32 



prove the existence of even small quantities of the poison, 

 when unmixed with other substances. 



The test of reduction, however, has, in more recent 

 times, been brought to a still greater degree of delicacy, 

 not to say of certainty. I allude to the test called Marsh's. 

 This test depends upon the property of arsenic to combine 

 with hydrogen, and form a gas which burns with a bluish 

 flame. The flame when brought in contact with a cold 

 surface, such as white porcelain or glass, causes a deposit 

 to take place, black, or chocolate brown in colour, readily 

 volatalized by heat, dissolving in cold nitric acid, which solu- 

 tion yields a residue when evaporated, dissolving in water, 

 and giving a brick-red coloured precipitate, with nitrate of 

 silver. 



Can any metal besides arsenic give the same results by 

 Marsh's test ? None, with which we are acquainted. Very 

 much has been said respecting the delicacy of Marsh's test, 

 and it appears to me, that much misconception has arisen 

 from the want of sufficient explicitness on this point. Thus, 

 it has been stated, that ,,ooo,ooo, of a grain of metal may 

 be rendered visible. It is by no means improbable, that so 

 small a quantity as that just referred to, may be rendered 

 visible on white porcelain, but it by no means follows, that 

 if 1,000,000 P art °f a grain of arsenic were to be placed 

 in Marsh's apparatus, its presence could be detected. 

 Again, in speaking of the extreme delicacy of Marsh's test, 

 experimenters have compared the quantity of arsenic placed 

 in the apparatus, with the quantity of fluid also present; in 

 such cases indeed, as might be expected, the relative dif- 

 ference in weight may be enormous. Thus Danger and 

 Flandin have asserted, that they have obtained metallic 

 films when the arsenic formed only the j^ooo P art °f 

 the fluid employed. Now, supposing this liquid to be 

 pure water, the arsenic would be in the proportion of about 

 1 grain to more than 28£ gallons. But another experi- 



