ADVISORY COMMITTEE ON ASTRONOMY 107 



with the questions (a) and (6) , relating to the amount and supposed 

 variability of the "solar constant," for it is impossible to know 

 how much of this total is lost before the measuring instrument is 

 reached. Experiments near sea-level can never determine this 

 correction with entire certainty, though made wath the widely dif- 

 fering thicknesses of intervening air corresponding to different 

 altitudes of the sun, for the atmosphere is almost unbelievably 

 variable in its absorption from day to day and month to month, 

 and even in different parts of an apparently clear day. Especially 

 is this the case with its lower layers, and, besides this, no two 

 layers of air at different altitudes are alike in their absorption. 

 What is reouired is to get high up in clear air, so thai the absorp- 

 tion will be much smaller (it amounts at sea-level to something like 

 50 per cent of the total radiation) > and where thc't which is left is 

 more constant in amount. Furthermore, dependence ought not to 

 be placed exchtsively on observations taken at tlie same elevation 

 above sea -level, for, while the higher layers of air are less variable 

 in ab.sorption than the lower, and while the absolute amount of 

 absotption is less the higher the observer, yet it is necessary also 

 to take into account the variation in quality of the absorption with 

 the elevation. A second lower station ought, therefore, to be occu- 

 pied and simultaneous observations made at the high and the low 

 station. This ought not to lead to dispensing with a lower observ'-- 

 ing station equipped equally as well for work as the higher, and 

 where, indeed, it might prove feasible, after the study of the air had 

 progressed satisfactorily, to do much of the observation on the ' ' solar 

 constant." It is clear that the extreme difeculty of observation 

 upon the high station ought thus to be avoided just as far as 

 possible. 



Methods of Determining Absorption of the Air and ike Value of the 

 Solar Constant. — These problems are naturally to be studied simul- 

 taneously, la this general statement I will not describe the pro- 

 cedure special to each, but only indicate several kinds ot data that 

 ought to be used. First, solar energy curves taken on the same day 

 at the high station through different air masses to be computed from 

 the altitude of the Sun and the height of the barometer at the .several 

 times of observation ; .second, solar energy curves taken at the low 

 station through similar differences of air mass ; third, similar curves 

 taken with the solar beam reflected on an approximate level betv.'een 

 distant mirrors both at high and at low stations ; fourth, actinometer 

 observations coincident with these several kinds of holographic work ; 



