CHEMISTRY OF INSECT COLOURS. 165 



own. I will be as brief as possible in stating the nature of my 

 new discovery, and the more so since, beyond recounting the one 

 or two startling facts, little can be said until a thorough investi- 

 gation shall have been made by means of an extended series of 

 experiments in this new direction. 



Briefly, then, the cyanide bottle, referred to above, after its 

 eight days' sojourn in the water-oven, had been placed on one 

 side for several months. At the beginning of November — in fact, 

 just after finishing the MS. of the foregoing pages — I thought it 

 as well, by way of special precaution, to give a final look to the 

 wings before throwing them away. The ediisa was brittle and 

 browner (dull gingerbread colour), as in the experiments made 

 on the gas-stove (supra) ; but, to my intense astonishment, I 

 found on the rhamni wing what looked by gaslight like red spots. 

 A subsequent examination by daylight failed to confirm this, but, 

 of course, the matter was far too important to rest here. Accord- 

 ingly, I started a fresh set of experiments by placing wings of 

 edusa and rhamni in (1) an ordinary (collector's) cyanide bottle, 

 and (2) in a bottle containing only a mass of potassic cyanide, on 

 to which I poured some water, so that the cyanide was, in places, 

 in a sloppy condition. (I should state that similar simple, not 

 shop-made, cyanide bottles were used in the previous experi- 

 ments.) On examining the cyanide bottles after sixty hours I 

 found that in the wet one the rhamni wings were largely reddened, 

 the red being, perhaps, most like a fresh blood stain, — a most 

 indubitable red certainly. Microscopic examination showed that 

 there was no possible delusion about the matter ; each individual 

 scale could be seen bright red, whilst the additionally interesting 

 fact was disclosed that the red shaded off into orange at its 

 border. Here, then, we have actually in progress, before our 

 very eyes, the (progressive) evolution of red from yellow via 

 orange. I may add that these metamorphosed wings were 

 removed to a dry glass slide, and have already retained their 

 colour for four weeks.* 



The edusa wing was less satisfactory ; it was chiefly a sort of 

 brownish, but I could not thoroughly decide whether or not there 

 was a purplish tint also. It was accordingly replaced in the 

 cyanide bottle, and after another forty-eight hours there could be 

 no longer doubt as to the red blotches on the wing, which, how- 

 ever, was to a great extent transparent. The effect was not by 

 any means so marked and striking as in rhamni, but I wish it to 

 be definitely understood that a red ivas produced. Seeing, how- 

 ever, that the rhamni effect was so much better (and remembering 

 also that Mr. Edwards' result was not obtained on edusa, but on 

 eurydice), it seemed to me likely that better effects might be got 

 with such a tint as that of hyale, with which accordingly I experi- 



♦ They are still red, June 24th, 1891. 



