4 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 68 



lack of sufficient sensitiveness of the apparatus for measuring the 

 feeble radiation of the longer wave-lengths, certain measures were 

 made connecting the amounts of absorption in the bands of the upper 

 infra-red, 0.7 fx to 2.0 /x. with the quantity of water vapor producmg 

 them/ These bands are those affecting the incoming radiation from 

 the sun. 



In the early part of the summer of 1913 attempts were made to use 

 a vacuum bolometer which Dr. Abbot had meanwhile prepared. The 

 air-pump, then in the possession of the Observatory, was not suitable 

 and the summer became so far spent that the experiments were finally 

 carried on without a vacutun. When the results were reduced the 

 next winter a serious discrepancy Avas found, the cause of which was 

 so obscure, that it escaped detection for some time. It necessitated 

 the repetition of the experiments in the summer of 1914. This source 

 of error lay in the circumstance that in the form of lamp then used 

 for the radiation source, the bolometer was exposed to a source of 

 somewhat different effective temperature when the radiation passed 

 through the water-vapor tube than when it passed through the 

 spectroscope alone. The comparison of the observed energies in 

 the two cases was to serve as a measure of the absorptions due to 

 the water vapor and thus the change in quality of the rays just 

 explained caused error. This error was avoided by the use of proper 

 diaphragms and the construction of a more suitable and far more 

 effective form of lamp. Some doubt as to the matter of field light 

 required further observations during the summer of 1916. Because 

 of complications resulting from the absorption of the radiation by 

 carbonic-acid gas, further observations were made during some very 

 cold days of the winter of 1916-17. On such days the losses due to 



^ The results of these oljservations were published in a series of articles in 

 the Astrophysical Journal discussing the transmission of radiation through 

 moist and dry air and water vapor between the wave-lengths 0.35 and 2.0 fi. 

 The first (1. c. 35, p. 149, 1912) gave the laboratory calibration, with known 

 amounts of water vapor, of the intensity of energy in the bottom of certain 

 absorption bands, the depths of which could be accurately measured bolo- 

 metrically. The second (37, p. 359, 1913) gave applications of the first paper 

 to the spectroscopic determination of the water vapor above Mount Wilson 

 ^nd a comparison of these values with determinations by Hann's formula. 

 The third (38, p. 392, 1913) treated of the non-selective scattering of dry air 

 and water vapor for the spectrum region 0.35 and 2.0 \i. The fourth (40, 

 p. 435, 1914) was concerned with the application of the dry-air transmission 

 coefficients to the determination of Avogadro's constant, the number of mole- 

 cules in a gram-molecule of any gas. The fifth (42, p. 394, 191 5) gave the 

 corresponding selective absorptions in the spectrum region 0.35 to 2.00 \i. 



