642 SIR ARTHUR H. CHURCH ON TURACIX. 



ninst be mentioned that this colouring-matter, when suddenly and 

 sti'ongly heated, yields a volatile, red, copper-containing derivative, 

 v^rhich, undissolved by weak ammonia-water, is not only soluble 

 in, but may be crystallized from, ether. It is the vapour of this 

 substance catching fire which gives rise to the green flame so con- 

 spicuous when a particle of turacin is heated in the air. If I 

 were discussing the nature and relationships of turacin from a 

 chemical view-point, much would have to be brought forward as 

 to its composition and probable formula. Here let it suffice to 

 say that analysis gives these percentage-numbers : — 



Carbon 53-69 



Hydrogen 4"60 



Copper 7'01 



Nitrogen 6'96 



Oxygen 27-74 



These figures, though deduced from many careful determina- 

 tions, do not lead unmistakably to one definite empirical formula. 

 With a colloidal compound like turacin, which does not admit of 

 purification by distillation or crystallization, there is always the 

 chance of a disturbing factor being present in the shape of a trace 

 of some impurity. Possibly such an expression as C^gHgylSr^^CuO^g 

 is worth suggesting. In this the ratio of nitrogen to metal is as 

 4:1, the same ratio which occurs in haematin from blood between 

 nitrogen and iron, and between nitrogen and magnesium in some 

 chlorophyll constituents and derivatives. With all these bodies 

 turacin is also brought into relation when its absorption -spectrum 

 is considered, especially in regard to that broad band in the violet 

 and ultra-violet (between the lines A and M) which is common to 

 all these colouring-matters, and which was figured and described 

 by the late Prof. Arthur Gamgee in a paper read before the Royal 

 Society in 1896 — a paper in which the author confirms my results, 

 especially those relating to the discove3y of turacoporphyrin. 



The spectrum of turacin may now engage our attention : in 

 the feather itself the pigmented web shows two broad bands, one 

 with its centre at wave-length at or near 585, the other with its 

 centre near 538. A solution in water containing a trace of 

 ammonia shows the same two bands somewhat shifted towards 

 the violet end of the spectrum ; there is also seen a very faint 

 band about the solar line F, but nearly as broad as band No. 2. 

 When turacin which has been precipitated from its alkaline 

 solution by strong hydrochloric acid is again dissolved in weak 

 ammonia-water, and the solution spectroscopically examined, a 

 narrow and faint fourth band situated on the less refrangible side 

 of D makes its appearance. It is almost certain that this band is 

 due to an alteration-product of the original turacin. The two 

 well-mai-ked bands, as they are shown when examining a feather 

 with the spectroscope, closely resemble those of the oxy-hfemo- 

 globin of the blood ; while the corresponding bands seen in the 



