34 



SMITHSONIAN MISCELLANEOUS COLLECTIONS 



VOL. 71 



percentage of variation was probably considerably less than the per- 

 centage of variation in any particular portions of the sky. 



With such wide ranges of variation definite results are out of the 

 question. Nevertheless, it seems reasonable to state that from 50 to 

 60 per cent of the radiation from the clouds was reflected by the 

 snow on this particular day. 



February 23. Means of Sets 



Set 



Hour angle 



— i'':i4'":iS' 



02 :oo 



o :53 :oo 



46 -30 



38 -.30 



— 31 :iS 



+0 :29 :30 



40 :oo 



50 :oo 



Ratio to sun 



H- 30° sky 

 around 



I .00 

 . 127 

 .128 

 . 112 

 .126 



I. 00 



.143 

 ■035 

 .010 



There was far better agreement among the readings than on the 

 previous day, as will be seen by inspection of the results for a and b, 

 which are typical. 



It is remarkable that the ratio of the radiation measured from 

 the snow is of the order of one-eighth the amount coming from the 

 Sim and a cone of sky 30° around it, while with the same cone, on the 

 preceding day, the radiation from the snow was found to be rather 

 more than one-half that coming from the clouds. This discrepancy 

 can be accounted for as follows : On February 22 we observed a 

 cone of light from a certain part of a totally cloudy sky and a similar 

 cone from the snow such that a directly reflected beam from the 

 center of the observed clouds would strike the surface of the strips 



