300 REPOKT— 1880. 



and even furtlier tlian this he has i^hotogTaphed the regions including 

 E, D, C, B, a, and A, together with ultra-red rays.' He used a ruled 

 speculum-plaue, and in certain cases a concave speculum-mirror. The 

 plate generally employed vras of glass ruled with 6,481 lines to the inch, 

 made by the beautiful machine constructed by Mr. L. M. Rutherford of 

 New York. The ruled surface is 1-yf o inch long and -j^'^y of an inch 

 wide. It appears to be unquestionably more perfect than any similar 

 grating made by Nobert and others. The grating being on glass gives a 

 bright transmitted spectrum which was generally used, the remaining 

 optical part of the apparatus being of glass achromatised according to the 

 plan of J. W. Draper. The slit was f*V in length and yj^ of an inch in 

 width. The jaws of the slit made of steel were provided with a micro- 

 meter screw for sepai-ating them, and another for placing them at an 

 angle so that occasionally photogTaphs were taken with the slit opened to 

 -^■ff inch at the top and only y^-y at the bottom, so aa to obtain a diflPerent 

 intensity at the two edges of the spectrum. Most of the photographs 

 were taken from spectra of the third order, which possesses the following 

 advantages : first, it is dilated to such an extent as to give a long image, 

 and yet is not one too faint to be copied by a reasonable exposure of the 

 photographic plate ; and secondly, the spectrum of the second order over- 

 laps it lin such a way that D falls nearly upon H, and T upon 0, and 

 these coincidences serve to determine the true wave-length of all the rays. 



In order to obtain a spectrum of uniform character for rays of all 

 refrangibilities, parts of the sensitised plate were protected by a series of 

 diaphragms during exposure for faint groups of rays ; by the removal of 

 these at intervals the strong rays were photographed with a distinctness 

 which could not otherwise have been attained. 



The region from Avave-length 400'0 to 435'0 only required about 

 1-lOththe exposure reqiiired by that from 344'0 to 351'0. In the photo- 

 graph published in ' Nature,' the line O had 15 minutes' and G 2^ 

 minutes' exposure to a wet bromo-iodised collodion plate, and still the 

 former is under-exposed. 



After the production of spectra which were in focus from end to end, 

 it was necessary to attach a scale to them by which wave-lengths might 

 be read. 



Plate X. is a copy of Dr. Draper's photograph. 



Using as a basis the numbers given by Angstrom for the rays, D,, &4 

 and G, the wave-lengths of the principal rays on Dr. He;iry Draper's pho- 

 tograph were calculated. Taking advantage of the fact that the second 

 spectrum overlaps the third, the ray D being near H of the third, and F 

 of the second being near O of the third, it is obvious that wdve-lengths of 

 three points, one at each end and one in the middle of the photograph, may 

 be readily ascertained. . As the rays D and b were too feeble to be easily 

 photographed, the following device was resorted to to indicate their 

 position, with regard to lines on the spectrum of the third order. In 

 front of the sensitised plate and close to it, were placed two very fine 

 steel points, one carefully adjusted to Dg of the second order, and the 

 other to F^ of the second order. On developing the picture after expos- 

 ing the plate to the ultra-violet spectrum of the third order, two sharply 

 defined images of the steel points were superposed on the spectrum. The 



' ' On DiflEraction Spectrum Photography, and the Determination of the Wave- 

 lengths of the Ultra-violet Eays, Natitre, 1874, p. 224 ; also American Journal of 

 Science and Art, Dec. 1873. 



