( 203 ) 
COLLOIDAL GOLD AND " PUEPLE OF CASSIUS." 
By James Moir, M.A., D.Sc, F.C.S., F.B.S.S.Af. 
(Read March 15, 1911.) 
A number of experiments have been done to test the generally-accepted 
view that "purple of Cassius " is a solid-solution of colloidal gold in 
hydrated stannic oxide. When once formed, the purple resists cold pure 
uqua regia, in this respect differing from pure finely-divided gold. Nearly 
all the experiments were made with a very dilute solution of HAuCl^ in 
distilled water, made so as to contain 5 parts gold in 1,000,000 of water. 
Acid solution of SnCls alone produces a pale yellowish-brown colora- 
tion of remarkable stability, the relationship of which to purple of Cassius, 
from which it is obviously different, is obscure. This yellow coloration is 
unaffected by caustic potash even when used in quantity sufficient to 
redissolve the tin precipitate first formed ; nor is it affected by boiling or 
the addition of neutral salts or of hydrazine, which would be expected to 
coagulate it if it were merely due to extreme fineness of division of precipi- 
tated gold. It even resists chlorine water and nitric acid (when not in 
excess), but strong HCl alters it to brown and then to a pink shade, and aqua 
regia slowly develops the purple of Cassius shade. Curiously enough, in this 
process the absorption-spectrum scarcely alters, beyond a slight increase 
in the amount of violet transmitted. When stronger solutions are 
employed the brown solution obtained by using SnCla alone has an olive- 
green pseudo-fluorescence in reflected light, whereas purple of Cassius 
(use of SnCl^ -|- SnCls) has an orange-red reflection, probably due to 
metallic gold. When the gold solution is made strongly acid before 
adding SnCla, the shade produced is pink, i.e., metallic gold is liberated. 
In order to obtain purple of Cassius from HAuCl^ and SnClz, the 
presence of an oxidant with loosely-bound oxygen is necessary. Dilute 
HNO3 (free from HNO2) has no effect, whereas even a trace of aqua regia 
which has been heated so as to form HNO2 and chlorine acts at once. 
Chlorine water and euchlorine behave similarly. Aqua regia from pure 
acids, and not heated, has remarkably little action. Addition of a trace 
of sodium nitrite before the SnCla gives an intense pink shade of the 
purple of Cassius. An excess of nitrite, however, prevents the formation 
14 
