12 FREER. 



from year to year,"^ but such variations are not frequent enough 

 or of great enough intensity to alter the picture as a whole. 



Although the maximum radiation at the various points men- 

 tioned is very similar in all, yet if we take the annual totals, we 

 find differences for such points as have been compared. Kim- 

 ball 22 has calculated the average monthly totals for Washington 

 and Warsaw for normal surfaces, and from them we can obtain 

 the yearly totals, which for Washington are 254,026 and for 

 Warsaw 216,200, so that Warsaw actually has 85 per cent of 

 the radiation received at Washington, although it is 14° farther 

 north.^s Unfortunately, pyrheliometer readings for places in 

 the Tropics are not at hand. We have ordered an Angstrom 

 pyrheliometer nearly a year ago, but the instrument has not yet 

 arrived. When it does, we will begin readings in Manila and 

 thus obtain comparative data. Enough has been shown already 

 to demonstrate that meteorological phenomena, percentage of 

 possible sunshine, and varying atmospheric transmissibility have 

 more influence than variation in the actual solar insolation on 

 perfectly clear days, and it is evident that such factors can 

 just as readily be introduced in the Tropics as in other regions 

 of the earth. 



Because data with the Angstrom pyrheliometer in the Tropics 

 are lacking, we attempted to solve the problem, for the present, 

 by having recourse to animal experiments. 



In considering this second phase of the question, a few funda- 

 mental facts must be borne in mind. The air surrounding the 

 earth absorbs the rays of the sun in a certain proportion, and 

 another part is reduced by reflection, molecular scattering, and 

 dispersion; this takes place in a greater proportion with the 

 more refrangible than with the less refrangible rays, so that 

 the light reaching the earth contains relatively a greater amount 

 of the rays of the upper range of the visible spectrum and infra- 

 red than are in the sunlight before it strikes the atmospheric 

 layer; indeed, all ultra-violet rays up to 291 jxix. disappear. On 

 the other hand, the dark heat vibrations of great wave length, 

 radiating from the earth, are absorbed in great measure by the 

 atmosphere. The coefficient of absorption of the air increases 

 with increasing density, but it never reaches that of a solid or 



"Kimball, /bid., 114, 115. 



"Ibid., 103. 



" These computations for Warsaw cover the period from July 1, 1904, 

 to December 31, 1906; the period for Washington from June, 1905, to 

 March 21, 1910. 



