10 MONTHLY WEATHER REVIEW JANUARY, 1925
days, I have not been able to find any very striking differ- ences between the weather of the two halves of the month. Tnere was greater cyclonic activity during the second
half than in the first, the weather was more changeable, and there was a pronounced southward and eastward
HAWAIIAN
By ALFRED
The recent publication by the Weather Bureau of statistics of rainfall and other climatic data for Hawaii'
puts in convenient form for study the individual monthly
amounts of rainfall as recorded for each month of the year at 59 stations in Kaui, 120 in Hawaii, 72 in Maui,
and 77 in Oahu, 2 in Lanai, and 5 in Molokai, a total of
335 records, but not necessarily for that number of indi- vidual stations, there being in a number of cases several records of observation for one and the same place. An at-
tempt has not been made to consolidate the several s s i w just mentioned into a single consecutive series.
The object of this pa er is to draw at,tention to the
touch uDon some of the best-known characteristics of
statistical data thus mac 7 e available and to very briefly
.~ ...~~
Hawaiiak rainfall. The outstanding feature of the rainfall distrihubion in
Hawaii is the ve? wide variation from one place to
also by the wide variation in the amounts for the mnic
months in different years. As much as 31.95 inches of rain have been registered
as fallin on a single day and 102.46 inches in a single
day of the month. The annual amounts range from a
maximum of 562.00 to a minimum of 2.46 inches on the
another separated s rom each other by only a few miles,
month o 9 31 days, or an average of 3.30 inches on each
island of Maui.
Nean monthly and cinnual rainfall, tq islands.-'l'able 1
below gives the monthlv and annual means for each of
the fo0U"r large islands of"Hawaii, Haui, Oahu, and Maui.
The monthly means were computed from the records of
10 ears or more in length on each of the several islands an were taken from the regular monthly ublication
Climatological Data, Hawaiian Section, 1923, y Thomas
A. Blair. The small number of stations on both Lanai and Molo-
kai and their geographic distribution make it inadvis- able to attempt to compile a mean for those islands.
TABLE 1 .-Average mon!hly preci pilation, Ha ILW iinn yro up, hg
idm ds
R B
-__
Hawaii ....... ._..._._ 9.158.65 10.58
Kam ___________._.___ 5. &I 4. ti0 7.66
Maui ________
~
___._._ 5.81 7.51 8.65
Oahu _.._________.__. 6.015.9i 6.31
I Summary of the Climatological Unta for the United States, by Sections, W:tshing- ton, N22. In Figure 3 of this publication, the block with the title "Oahu (Mountainsi Leeward" needs a word of explanaton in view of the well-known decrerrce of rainfall on tho leeward side of the mount.ains. The author of the publicatim in an iinpublished letter indicates that the hlodi as
above indicated refers exclusively to that part of the KO-o-lau Range extending from its crest a few miles to windward. It is recalled that the average altitude of these mount:rins is not more than 2,000 feet, not high enough to block tho pnssage of the northeast trades. The rain that falls immediately in lee of the crest should be considered in the nature of an oversplash from the rain that falls on the windward SidQ.-EDITOE.
flow of polar air on the 26th-28th. On the whole, the differences were not important. The temperature and precipitation were both rather close to the normal, so
close that we feel justified in characterizing it as an
average winter month.
RAINFALL
J. HENRY
The monthly means of the above table are to be con- sidered as provisional rather than definitive. I n comput-
ing the monthly means regard has been had primarily to
length and reliability of the record rather than to
geographic position on the islands.
I t is perhaps well known that the majority of rainfall stations on the islands are a t low levels and on the wind-
ward side, whence it follows that the island mean will be increased when a majority of the stations thereon have
a windwitrd exposure. It is believed that the means of Table 1 for both Hawaii and Maui are too large by reason
of the disparity between the number of rainfall stations on the resiective windward and leeward exposures. In
of which is on the windward coast, the other being in the
interior a t iin elevation of 800 feet. The annual average
a t the first named is 67.03 and a t the last 31.82 inches. Lanai has but a single reporting station, situated nearly
in the middle of the island a t an elevation of 1,780 feet.
The :~nnual avernge is 35.70 inches.
the case o 'f Molokai but two records are available, one
3:LV
4.00
JlonthZy dintribution.-The group month1 means show
three rather distinct mnsima, one each in d r c h , August, and December and three minima, one each in February,
June, and Ocbober. Figure 1 presents these facts graphically. The chief maximum occurs in December and
the chief minimum in October. These are most likely
the more stable of t,he phenomena of monthly distribu-
tion. The August maximurn is largely dependent upon the intensity and position of the North Pacific area of
high pressure, which usually is centered about north
latitude 40' and west longitude 145'; the spring maxi-
mum, which on Hawaii is deferred until April, is probably in some way related to the increased wind movement of
that season coupled with the frequency of occurrence of
southerly storms.
The chief minimum, which on Maui is deferred until October, comes in early summer at a time when c clonic activity to the northward of the islands is at a low e g b and the pressure graclieut for northeast winds is at the lowest
point for the year. This minimum is most pronounced
on Kaui and Oahu and on the leeward slo 3s of Maui,
t Molokai and in a much less degree in the %au district of Hawaii.
JANUARY, 1925 MOKTHLY WEATHER REVI'EW 11
The difference in latitude between the southern tip of
Hawaii and the northern coast of Kaui is 3 degrees, a comparatively small distance when temperate latitudes are considered but probably too great to be disregarded
in the Tropics. The February minimum is pronounced on the most northerly islands of the group, Kaui and Oaha, and is also in evidence on windward Hawaii,
windward Oahu, windward Maui, and leeward Oahu.
Dr. B. M. Varney tells me that a weak secondary
February minimum is noticeable in central and southern
California rainfall. I have also found a decided minimum for a mountain station in Porto Itico, although the record
is for but nine years' observations. The geographic
situation of that island is somewhat analogous to the
Hawaiian group, viz, in about the same latitude and southwest of a great oceanic area of high pressure, thc Azores HIGH.
The autumn minimum in Hawaii conies a t a season when the atmospheric processes that cause rain are
inactive, and in this respect the conditions in the Hawaiian group are not materially different from those which
obtain over continents.
TnR daiZy ruinfoZZ.-I have computed from the daily rainfall records of Hawaii, 44 stations in all, the average
amount of precipitation for each clay of January. 1923,
January /923
1 6 16 3f
FIG. 2.-Perccntage frequently of rainy days anti average daily amounts of rain at 41
stations, island of Hinwnii, during a wet month, January, 1923
a very wet month in the whole group; also the percentage
of frequency. The latter was determined simply by counting the number of stations on which a measure-
able amount of rain was recorded for each clay of the month. These percentages have been plotted to form the irregular line at the to of Figure 2 and the series of
amount of preci itation for each dny of the month.
amount of rain that was recorded a t the five stations in
the leeward districts of Hawaii-north Kona and south
vertical bars immediate P y below represrnts the averago
The lower set o P vertical bars represents the average
Kona, respectively. The rainfall of these two districts on the leeward side of Hawaii is essentiallv diffmmt from that of similar
sides on the remaining islands, the chief clifference being that the maximum aniounts for the ycar come in the months of May, July, and August, as will be shown
later. (See fig: 4). -
-
The above diagram shows that considering the island
as a whole a mcisureablc amount of rain fgll on every day of the month; in the leeward districts, however,
ths was not the case. The diagram further shows that
there were a t least three periods of three to five days
each with some rain a t 90 per cent or more of the stations
and that the average amount on these cla s was several times greater than on the other days. G he lower set of bars shows that in leeward districts there were re- latively greater rains on the clays of high frequency and
amounts in all Hawaii. The causes o f rainfall in Hawaii.-In general the
primary cause of rainfall in the Hawaiian group is the pressure distribution over t.he North Pacific that con-
trols the direct,ion and force of t,he winds. The northeast trades, which some, one has said blow " persist.ently and
mo:iotonoi!sIy" over nearly all parts of the grovp, are the immediate cause of the rainfall. These winds, as is well known. pass over a very considcrable extent of water
stirface and for the most part must be rather highly charged with water vapor. Meeting the high cliffs and
other obst,ruct,ions that abound on the windward shores
of the several isla.nds, they are forced upward, their moist.ure is condensed by t,he cold of elevat.ion ana falls
as rain. Thus the rainfall must be classed as almost wholly orographic. Average pressure charts a d the rainfa.11 of Hawaii.- When considcring average pressure charts a mental res-
ervation about as follows shov.ld always be made: While
avcrnge pressure charts are good to look at, they very rarely represent the real conditions which prevail on any
single day. month, or year; hence when we speak of the
North Pacific HIGH as a distinct entity which has form and movement we, t.0 a cert.ain extent, create B false
impression. What we really wish to express is the
thought that' the pressure over the North Pacific, as else- where over the oceans, is constantly changing and never remnina a t the same level and in precisely the same forma-
tion for more than a day or two.
Heavy rains in Hawaii, part,icularly in the months November to April, me associated with changes both of
geogra hic position and intensity, of this North Pacific
osition of t'his pressure formation in February, accor8ng to Bartholomew's At.las, is close t o the intersection of the 140th meridian of west longitude
with the 35t,h parallel of north latitude. This HIGH has a
sensonal shift as follows: I n the cold season it approaches
and a t times merges with the continental area of high
pressiire over the Canadian Nort,hwest. As t,he warm
season advances the center of the HIGH is found farther
and farbher west and north; thus in August and Septem- ber, which mont,hs mark its greatest development, its center is npprosimately in west longitude 150' and north
latitude 40'; it then begins to recede t,oward the North
American Contiimit to again begin its annual cycle of movement. The level of the barometer in this HIGH is
lowest, 30.10 inches, in Oct,ober; it increases to 30.20 in
November nnd cont8iiiues a t that level until the end of
March, the,n rises to 30.30 and maintailis that level, April to September, both inclusive.
As indicated in the abow parawaph, the North
Pacific HIGH is not so yronounced in tTie winter as in the summer; he.nce. it follows that the trade winds of winter
are less stable than in summer when the HIGH is best tleve.loped ; moreover, the frequency and iiitelisity of
cyclonic systems that pass eastward along the llleutians in winter is a second factor which tencls t o increase and
a t t,inies reverse the, direction. of the prevailing winds of the Hawaiiaii Islands. Owing t o the low latitude of the islands, ver few, if any, fully developcd cyclonic systems
rarely occurs. Nevertheless, in the months from Novem-
ber to April, the procession of cyclonic storms or areas of
area o P high pressure or, for short, the North Pacific HIGH.
The average
pass direct f y over them, and cyc.lonic rainfall, as such,
12 MdNTHLY WEATHER REVIEW JANUARY, 1925
low pressure is moving across the North Pacific from west
to east in the neighborhood of the 52d parallel of north latitude or along the Aleutians. Each of these LOWS is attended b a fall in pressure, which on the average
its center to, say, one-quarter of that amount in the
region occupied by the North Pacific HTQH; thus pressure in the high suffers a reduction whenever a LOW nioves
aast along the northern path. As the pressure is reduced
the gradient for northeast winds over Hawaii is also cli- minished, and when the ressure reduction is great
even SW. winds, which characterize the so-called “Kona” or southerly storms a that are a feature of the cooler sea-
son in the islands. Thus it results thnt those slopes which were leeward slo ies during the prevaleiice of the
windward slopes for northeast. winds, now receive
generous rains. storms in a
several fold as compared with a dry month. During
the vears 1905 to 1933. 38 months out of the 228. or 17
shades off P rom four-tenths t80 five-tenths of an inch at
enough that gradient is rep Y aced by one for SE. to S. or
trades now become win d ward slopes, aiid, like the former
single month will increase the rainfall o 9 that month
The occurrence of one or more souther1
and increasing the rainfall. The fact that the wind has been southerly for a few days increases the likelihood of heavy falls of rain by reason of the presence over the
islands of a mass of warm southerly air of high moisture content.
High pressure at Honolulu is associated with li h t rainfall and conversely low pressure with heavy r a i d l l ,
the explanation being that with hi h pressure there are
versely low pressure is indicative of frequent interrup- tions. A normal or weak trade is indicative of moderate rains on windward slopes and little or no rain on leeward
slo es.
gates on Hawaiian railgall basad on the Northern Hemi- sphere Weather Mup.-This map for the last year or so
has been prepared from radio reports of meteorological observations made on board slii s navigating the Pacific.
Together with reports from %onolulu and Midway Island it serves to illuminate the character of the pres-
sure formations that control the weather of the islands,
and especinlly the relation of the pressure distribution
to heavv rains on the islands. I invite attention to the
few, if any, interruptions to the If E. trades, and con-
two gro’u s of days January 12-15 nnd 19-21, 1923, as per ;cent, had a rainfall of 10 inches or more. illustratei in Figure 2 . The pressure distribution tribution of these months throughout the year wns as.d associated with these two groups was in the beginning The dis-
.J a 13. 1923 J a n 1.1. 1573 Llil:l. I:. IC.?? J a n 16. c___ ____ Jan. IO. 1923 Jan. I I . 1923 Jan. 12. 1723 -- -____
FIG. 3.--Barograph trues at Honolulu. January, 1923
follows: Winter, 18; spring, 11 ; summer, 3; and autumn,
6. June and October had no rainfall of ns much as 10 inches. The months of heavy rains were irregiilarly rlis- tributed throughout the poriod, the first half having
but 12 as against 26 in the last half. In t,lie latter t.lier.e, were three consecutive years with rainfall uhove the
avera e, 1921-1923. The islands have thwefore just
the data does not indicate fall. Casual inspection 07
that any definite cycle in the rainfall is operative. Other studies, however, as indicated in the liest paragraph,
point to a different conclusion.
Mr. Joel B. Cox in his paper on ii Periodic fluctuations of rainfall in Hawaii” postulates the existence of a
3.7-year period ancl submits the evidence therefor. He also finds evidence that cycles of 11.1 years ancl 3 3 years are present in Hawaiian rainfall. Rising pressure is also an important factor in the rnin-
fall regme of Hawaii. St, soon as’ pressure begins to
rise after a fall the winds become northerly and north-
easterly, thus temporarily intensifying the NE. trades
asse ii through a time of al.ent8er than the, avernge rain-
-
8 Cf. Dalngerfield, L. A., Kona Storms. Mo. WEATHKB REV., 49: 32i-329.
8 Transactlow of Amerlcan Soc. of C. E. 87:W (19%).
Tllcir, Thomas A.. Trade Wlnds of Hawaii. hfo. WEATUER REV.. 51: 525. 326.
one that gave a gradient for fresh northeast winds. This grndient weakened on the 13th and was considerably
intensifiecl on the 14th and 15th with maximum wind velocities a t Honolulu of 56 and 55 m. p. h., respectively.
This change in gradient was induced by changes in pressure of which the only example available is that for
Honolulu. The barograph traces for that station for the periods in question are reproduced in Figure 3 above.
Exnmining these close1 it will be seen that the diurnal
evidence throughout the period, but that it is consider-
ably disturbed on several days. The 10 a. m. and 10 p. m. diurnal maximum is clearly shown in the traces for the
10th and the l l t h , but on the 12th, while the 10 a. m.
mnsimum is plainly shown, the p. m. minimum for that
date is perceptibly lower than for the revious day; the immediately following mLximum is a E o a shade lower than it was on the 12th. The fall thus initiated con-
tinued slowly and reached its lowest level on the p. m.
of the 14th. Compare thnt minimum with the one of
the 12th 48 hours earlier. This change is unlike those which occur in higher latitude on the approach of a
barometric disturbance, being much slower in com- pleting the change from high or norinal to low and back
pressure variation so c 7l aracteristic of the Tropics is in
JANUARY, 1925 MONTHLY WEATHER REVIEW 13
to its initial level; the amplitude is also much less. The
pressure did not a ain reach the level of 30 inches until
for several days i t again began to fa 1 on the 19th, and, as in the preceding case, the fall is first apparent in the
10 a. m. maximum of the 19th. The lowest level was
reached on the 20th, but the barometer remained low
throughout the remainder of the month, with va,riable winds and moderate rainfall. (See fig. 2, as to rainfall).
The c cle of changes described above is most likely due
to a fa1 9 in pressure that advances from west to east in synchronism wit,h the west-east movement of estra-
tropical cyclones over the path which gasses along the Aleutians or close to the parallel of 52 north latitude.
I t is probable that there is a slight southward sensonal shift in the latitude of North Pacific cyclone pablis in
the cold season and that on occasions these st>orms pass
nearer to the Hawaiian group of islands than at others.
of t,he North Pacific.
HIGH, whether west-east or north-south, is also an inipor- tant factor in the weather control of bhe islands. If the
longer axis takes a north-south or a northwest-southeast direction, southerly winds over the Hawaiian ioup and
the resulting weather conditions will follow. %ie speed of movement of traveling areas of high pressure thnt
advance from the h i a t i c mainland across the North
Pacific is also important; rapidly moving HIGIIS will bring more frequent changes in wind direction and conse- quently in the weather, than slow moving ones.
be incomplete without special reference to thqt of eeward
Hawaii, or the north nnd south Kona districts, respec-
tively.
These districts are in the shadow of the great mountain masses of Mauna Loa and Mauna Kea, whose summits
rise far above the upper limit of the northeast trades.
These latter therefore are not in pny way an importa.nt factor in the rainfall of the Konrt districts.
I have ‘ust come across an account of the summer
by the late Ma’ Clarence Edward Dut,ton, Ordnance Corps, United h a t e s Army, who made a geological reconnaissance in the islands in the enrly eighties.‘ Major Dutton snys in connection with his observations
on the weather of the islands:
I waa much impressed with the fact. that the trade wind is not felt on any of the high mountains above 7,000 t o S,OOO feet. The upper part of this island [Mauna Loa] is in a region of comparative
calm, excepting t h e uppermost 2,000 feet., where a gent.le wind blows in a direction opposite t o the t.rade wind.
His account of the w-eabher of the Konn dist,rict: was
evidently based on conversations with the inhabitnnts.
It is summarized in the next paragraph.
In the morning the sky is clear and the sun shines gloriously; after sunrise the nir is dead calm, but about
10 o’clock the sea breeze set,s in, blowing from t,he west,
and ascends the mountain slopes. Quickly the clouds
gather, and at length rain falls steadily throughout t,he afternoon and into the night. At 9 or 10 o’clock the wind
ceases and the land breeze sets in. As a rule there is no rain at the lowest levels upon the margin of the sea.
There is a narrow belt of land close to t,he ocean ranging in width from a mile to a mile and a half or even 2 miles
where rain seldom falls. Here the slope is comparat,ively
K
10 p. m. of the 15t % ; after remainin close to that level
magnitude of pressure changes at Honolulu on the closeness of approach of the estra-
of the longer axis of the North Pacific
P
Any sketch of the rainfall of the Hawaiian grou
showers o 1 leeward Hawaii written some 40 years ago
___
6 C. E. Dutton, Hawallan Volcanoes. Annual report. Director of U. S. Geological Survey, 113, pp. 81-219.
gentle and as the sea breeze blows inland the effect of the ascendin current is not felt until the stronger slopes a
In the main this account faithfully describes the
weather of Konn districts. A more recent statement is that supplied by Mr. Lawrence H. Daingefield, who was
in charge of the Hawaiian service from 1918 to 1922 and visited the Kona re ion in July 1922. In an unpublished
letter to the editor hr. Daingerfield says:
From personal observations when in Kona and especially when
ascending Mount Hualalai, S,269 feet, in north Kona, I can sax that I saw that thereoccurs a distinct strengthening of the south-
westerly wind in the day time which brings in the clouds (forced cumuli I called them) and frequent showers. At the same time the
trades brought in clouds from the northeast, and I observed the “battle of the clouds” on Hualalai Summit, a wonderful although
not unusual occurrence at t h a t point.
couple o B miles inland from the sea are encountered.
FIG. 4.--Monthly rainfall distribution in Kana districts of Hawaii
It has been definite1 established that there is - tically no wind in the zona districts except the lan grac and sea breezes, a condition that, faditates vertical convection
in the dayt<ime. The 0ccurrenc.e of t,hunderstorms in this region is also evidence of instability in the lower air
levels. On the whole i t is quite probable that vertical convection plnys no unimportant part in the rainfall of
the leeward districb of the Island of Hawaii. In Figure 4 I present curves to illust,rate the monthly
rninfall distribution in south Kona. The top curve IS
that for Nn oopoo, a station on the beach, as indicated by
increase in the rains of May with a diminished fall in
June. The latter, however, is a month of minimum rain-
it,s geograp E ‘cal coordinates. Here there is a distinct
14 MONTHLY WEATHER HE:\:TE:\\‘ JAXUAKY, 1925
fall in most of Hawaii. The next three curves are pre- sented to show the marked influence of altitude, beein-
ning with the Kalahiki station a t 750 feet and followezby
h a l l y the curve a t the bottom is a cornposite one made from all the available rninfall records of both north and
south Kona. The bottom curve shows that the months
of greatest rainfall coincide with the growing seaSon of sugar cane in the Eona districts.
’ altitudes of 1,500 and 1,800 feet, respectively; and Mauna Loa are, as noted above, not subject to trade
winds, and thus less moisture is lost by evaporation than
in districts having a northeast exposure.
These districts bein two other records in the immediate neighborhood at in the shadow of the mountain masses of Mauna Kea an %
SOME OUTSTANDING AEROLOGICAL PROBLEMS
By WILLIS RAY GREGQ
[Presented bolorr the American Meteorological Society as a part or the “hleisinger Memorinl.” January 3, 1925, Wnshington]
It is a very ood thing to pause occasiontzlly in our
determine what new lines should be taken up and what
changes, if an , in the old should be made. It is, I
considering the work of Doctor hleisinger, for he was, to a greater degree than nny one else I hnve known,
absorbed with the fascination of ferreting out the secrets
of the atmosphere. Less than a week before he left
for his work nt Scott Field we talked for an hour of the problems that lie before us, some of which he planned
to attack upon his return, and in all of which he wns
ready and eager to give his assistance and advice. And so this brief talk constitutes, in a sense, n recapitulation
of what we discussed at thnt tinie. I shall be glnd in- deed if it leads some aspirina student, by n contribution
along one or more of the fines suggested, to earn an award under the Meisinger Aerologic,al Resenrch Fund, concerning the purposes and present status of which
you are shortly to hear.
The roblenw I shall name (I can do little more than
order of relative importance, except that the last one is, to my mind, the most important.
1. The dillmu1 variation qf meteorological elements at
diferent hei@ts.-This is, erhaps, a subject of greater
may be pointed out that the diurnal variation in wind speed a t moderate liei hts
phase characteristic of the surface, viz, maximum veloc- ity in the afternoon, minimum in the early morning, is
reversed above 100 meters; that, between about 300 nnd
1,000 meters the diurnal amplitude is large, amounting
to 3 or 4 m. p. s., on the averwe; and thnt, above about
1,500 meters there is essential?y no variation. Studies of the temperature variation agree in showipg
a diminishing amplitude from the surface to about 1 to 2
kilometers, where it is very small, less then 1’ C., but these studies give contradictory results a t greater heights.
Some investiuators claim a reversal of phase in these up-
er levels. &ore recent and more abundant data a t brexel and other stations in this country, however, indi-
cate that the phase is essentially like thnt a t tlie surface and that the ninplitude is about lo C. These later dnta
represent practically d l varieties of weather conditions, and the same results are shown for all parts of the year.
The investigation is, however, in a preliminary stage only.
It should be ushed to completion, and extended to the other meteor0 ogical elements, such as humidity, ressure, density, wind, etc. The data we now have, as t ie result
of numerous series of successive kite flights covering
periods of 24 to 30 hours, are ample for a complete and
authoritative analysis of this problem.
2. Winds and weather along airways.-This is a problem
of immediate practical value. I n order intelligent1 to
determine regular flight schedules for commerciar or
work, take stoc a of whnt we are doing and endeavor to
think, peculinr 9 y fitting that vie do this now when we are
name t Yl em, in the time nllotted) are not given in any
theoretical interest than o P practical value, although i t
Yt’ in determining regular flig ?l t sc P’ iedules. We know t iat the aps an important
P s
other purposes it is essent,iul to h o w , for different sec-
tions of an airway, the resultant wind a t flying levels,
the perc.entage frequency of wind speeds above certain limitmg values mid the percentage frequency of occur- rence of weather conditions, such as widespread heavy
precipitation, blizzards, etc., t’liat, mike flying impossible
or a t least, hazardous. Studies along this line have al- ready been completed for Oklahoma and east Texns
and for the New Pork-Clicago airway. They should be extended to a11 secbions of the country where regular flying, coniniercial or otherwise, is likely to develop, and
preferably this should be done before such development
takes place. For much of tlie eastern and central por- tions of the country suit,able data for this problem are now nvailable.
3. The free-air .in thz1nderstorms.-T1iundeist.orms con-
stitut,e t,he most difficult condition of the atmosphere to explore. For ninn-carrying airc.raft to attempt this is suicidal; pilot balloons quickly disappear in the clouds,
and even while in sight yield data of doubtful value, owing to vertical air movements, unless observed with
two theodolit,es; and Bite flying frequently results in destruction of the kite line and damage to tlie equip-
ment, besides involving the possibility of injury to per-
sonnel. Nevertheless, a considerable mass of data has been collected, mostly perhaps immediately before and
after the thundeistornis, but partly also within them.
It is believed that an anal sis of the data in hand would
be fruitful, arid would a K mirably supplement some re-
cent statistical studies.
4. Ciouds.--Notable contributions to this subject hnve been made in the past 30 years or so, and much of our knowled e of the larger features of the planet.ary circula-
exhausted, however. Little is known, for example, as to the average and extreme thickness of the different
types, the variations in height, thickness and speed with latitude, in diflereut parts of c. clones and anticyclones
need to know more than we do now with reference to lapse rrttes in clouds and their relnt,ion to precipitation.
Data for the,se studies, alt,hougli perhaps not adequate for a final settle,nieat of them, are nerertheless quite
abundant, through the observnt’ions we liavo made with
kites and pilot balloons, and would, it is believed, well repay earnest effoi-t on the part of those who can under-
take their analysis. These data should of course be
considered and combined with the results of earlier in- vestigations, such as those of the International Cloud
Years. The latter yielded information of immense value. It seems certain that the more recent data will sup le-
nient those where they were weak and bring the w I! ole
subjec,t to a inore satisfactory status than has heretofore been possible.
5. The free air i n the Tropics.-The planetar circula- tion more close1 approaches its ideal state in t&s region
tion are % ased upon these studies. The field is far from
and in elevated as compared wit t low-lying regions. Wo
than in any ot z er. It is here comparatively free from