The colouring-matter of cochineal, to which alum, potassic carbonate, and tartrate have been added, gives a spectrum very similar to that of blood (see “Foods,” p. 82); but this is only the case when the solution is fresh. The colour is at once discharged by chlorine, while the colour of blood, although changed in hue, remains. The colouring-matter of certain red feathers, purpurin-sulphuric acid, and a few other reds, have some similarity to either the hæmatin or the hæmoglobin spectrum, but the bands do not strictly coincide; besides, no one would trust to a single test, and none of the colouring-matters other than blood yield hæmatin. The blood in CO poisoning has also other characteristics. It is of a peculiar florid vermilion colour, a colour that is very persistent, lasting for days and even weeks. Normal blood mixed with 30 per cent. potash solution forms _greenish_ streaky clots, while blood charged with CO forms red streaky clots. Normal blood diluted to 50 times its volume of water, and then treated successively with yellow ammonium sulphide in the proportion of 2 to 25 c.c. of blood, followed by three drops of acetic acid, gives a grey colour, while CO blood remains bright red. CO blood shaken with 4 times its volume of lead acetate remains red, but normal blood becomes brown.[50] [50] M. Rubner, _Arch. Hyg._, x. 397. Solutions of platinum chloride or zinc chloride give a bright red colour with CO blood; normal blood is coloured brown or very dark brown. Phospho-molybdic acid or 5 per cent. phenol gives a carmine-coloured precipitate with CO blood, but a reddish-brown precipitate with normal blood (sensitive to 16 per cent.). A mixture of 2 c.c. of dilute acetic acid and 15 c.c. of 20 per cent. potassic ferrocyanide solution added to 10 c.c. of CO blood produces an intense bright red; normal blood becomes dark brown. Four parts of CO blood, diluted with 4 parts of water and shaken with 3 vols. of 1 per cent. tannin solution, become at first bright red with a bluish tinge, and remain so persistently. Normal blood, on the other hand, also strikes bright red at first, but with a yellowish tinge; at the end of 1 hour it becomes brownish, and finally in 24 hours grey. This is stated to be delicate enough to detect 0·0023 per cent. in air. If blood be diluted with 40 times its volume of water, and 5 drops of phenylhydrazin solution be added, CO blood strikes rose-red; normal blood grey-violet.[51] [51] A. Welzel, _Centr. med. Wiss._, xxvii. 732-734. Gustave Piotrowski[52] has experimented on the length of time blood retains CO. The blood of dogs poisoned by this agent was kept in flasks, and then the gas pumped out by means of a mercury pump on the following dates:-- [52] _Compt. Rend. Soc. de Biol._, v. 433.