for
classification and should not be made the basis for taxonomic difIerentiation.
By keeping this criterion in mind, we may hope to arrive at some kind of natural
classification in which the different clones receiving taxonomic rank actually
represent well established biotypes.  These should correspond as much as
possible to "adaptive peaks" in the almost continuous array of genetic com-
binations on which natural selection is at work (53).

A complex metabolic process, in particular a certain "type of metabolism"
involving elaborate chains of reactions should be a valuable taxonomic criterion
provided the differences between phenotypes cannot be traced to a change in one
single link in the reaction chain. The same may be true of important differences
in cell shape.

It is interesting to notice that in many current systems of bacterial~classification
"species" and even "genera" or "tribes" are often separated on the basis of
character dilTerences that may be brought about by a single mutational step:
for example, the tribes Eseirerichiae and Prohae, the genera Salmonella and Eber-
t.heUa, the species StuphyZuc~m aurem and Stuphykmcma albw in the classi-
fication of Bergey's Manual (18). Genera (for example, Phgdommas) are
separated from closely related groups (Pseudomrmas) on the basis of plant
pathogenicity, a character that may well arise or disappear by mutation (97).
It is obvious that such mutable properties can be used only in practical determina-
tive keys without claim to any taxonomic significance.  Even then, the greatest
caution should be observed, since variable characters may prove too elusive to
permit recognition of organisms of practical importance. In many cases,
description of a variability pattern might prove a much better taxonomic criterion
than description of any one or more of the variable phenotypic traits themselves
(131).  Similarities in the mutability patterns of different strains are likely to
indicate important genetic similarities, because they must depend on the pos-
session of a common set of mutable determinants.


19471          RECENT ADVANCES IN BACI'ERLiL QENEl'ICS          33

REFERENCES

1. ALLOWAY, J. L.  1932 The transformation in vitro of R pneumococci into S forma of
   different specific types by the use of filtered pneumococeua extra&e.  J. Exptl. Med.,
   55,91-99.
-

la. ABB~BAM, E. P., CALLOW, D., AND GILLWEB, K. 1946 Adaptation of StuphyZococcus
aureua to growth in the presence of certain antibiotics.  Nature, 158.818-821.
2. ALMQUIST, E. 1922 Variation and life cycles of pathogenic bacteria. J. Infectious
  Dine-, 81,483493.

. 3. ALMCJUIST, E. 1924 Inveetigatione on bacterial hybrids. J. Infeotiow Dieeases, 35,
    341-340.

,4. ANDEBBON, E. H. 1944 Incidence of metabolic changes among virus-resistant
    mutants of a bacterial strain. Proc. Natl. Acad. Sci., U. S., 30,397-403.
. 5. ANDEBSON, E. R. 1946 Growth requirements of virus-resistant mutanta of Esch-
    erichia coZi strain "B".  Proc. Natl. Acad. Sci., U. S., 32, X&128.
6. ARK, P. A.  1946 Mutation in certain phytopathopienic bacteria induced by acenaph-
   thene. J. Bact., 61,899-701.

7. ATKIN, E. E. 1925 The significance of Berologicel typea of gonococcus. Brit. J.
  Exptl. Path., 6,235-248.

3 AVDVBXAV, A. 1942 Mutationa additives de Moroaslha lwo$. Ann. Inst. Pasteur,
   wrb2s-s37.

9. AUEBBACE, C. 1945 Chemically induced moaaiciam in Droaophik, melanogader.
  Proc. Roy. Sot. Edinburgh, 83.211-222.

10. A~BI, 0. T., MACLBOD, C. M., m MCCARTY, M. 1944 Studies on the chemical
  nature of the substance inducing transformation of pneumococcal types.  Induction
  of transformation by a deeoxyribonucleic acid fraction ieolated from pneumococcue
  Type III. J. Exptl. Med.,79,137-158.
11. B-IAN, J. 1933 .Eiie cytologieche Untersuchung aber daa Chromatin und den
  EntwicklungsrykIue der Bakterien. Arch. Mikrobiol., 4,409-418.
12. B-IAN, J. 1936 Ueber die zytologiwhe Struktur und den Entwicklungsryklue der
  Actinomyceten. Acts Sot. Botan. Poloniae, 13,105-l%
13. BEADU, G. W. 1945 Genetics and metabolism in Neuroepora. Physiol. Rev., 25,
   643-663.

14. BEIU)LE, G. W. 1945 Biochemical genetics. Chem. Rev., 37,15-96.
16. BEAD=, G. W., in, CONRADT, V. L. 1944 Heterocaryoeie in Neurorporcr coma.
  Gewtiw, 29.2!21-3oS.
16. BEADLE, G. W., AND TATUM, E. L.  1941  Genetic control of biochemical retactiona in
  Neuroapora. Proc. NatJ. Acad. Sci., U. S., 27,499-S&.
17. BENECPE, W. 1912 Bau und L&en der Bakterien. Teubner. Leipzig.
18. BEBOEY, D. I-I., et al.  1939 Bergey'e manual of determinative bacteriology.  5th Ed.
  Williama & Wilkins Co., Baltimore.
19. BERRY, G. P.  1937 Rabbit fibroma to myxoma transformation with heat-inactivated
  myxoma elementary bodies.  J. Bact., 34.349-350.
20. BERBY, G. P. 1938 A non-lethal "mutant" &rain of virus myxomatoaum derived
   from fibroma virus.  J. Bact., 3&2&X&3.
21. BEBRY, G. P., AND DEDRICK, H. M.  1936 A method for changing the virus of rabbit
  fibroma (Shape) into that of infectious myxomatoeis (Sanarelli). J. Bact., Sl,
   50-51.

22. BOIVIN, A., DE~~JNAY, A., VENDEELY, R., AND LEHOUY~, Y.  1945 L'acide thymo-
  nucleique polym&is& principe paraiasant susceptible de determiner la @cificitt?
  e&ologique et l'equipment enaymatique dea bact&iea. Signification pour la
  biochimie de l'hbr&litb. Experientis, 1.834-335.
23. BOIVIN, A., DELAUNAT, A., VENDRELY, R.,`MD LEHOVLT, Y.  1945  Sur les modalit&
  dea interactions bactiriennee: effeta antagonietee et induction de transformations
  dana lee propri&& dea germea.  Compt. rend., Z&718-719.


34                         8. E. LURIA            [VOL. 11

24. BOIVIN, A., VENDRELY, R., AND LEEOUIIT, Y. 1945 L'acide thymonucl&que haute-
  ment polym6rie4, principe capable de conditionner la ep&ificiti ekologique et
  l'dquipment enzymatique dee bact&iee.  Coneequencea pour Is biochimie de
  l'h&&lit& Compt. rend., 441, k0-643.
25. BORDET, P.  1941 Transformation rkereible du type "emooth" en type "rough" par
  abaieeement de la temperature de culture; see conedquencee au point de vue dec
  charactka antigeniquee et de la eeneibilit~ B l'akine. Bull. acad. roy. m&d.
  Belg., 6,~288.

26. BBAUN, W. 1946 Dieeociatioa in Brucda ubcrtur: a demon&ration of the file of
    inherent and environmental factors in bacterial variation.  J. Bact., 61,327-349.
?27. BFLAUN, W.  1946 The effect of serum upon dkociation in Btwcdla abortuc a demon-
    &ration of the r8le of selective environmente in bacterial variation.  J. BaOt., 60,
    243449.
- 28. BBIEBLEY, W. B.  1929 Variation in fungi and bacteria.  Proc. Intern. Congr. Plant
    Sci. II. 1629-1&k

29. BBON~~ENBRIONNZZ, J. J. 1928 Virus dieeaeea of bacteria. Bacteriophagy. In:
   Rivers, Filterable viruses. Willisma 6 Wilkins, Baltimore.
. 30. BUNTING, M. I.  1940 A description of some color varianta produced by Serratia
   marwacma, &rain 374.  J. Beet., 40,57*.
__ 31. BUNTINQ, M. I. 1940 The production of stable populationa of color variants of
   Sew&a marceaccna $974 in rapidly growing culturea.  J. Bact., 40,69-31.
32. BTJNTINQ, M. I.  1949 Factors affecting the distribution of color varisnta in ageing
    broth cultures of Serrak marceweru 1274.  J. Bact., 4&693-&l&
33. BUNTINQ, M. I., MD haaarrau, L. J. 1949 The dietribution of color variants in
   ageing broth culturea of Serrati murcaum $974.  J. Bact., 43.5%591.
34. B-T, E. 1925 Actiona d'entrainement entre racea et eeficea microbiennea.
   Compt. rend. 800. biol., I, 142%14!25.
35. B-T, F. M.  1929 "Smooth-rough" variation in bacteria in its relation to bac-
   teriophage.  J. PAth. Bsct., 8& 15-@.
30. COLID, L. J., AND WBIQET, W. II.  1916 Application of the pure-line concept to bac-
    teris. J. Infectious Diceseen, 19, !ZO9-2!21.
30a. CBOLAND, R. 1943 Action dee rayons X aur IA fdquence d'une mutation bact&
   rienne. Compt. rend. 416, 616-018.

37. DAELINOTON, C. D. 1944 Heredity, development, and infection. Nature, 154,
   164-169.

38. DAWBON, M. II., AND SIA, R. H. P.  1931  In vitro traueformation of pneumococcal
   typee.  I. A technique for inducing trsnaformation of pneumococcsl typee in vitro.
   J. Exptl. Med., 54.531-599.
`39. DBEIU, C. J.  1939  On the "~tiv~tion" of the hsctaw of Emherkhia co&mukabik.
   J. Bact., 87.473-4S3.

40. D-BID, C. J., DUTY, A. D., AND hfICEStLBoN, I. D.  1939 The lactaee activity of
   Ewharichia wli-mutab%. `J. Bad., HP $56-363.
- 41. D~LBB~~K, M. 1945 Spontaneous mutations of bacteria. Ann. Missouri Botan.
   GA&II, 39,~233.

41a. DE~B~~K, M. 1945 Discueeion of paper by Banner. Cold Spring Harbor Sym-
  poeis Quant. Biol., 11, 23-24.

42. D~LBR~~CCI, M., AND BAILBY, W. T., JB. 1946 Indwd mutations in bacterial viruses.
  Cold Spring Harbor Sympoeia @ant. Biol., 11,33-37.
43. DEYEBEC, M. 1941 U&able genee in Drosophila. Cold Spring Harbor Sympoeia
  @ant. Biol., 9,145-159.

44. DIWBIBEIZ, M. 1945 Production of rtaphylococcm rtraine reeietant to earioue con-
   centratione of penicillin.  Proc. N&l. Acad. Sci., TJ. S., 3l, 10-24.
_ 45. DE--, M. 1846 Genetic a&ecta of &a~ in Siaphylocecw aurcua producing
   etraine reaietant to various concentrationa of penicillin. Ann. Mieeouri Botan.
   Garden, 39,131-133.


19471         RECENT ADVANCES IN BACTERIAL GENETICS         35

46. DEIIEBEC, M.  1946 Induced mutations and possible mechanisms of the transmission
  of heredity in &chcrichia cow.  Pm?. Nat]. Acad. Sci., U. S., 32,36-48.
47. DQYEBEC, M., AND FANO, U. 1943 Bacteriophage-resistant mutants in Eachsrichia
  wli. Genetics, 30, 119-136.
48. DEMEBEC, M., AND LATARJET, R.  1946 Mutations in bACteriA induced by radiation.
  Cold Spring Harbor Symposia Quant. Biol., 11, 33-50.
49. DESILOWITZ, M. W.  1937  Bacterial variation as studied in certain unstsbie variants.
  J. Bact., 33,349367.

-60. DESICOWITZ, M., AND SCHAPIRO, A.  1935  Numerical relations of an unstable variant
   of SoZmoneZZa o&r&. Proc. Sot. Exptl. Biol. Med., 32.573577.
51. DIENEB, L. 1946 Complex reproductive processes in bacteria. Cold Spring Harbor
   Symposia @ant. Biol., 11,51-59.
52. DIENEI, L., AND SMITH, W. E.  1944 The significance of pleomorphism in Bacteroides
   strains. J. Bact., 46.125-153.

53. DOBZBANBKY, T. 1941 Genet.ics and the origin of species. Columbia Univ. Press,
   New York.

54. DOBZEANSKY, T., AND HOLZ, A. M.  1943 A re-examinat.ion of the problem of mani-
  fold effecta of genes in Drosophila melanogaslsr.  G netica, 26.295303.
55. DOUDOROFF, M. 1936 Dynamics of dissociated % acterial cultures. Proc. Sot.
  Exptl. Biol. Med., 35, 339341.
66. DOUDOROFF, M. 1946 Experiments on the adaptation of Escherichiu coli to sodium
   chloride. J. Gen . Pbysiol., 23,535-611.
57. DUBOB, R. J.   1945  The bacterial cell. Harvard Univ. Press. Cambridge,
   Massachusetts.

5% EASTWOOD, A. 1932 Combining affinities in bacterisl variation and carcinogenesis.
  J. Hyg., 33,391-331.

59. EDWARDS, P. R., AND BRUNER, D. W.  1942 A description of an unusual Salmonella
  type with special reference to the evolution of Salmonella species. J. Bact., 44,
   289-300.

66. EIYIEBBON, 8. 1944 The induction of mutations by antibodies. Proc. Natl. Acad.
  Sci., U. S., 30,179133.

c 61. EK~BBON, S. 1945 Genetics as a tool for studying gene structure. Ann. Missouri
   Botan. Gsrden, 3% 243-249.
62. EPBBUBBI, B. 1942  Armlysie of eye color differentiation in Drosophila. Cold
   Spring Harbor Symposia Quant. Biol., 10,49-43.
63. F-o, U., bw~ DIYEBEC, M. 1944 Genetics: physical aspects. In: Glasser,
   Medical Phydcs, Chicsgo.

64. F~LKIES, P., GLADBTONE, G. P., AND Hwrouv, B. C. J. G. 1933 The nitrogen and
   vittunin requirement8 of B. Lyphouw. Brit. J. Exptl. P&h., 14,189195.
4. GALB, E. F. 1943 Factors influencing the ensymic activities of bacteria. Bact.
   Rev., 7, 139-173.
658. GARROD, A. E.  1923 Inborn errors of metabolism. 2nd Ed. Oxford Univ. Press.
66. GILMAN, A., AND PEILIPE, F. S.  1946 The biological actions snd therspeutic applica-
   tions of the ,!%chloroethyl amines And sulfidee.  Science, 103,499-415.
67. GOWISN, J. W. 1941 Mutation in Drosophila, bacteria and viruses. Cold Spriug
   Harbor Symposis Quant. Biol., 9,187-192.
_ 63. GOWIBN, J. W., m LINCOLN, R. E. 1942 A teat for sexual fusion in bacteria. J.
   Baet., 44.551-554.
. 69. GOAT, C. H., AND TATUM, E. L.  1944 X-ray induced growth factor requirements in
    bacteria.  Proc. Natl. Acad. Sci., U. S., 80,494-410.
70. GBIFRTE, F. 19% The significance of pneumococcal types. J. Hyg., 27. 113-159.
71. HABIBYAN, 6.  1941 Dissociation of etlrphylococci by x-rays: toxin-producing sbili-
   ties of x-ray variants.  J. Beet., 42,143-149.
72. HABEBMAN, S., AND E-WORTH, L. D.  1940 Lethal and dissociative effects of x-rays
    in bacteria.  J. Bact., 49.463593.


86                        6. lu. LURXA            [VOL. 11

73. HAD-Y, P. 1927 Microbic dieaocirrtion. J. Infectious Diseaaee, 4&l-212.
74. HALDANB, J. B. 8. 1939 The biochemistry of the individual. In: Needhsm and
  Green, Perspectives in Biochemistry. Cambridge Univ. Press.
76. HERBHEY, A. D. 1946 8pontaneoue mutations in bscterial virueer. Cold Spring
  Harbor 8ymposia Quant. Biol., ii, 67-77.
76. HINWLWOOD, C. N. 1944 Bsotericrl growth. Biol. Rev. Cambridge Phil. Sot.,
  1@,1!%&163.

77. HOBO~XT~, N. H. 1saS On the evolution of bioch&nical syntheeee. Proc. Nat].
  Acsd. &i., U. 8., 6l. lb%167.

78. JOLLOS, V. 1921 Experimentalle Protirtenatudien. I. Untereuchungen iiber Veria-
  bilitit und Vererbung bei Infu8orien. Arch. Protistank., 4S. l-222.
79. KQLNEB, A.  1947 Second8ry colon& of b8cttwi8 induced by salta of alkali metals,
  with epecid reference to C~romoba&~fium uiofuccum and other b8C@i8 on lithium
  chloride eg8r.  Am. J. Botany, In preae.
60. KLIQNBBEXOER, E. 1942 %me new observations bearing on the nature of the
  pleuropneumonie-like orgsniem known (LB Ll aaeociated with S~rsptiacilZua monili-
  formir. J. Hyg., 42, -96.

81. KLIENEBCBOBB-NOBEL, E, 1945 Chsngea in the nuclear etructure of bacteria,
  particularly during epore fornmtion.  J. Hyg., 44, 99-108.
82. KNAYSI, G. 1923 Morphological and culturel studies of Bacillus msgathcrium with
  special reference to di8mocirtion. J. BIMt., W,62&644.
83. KNATEI, G. 1942 The demon&ration of a nucleus in the cell of a etaphylococcus.
  J. Bact., 46,666-%4.

84. KNIQET, B. C. J. G. 1966 Bacterial nutrition. Material for a comparative physi-
   ology of bacteria. Med. Rsserrrch Council (Brit.) Special Rapt. &riea, 210.
-66. KOEN, H. I., AND HARRIS, J. 8.  1942 Methionine made an essential growth factor
   by cultivation of E. coli in the pretsence of methionine and eulfanilamide.  J. Bed.,
    44,717-718.

86. KOBER, S. A., AND WRIQHT, M. H. 1943 Experimental variation of nicotinemide
  requirement of dysentery bacilli. J. Bad., 46,23%249.
87. KRIBTENBEN, M. 1940 Rechercheu cur Is fermentation mutative de8 bactdriea.
  Acts Path. Microbial. #csnd., 17, 193-231.
88. LANDY, M., AND GERBTUNQ, R. B.  1995 p-Aminobenroic scid synthesis by Neiaasria
  gonorrti in relation to clinical and cultural sulfonamide reeietance. J. Immunol.,
   61,~277.

&I LANDY, M., LABIUY, N. W., OBWALD, E. J., AND &TBQIOETOFF, F.  1943 Increeeed
  eyntheaia of p-uninobensoic scid 88socisted with the development of rulfonemide
  reuietance in i%ph&xoaw aurew.  Science, 97.266407.
90. Lma, D. E. 1946 Action of mdi8tione on living cello. Csmbridge Univ. Preen.
91. LZD~BP~EBO, J., AND TAT~M, E. L. 1946 Gene recombination in lklisrichiu cdi.
  Nature, 166,656.
92. LBDEBBBEQ, J., AND TATUM, E. L.  1046 Detection of biochemical mutanta of micro-
  organisms. J. Biol. Chem., 166, 381-262.
i43. LEDEBBERQ, J., ANY TATUN, E. L. 1946 Novel genotypes in mixed cultures of bio-
  chemical mutanta of, bacteria. Cold Spring Harbor Symposia Quant. Biol., 11,
   113-114.
94. LEQBOUX, R., m G~NWBAT, J.  1936 &ude comparative entre le b8cille de Whit-
  more et le b8cille pyocy8nique.  Ann. inet. P~teur, I, 24Wlk
96. Lmwm, I. M.  1933 Becondary coloniee of bscteria with special reference to Badus
  mycoida. J. B8ct., 26,6@-367.

96. LEWIS, I. M. 1934 B8&kl variation with qwcial reference to behavior of some
   mutable strains of colon b8cteris in synthetic media.  J. Bsct., 2&61!+636.
- 97. LINCOLN, R. E. 1940 Bacterial wilt rwiat8nce snd genetic host-parasite intar8ction.m
    in mrire. J. Agr. Reae8rch, 660,217-26!3.


19471        RECENT ADVANCES IN BACTERIAL GENETICS         37

98. LINDFJQREN, C. C. 1942 Nuclear spparatuz and sexual mechanism in a micrococcus.
  1OW8 8t8te College J. Sci., l&307-318.

99. LINDEQBEN, C. C. 1945 Mendelian and cytoplaemic inheritance in yeasts. Ann.
  Mizzouri Botan. Garden, 1,107-123.

100. LINDEQREN, C. C. 194!I Yeast genetics. Bact. Rev., 9.111-170.
101. Lm~orrrrn, M., N~QRE, L., SEIQNPVBIN, R., AND ROMAN, G. 1033 Interactions
   microbieunea Btudidees par la m&hode de la parabiose. Ann. inst. Pasteur, 61,
   822-m.

-102. L~ENIS, F. 1921 Studies upon the life cycles of bacteria. Memoirs Natl. Acad.
   Sci., U. S., 16, (II Mem.) l-232.
103. Lumn, S. E.  1946 Mutationa of bacterial viruaea affecting their hoat range.  Genet-
   ica, #I, 84-99.

104. LIJBIA, S. E. 1945 Genetics of bacterium-bacteriai virus relationship. Ann. Mis-
   souri Botan. Garden, 32,~242.

105. LUBIA, S. E. 1946 @mtaneoue bacterial mutations to resistance to antibacterial
    agenta.  Cold Spring Hstbor Symposia Qusnt. Biol., 11,136-137.
,-X6. LURIA, S. E., AND DELBBUCK, M.'  1943 Mutations of bacteria from virus sensitivity
    to virus resistance. Genetics, 26,491-611.
107. LWOFF, A. 1932 Recherches biochimiquee aur la nutrition des protozoairee. Mas-
   eon, Paris.

108. LWOFF, A. 1943 L'bvolution phyeiologique. Etude dea pertez de fonctiona chez les
   microorganiames. M-on, Parie.
-109. LWOFF, A.  1946 Some problema connected with epontaneoue biochemical mutations
    in bacteria. Cold Spring Harbor Symposia Quant. Biol. 11, 139-155.
110. LWOFF, A., AND AUDUREAU, A.  1941  Sur une mutation de hforoxelh lwoji apte 8 Be
   d&elopper dana lee mitieux B l'acide succinique. Ann. inst. Pasteur, 67,~111.
111. LWOR, A., AND AUDUBEAV, A. 1944 Recherchea enzyrmrtiquee aur lea mutations
    bactt%enuea.  I. L8 cocsrboxyke de l'acide ox8lockique chez la forme normale et
   le mutant "succinste" de MorozeZlo lwofi.  Ann. inat. Pasteur, 70,51-54.
112. MCCARTY, M.  1946 Chemical nature and biological specificity of the substance in-
   ducing transformation of pneumococcsl types. Bact. Rev., 10.63-71.
113. McCAE~~, M., AND AVERY, 0. T.  1946 Studies on the chemical nature of the sub-
   etance inducing transformation of pneumococcal types.  II. Effect of desoxyribo-
   nuclerree on the biological activity of the transforming subetsnce.  J. Exptl. Med.,
    03,&06.
114. MCCABTY, AND A-BY, 0. T.  1946 Studies on the chemical nature of the subetance
    inducing transformation of pncumococcal types.  III. An improved method for the
   i~o~stion of the transforming aub&ance and ita application to pneumococcua Typea
    II, III, and VI.  J. Exptl. Med., 63.97-104.
115. McCaam, M., TATME, H. E., AND AVEBY, 0. T. 1946 Biochemical studies of environ-
   mental factors ervwntiel in tranefonnation of pneumococcal types. Cold Spring
   Harbor Symposia f&ant. Biol., 11, 177-183.
116. MCILWAIN, H. 1943 Nutritional rtudi~ of bacterial variation. II. The derivation
   of drug-resistant straina in the absence of any inhibitor.  Brit. J. Exptl. Path., 44,
    212217.

-1168. MCILWAIN, H. 1946 The magnitude of microbial ra8ctionn involving vitamin-like
   compound& Nature, 166,693-902.

117. MABSINI, R. 1907 Veer einen in biologiacher Beziehung intereasanten Holistemm
   (Baderium coli-mutubile).  Arch. Hyg., 61, %6-292.
118. MELLON, R. R. 1925 Studies in microbic heredity. I. Gbeervationz on 8 primitive
   form of sexuality (zyogoepore formation) in the colon-typhoid group.  J. Bact., 10,
    431-601.

11% MELLON, R. R. 1926 Studies in microbic heredity.  II. The eexuel cycle of B. coli
   in relation to the origin of variants with special reference to Neiazer and Maaaini'e
   B. coli-mutab&. J. Bsct., 10,57@488.


38                       8. E:. LURIA            [VOL. 11

129. M~Lu)N, R. R. 1942 The polyphaaic potencies of.the bacterial cell; general biologic
   and chemotherapeutic significance. J. Bact., 44,1-25.
121. MIR~KY, A. E. 1943 Chromoeomes and nucleoproteins. Advances in Enaymol., S,
    l-34.
122. MONOD, J. Quoted by Lwoff (199).
123. MOREL, M.  1945 Chara~tire~ phyaiologiqwr ditrerentiels de deux variantes 8 et R
   de Prokw vulgar-k Ann. inat. Patrteur, 71,225232.
124. MOBQAN, T. H. 1923 The theory of the gene. 2nd Ed. Yale Univ. Press. New
   Haven.
125. MULLEB, H. J. 1928 The problem of genie modification.  Verh. V. int. Kongr. Ver-
   erbungew., 234-m.

1%. MULWB, H. J. 1939 Bad&ion and genetice. Am. Naturalist, 84.2~251.
127. MULLEB, H. J. 1939 Reversibility in evolution considered from the standpoint of
    genetics. Biol. Rev. Cambridge Phil. Sot., 14,251~239.
123. NYBDBO, C. 1933 Mutationmuncheinungen bei einem Aerobakterstamm. Zentr.
    Bakt. Paraaitenk. Infek.  I, Orig., 148,178199.
129. OAPLB~BQ, E. F., AND LVEIA, S. E. 1947 Mutations to sulfonamide resistance in
   Staphylococcw atbr6u6. Genetics, In press.
139. OPEN, A. I.  1933 The origin of life. Macmillan, New York.
131. OBLA-JENBEN, 5. 1942 The lactic-acid bacteria. 2nd Rd. Munksgaard, Copen-
    hagen.
-132. PAXX, L. W.  1933 A new "mutation" in the coliform group of bacteria.  J. Heredity,
    29,~334.

, 133. Pm, L. W., AND ROBBINS, M. L. 1942 The concept of stability and some of ite
   implicatione. J. Bact., 43,661-W.

.c 134. PABB, L. W., AND SIYPBON, W. P. 1949 Coliform "mutantB," with respect to the
   utiliration of citrate. J. Bact., 49,407-.
135. PAWING, L.  1949 A theory of the structure and process of formation of antibodies.
    J. Am. Chem. Sot., 64,25432957.
1313. PESHIEOPF, M. A. 1949 Paranecrotic alterations caused by light in Protsue uulgutis
   awarmem. Mikrobiologia (U.S.S.R.), l5, 331-393.
137. PIE~IU, G. 1937 Cytologieche Untereuchungen an Paratyphm- und Colibak-
    t&en. Arch. Mikrobiol., 8, M.
138. POCHON, J. 1941 Metabolisme et evolution dee bactdries celiulolytiques ana&obies
    librea et parasites. Ann. inst. Pasteur., g& 57-77. Quoted by Lwoff (108).
139. PONTECORVO, G. 1949 Genetic eyetemn baaed on heterocaryosis. Cold Spring Har-
   bor Symposia @ant. Biol., ll. 198-201.

139s. Po~~~~onvo, G., AND GXMMELL, A. R.  1944 Colonies of Penicillium nohtum and
  other moulde an modele for the study of population genetice. Nature, 184,5X&534.
140. RAOUX,, Y., AND COBDIBB, P.  1948 Comportement biologique de quelquea varidtda de
  Protutu Xl, notamment en ce qui concerne la synth& des co-enrymee. Compt.
   rend. sot. biol., l37,37&376. Quoted by Lwoff (193).
141. REED, G. B. 1933 A hypothetical view of bacterial variation. J. Bact., 88,530-588.
142. R~IED, G. B.  1937 Independent variation of several characterietice in-s. mur1~~8ns.
   J. Bact., 80, 255-263.

143. R~IYINN, H. A. 1937 Baoterial type transformation. IV. M&r- t#trag6nu8
    infection. J. Bact., 88,499-511.
14. REIMANN, H. A. 1937 The eignigcance of bacterial variation.  V. Micrococcw t&a-
    genus infection.  J. Bact., 33,613-523.
145. RHOADEB, M. M. 1941 The genetic control of mutability in maize. Cold Spring
    Harbor Symposia Quant. Biol., 9.139144.
146. ROBINOW, C. F. 1945 Nuclear apparatus and cell etructure of rod-shaped bacteria.
     In Dubos: The bacterial cell.  Harvard Univ. Press.  Cambridge, Mace.
/147. ROEP~, R. R., LIBBY, R. L., AND SIiALL, M. H.  1944 Mutation or variation of
    E6charichia coli with respect to growth requirements.  J. Bact., 48.491-412.


lQ47]          BEcmNT ADVANCE8 IN B QENETIC8         39

`148. RTW, F. J. 1946 Bask-mutation and adaptation of nutritional mutants. Cold
   Spring Harbor Sympo6ia Quant. Biol., ll,216-227.
149. RTILN, F. J., AND L~D~IUIIBQ, J.  1948 Reveme-mutation and adaptation in leucine-
   le6eNeuroepora. Proo. Natl. Acd. Sci., U.S., 82,183-173.
`150. RYE, F. J., &ENIEID~, L. K., AND BALLENTINE, R.  1948 Mutation6 involving the
   requirement of uracil in Cloetridium. Proc. Natl. Acad. Sci., U.S., 82,281-271.
15Oa. SOJTT-MONC~F, R. 1937 The biochemhstry of flower colour variation. In Need-
   ham and Green: Perapectivee in biochemi6try. Cambridge Univ. Press.
181. SWAQ, M. G. 1948 Enzyme problem6 in relation to chemotherapy, "adaptations",
   mutations, re&tance, and immunity.  Advance6 in Enzymol., 0.33-127.
182. SNVAQ, M. G., WD GBEEN, M. N.  1944  The me&a&m of reaietance to eulfonamide6.
   II. Abeence of correlation between re&tance~ and the formation of a&mine by
   Staphylococcw aur6uI. Noninterference with the utili6ation of glucose ae a critical
   factor in the development of reeietance to sulfonamidea.  J. Bact., 48,823-830.
163. 8XWmUtNt3, J. M., m TANNER, F. W.  1946 The inheritance of environmentally in-
    duced character6 in bacteria. J. Bact., 49,383-393.
164. SEUIRO, A., 8PIIEGBLWAN, S., by) KOB~B, H.  1937 A rapid method for the study of
   genetic6 in large populations. J. Genetics, 84,237-245.
,165. SEINN, L. E. 1939 Factor6 governing the development of variational atructuree
    within bacterial coloniee.  J. Bact., 88, S-12.
188. SEI~ZI, W., AND ROBIBTE, E. C.  1948 Biochemical trsneformatione a6 determined by
   competitive analogue-metabolite growth inhibitions. II. Some trsneformation6
   involving p-aminobenxoic acid.  J. Biol. Chem., 161,4&l-471.
167. SILVEBMAN, M., AND W~BKMAN, C. H.  1939 Adaptation of the propionic-acid baa-
   teria to vitamin BI eyntheeis including a method of a66ay.  J. Bact., 88.25-32.
188. SONNIBOBN, T. M. 1943 Gene and cytoplaem. I. The determination and inheri-
   tance of the killer character in variety 4 of Paramecium aurelia.  II. The bearing of
   the determination and inheritance of characters in Paramecium aurelia on the prob-
   leme of cytoplaemio inheritance, pneumococcw transformation, mutation6 and de-
   velopment. Proc. Nstl. Acad. Sci., U.S., 29, 329-343.
159. SONNEBOBN, T. M.  1945 The dependence of the phyeiological action of a gene on a
   primer and the relation of primer to gene.  Am. Naturalist, 79,318-339.
180. SONNEBOBN, T. M.  1946 Diecueeion of paper by Luria, in Cold Spring Harbor Sym-
   poaia Qusnt. Biol., 11,138.
181. SPIEOE~, S.  1945 The phyeiology and genetia significance of enzymatic sdapta-
    tion.  AM. Mkwouri Botan. Garden, So, 139-183.
182. SPIEOELMAN, S., AND &MJCN, M. D.  1948 Gene6 and nucleoproteins in the eyntheeie
   of enzymea. Sciencet, 104,681-584.
183. STANIEB, R. Y., AND VAN NIEL, C. B.  1941  The main outlinea of bacterial clamifica-
   tion. J. Beet., 42, 437-488.
164. STANLEY, W. M.  1938 Biochemietry of viruses.  In Doerr and Hallauer: Handbuch
   der Vinwforechnng, Vol. 1. Springer, Wien.
- 185. 8TEWABT, F. H. 1927 Segregation and sutogamy in bacteria. A contribution to
   cellular biology. Adlard L Son. London.
108. STIEWABT, F. H.  1948 A review of some recent work on papillary variation in bacteria
   and bacterial cytology. J. Hyg., 43, 138-141.
167. STUB~VANT, A. H.  1944  Can specific mutation6 be induced by serological method67
   Proc. Natl. Acad. Sci., U.S., 89,178-178.

163. TATVH, E. L. 1945 X-ray induced mutant strain6 of Ewhtwkhia coZi. Proc. Natl.
    Acad. Sci., U. S., 31.215-219.

-189. TATUU, E. L. 1946 Induced biochemical mutation6 in bacteria. Cold Spring Liar-
   bor Sympo6ia Quant . Biol., 11 ,278-284.
170. TATVM, E. L., AND BEAD=, G. W.  1948 Biochemical genetior of Nsuro6pora.  Ann.
   Missouri Botan. Garden., 82,125129.

171. ~YO~~R-~OVBKT, N. W., ZI-, K. G., mu D~LBB~~cK:, M. 1938 Uebr die


40                        8. 1. LURIA            [VOL 11

    Natur der Genmutmtion nnd der Genetruktur. Nschr. Gee. Wiee. Gottingen, Fach-
   gruppe, VI, N. F. 1, 139-217.
172. VAN NIBL, C. B. 1944 The culture, general phyeiology, morphology, and claeeifica-
   tion of the non-sulfur purple and brown baateria.  Bact. Rev., 8.1-118.
173. VANDBELY, R., AND Lnaonrrr, Y.  1946 Lee acidee ribo- et d6eoxyribonucl6iquee de la
   cellule bacterienne et leur signification. Compt. rend., 222,X%7-1369.
174. WADDELL, A. H. 1945 Curvea formed by colonies of microorganieme growing on a
   plane surface. Edinburgh Math. Notee 35, 14-19.
175. W~L, R. 1946 Modification profonde (mutation?) d'un Staphylocoque eoue l'in-
   fluence du bact&iophsge.  Ann. inst. Paeteur, 72,277-m.
176. WALL, A. W., AND YOUMANB, G. P.  1940 Growth of bacteria in media aontaining
    colchicine. Proc. 800. Exptl. Biol. Med., 44.271-273.
,177. WEIT~~, P. B. 1925 Further studies of the Salmonella group.  Med. Reeearah Coun-
   cil (Brit.) Special Rept. Seriee, 103.
178. WITKIN, E. M.  1946 Inherited differences in eeneitivity to radiation in Ettchtwiohh~
    coli. Proc. Natl. Acad. Soi., U. S., 82.59-53.
179. WITKIN, E. M.  1947 Genetice of reeistawe to radiation in &heri&a coli. Gene-
   tic+ In press.

130. WOLLMAN, E., AND WOLLMAN, E. 1925 Sur la tranemieeion "parah&r&-iitaire" de
    charact&ee ches lee bactirieu. Compt. rend. eoc. biol., 98, X53-1569.
181. WOLIXAN, ELIE Quoted by Lwoff (109).
132. WOOD, H. G., ANDEBBON, H. A., AND WEBIUUN, C. H.  1933  Nutrition of the pro-
   pionic acid bacteria. J. Bact., 86,201-214.
133. WBIOIXT, 8.  1941 The physiology of the gene. Phyeiol. Rsv., 41.437-527.
334. Z~NHOP, 5. 1948 Studiee on bacterial mutability: the time of appearanae of the
    mutant in ~6chetichia c&. J. Baot., 6l, 351-361.