164 W. M. DOHERTY. 



in the immediate vicinity, especially those with a "rose," 

 became coloured more or less intensely green, and if the 

 quantity of chloroform was large, electric blue. This 

 coloration was due to the volatile chloride of copper formed 

 from the metal of the burner itself, and the chlorine, one 

 of the products of the decomposition of the chloroform, or 

 tri-chloromethane. I found that if the "rose" of the 

 burner was cleaned so as to expose the metal, and if the 

 flame was carefully adjusted to a certain point, it was 

 possible to detect extremely small quantities of chloroform 

 in relatively large quantities of water by simply putting 

 the mouth of the bottle containing the liquid in juxtaposition 

 with the air orifice of the burner. 



An improvement of the test is in the use of a type of 

 spectrum burner, designed for the detection of boric acid, 

 which answers admirably, provided a piece of clean copper 

 wire (not too thin or it will melt) is placed in such a position 

 on the burner as to be outside the inner cone of the flame. 

 If a solution containing one drop of chloroform to 100 cubic 

 centimetres of water, after shaking well, be placed in the 

 glass arm attachment of the burner a distinctly green flame 

 is produced. Any volatile halogen compound will, how- 

 ever, give the colour, for example hydrochloric acid, 

 though it is not nearly as active as chloroform. Applied 

 as a test for chloral hydrate it was found most satis- 

 factory. If a solution of this substance be placed in the 

 glass portion of the spectrum burner previously men- 

 tioned, there is no coloration of the flame. Immediately 

 on adding caustic alkali, the flame becomes coloured with 

 an intensity depending upon the amount of chloral hydrate 

 present, the effect being produced by the well known 

 reaction : — 



CCI3CHO + NaOH = OCI3H + NaCH0 2 



