130 Prof. Arthur Smithells on the 



Flame-Spectrum of Gold Trichloride. 



Owing to the easy decomposability of gold trichloride and 

 the desirability of ascertaining more accurately the source of 

 the spectrum commonly attributed to it, I have examined it 

 with the separator. 



When gold trichloride is introduced on the end of a pla- 

 tinum wire into a Bunsen-flame, a bright green flash of light 

 is produced accompanied by some bright sparks. The 

 coloration only lasts a very short time, and a considerable 

 residue of metallic gold is left on the wire. These facts are 

 noted and a map of the spectrum is given by Lecoq de Bois- 

 baudran (Spectres Lumineux). 



A moderately dilute solution of gold trichloride, introduced 

 into the separated cones by means of the apparatus used for 

 copper salts (p. 123), does not colour either of them ; but 

 when a piece of asbestos moistened with strong hydrochloric 

 acid is held in the upper cone, a brilliant green colour is pro- 

 duced and the green lines are seen in the spectroscope. When 

 the coal-gas is previously passed through the saturator con- 

 taining asbestos moistened with strong hydrochloric acid, the 

 outer cone becomes slightly green, especially at the edge. If, 

 however, the saturator is jacketed with steam, so as to increase 

 the volatilization of hydrochloric acid, the upper cone becomes 

 intensely green. The same effect is obtained by using chloro- 

 form in the saturator. 



If when, in either of the above ways, the upper cone has 

 acquired a bright green colour, the outer tube of the separator 

 is slid down over the inner one, the colour diminishes in 

 intensity, and when ultimately the orifices of the two tubes 

 are level, it disappears almost entirely except at the extreme 

 edge. The only alteration which the outer cone can suffer by 

 this approximation to the inner one is that its average tempe- 

 rature must increase, whence it appears that the coloration by 

 the gold salt depends upon two circumstances — abundance 

 of hydrochloric acid and a low average temperature. 



This conclusion is easily confirmed. If, instead of bringing 

 the cones nearer, the upper one be surrounded by oxygen 

 instead of air, it becomes smaller and of higher average 

 temperature, and its green colour disappears except at the 

 extreme edges. 



The fact that a chloride of gold can under any circumstances 

 produce its individual spectrum is at first thought remark- 

 able when we consider the extreme ease with which the 

 salt is decomposed, and though the lower chloride or chlorides 

 might be formed, they are likewise easily decomposed, leaving 



