order of intensity:—5163⋅0, 5588⋅2, 5449⋅4, 4725⋅4, and 4687⋅8; this was on April 1 and 2, 1892. But Campbell examining the same object on April 5 saw[1] quite a different spectrum, consisting of the characteristic bands 5630, 5170⋅3, and 4723, with the relative intensities 1, 6, and 2. The wave-length of the middle band appeared to vary, and Campbell suggested that three bright lines formed at wave-lengths 5170, 5164, and 5157, and disappeared in that order. If this were a real change between April 1 and April 5, it is significant to note that the perihelion occurred on April 6, but Campbell's measures were not sufficiently delicate to detect any such change about that time. Von Gothard photographed the spectrum of this comet with the spectrum of a Bunsen flame on the same plate, and found the two to be identical as far as the fourth band (473–464)[2]. Beyond that, new unknown lines and bands appeared and the band at 389–387 was faintly represented. The new bands were similar in appearance to the hydro-carbon bands, and this led Von Gothard to the conclusion that the hydro-carbon in comets is different, or exists under different conditions, from that appearing in the Bunsen flame.
The next comet of importance to which reference must be made was Holmes's, discovered on November 6, 1892, and described generally elsewhere.[3] This was a remarkable body, a diffuse, nebulous, tailless mass easily confounded with the Andromeda nebula; and its spectrum was just as peculiar. Instead of the usual display of bands Keeler[4] saw simply a fairly bright, continuous spectrum extending from D to about half-way between F and G, the maximum brightness being a little below h. Careful scrutiny failed to reveal any lines, and the nucleus was seen as a brighter streak running through the whole spectrum. Observations made by Campbell[5] on November 8 and 9, 1892, showed a spectrum "of an extreme type and probably unique", a continuous spectrum, extending
