4 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 65 



Our determinations rest on the assumption that for all excellent 

 days the atmosphere may be regarded without sensible error as made 

 up of layers, concentric with the earth, which may differ in trans- 

 parency from layer to layer in any gradual manner, but which, within 

 the time and space covered by a solar beam during a single morning 

 of observation, are for each layer by itself sensibly of uniform 

 transparency. As the relative transparency of the several layers is 

 not assumed to be known, it is convenient to limit the duration of a 

 single series of observations to the. time interval during which the 

 solar zenith distance is less than j?° . During this interval the rate 

 of decrease of path of the solar beam in the atmosphere, with decreas- 

 ing solar zenith distance, is sensibly the same in all the supposed 

 atmospheric layers, and is proportional to the change of the secant 

 of the zenith distance. For greater zenith distances than these this 

 proportionality does not hold, because of the influences of curvature 

 of the earth and of atmospheric refraction. 



Figures I and 2, and table 2. show something of the variety of 

 conditions of observation encountered ; first, as regarding the inten- 

 sity of sunlight at the observing station ; second, as to the effect of 

 atmospheric humidity on the infra-red spectrum ; third, as the effect 

 of dust upon the visible spectrum. We draw attention to the close 

 agreement of the solar constant values obtained in these contrasting 

 circumstances of observation. 



Table 2 — J'arictx of Conditions of Obserzutioii 



1 This value is corrected as suggested in note 2, Annals III, page 113. 



1 Determined bv Fowle's spectroscopic method, and gives the depth of liquid water which 

 would result if ail the atmospheric water vapor above the station should be precipitated. 

 Experiments of ion show close agreement of this method in its results with those obtained 

 for the same days by integration of humidity observed at all altitudes by sounding balloons. 



