410 MONTHLY WEATHEE REVIEW. SEFTEBIBEB, 1898
to do much damage throughout this entire region, but that
the greatest destruction occurred about 4 miles east of Clear-
mont, over a region of about 4 square miles, as indicated on
the accompanying map. Mr. Brink states that a t one point
in'this region the fall of bail was so heavy that a drift un-
protected by any artificial means remained lying on the
ground for four weeks after the storm, and this too at a sea-
son of the year when the soil had a comparatively high tem-
perature. Mr. Brink, leaving Marysville, visited the region
where the storm occurred, just four weeks after September
5, and aecured several excellent photographs of the scene.
Even a t that time he found the people of the neighborhood
gathering the hail for the purpose of making ice cream.
Pieces of ice had been picked up a t the time of the storm of
a cylindrical form of about four inches long by about two
and a half in diameter. We reproduce on Chart SVI, Fig. a,
a photograph showing how complete had been the destruction
of growing corn. The stalk seen in the photograph was quite
mature, and is the only one left standing in a field of 10 acres.
Had itnot been for the hailstorm, this field would have yielded
60 bushels to the acre when harvested in the middle of Septem-
ber. The half tone, Chart XVI, Fig. b, shows the complete
destruction of the foliage in a grove of young trees. Of
course, the falling cakes of ice destroyed the roofs and even
the sides of the houses, eometimes the north and sometimes
the north aud west sides a t Clearmont ; when its force was
not sufficient to make a fresh hole through the boarding, the
hail would generally, a t least knock out the knot holes. The
northern edge of the storm seems to have passed about ten
miles north of Maryville; but 2 or 3 miles farther north a
narrow strip of country also suffered severely; see Chart
SVI. The statement that in some places the earth was cov-
ered to the depth of a foot, while everywhere i t was white
with ice, seems to justify the conclusion that there must have
been an average depth of from 4 to 6 iriches over the central
region of, a t least, 4 square miles in area.
THE TELEGRAPH SERVICE WITH THE WEST INDIES.
BY J. H. RQBIXSOX, Chief of Division.
Daily telegraphic weather reports from the West Indies,
Central and South America, and the Mexican Gulf Coast are
now received a t the Central Office of the Weather Bureau, i. e.,
from Nassau, Habana, Santiago de Cuba, Kingston, San to
Domingo, San Juan, St. Thomas, St. Kitts, Dominica, Marti-
nique, Barbados, Trinidad, Curacoa, Colon, Coatzacoalcos
Vera Cruz, Tampico, and from Merida when the conditions
are threatening. Daily reports are also received from Ber-
muda. For details a9 to routes uf submarine cables over
which reports are transmitted, see Chart, SIII, showing po;nts
a t which the cables connect with the land lines.
During the entire duration of the recent war with Spain
daily weather reports were regularly received a t the Central
Office of the Weather Bureau from Habana. The observa-
tions were taken by the observers of the Spanish Meteoro-
logical Service for the Antilles, and were telegraphed by their
operators to Key West, where they were retelegraphed to
W-ashington. The submarine cable from Key West to Habana
remained intact during the entire war, ae did also the cable
from Santiago de Cuba to Kingston, Jamaica. As the Key
West-Habana cable had its outlet through the United States,
it was not deemed wise to disturb i t ; the Santiago de Cuba-
Jamaica cable, however, was grappled for, but not caught.
I n connection with this article, perhaps i t would be of in-
terest to know that the Weather Bureau observers on the sea-
coaet in many instances rendered efficient service to their
country by reporting passing vessels, one of the most impor-
tant being the arrival of the Oregon off Jupiter, Fla.
WEATHER TPPEEI AT HAVREI, MONT.
By C. W. LIXQ, Observer, Weather Bmeaa (dated October 10,1808).
I forward herewith a few generalizations relative to the
marked pressure conditions that produce certain weather in
this vicinity as deduced' from a study of the daily weather
maps for the past two years. The cause of the warm sonth-
west chinook winds that prevail here is easily explained to
visitors as due to mechanical or dynamic heating, and the
old Japan current theory, still held by many, is easily obliter-
ated. The conditions preceding cold waves, or colder weather,
are very well marked and invariably produce the expected
cold weather, with high northwest winds.
Chinook conditiona.-A high over Wyoming and southern
Idaho, with a low pressure over northern Montana, invariably
brings to Havre a dynamic rise in temperature, Le., a chinook
wind in winter and a warmer spell of weather in summer, and
accompanied by high southwest winds.
Cold ,wave conditions.-A low barometer over Wyoming,
Utah, and southern Idaho, with high over northern Mon-
tana, Alberta, and Assiniboia, indicates intensely cold weather
for Havre in winter and much cooler in summer. A high
pressure area, with the low in Wyoming, by its vortical action,
draws down intensely cold air from above (see journal of
December 1 and 2,1597, and of March 25,26,27,1598). These
conditions produae cold, high, northwest winds.
Tl~c sunmer t?ype of weather conditions in Oregon, Le., a high
over \Vashington and Oregon with a low over the northern part
of Montana and southern Alberta, indicates a warm spell for
Havre and vicinity (see data July 13 to 17, 1598).
d jtrlling barometer after a dry and warm spell of weather,
barometer falling below the normal, with falling temperature,
with the forecast for warm, indicates a heavy June or July
rain in a day or two ; also a cool atmospheric wave, followed
by a warm wave (see thermograph sheet July 18 to 19, 1898,
and the precipitation on those dates).
d low prcusrcre area of 29.7, or lower, extending west to the
coast and far south, indicates for Havre two or three days of
rain or snow. P o --
ANEROLD BAROMETERS.
By C. F. MARVIN. Profmsor of Meteorology (dated November 8, i8W).
The unreliability of aneroid barometers when anything
like accurate nieasures of pressure are required is almost
proverbial. I t i s evident that much work remains to be done
in order to ascertain the laws governing the irregular action
of this most convenient and, in some cases, indispensable
type of barometer. It will be still better if we can ascertain
and eliminate the real causes of the anomalous behavior.
The face reading of an aneroid even when not compensated
for temperature is often thoughtlesely accepted with every
confidence as to its infallibility, but it is pretty generally
understood among the more critical observers that aneroids
require to be frequently checked and verified by comparison
with standards, aud that a slow change goes on within them
after they have been subjected to a considerable change in
air pressure.
It may be stated that, broadly classified, the inherent
errors of all aneroids are of two kinds. When the instru-
ment is exposed over long periods to only such pressure
changes as occur a t any station a t the earth's surface from
day to day, its error will remain sensibly constant for a considerable period of time, but from time to time rela-
tively large changes may and generally do take place in this
error without any apparent cause and can not, therefore, be
duly allowed for. The other inherent error is an effect con-
nected with any considerable change in pressure. Suppose,
for example, the pressure is changed in a short space of time
from 30 to 2.5 inches; a t the instant the pressure becomea
atationary the index of the aneroid will show a certain read-
L~EFTEMBEB, 1898. MONTffLY WEATHER REVIEW. 411
ing, but instead of remaining stationary with the pressure, it
continues moving slowly downward over the scale showing a
lower and lower reading as time progresses. This after effect
is called “creeping;” it is aI;alogous to the slow change in the
zero of a thermometer and varies greatly under different
circumstances and has been observed to increase during
several days or even weeks. Its amount in a given barometer
depends among other things upon how much asd how rapidly
the pressure has changed and whether the change was con-
tinuous or not. Finally, i f the pressure is restored to its
initial value, the index will fail to return immediately to the
original reading, but will slowly creep toward it. These errors
are separate and distinct from errors of graduation, or the
effects of temperature and friction, or those depending upon
what position is given the aneroid, all of which latter can be
more or less perfectly disposed of. From the foregoing it
necessarily follows that the readings carried up and down
mountains or sent up with kites and balloons are subject to
a series of errors which it is practically impossible to deter-
mine with any degree of exactness.
A most valuable series of experiments upon the aneroid
has been published by Dr. C. Chree, Superintendent of
Kew Observatory, whose paper on this subject, entitled Ex-
periments on Aneroid Barometers a t Kew Observatory and
their Discussion, has recently appeared in Philosophical
Transactions, Series A, vol. 195, 1895. The author’s object,
as he states, is “t o acquire knowledge likely to increase the
usefulness of the aneroid under the conditions in which it is
actually employed .” In the execution of this object the tests of hundreds of
aneroids, made during many years for the makers, a t the Kew
Observatory, were exhaustively discussed and other special
tests were made to develop the law of the after effect or crety.
Dr. Chree’s analysis is most comprehensive and complete,
and leaves little to be desired, but in the present writer’s esti-
mation, the fair conclusion to be drawn from the investiga-
tion is that the errors of all ordinary aneroid barometers are
hopelessly involved in a complex law, so that under the com-
plex conditions of use i t is impossible, or a t least imprac-
ticable to satisfactorily predict ur determine the errors of a
given instrument. Among other writers on this question
may be mentioned Lovering, Proceedings American Academy
of Sciences, 1849, Vol. I1 ; also American Journal of Science,
2d Series, Vol. IS, p. 249: Balfour Stewart, B. A., Report,
1867; also Proceedings Royal Society, Vol. S V I , 1869; Philo-
sophical Magazine, Vol. SSSVII, 1869 : Whymper, How to
Use the Aneroid, John Murray, 1891, and Carl Barus, Ameri-
c a ~~ Journal of Science, Vol. I, 4th series, 1896.
Lovering, Stewart, and Whymper may be said to have ob-
served and recorded the facts relative to the errors of ane-
roids. Dr. Chree has gone a step further and formulated the
approximate law of error. Unfortunately this law appears
to be too complex and iiivolved to become practically useful.
What still remains to be done is to definitely ascertain the
cause of the difficulty, and if this can not be quite largely
eliminated by an improved construction we must then invent
some new and better type of instrument. A step in this
direction has been taken by Prof. Carl Barus, whose counter-
twisted curl aneroid, he claims, is practically free froin the
after effect and similar errors common to other forms of
aneroid. The experimental forms investigated by Professor
Barus are very delicate pieces of laboratory apparatus, and
are scarcely available for practical work. The “ counter-
twisting” is, however, a new feature, and may yet prove to
be of great value in practical instruments.
The present writer has given some thought to this problem
from time to time, but no opportunity has been offered for
making any new investigations. It seemed to me that the
real seat of the greater part, if not all of the after effect or
creeping, is in the corrugated aneroid vacuum boxes them-
selves, as distinguished from the tempered steel springe that
are employed to keep the box from collapsing under the pres-
sure of the air. This conviction ‘was forced upon my mind
after reading Mr. Whymper’s valuable paper on the errors of
the aneroid, and in 1892 I made the following simple experi-
ment, which greatly confirmed this supposition: The vacuum
box of an old aneroid was removed, and a heavy weight (a
trifle over 50 pounds was required) was applied directly to
the steel spring, thereby straining it as nearly as possible to
the same extent as did the air pressure exerted through the
medium of the corrugated vacuum box. Any desired changes
in the position of the index were made by appropriate
changes in the weight. No after effect comparable in magni-
tude with that exhibited by ordinary aneroids was ever ob-
served. In other words this tempered steel spring behaved
to all intents and purposes as i f it were a perfectly elastic
body. Readings of the pressure scale could be made corre-
sponding to about 0.005 of an inch on the barometer. A care-
ful or full investigation was not attempted. I believe never-
theless that the tempered steel springs employed in all ane-
roids are or may easily be made to be highly trustworthy. On
the other hand the process of constructing the vacuum boxes
is well calculated to develop therein irregular and imperfect
elastic properties in the highest degree. The top and bottom
surfaces are each formed of a thin circular sheet of metal,
with a narrow rim bent up around theedge. In order to give
flexibility several concentric corrugations are formed over
quite the whole face of the disk. The crimping and bending
operations necessary in the manufacture of these corrugated
disks have a marked eflect upon the elastic qualities of the
metal, which to make matters worse is generally of brass,
german silver, or some similar alloy well known to be only
imperfectly elastic under the most favorable conditions. The
metal must originally be, more or less, in a soft pnd annealed
conditlon in order to withstand the corrugating and bending
operations. Those portions which are stretched and com-
pressed by the process become stiffer and more elastic, and
the finished disk is permeated by a most complex and irregu-
lar system of internal strains. The arrangement of mole-
cules is undoubtedly a highly unstable one, and i t is not eur-
prising that large, discontinuous, and unexpected changes
take place in the readings of the finished instrument.
In 1896 I had occasion to design a meteorograph suitable
for use on the kites recentlyemployed by the Weather Bureau
in making aerial observations. The lightest possible form of
barometer of the aneroid type was an obvious necessity.
The considerations outlined above led me to eubstitute steel
in place of brass for the vacuum boxes, and a preliminary
efl’ort was made to profit if possible by the reeults brought
out by Professor Barus. Mechanical difficulties in securing
a satisfactory form of recording aneroid on these lines forced
me however to abandon this attempt a t the time and adopt a
type of instrument similar to the well known Richard baro-
graph. The vacuum boxes were made of steel instead of
composition metal. The performance of these barometers,
although better generally than brass instruments, has dis-
appointed my expectations, as they have exhibited consider-
able after effect, which I am still convinced is due primarily
and largely to the imperfect elastic properties of the boxes
themselves. Boxes of small diameter were necessarily adopted,
but these are inferior to those of lrtrge diameter. In support
of this statement I find a very significant note in Dr. Chree’s
paper, p. 446, stating that the after effect in 30 barometers of
special make and large‘ size (how large is not stated) as tested
a t Kew, was so exceptionally small that he found i t necessary
to multiply the actually observed effects by seven in order to
make the figures similar in size to corresponding reeults from
ordinary aneroids.
412 MONTHLY WEATHER REVIEW. SEPTEMBER, 1898
I n concluding these remarks attention iscalled to a certain
source of error in testing aneroids that does not appear to
have been mentioned hitherto, and if not noticed or guarded
against may have an appreciable effect on results. Any sud-
den change in the air pressure under the receiver of a pump
inevitably heats or cools the gas dynamically, in consequence
of which there is a most pronounced and real " after effect "
in the pressure of the air within the receiver. I n niy own
experience I have been very greatly surprised a t the slow-
~ ~
ness with which the gas acquires its stationary tempera-
ture and the magnitude of this effect on the resulting station-
ary pressure. The slow rate of pressure change adopted
by Dr. Chree, namely, one inch in five minutes, in all
probability eliminates any error of this kind, but the point
is not mentioned, and it is just possible that the results of
the older observations and of investigations made without
due regard for this effect niay be somewhat in error in conse-
quence.
NOTES BY THE EDITOR.
THE OMAHA CONVENTION OF WEATHER BUREAU
OFFICIALS.
On several previous occasiohs conventions of Section
Directors of State Weather Services have been held, to the
great advantage of the individuals and the Service, and i t
was, undoubtedly, a wise innovation when the Chief of the
Weather Bureau decided to expand this idea and call for a
general convention of Weather Bureau officials of every
grade. The convention was of a thoroughly cosmopolitan
character, every section of the country was repreeented, and
every class of men. There was a large sprinkling of volun-
tary observers, an encouraging number of the younger
employees, and several of the oldest and most venerable.
Three men were present from the class of 1571, but the
classes that were most prominently in evidence were those of
1851-83. The official report will show that the long pro-
gramme was attacked and faithfully followed up, although
the work had to be done too rapidly for comfort, owing to
the loss of a day. The photograph of the group of seventy
members remains as a visible embodiment of the fraternal
intercourse, the social pleasures, and the intellectual profit of a
meeting that will always remain vividly impressed upon the
memories of all who were present as one of the most
delightful events of oEcial life. If it were not for the
expense we are sure that every one would attend such a con-
vention every year. Many inquiries were made for those
who could not be present; both we and they lost niuch by
their absence. The enthusiasm of all who took part in the
discussions was remarkable ; every one had some positive
results of his own local experience to communicate for the
benefit of the others. The diversity of ideas impressed one
with the conviction that everywhere the work of the U'eatlier
Bureau is being adapted to special local conditions and that
a hard and fast rule for the whole country would, oftentimes,
work inconvmience or injury. One learned not to be so
intolerant of the views of others and so positive that his own
ideas will suffice for all occasions. The new devices sub-
mitted hy Townsend of Philadelphia and Sims of Albany a t
their own expense and the new principle in meteorology
brought forward by Hammon of San Francisco excited deep
interest.
By its rather early adjournment the convention, unfortu-
nately, missed the telegram inviting u s to a special excursion
to Lincoln, Nebr., where we should have inspected the rela-
tions of the Service to the &ate University. May we b
more fortunate next time I I n a few cases some general
expression of opinion was uttered by the convention but, as
a whole, the sentiment that pervaded i t seemed to be to the
effect that no business. properly so-called, need be transacted,
as we were brought together a t the call of the Chief to confer
with him. Consequently, no vote was taken as to the time
and place of the next nieeting. that being a matter that can
be left with Professor Moore ; nevertheless, a hearty accla-
mation followed the pleasant rivalry between Hainmon and
Pague in advocacy of San Francisco, Cal., and Portland,
Oreg., respectively. On the whole, the general conclusion
must be that such conventions are essential to the welfare
and strength of our meteorological service. Scattered as we
are, widely over the whole country, we get but little oppor-
tunity for personal intercourse, we pursue our studies alone
and with difficulty, little items of daily practice and of me-
teorological theory that would be quickly settled by con-
ference with some neighboring observer, give us unnecessary
trouble. The annual convention is a clearing-house, where
we may balance accounts, discuss ideas, settle perplexities,
dissipate the troubles of official life, burn our bridges, and
take a new start.
THE WEATHER AND THE SUGAR, CROP.
In the MONTHLY WEATHER REVIEW for August, 1897, page
354, we have given the general relation between annual ram-
falls and sugar crops in the Island of hlauritius for the years
18SO to 1895, as quoted from the annual report of the Royal
Alfred Observatory for the year 1895, by Rlr. F. F. Claxton,
who is now the director succeeding Dr. Meldrum who resigned
September 30,1S96, on account of failing health, after a term
of twenty-two years in the service. Since that date tlie re-
ports for 1896 and 1597 have been received, from which we
extract the following table showing the relation between the
annual sugar crop of the whole ieland and the rainfall. The
sugar crop is the result of the gr0wt.h of the previous fifteen or
eighteen months, beginning with the planting in September
of the second year previous. The following table gives the
total rainfall for those months during which the cane of the
respective crops has been growing. It is an average for four
stations, viz, Pamplemoussss, Gros Bois, Cluny, and Union
Bel-Air, which fairly represent the sugar districts:
Years of harvest.
1830 .......................................... 1AR1.. ......................................... ISH% ......................................... 1R83 .......................................... 18%. .........................................
1896 ........................................... 1%;. .......................................... 18%. .......................................... 1 %w9. ........................................ lm!! ........................................... lKYl .......................................... 188'2'. ........................................ 1893. ....................................... 1894 ........................................... 18!>5. ......................................... 18Y6 ...........................................
'RR?. ........................................
Total sugar crop.
Ziilogi-aim. 119.:31,402 117, tXl9.610
116,719,997 lX3.396.853 IS? 784,339 ll5,299,030 100,876 &I
124,073: 140 lS2?,172,988 124,564,951 130,220,%3 118,813,075 63,; 18,573 139,751,810 113,'i93.319 142. W5.7.22
151,677,973
Rainfall durlng growth.
Inches. F.39
18.88
118.37
84.09 75. I 77.13 57.25 86.18 125.40 108.71 88. [u
96.61 98.78 80.39 88.11 9n. 11
1oR. 58
*Destructive hurricane.
For the crop of 1897 the corresponding rainfall was tlie
lowest on record, and in fact, scarcely one-half of the nor-
mal amount, and the sugar crop was osceedingly poor; but
the exact figures are not a t hand to be inserted i n the above
table. If we rearrange the above figures in the order of the