52 SlXTfl RKPORT -I8;3f>. 



When, however, this test is employed, we must first assure 

 ourselves that no chlorides exist in the solution ; for chloride of 

 silver, which would be formed, is blackened by the sun's rays, 

 even though no organic matter be present. 

 Gases. For determining the quality and amount of the gases chemi- 



cally combined with a mineral water, Mr. Walcher* suggested 

 a modification in the common apparatus, witli a view of obviat- 

 ing the error likely to arise from a portion of the water being 

 driven over by the ebullition. 



In his experiments, the glass globe containing the water to 

 be boiled was connected, air-tight, to a little phial, from which 

 proceeded a sigmoid tube, passing under mercury, or into the 

 vessel containing the substance intended to absorb the gas. 



Let us suppose, for instance, that our object is to ascertain 

 the amount of nitrogen and oxygen which a water contains. In 

 that case we fill the phial with carbonic acid, and the graduated 

 tube with solution of potass. The air expelled by ebullition, 

 together with a portion of the water itself, entering the phial, 

 expels the air, which passing into the tube, is robbed of its car- 

 bonic acid by the potass. 



After the experiment is over, the air remaining in the phial 

 may easily be transferred into the jar, and the water which 

 came over uic-iy be passed back again into the glass globe, in 

 order that it may be treated like the rest. 



In this manner, perhaps, a somewhat greater degree of ac- 

 curacy may be attained, than where a glass globe with a sigmoid 

 tube alone is employed. 



But I conceive that the utility of Mr. Walcker's plan will be 

 chiefly felt where the object is to ascertain the amount of sul- 

 phuretted hydrogen, or of carbonic acid in a mineral water, by 

 boiling it, and passing the gases over into a solution calculated 

 to absorb them. 



In such cases, if any portion of the water comes over with 

 the gas, the result is entirely vitiated ; and to prevent this, 

 there seems to be a convenience in the intervening bottle, which, 

 however, where sulphuretted hydrogen is expected, should be 

 filled with some gas not containing oxygen. 

 Suiphuret- After all, however, the simplest mode of ascertaining the 

 tea hydro- amount of sulphuretted hydrogen is by adding directly to the 

 ^^"" water some reagent, which precipitates it in a state of combi- 



nation. 



Mr. Richard Phillips, in his analysis of a spring near Wey- 

 moutht, has employed the nitrate of silver, which appears to be 



• Brwode's Journal of Science for 1828. 

 t Phil. May., vol. iii. p. 158. 



