8 4 



SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 65 



space. The problem is important for the knowledge of the loss of 

 heat from the oceans, and would probably be worth a special inves- 

 tigation in connection with an elaborate discussion of the quantity 

 of heat absorbed from the incoming sun and sky radiation by water 

 surfaces. Here I propose only to give a short preliminary survey 

 of the question, giving at the same time the general outlines of the 

 probable conditions. 



Zenith distance. 



Fig. 15. — Radiation from water surface to sky. Lower curve for water 

 surface. Upper curve for perfect radiator. From Bassour observations 

 (p — 5 mm.). Ratio of areas 0.937. 



In figure 12 I have given some curves representing the relative 

 radiation from a black surface in various directions toward rings of 

 equal angular width. " The total energy emitted is represented by the 

 areas of these curves. Now, if every ordinate is multiplied by the 

 factor (i— R<p), where R<p can be obtained from Fresnel's formulae, 

 if we know the index of refraction, the area included by the new 

 curve will give us the radiation emitted by a water surface under the 

 same conditions of temperature and water-vapor pressure. In figure 

 15 such curves are given. I have here assumed the mean refrac- 



