31 RESEARCH FOR ARSENIC AND ANTIMONY. 
the readers of these f Reports/ it is necessary to make a few 
observations upon them before proceeding to describe that 
method which, while it appears to me to combine the advan¬ 
tages of the two, is attended with less risk of fallacy than 
either process taken separately. 
Marsh’s Process. 
Marsh’s process consists in the conversion of the ordinary 
compounds of arsenic to arseniuretted hydrogen gas, in a 
properly constructed apparatus. In an early volume of Guy’s 
Hospital Reports (vol. ii, p. 70, 1837), I published an ac¬ 
count of Mr. Marsh’s discovery, with a drawing of the appa¬ 
ratus used and recommended bv him. Mr. Marsh confined 
his results to the combustion of the gas, the appearance of 
the metallic deposits, and the production of arsenious acid as 
one of the results of combustion. Since that date numerous 
forms of apparatus have been suggested; and the means of 
distinguishing arsenical from other deposits resembling them 
have been described in most works on chemistry and toxi¬ 
cology. 
It is conceded by all those who have experimented on the 
subject that, in the absence of disturbing causes, this is a 
most delicate process for the detection of arsenic. Mr. Marsh 
himself does not claim for it a higher degree of sensitiveness 
than that of enabling him to detect the 1- 120th part of a 
grain of arsenic (the quantity contained in one minim of 
Fowler’s solution); while, with regard to the effect of dilution, 
he found that one grain of arsenic in twenty-eight thousand 
grains (or four imperial pints) of water, furnished, when 
placed in a larger apparatus constructed for the purpose, up¬ 
wards of one hundred distinct metallic arsenical crusts.* 
MM. Danger and Flandin assert that metallic deposits may 
be procured when the arsenic forms only 1-200,000th part 
of the liquid examined ( e De l’Arsenic,’ p. 83); while M. 
Signoret states that he has procured metallic deposits with 
only the 1-200,000,000th part of arsenic in the liquid. 
This is equivalent to one grain of arsenic dissolved in three 
thousand gallons of water. We must regard these results 
either as greatly exaggerated, or depending upon some pecu¬ 
liar mode of computation. In general, either the degree of 
dilution with water or the supposed weight of the deposit 
derived from combustion, has been taken as the standard of 
delicacy; but the real question in practice is, what weight of 
* On the Separation of Arsenic, ‘ Transactions of the Society of Arts/ 
vol. li, p. 10. 
