i8 



SMITHSONIAN MISCELLANEOUS COLLECTIONS 



VOL. 



65 



As appears above, the computed transmission for wave lengths 

 exceeding 1.37 jx is approximately unity, and the computed atmos- 

 pheric extinction coefficient, as given in line 3, sensibly zero. Line 4 

 gives the general extinction for 0.5 centimeter of precipitable water 

 vapor, corresponding to the humidity of September 20, 1914. 



We are now in position to determine a correction to F(z) as given 

 by Bemporad. If the extinction were all molecular scattering, his 

 values would be the true ones. If it were all due to water vapor, 

 we ought to employ approximately secant z, because of the low level 

 of water vapor. We have therefore determined for each wave- 

 length the weighted mean between Bemporad's F(z) and secant s, 

 giving weights in proportion to the numbers (1 — a a \) and 

 (1 — dw x ) for wave lengths less than 0.764 fi, and in proportion to 



the numbers ( 1 — a c \ ) and ( 1 — aw-, ) for wave lengths exceeding 



2 

 0.764 p.. In one case we have made an exception, namely, for wave 



length 2.348 jx, which is within the band of carbon dioxide absorption. 

 As this gas forms a nearly constant percentage of the atmosphere up 

 to a level of more than 10,000 meters, we have used Bemporad's 

 F{z) at this wave length. In figures 3 and 4 the reader will see 

 plotted the air-masses as used, and also the lesser air-masses corre- 

 sponding to Bemporad's F(z) . 



The following are the circumstances of the spectro-bolometric 

 observations of September 20, 1914: 



Extent of spectrum observed (in arc) 270'. Bolometer subtends 17" . Slit 



subtends 50". 

 Extent of spectrum observed in wave lengths : X — 0.342 n to 1 = 2.348 \i. 

 Time elapsing after start o m 30 s to 7 m 15 s . 



Bolograph No. 



Time of start; 120th me- 

 ridian mean time 



h. m. 

 9 40 



Latitude, 34 12' 55" N. Longitude, 118 3' 34" W. Altitude, 1,727 meters. 

 1 Bolograph 9 omitted because of interference of a guy wire. 



In accordance with our usual course, described in Vol. Ill of 

 our Annals, we measured the ordinates of smoothed curves on all 

 the holographs at 38 wave lengths. These were equally spaced in 

 prismatic deviation, excepting that in a portion of the infra-red 

 spectrum we observed at points twice as close together as in the other 

 parts of the spectrum. Table 8 includes the measured ordinates of 

 the smoothed curves (unit 0.1 mm.) and corresponding air-masses, 

 according to Bemporad, for September 20. Our corrected air-masses 

 appear only on figures 3 and 4. The third column of the table gives 

 the factor to reduce to uniform scale throughout the spectrum. 



