MONTHLY WEATHER REVIEW Editor, ALFRED J. HENRY Assistant Editor, BURTON M. VARNEY
CLOSED JULY 2,1926 ISSUED AUGUST 6, 1926 MAY, 1926 Vol. 54, No. 5
W. B. No. 895 - -__- -
A NEW PROOF OF THE VARIABILITY OF THE SUN, BASED ON MOUNT WILSON OBSERVATIONS
55/. 5=L/ By C. G. ABBOT, Assishalit Secretary
[Smithsonian Institution, Washington, June 21,19261
On returning from abroad, I find that a number of
critical articles have appeared which seem to indicate great doubt, not only R.S to the existence of short-inherval
variations of the sun ’s radiation, but, even of t,he longer-
interval swings which seem to be esposed by tlie work of the Shithsonian Astrophysical Observatory. It is
said that traces of yearly periodicity are fouiid which lead to the hypothesis that the variations are clue to terrestrial influences. It is said t,lirtt t.he pyrheliometer
observes so much sky around tlie sun t.hat, variations of
tlie haziness of the sky introduce variations of solar
constant det,erminations by no means neglivible. I t is
said t,hat the work a t Mount yilson, whicri cont,inuecl
over the summer and aut,unin months of manv years, is
so inaccurat,e, owing to terrestrial influences, that it mag
be discarded from consideration, and with it all ap li-
conditions.‘ It is said that the com utations and as-
servations are so complex and doubtful that the variations which are found in the solar constant are
more probably clue to t,hese than to real variations in
the sun.
It would be easy to collect very extensive inatter in
rebuttal, after which the authors of the c.riticisms would doubtless again bring in a rejoinder, and so the contro-
versy might go on indefinitely without convinc.ing the
authors on either side and leadin- the readers int,o n
hopeless state of uncertainty. fiortumtelv, a very
sim le consideration has occurred t,o me which, it seems pro B able, may convince the critics of the variation of
the sun that both long and short, period fluctuations
real1 exist, and that the work of Mount Wilson should
The new consideration may be understood by reflecting that if the observer could locate himself upon the moon he
would need there only t>he pyrheliometer to follow the
variations of the sun,? because there would be no screen interposing between him and its rays, and if inst,rumen-
t,ally accurate his results would be real indications of the
constancy or variability of t,lie sun itself. In the actual conditions the observer is hindered by the presence of the
ocean of atmosphere above him, whic,h contains variable
elements. If it were possible t,o c.onfine observing to conditions in w1iic.h these variable ele,nient.s remained
sensibly constant, the presence of the at,mosphere would
cations whic.h have been made of those results to went P ier
sumptions macle in conneckion with t Y ie bolometric ob-
not g e rejected.
1 The reader may here profitably refer to the MoNTaLY WEATHER REVIEW for July,
9 See MONTHLY WEATHER REVIEW, July, 1925, p. 250, the flrst two paragraphs under
1925, p. 286, quotation in second column, not omitting the seventh sentence.
“Analysis of Pyrheliometer Readings” et?.
second column. See also MONTELY WEATHER REVIEW. Dekmher, 1825, p. 527, flrst paragraph of
637-26-1
no longer be a hindrance to determining the variability
of the sun. It would then become like a partially trans-
parent screen, reducing, it is true, the solar rays, but recluoing them in the same proportion on every day of
observat,ion. Let us consider whet,her this device can not
be employed. Between the ‘yeais 1910 and 1920 the intensity of solar
radiation was observed on many days of the summer and
autumn months by means of two pyrheliometers mounted upon a single stand and read alternately, and usually by one or the other of two observers, C. G. Abbot and L. B. Aldrich. On numerous occasions durin this interval
water-flow pryheliometer, and with various secondary
pyrheliometers, which were also intercom ared between
the Astrophysical Observatory, this series of intercom-
parisons of pyrheliometers does not indicate fluctuations
of come uence in the scale of readin of these instru-
tions were made on the same days with the pyrheliometric
work, and there were determined from these spectrobolo-
iuetric curves, by the method of Fowle, the total uantity
of the atmosphere. This method, as is stated in Volume IV of the Annals of the Astrophysical Observatory, has been checked by Mr. Alclrich and Mr. Fowle with favor-
able results. In the use which I shall make of it here it is not, however, necessary to su pose that the actual quantities of precipitable water 9 etermined are strictly
correct, because the npplication is limited to the deter- mination of days of equal quantities of precipitable water.
This merely means that the depth of certain great water vapor bands in the bolographs of the infra-red spectrum
were substantially identical for equal air masses, and this
is surely an indication that the total uantity of water in
Taking, then, this homogeneous body of pyrheliometric
observations, I have limited myself in the resent article
of the sun is so nearly unifoim that it is not necessary
for the present purpose to make corrections for solar
distance. Any question of the early periodicity in the
yrheliometric work of July-s from 1910 to 1920 indicate t i! ah on some of these Julys the sun’s radiation was more intense than on
others 8 To solve this question, I divided the observations as re orted in Volume IV of the Annals of the Astrophysiwl
Ogservatory into four groups. Group 1 cogtaiaed ‘Gnly
these pyrheliometers were compared wit % the standard
themselves. iis indicated in Volume IV o P the Annals of
ments. !During all this interval, bo T ometric observa-
of precipitable water between the observer and t B e limit
the atmosphere was for these days su ‘t stantially equal.
mainly to the month of July. In this mont I: the distance
solar constant values is obvious I y eliminated. The first question, then, is: Does the
191
192 MONTHLY WEATHER REVIEW
.5.---- ...................................... 10 ...........................................
17 ...........................................
*6 ...........................................
27 ...........................................
28 ...........................................
:!Y ...........................................
MAY, 19%
Nni .
4.9 5.5
3.6
7.9
4. I
I .(] $5
those days in which the a parent atmospheric trans-
getween narrow limits, whose mean was approximately
0.904, and when the precipitable water lay between
narrow limits averaging aprxjmately 5 mm. Group 2
contained only days in w ich the atmospheric trans-
parency, still within narrow limits, was somewhat less nntl the precipitable water, also within narrow limits, was
approximately 12 mm. Group 3 contained only those
days in which the atmospheric trans nrency, still between
also between narrow limits, was approximately 20 m1n. Group 4 contained the days which were rejectecl, either because the precipitable water much exceeded 20 mm., or
the transparency, if fnllinm within one of the groups, WRS
accom anied by precipitnzle water conditions which did
re'ection, numbers of days which are given in Tnhle 2,
wkch follows. ive in Table 1 the iirrnnge-
been printed in three types for R purpose which will
appear later.
arency, as determined by t Tl e pj-rheliemeter alone, lny
narrow limits, was again less, and t E e precipitable water,
not fa1 P in the same group. There remained, after this
ment of data for the niont V P o July, 1915. Values have
To illustrate the groupin
2.655
?.w
2.935
2.940
2.923
TABLE l.--Saniple grouping
1 .9 ~0
1.950
1.976
1.979
1.943
Pyrh.1 I Solarcon- m=1.5 1 stant E'o
I O ...........................................
11 ........................................... I ? ...........................................
13 ...........................................
14 ...........................................
13 ...........................................
Mean ..................................
9 ...........................................
i~ ..........................................
Y~~~ ........................................
9 ...........................................
14.8
11.7
12.4
11.4
8.6
10.2
11.5
1 x 5 21.9
21.2 15.6
?.SO?
2.690
2.710 2.i10
1.764
1.CAS
1.915
1.968 1.955
1.934
~~
~~
0.070 0.071
.......
meau ....
l;t% I*% -1.02 -2.01
-1 .0 6 -1.34 -0.08 +0.41
-0.Y5 -1.70
~~
~~
~___
5, IF. 17, 26, 27, 23. 29, 31 ...................... 1 0 ,l l . 12,13,14, I8 ...........................
3 ,7 ,3 ,5
Tutal.. ................................
3,4,5,6,17.18,25,25 ......................... I , 10, 12,15 ...................................
1 1 , 31
Totnl. .................................
......................................
.........................................
.'., .>-
IS
S
5s
32
11' 4
45
-~
O.Oi3 0.031 0.053
mem ..-.
~~
-0.14 +0.31 +1.56 +0.6i +2.06 -0.62
+0.69 +0.30
~~
~~
~___
. b'JS
.Si8
,006
.0l7 ..............
Totill.. ................................
11. 24. ?5, ?ti, ?i, 2 S -. .........................
6,14,15,?1,32,23 ............................
4,5,9,3Y .....................................
................................. Tots.
1,2Y,30, 31 ................................... 2 ,3 ,6 ,7 ......................................
10,11,14 .....................................
1ti
24
15
Y
50
1 Ij
I:!
ti
___
-~
~- ~-
mean ....
0.104
0.051
0.114
mean ....
+I. 21; I +I. 11 ___~ ~~
+I . I U +].I 4
-0.66 +0.20
+O. 44 +O. i 3
+O.?l +0.41
___~
~___ ___-
,902
.SY5
.Si7
,023
,011;
,026
,901
.%5
.Siti
,015
.01Y
.Or?
Totiil.. ................................
9, 11, 12. 13, 19 ................................ 8, 10, 14. l i . 13, 2.5, 26 .........................
24.29 .........................................
Tot31 ..................................
34
Y J
I1 4
45
.YO3
.SS
.a 1
,019
.015
.WG
Date ,pp. *'a'
0. .S96
,901
,906
,904
,906
,908
. Yu(I
.w
, VI13
.8Yb
. S Y i
. s91
. SSI
,895
,909
~~
2. SS.4
2.797 1.929
2.800 I .#?
2 .m 1.Ni
2.765 1.959
2.860 1.949
2.857 1.960
.395
,394
. 8% ,898
,897
hlrau .................................. 1 I!#.? ,858
1 To reduce to culories, multiply by 0.511.
-
Pyrheliometer at air mss 1.5 Ppt. Ha0 Num- ber of days froup July dates
~
isnge
Year __
Range hleou hIenn Range I Pyr. I S. C.
.U!Il.
4. !I
11. i 2. 0
4 .8
Y. 7
22.0
6. Y 13.2
IS. G
5.4
11.5
18.2
5 .6
11. ti
17.0
4. a
Mlll.
4.2
1;. I;
.....
3.4 1. 2
1. Y
3. 8
7.6 3.2
4.4 6.2
6.3
3.2 4. 3
0.7
0.8
6.0 ......
3.0
4.5
3.6
4.9 3.3 4.6
4.2 5.0 2.1
0.902 0.011
,597 .UOO
. S i Y .......
1910
1811
' 1914
1015
1916
1917
1 2
3
1 2
3
1
2 3
1 2
3
1
2 3
1
2 3
1
2 3
1
2
3
1
2
3
6
3
1
6 2 2
3 4
5
3 6
4
8
4
2
2 0
4
6
6
4
4
4
3
5
2
1 461
1.395 1.360
1.463
1.383 1.333
1.462
1.39s 1.35i
1.474
1.43'
1.389
1.480 1.39Y
1.367
1.4i5
I . 398 .........
1.492
1.413 1.332
1.493 1.414 1.332
1.475
1.414
1.338
0. iin 0.113 1. 899 1 . UO!I
1. 963
W'righte
1.9%
1. S95
1. YS.9
Weight?
1.347 1.949
1.9B7
Weighte
1. 944
1. M8 1.943
Weighte
1.945
1.908
1 .9 3
Weighte
1. Y29
2.007
Weighte
1. Y60
1.943 1.959
Weighte
1.955
1. Y 4 l
1. YS!
Weighte
1.942
1. 934
1. Y34
Wei:bte
........
2 ,3 .4 . 5.3,''s ................................
ti, i, 15 .......................................
19 ...........................................
Total ...... -..% ....................
0.082 0.009 0.024
0. 041 0. 046
0. 066
0. @;(I
0.063 0.057
0. MO 0.064
0.025
0.033
0.02s . -. . - -. -
0. M;6 0. 027
U. O i l
0.060 0.013
0. OiS
0. 0Zi
0.023 0. uti4
0.085 -0.51 -0.57
0.OYS -1.56 -?.06 0.0% -2.06 -2.44
.w .011
.YU1 ,009 .ss5 .U l G
1,4,S,24.30,31 .............................
7. I? ..........................................
13, 20 .........................................
Total.. ............................... mean..-. -0.133 -1.17
~~
0,039 -U. S5 +O. 72 0.12i I -U.30 +0.6?
5, 2S, .Y. ....................................
1. 2, li, 26 ....................................
18, 19, 20, 21, 30 ............................... 1 11)
Tutnl .................................. 1 34 meau----l -0.72 I +0.60
.903 ,012
SY5 .03
:s s
~
,010
0.022 f O .2 7 +0.36 0.01s I -0.80 1 -1.3Y 0.027 +O. 44 -1.33
I l i . S
3.4 12.3 20. 1
5.0
11.9
?3. Y
6. 3
13.6
21.0
1913
1919
1920
meon .... ______ +O. 13 +O. 55
0.022
0. OFIO +O. 2s -0.0,;
0.00s -1.io -1.29
On each of the days included in groups 1, 2, and 3
the reading of the pyrheliometer as it would have been
found at air mass 1.5 was deteimined by logarithniic
interpolation from the series of observations reported in
Volume IV of the Annals, and on each of the clays the value of t8he solar constant of radiation, E'o, as given in
MAY, 1926 MONTHLY WEATHER REVIEW 193
1910 ....................................................
1911 ....................................................
1914 ....................................................
1Yl5 ....................................................
1916 ....................................................
1917 ....................................................
191s ...................................................
191Y ....................................................
lB'ZO...- ................................................
Volwne IV of the Annals, was taken out. Mean values
for the pyrheliometry a t air mass 1.5 and for the solar constant were obtained for eac.h group in each month.
It was necessary to omit the years 1913 and 1913 in this analysis, because the volcanic dust t,hrown up in
the great eruption of Mount Katmai had rendered t,he atmosphere so turbid t,liat, although precipitable water
conditions suitable to t,he groups were available, t,hey were not accom anied by transmission c,oeffic.ient,s
which fitted a t aly. Accordingly, t,hese years were re-
jected from the study. Readers will erceive that errors due to extraneous
sun seen by the pyrheliometer are minimized in this method of treatment because the brightness ancl trans- arency of the atmosphere were practically ident,ical
for all days and years wit,hin the same -roup, and be-
cause sinc,e all observations are taken in J d y the out, oing
regions of nearly the same temperature from year t,o
year. Also errors from changing t,ransparency during a day are minimized because, first, no esce,pt,ional days were included, and, second, because all observat,ions
relate to one and the same moment of t,he day and not, as
in t,he long met,hod bolometry, t,o a eriod of several hours.
It was considered t.hat the indivi B unl days of t.he several groups should receive different weights, and after con-
sideration, the convention was rttfopt,ed that individual days falling in group 1 should receive weight 4: t,hose
in group 2, we,ight 3: and t,hose in group 3, weight 3,
respective,ly. With this convention, and in considern- tion of the numbers of days in t,he several groups, there
were obtained for the whole int,erval 1910 to 1920,3
group means both of the pyrheliometry at air mass 1.5
and of the solar constant Elo. See Table 4 below. From
these group means were taken the percent,age departures of the group means for t,he seve,ral Julys, as indicabed in
Table 2. Finally, in order to get single values repre-
senting the pyrheliometric and bolometric result of each July, the weighted mean of the percentage departures
for the three groups was detemined, as n0te.d in Table
2, and which are summarized as the results of the in-
vestigation in Table 3. The results are also shown
graphically in the accompanying illustrat'ion.
radiat,ion from t K e skv and toward t,he sky around t,he
radiation fro& the instrument reached at.mosp a eric
-0.95 -1.70 -0.60 -1.17 -0.72 +0.60
+0.69 +0.30
+o.oa -0.2? +1.26 +l. 11
fO.44 fO.73
i-0.13 +0.55 -0.52 -0.49
....... .Fun Spotz
-
Solar constant E'o
I t will be perceived that with the exception of July.
1914, there is it very close agreement between the bolo-
metric and pyrheliometric results, and that both are in strong correlation with the variation of the sun-spot numbers. I have also tried the month of August for
the same interval, in the same manner, and have ob-
tained results of the same general import, thou h they also indicate a departure in the same sense for t % e bolo-
: The Mount Wilson solar constant values for 1919 and 1920 are taken here as they were in Tables 49 and 50 of Volume IV of the Annals of the .4strophysical Observatory
of the Smithsonian Institution, for leasons stated on page 1 7 i thereof.
Pyrheliomctry hI=1.5
metric result of 1914.4 Thus, this new and simple method confirms the result formerly obtained from Mount
Wilson solar constant determinations to the effect that in a term of years the intensity of solar radiation in-
creases ,with sun-spot numbers. The new work supports the solar constant values both as to the times and magni-
tudes of the change.
TABLE 3.-Long-interval solar uariutioiis, J u l y
(hrount WIISOU Data)
Percentage depar-
I tures
Year Sun-spot numbers
14 3 5 71
a 53 117
105
fA 26
In the following Table 4, are given the grou means f u r the solar constant (E'~) and pyrlieliometry &I= 1.5)
for tlie whole series of Julys and Augusts. In obtaining all the pyrheliometric ralues for August, I allowed for
the variation in solar distance and computecl the ralues for the same solar distance as of July 15, in order to fit
them for use in Table 6 , below.
TABLE a.-G'roitp nietru aolar constarit nntl pyrhc1ioi)ictric d u e s
Mean values: July ............................ 1
1.935
1
1.935
I
1.955
1 :::A; I ::$Ai 1
1.361
August .__.. .................... 1.937 1.942 1.957 1.318
It will be noted that the aweemexit ofisolar constant
values for Groups 1 and 2 for 3nly and August is as close
as c ~u l d he hoped for, but that in each case the mean for
Group 3 is about 1 per cent, higher than for the others.
This, I am inclined to think, is a real indication that the solar constant values obtained on Mount Wilson on days of excessive haziness and humidity were made too high
by reason of the ilafluence of sky radiation. Doctor Dorno points out this source of error in the MONTHLY WEATHER REVIEW for December. We hac1 become aware of it IL good while before, and had taken measures
to evaluate its mngnitude and to eliminate it in future
work. I hope to treat this matter more extensively in a
forthconiinu publication. rt is to Be noted that in taking the mean monthly depurtures of the solar constant in Table 3 of this present
comniunication, they are not comparable with the depar-
tures which could be taken for the mean monthly values published by Mr. Clayton in Smithsonian Miscellaneous
C'ollrctions, vol. 77, No. G , and in Table 53 of Volume IV of the Annals of the Astrophysical Observatory, for
two reasons.
4 One is inclined to think from the resultsofboth Julyand dugust that forsome reasons
the bolometric results of 1914 are about 1 per cent high throughout that year a con- clusion a.hich is quite in line with their departure from what would he expecteiat that time of the sun-spot cycle. I have made n partial investigation of this question and find that tlie atmospheric transmission coeficients for the different wave lengths for 1514 mere lower than a.ould he expected for days of equal precipitrrble water for other years, and it is possible that me may be able to Bnd why the results of this year are thus
out of line
194 MONTHLY WEATHER REVIEW BbY, 192%
Pyrh. mean
____
-2.45 -2.40 -1.41 -1.67 -1.07
-1.31 -1.29
-?. 65 -1.10
-2.67 -1.02 -2.49
-1.12 -1. ltj
-1.55
-1 .5
-1.3; -0.67
-1.47
I n the first place, present departures are given in per-
centages calc,ulated severally from t’he general means of the three groups above described. Thereby tjhe c.orrec-
tion for sky brightness which I have just explained is
eliminated. In the second lace, t,he mean mont,hly values given by
Mr. Clayton a n i published in Volume IV of the Annals
include all of the days, among them t,liose of Group 4
which have been rejected in t,his discussion, for tlie reasons
given above. As stated by me in Smiths0nia.n Miscellaneous Col-
lections, vol. 77, No. 5, p. 3, I had su posed that the range
Wilson was erhaps twice as great as the true one on
This ex ectation is now confirmed, for it, is se,en that the
ran e o F results for the month of July given in Table 53
of folume IV of the Annals is about double the range
which is given in Figure 1 of the present paper.
of solar const,ant values given in t, K e Annals for Mount
acc.ount of t E e, s0urc.e.s of e.rror which I cliscussed tliwe.
TABLE 5.-Prooj of short-interval s o h variation
(Percentage deviations from monthly means for individual days)
Solar
con- staut mean
-1. 05 -1.74
+o. 10
-0 .q
+0.14 -0.42 -0.w
-0.41
-0.62 -0.46
+1.50 -0.45 -0.m -0.13 -0.19 -0.10 -0.61
+o. 32
-0.42
High values
I
1810, July ___._
1911 ........... 1914 ........... 1915 ........... 1818 ........... 1817 ........... 1918 ........... 1918 ...........
1820 ...........
1810,August.-. 1911 ........... 1814 ........... 1815 ...........
Month
IF-1
Pyrh.
1 mean stnnt days mean
2 2.58 1.68
2 2.3? 1.63 1 1.38 0.10 4 1.69 0.52 3 0.82 -0.09
1 1.31 0.48
3 1.28 0.82 3. 0.67 -0.10
2 0.59 0.16
5 2.11 0.51
6 1.71 0.91
2 1.10 0.15
6 0.87 0.16
...... I ...............
2 i-0.33 ..................... ......................
......................
2 -0.02
3 +0.02 2 +o. 11
3 -0.21
1 +0.20 2 -0.05
......................
......................
...................... ...................... -
20
fO.03
Medium vnlues
-0.11
+O.lO
-0.24 -0.93 +0.33
-0.36 -0.39
-0.34
days mean
-1-1-
2 -0.09 -0.47 2 +o. 18 -0.92 1 I 1 0.00 -1.32
__
Num-
her
of days
2 2 1 4
3 1
3 1
1
4
10 1 5 5
; 1
3
51
__
I have used this new pyrheliometric method of con-
sideration not only as furnishing evidence of lone interval fluctuations of the solar radiation, but to Xetermine
whether short interval solar changes are also probably
real. For this pur ose I have confined myself to the valuesinGroup 1 as Raving greater weight than the others. These values I have divided, in each nionth (July and
August, 1910 to 1920) as between high, medium, and low,
Medium values, however, are frecpentljr absent. A11 of
the days included in Table 1 are thus indicated by dis-
es, but thou h doubtless Grou s 2 and 3 would
on1 Group 1 in this study. I have set over against the
the same identical days. Obviously the range of
pyrheliometry includes its errors and differences of conditions. Hence it must esceed the range of solar
constant values for identical da s whose errors may tend
de arture from the mean of tlist
metry a t air mass 1.5 and of the solar constant value E6.
show the tP% p enomenon,$ have, as stated a%ove, employed
pyr 1 eliometry the solar constant values, E’o, found on
In eac l instance I have deter-
P month, both of the pyrlielio-
I n general, the two sets of data agree as showing which
are the days of high ancl days of low solar constant. The monthly values arid t,he mean of all the results are as
given in Table 5.
From this, it seems to be indicated that not only did bhe solar constant vary in a close relation with the sun-
spot numbers during the months of July and August from
1910 to 1920 (excluding t’he years 1912 and 1913, which
were not capable of treat,ment by the new method) but
also that during this long eriod of t,ime the high and low
selves in the pyrheliometry quite as plainly as in the solar constant values published in Volume IV of the
Ann&, and on the whole in harmony therewith.
The pj-rhe.liometric me.thod w1iic.h I have explained has some valuable applications and certain limitations,
as shown in Table 5.
Bdva.nta.ges qf thr method.-1. It is direct, for it simply employs measurenienhs of total radiation, without spect,rum work except as an indication of atmospheric
humidity.
2 . It is competent to confirm the existence of solar variability both of long and short interval.
3. It furnishes means of testing whether the eneral
over a period of years. 4. I t will give new testimony as to the reality of certain
a parent prolonged depressions of the solar constant,
Disad.ca.ntage!s o the method.-1 . The pyrheliometric method c.an not e applied c.onvincing1 to treat long interval variat,ions in years like 1912 an 1913 when the
atmosphe,ric t,ranspare,ncg for equal humidity was
ahnorma,lly low on account of volcanic dust.
2. It is unsatisfactory escept for stations of very
escellent and uniform conditions.
3. It is applicable to only a pa.rt of the cloudless days
a t any station, because on some days the relations be- tween atmospheric humidity and t,ransparency are so
abnormal that such d a p fit none of the groups.
Dif- ferences between sky conditions of different days, even if small, produce daerences of pyrheliometeric readings.
These must be eliminated by taking means for many da s.
5. I n short, the method is not a substitute for the so T ar
constant methods, but only supplementary to them. We are uoing on to ap ly tlie new pyrheliometric
met,hocl to bontezuma a n 8 Harciua Hala results from
1920 to 1925. In each of these stations the pyrhelio-
meters were repeatedly compared with other instruments
and seemed to furnish a perfectly homogeneous series
of observations for discussion. Furthermore, the char- acter of the sky, es ecially a t Montezuma, is so much superior to that a t b o u n t Wilson that we ma expect
thus far. It will be exc.eeding1 interestin when this
discussion is completed, to see J t h e great ckpression of
t,he solar constant from about March, 1922! to the
present t,ime is x-erified, ancl it will be exceedingly valu- able to assure ourselves that the scale of observations throughout t.he recent period has remained unchanged.
I hope soon to publish t,he results of such a study in the
Smithsonian Miscellaneous Collections, and a t that time to discuss more fully the influence of radiation from and toward the sky near the sun, and the influence of vol-
canic dust in solar constant values.
clays for the months of Ju P y and August indicated them-
scale of solar constant determinat,ions remains unc % anged
wiich, P if real, are iniportant.
B sf
4. It is incapable of giviug individual results.
even better results than have been found in t x e work