NO. 19 ENERGY SPECTRA OF STARS ABBOT AND ALDRICH 3 



the best wave lengths to observe in the stellar spectra, and we chose 



eight places more or less by guess. After completing our observing 



on September 7, 8, 9, Abbot took the prism to our station on Table 



Mountain, Calif., and with A. F. Moore made holographs of the 



solar spectrum. From these holographs, combined with the work of 



Hoover and Greeley, the curves shown here in figure i were 



computed. 



The combination prism actually prepared is not as uniform in its 



dispersion as that which we hoped to obtain, whose characteristics are 



shown in Publication 3843, figure 2. The actual prism has a range of 



dd 



-7— of fivefold in dispersion between wave lengths 3300 and 22,000 A. 



Still it is more than twice as uniform in dispersion as a simple 60° 

 Jena Crown glass prism, such as we use for solar-constant work, and 

 it is almost completely uniform in dispersion between the wave lengths 

 of the D sodium lines and 22,000 A. 



Not knowing the dispersion of the prism before the observing of 

 September 7, 8, 9, we chose places which afterward were found to 

 have the following wave lengths : 



Table i. — Places observed in stellar spectra 



Prismatic deviation from 



the D lines — 13'5 —io'.8 — 8^4 — 514 —i'.8 +i!8 +7'.! +9!5 



Wave lengths, microns. . 0.423 0.448 0.471 0.505 0.559 0.622 0.750 0.817 



Exposure range, microns 0.0140 0.0154 0.0172 0.0206 0.0248 0.0324 0.0416 0.0444 



The radiometer vanes, as stated in Publication 3843, were each 

 0.20 millimeter high and 0.44 millimeter wide. The spectrum, as 

 it fell upon them, extended vertically, and was intercepted by the 

 dimension 0.20 millimeter. With the spectroscope as about to be 

 described, this corresponded to an exposure of 2^07 in the spectrum, 

 and the exposures range in wave length as given in line 3, table i 

 above. Thus within the wave-length interval observed in 1947 the 

 range of dispersion was about threefold. 



With a prism of such small total dispersion it was important to 

 avoid stray light scattered from one region of spectrum to another. 

 This we accomplished by greatly lengthening the travel of the beam 

 after its dispersion by the prism, before focusing it by the image- 

 forming lens. The graph, figure 2, shows the optical path schemati- 

 cally. It will be perceived that the range of spectrum which could 

 fall within the radiometer is limited by the angle subtended by the 

 lens h at a distance of 61 feet from the prism. This angle is 12'. At 



