168 



SCIENCE 



[N. S. Vol. XLIV. No. 1127 



Dr. J. W. Shipley, who during the last two 

 years has been assistant professor of analytical 

 chemistry at the Ohio State University, is 

 going to the Agricultural College of the Uni- 

 versity of Manitoba, Winnipeg, as assistant 

 professor of chemistry. 



Mr. F. S. Nowlan, of Columbia University, 

 has been appointed instructor in mathematics 

 at the Carnegie School of Technology, Pitts- 

 burgh, Pa. 



At Lehigh University, R. L. Spencer has 

 been promoted to be assistant professor of 

 mechanical engineering and S. J. Thomas to 

 be assistant professor of biology. 



DISCUSSION AND CORRESPONDENCE 



ATMOSPHERIC TRANSMISSION 



To the Editor of Science : Replying to the 

 first point in Mr. Abbot's communication in 

 Science for February 18, 1916, page 240, in 

 reference to the variability of atmospheric 

 transmission of solar radiation during a single 

 day, I have never denied that occasions may be 

 found when the diurnal transmission is sub- 

 stantially constant, but have distinctly averred 

 that such uniformity sometimes exists. "What 

 I must deny, however, is that the Mount 

 Wilson observations of September 20 and Sep- 

 tember 21, 1914, are in the category of meas- 

 urements unaffected by diurnal changes of 

 transmissivity. The trifling variations from 

 minute to minute on these dates may indeed 

 have been small, but these are not now in 

 question. They may be eliminated for our pur- 

 pose by passing a mean curve through the 

 plotted observations ; but when thus smoothed, 

 the mean curve shows peculiarities which can 

 not be neglected. I have drawn such curves 

 and find the following significant features : 



Concerning ourselves simply with the trans- 

 mission of solar radiation by a unit of atmo- 

 spheric mass, equivalent to a single vertical 

 transmission, if the rays presented for trans- 

 mission were of unvarying quality, and if the 

 transmissive properties of the atmosphere re- 

 mained likewise unchanged through the day, 

 we should have a perfect day for the purpose 

 of the deduction of the solar constant from a 

 comparison of high-sun with low-sun meas- 



ures. But, in general, neither of these desid- 

 erata exist. Por example, on September 20, 

 1914, between air masses 2 and 3, the radia- 

 tion fell off from 1.437 to 1.311. Transmis- 

 sion by unit mass, 



T (2 _ 3) = 1.311/1.437 = 0.9124. 

 Between air masses 7 and 8, the radiation 

 diminished from 0.983 to 0.922. 

 T (T _ S) = 0.9378. 



Here it is as if the air had become more trans- 

 missive, although this undoubtedly means that, 

 for one thing, the rays which have penetrated 

 more deeply have become more transmissible 

 through the total loss of some of their more 

 absorbable ingredients. Be this as it may, we 

 can not discriminate between this source of 

 variability and another one which is always 

 present (and always potent except in times of 

 extreme cold) and which comes from the evap- 

 oration of water at the earth's surface and the 

 ascent of considerable masses of aqueous vapor 

 into the convectional layer of air in the middle 

 of the day, whereby the midday atmosphere be- 

 comes less transmissive, and the apparent 

 transmission deduced from comparison of high- 

 sun with low observations is illusory. 



For air masses 14 and 15, the radiation was 

 0.680 and 0.648; T (M „ 15) =0.9530. That is, 

 there was still a further increase of transmis- 

 sivity of unit air mass with this larger depar- 

 ture from midday conditions. Similar results 

 are found on September 21, 1914, namely, 



r (2 . s) = 1.297/1.437 = 0.9028, 

 T„_ s) =0.889/0.947 = 0.9390, 

 T (11 _ lH = 0.630/0.660 = 0.9545. 



M. R. Savelief, observing in Russia in very 

 cold weather, obtained between air masses 4.5 

 and 5.5 a transmission equivalent to that for 

 Mount Wilson between air masses 2 and 3, and 

 was able to match Mount Wilson T( 7 _ a) with 

 the interval between air masses 9 and 10. His 

 observations represent a much closer approach 

 to uniform transmission than those cited by 

 Mr. Abbot; and this is doubtless due to the 

 comparative absence of aqueous vapor whose 

 pressure at the earth's surface was from 0.7 to 

 0.9 mm. in the Russian measures, whereas the 

 Mount Wilson observations were made with 



