128 Prof. Arthur Smithells on the 



easily seen by holding a bead of cupric chloride well above 

 the tip of a small Bunsen-flame, and supporting above it a 

 porcelain basin filled with cold water. Though no flame is 

 seen the salt melts and volatilizes sufficiently to give a con- 

 siderable film on the basin, which, on examination, is seen to 

 be almost wholly cuprous chloride. 



Taking it as established that the blue part of the flame is 

 due to cuprous chloride, the question arises, How does this 

 salt become converted into oxide of copper to which the green 

 part is due ? This could be easily accounted for by the action 

 of steam upon the chloride. That this is the correct explana- 

 tion is established with something like certainty by experiments 

 with flames of'Combustibles containing no hvdrooen. In the 

 flame of carbon monoxide or carbon bisulphide, cupric chloride 

 produces almost exclusively the blue colour ; whilst in flames 

 of hydrogen in which, according to the hypothesis, the 

 existence of cuprous chloride should be precluded by the 

 presence of steam, the green colour greatly predominates. 

 Again, a small jet of hydrogen or steam impinging on a 

 carbon-monoxide flame coloured intensely blue by cuprous 

 chloride produces the green at the point of contact. 



The remaining point of interest is to ascertain which of the 

 oxides of copper produces the green colour. When carefully 

 purified CuO is dusted on to a Bunsen-flame, a green tint is at 

 once produced ; and the deposit obtained on porcelain from 

 the green part of a copper-chloride flame appears to be black 

 at the instant of deposition. These facts point to cupric oxide 

 as the substance which produces the green colour. On the 

 other hand, cupric oxide is a substance which loses oxygen at 

 a high temperature. The dissociation of cupric oxide was 

 studied by Debray and Joannis (Compt. Rend. xcix. p. 583, 

 1884). Heated in vacuo they found it to yield oxygen at 

 350° C. At the melting-point of silver the tension of this 

 oxygen amounted to 56 millim., and a little above the melting- 

 point of gold to 1000 millim. It w 7 ould appear to be im- 

 possible for cupric oxide to exist at the average temperature 

 of a Bunsen-flame, for this would demand an oxygen-tension 

 vastly greater than exists within or around such a flame. It 

 seems therefore necessary to ascribe the green glow to cuprous 

 oxide or some lower oxide of copper, but I have been unable 

 after several attempts to volatilize cuprous oxide by external 

 heating. 



It has already been stated that a cuprous-chloride flame 

 containing plenty of hydrochloric acid is surrounded by a dull 

 red margin. This appears to be due to cupric chloride. 

 When a super-aerated flame is obtained in the separator and 



