o f Edinburgh , Session 1884-85. 
131 
subsequently shaking ether with the strongest hydrochloric acid and 
then adding an ethereal solution of the colouring matter, a bluish 
aqueous solution was formed, and a brownish-yellow solution in 
ether. The general reactions indeed were those of blue chlorophyll, 
which has been described by Sorby as existing in the leaves of land 
plants and in certain sea-weeds. The yellow substance was one of 
the colouring matters which result from the alteration of chlorophyll. 
On heating the alcoholic solution with zinc dust in a sealed tube at 
100° C., the green colour became very much brighter, and the 
yellow being removed the tint was much purer. After four hours 
the liquid was partially decolorised. The spectrum of the solu- 
tion was carefully examined, the details of the examination here 
follow. 
Spectrum of a Green Colouring Matter derived from Ocean 
Shells . — The shells were preserved in alcohol, which dissolved the 
colouring matter. A bottle was received from Mr J. Y. Buchanan in 
January 1884, which bore the following label : — 
“S.S. ‘Dacia,’ 17th JSTov. 1883. 
Lat. 33°-42' N. Long. H°-7' W. 
Depth, 533 fathoms. Coral Patch,” 
The original liquid, which measured about three ounces, was 
examined just as it was poured from the bottle, and again after 
undergoing concentration by evaporation. It was of a pale green 
colour and slightly yellowish, but after being concentrated it became 
a dull olive-green. 
The spectrum was observed with a small spectroscope of the 
usual form, fitted with a compound prism capable of dividing the 
D lines ; the measurements were made upon a divided arc ; a pointer 
shaped like a broad finger was placed in the eye-piece. One side of 
the finger was perfectly straight and vertical, the other was curved 
and brought to a fine point like a pen nib ; when this was moved 
over an absorption band, a reading was taken as soon as the fine 
point ceased to be visible. The arc measurements were reduced to 
wave-lengths and oscillation frequencies by means of two interpola- 
tion curves obtained from those flame and spark spectra recommended 
as affording good lines of reference in the Eeport on Spectrum 
Analysis, in the volume published by the British Association for the 
year 1881 : — 
