course the negatives obtained with such an instrument are practically useless for the fine determination of wave-lengths, but they are capable of showing the qualitative variations and coincidences of cometary radiations.
The spectrum obtained, with one hour's exposure, by Chrétien and Senoque, showed the three characteristic bands 564, 518, and 472, and the cyanogen band at 389. Almost all the light of the comet was concentrated in the two bands 472 and 389, and Baume-Pluvinel remarks, parenthetically, that objectives intended for comet photography should, obviously, be especially corrected for this more refrangible region of the spectrum. The heads of the bands at 564 and 518 were bright visually, but were much fainter, relatively, on the photograph, although orthochromatic plates were used. It is most essential that this difference of photographic and visual intensities should be carefully considered when comparing the relative intensities of spectral lines.
The spectrum of Borelly's Comet of 1903 (iv.) was photographed by Deslandres[1] at Meudon, and found to be similar to that of Rordame's Comet obtained by Campbell in 1893.
Perrine[2] also found the same five bands in the spectrum of this comet as were recorded by Campbell in that of Rordame's Comet, but noticed that the band at 420 was much weaker relatively. Curtis also recorded the similarity with Rordame's, and noted a strong continuous spectrum.
In 1904, Pickering employed an objective-prism camera to photograph the spectrum of Brooks's Comet (1904, i.) which he found to be nearly continuous, with two slight condensations.
The success of the prismatic camera as an engine of research in the investigation of cometary spectra soon led to its more general application, so that on the appearance of Daniel's Comet of 1907 (iv.) several observers employed it to obtain photographic records of the comet's radiations. A great advantage of spectra thus obtained (in that they show monochromatic images of all the sufficiently bright parts of the
