158 PROCEEDINGS OP THE AMERICAN ACADEMY. 



measured quantity was added about oue tenth of the amount of con- 

 centrated nitric acid, and the whole evaporated to dryness over a free 

 flame. As the mass nears dryness, the flame is lowered. More acid 

 may be added if necessary, care being taken to have an excess, in order 

 to avoid carbonization of the mass at the end. Deflagration often 

 ensues, but loss of arsenic is not to be feared in presence of an excess 

 of nitric acid. The organic matter is nearly all destroyed by this 

 treatment, but, to eliminate all, the residue is transferred to a smaller 

 dish, a little more nitric acid and strong sulphuric acid are added, and 

 the whole heated until a clear white melt is obtained, which fumes 

 strongly. After cooling, water is added, and the contents of the dish 

 added to the reduction flask of the Marsh apparatus. The detection 

 and determination of the arsenic are conducted by a modification of 

 the Berzelius-Marsh method, published by me two years ago.* It 

 is only by such an optometric process that the extremely small 

 amounts of arsenic can be quantitatively estimated, the principle of 

 the modification being the comparison of the mirrors with those 

 obtained from definite amounts of a standard solution of arsenious 

 oxide. 



Thus the reagents used are sulphuric acid, zinc, distilled water, 

 and nitric acid, all of which have been repeatedly tested in larger 

 quantities than would be used in one analysis. The utensils : 

 evaporating dishes of Berlin porcelain, stirring rods, funnels, beakers, 

 and measuring cylinders have been used for this purpose alone. The 

 filter paper has been tested in quantity. Fearing that the glaze of 

 the dishes might contain arsenic, and that the glaze would be dissolved 

 by the acid sulphates, I kept a quantity of melted acid potassic 

 sulphate in a dish for some time, stirring also with the glass rod used. 

 No arsenic was found in the solution, showing that there was no dan- 

 ger from this source. Finally, blank trials with non-arsenical urine, 

 conducted exactly like the others, confirmed the purity of the reagents 

 and the cleanliness of the utensils. In collecting samples of urine, 

 care has been taken to guard against accidental introduction of arsenic 

 from unclean bottles. 



The reduction tube of the apparatus is drawn out so as to give 

 two heating places. Before introducing the solution to be tested, 

 the apparatus is run for one hour with the lamp under the first 

 heating place, and the absence of arsenic in the apparatus fully 

 determined. One half, or any aliquot part, of the solution is then 



* These Proceedings, XXVI 24. 



