NO. 4 SOLAR RADIATION — ABBOT, FOWLE, AND ALDRICH 2"] 



a curvature must occur in logarithmic plots of total radiation. It 

 will be seen that our observations fully confirm their view, which 

 depends upon the fact that the total radiation is composed of parts 

 for which the atmosphere has very different transmission coeffi- 

 cients. 



Referring to tables 2 and 11, and to Annals, Vol. Ill, table 47, 

 the reader will see that the atmospheric transmission on September 

 20 and 21, 1914, was distinctly above the average, and indeed was 

 as high as we have ever found on Mt. Wilson. Secondly, the 

 quantity of water vapor between the station and the zenith, as found 

 by Mr. Fowle's spectroscopic method, was unusually small and satis- 

 factorily constant. Hence, we may conclude that the two days in 

 question were, as they appeared to the eye, days of the highest excel- 

 lence at Mt. Wilson. When we compare the results obtained from 

 them on the solar constant of radiation, as given in table 10, with 

 those obtained in other years, as shown in table 1 and in Annals, Vol. 

 Ill, table 44, we see that the values were very close to the mean 

 results of all our observations. We see further, from table 10, that 

 the results obtained were very nearly the same, whether we used only 

 the later observations, taken between air-mass 1.3 and air-mass 4, as 

 in our usual investigations ; whether we employ only the observations 

 between air-mass 4 and air-mass 12, as recommended by Mr. Very; 

 or, finally, whether we take all the observations from air-mass 1.3 to 

 air-mass 20. In every case the result is the. same almost within the 

 error of computing. 



From this we feel ourselves fully justified in drawing the con- 

 clusion that our former work has not been vitiated by the employment 

 of too small air-masses, and that, in fact, hardly different results 

 would have been obtained had we observed from sunrise of every 

 day in which we have worked. On account of the uncertainty which 

 attends the theory of the determination of air-masses, when zenith 

 distances exceeding 75 ° are in question, we conceive that it will be 

 better to confine our observations hereafter, as we have generally 

 done in the past, to the range of air-masses less than 4, where the 

 secant formula applies in all atmospheric layers, irrespective of 

 optical density, refraction, or the earth's curvature. 



Third objection. — We attach very little weight to any determina- 

 tions of the solar constant of radiation which we have made hitherto, 

 except those made by the spectro-bolometric method developed by 

 Langley, as just employed for September 20, 1914, and which is the 

 definitive method employed by the Astrophysical Observatory of the 



