BUNSEN ON THE CACODYL SERIES: 309 
The correctness of this assumed atomic constitution, as far as 
respects the carbon, hydrogen, arsenic, mercury, and chlorine, 
is placed beyond a doubt, by the correspondence between the 
numbers found experimentally and those calculated. There is 
not, however, the same certainty respecting the oxygen. We 
can assume 1, 1}, and even 2 atoms, without passing the limits 
of experimental errors, which arises from the imperfect analytical 
means at our disposal. I do not therefore adopt the above for- 
mula in the following remarks as the only one which I conceive 
may be true, but because it most easily explains all the varied 
phzenomena attending the decompositions. Some results which 
I have obtained in experimenting on the higher oxides of caco- 
dyl, render it probable that this compound contains 1} atom of 
oxygen. Until I have finished the examination of the com- 
pound, which appears to afford the key to the solution of this 
question, I must hesitate in adopting this latter opinion, which 
presupposes the most extraordinary phenomena in the decom- 
positions of organic chemistry. I do not however bind myself to 
the first supposition, although it is the ground-work of the fol- 
lowing remarks, but expect to return to this subject in discuss- 
ing the higher degrees of combination of cacodyl. 
If we assume the empirical formula of this substance to be 
C,H, As, O Hg, Cl,, the rational composition may be expressed 
‘in two ways. We can view it as a compound of a higher chlo- 
ride of cacody] with suboxide of mercury, viz. Kd Cl, + Hg, O, 
or as a compound of oxide of cacodyl and sublimate, KdO + 
2HgCl. The following are the grounds which incline me to 
adopt this latter view. 
If we mix a weak solution of this substance with less hydrate 
of potash than is necessary to precipitate the whole of the sub- 
limate, yellow oxide of mercury is separated, which is converted 
in a few seconds into subchloride of mercury, while it acts as 
an oxidizing agent on the remaining oxide of cacodyl. This 
subchloride is afterwards decomposed, on adding a larger quan- 
tity of hydrate of potash, into suboxide of mercury, which again 
oxidizes the free oxide of cacodyl, and undergoes itself a partial 
reduction. Hydrate of potash therefore first produces a chloride 
of mercury, and the black precipitate which is sometimes ob- 
served must now be regarded as the result of a later decompo- 
sition. 
When hydriodic acid is poured over this compound, red iodide 
