284 BUNSEN ON THE CACODYL SERIES. 
periment than is absolutely necessary. It is also of great conse- 
quence, in order to avoid the most dangerous explosions, to fill 
the apparatus which is intended to hold these substances most 
carefully with carbonic acid. The smell which most of these 
substances diffuse is frightful. It immediately produces vomiting 
in delicate persons, and appears especially to affect the nerves 
when long exposed to its action. The only cases of rapid poison- 
ing take place from cyanide of cacodyl, which produces giddiness, 
stupefaction, and even fainting, when diffused in the smallest 
quantity in the atmosphere. I am myself however convinced, 
by a long and continued study of these compounds, that how- 
ever tedious it may be, in consequence of the numerous precau- 
tions necessary, every danger may by a little care be avoided. 
I. THE LOWER DEGREES OF COMBINATION OF CACODYL. 
A. Amphide Compounds of Cacedyl. 
1. Oxide of Cacodyl. 
The great difficulties in the way of the direct estimation of 
the oxygen of this substance, which I formerly designated by 
the empirical name alcarsin, induced me to believe at first that 
it contained none, and this supposition received considerable 
countenance from its remarkable property of spontaneous com- 
bustion and its reactions with potassium. The peculiar mode of 
its formation, from the acetic acid salts with alkaline bases, 
does not however favour this supposition, as it presupposes a de- 
composition equally complicated and unusual. Berzelius con- 
cluded that it most probably contained oxygen, which is easily 
explained by abstracting two atoms of carbonic acid from two 
atoms of acetone (hydrated oxide of cenyl) and one atom of ar- 
senious acid, which would leave one atom of alcarsin behind, as 
in the following equation :— 
gat. Acetone. . . . - Cy GH, O, 
1... Arseniousacid . . As, =O) 
3 Ao Catbonic acid...) aC, O, 
1 ... Oxide of cacodyl . C, Hg As, O 
In order to determine this question experimentally, I have 
instituted a new series of most careful researches on this sub- 
stance, and as the results of the same have meanwhile been 
