ASTRONOMY — HALE. 63 



first. It has been easy to change the exposure time by the use of a 

 cone pulley on the motor and by the provision of a worm gear for 

 very slow speeds. The convenience of experimentation thus afforded 

 has rendered possible a long series of tests involving considerable vari- 

 ations in slit- width, speed, and plate sensitiveness. These experi- 

 ments will be completed with the 5-foot spectroheliograph, which 

 offers many additional advantages. It may be said, however, that 

 the most satisfactory results (for H 2 ) have been obtained with Process 

 plates, high speeds (from 0.2 to 0.3 inch per second), and wide slits. 

 It must not be forgotten, as Newall has pointed out, that very wide 

 slits can not be used without producing an astigmatic effect in the 

 direction of dispersion, unless an even number of reflections (or no 

 reflections) occur in the optical train. The 5-foot spectroheliograph, 

 when used with two prisms, is designed to give either one or two 

 reflections, as desired. 



In my discussion of the results obtained with the Rumford spec- 

 troheliograph, I adopted a working hypothesis which assumes that 

 photographs taken with the second slit set on the broad dark band H, 

 represent the low-lying, dense calcium vapor in the flocculi, while 

 those taken with H, show the less dense calcium vapor at a higher 

 level. Mr. Evershed has advanced good arguments in favor of his 

 view that the H. 1 photographs represent the faculae proper, the 

 increased contrast being due to the dark background afforded by the 

 dark band. This assumes that the level of the faculae is above that 

 of the denser calcium vapor, but photographs of the spectrum of the 

 faculse taken on Mount Wilson show that the bright band represent- 

 ing their continuous spectrum is greatly reduced in intensity near 

 the middle of H, and K,. The strength of this counter-argument, 

 however, is lessened by the fact that this same evidence conclusively 

 shows the absence of such reversals of H; and K, as my working 

 hypothesis would seem to require. It is evident that further advance 

 must depend in large measure upon a series of careful comparisons of 

 faculae near the sun's limb (taken direct or with a spectroheliograph 

 having its second slit set on the continuous spectrum) and of the 

 same objects photographed nearly simultaneously with a spectro- 

 heliograph set on H,. 



A Zeiss stereocomparator (not available in the work at the Yerkes 

 Observatory) has afforded the means of comparing photographs with 

 the necessary precision. Numerous plates of the recent large spot 

 groups have supplied excellent material for the tests. Though the 

 comparisons are not yet completed, it may be said that they now 

 seem more favorable to Mr. Evershed' s view than to my own. In 



