120 Prof. F. L. 0. Wadswortli on the 



The same conclusion, however, does not hold when the 

 method o£ mounting the grating is changed. In 1895 the 

 writer suggested and investigated several new methods of 

 using the concave grating as a direct ohjective si)ectroscope, 

 with a parallel incident beam of light. At that time a 

 number of mountings were designed and constructed for the 

 purpose of testing the proposed new methods, and the results 

 obtained were very satisfactory^. The grating used in these 

 tests was, however, a comparatively small one of only about 

 3 cm. aperture, and 152 cm. radius of curvature. Used in 

 the ordinary way (Rowland mounting) its semi-angular aper- 

 ture measured at the centre of curvature ^vould therefore be 

 about yS = '01. The effect of aberration on gratings of larger 

 angular aperture than this was computed, and it was con- 

 cluded at that time f that semi-angular apertures as large as 

 ^ = ■05 could be used if the resolving-power did not exceed 

 30,000 units. Since this paper was written the concave 

 grating has been used in the way described by Poor and 

 Mitchell at Johns Hopkins, and later by Mitchell at Yerkes 

 Observatory, in the study of stellar spectra ; and by Frost, 

 Mohler and Daniel, Jewell and Humphreys, and the writer, 

 in photographing the flash spectrum during a solar eclipse. 

 The gratings used by Poor and Mitchell X were 100 cm. radius 

 of curvature and about 5 cm. and 14*6 cm. aperture, ruled 

 respectively with 5905 and 2842 lines per cm. The resolving- 

 powers in the first order spectra were therefore about 30,000 

 and 41 ,500, and the semi-angular apertures about /3= "025 and 

 /5 = -073 respectively. The results obtained on stellar spectra 

 were considered very good. 



The results obtained by eclipse observers have all been 

 much less satisfactory. Frost § used a concave grating of 

 4*4 cm. aperture and 150 cm. radius of curvature, ruled with 

 5684 lines per cm. The resolving-power in the first spectrum 

 was therefore about 25,000, and the semi-angular aperture 

 about /3=*015. The field obtained was small, but this is 

 ascribed by Prof. Frost to be in part due to the fact that flat 

 plates were used instead of curved plates. The grating used 

 by Mohler and Daniel || had an aperture of 10 cm., a radius 

 of curvature of 305 cm. and a ruling of 5684 lines per cm. 

 In the first order the resolving-power was about 57,000 and 

 the semi-angular aperture, /3, was about 0016. The definition 



* Astrophysical Journal, vol. iii. pp. 54-60, Jan. 1896. 



t Ibid. pp. 60-61. 



X Ibid. vii. p. ]57, March 1898; and vol. x. p. 29, June 1899. 



§ Ibid. xii. p. 85 ; p. S07; December 1900. 



II Ibid. vol. xii. p. 361. 



