CHEMICAL SCIENCE. 267 



carbon. It contracts no definite combinations with the bases nor with 

 sea-salt ; it remains unaltered when placed in contact with potassium. 

 It reduces (with the greatest difficulty) copper solutions; it is, how- 

 ever, easily operated on by azotic acid, with which it rjives the same 

 products as veritable sugars. It is not in the least influenced by 

 beer-yeast ; it exerts no influence on polarised light. In a word, phy- 

 cite belongs to that class of singular bodies, which, while presenting 

 the sweet flavor, lack altogether the characters essential to sugars, in 

 the scientific acceptation of that word. 



OX THE DETERMINATION OF AMMONIA CONTAINED IN WATERS. 



Since the discovery by Braudcs and Liebinj, of the occurrence of 

 ammonia in rain water, no complete investigations of the quantity of 

 ammonia in natural waters has yet been made. Bousingault has now 

 becjun to determine the ammonia in such waters by means of a distil- 

 latory apparatus. He regards it as certain, that water charged 

 with a small quantity of ammonia, will have given off the whole of this 

 with the watery vapor when two-fifths of the water have distilled 

 over. 



We may, consequently, by submitting large quantities of water, as 

 10 litres or more, to a preliminary distillation, obtain a concentrated 

 fluid, so as to treat this in the still set apart for the determination of 

 the ammonia. Where the water is not too poor in ammonia, it may 

 be placed in the apparatus itself. 



The apparatus consists of a retort capable of containing f litres. A 

 litre of the water to be examined is put into this, if necessary previ- 

 ously concentrated. The cooling of the distillate is effected by means 

 of a" glass worm tube. One-fifth is distilled off, and the quantity of 

 ammonia contained in it ascertained, according to Peligot's method of 

 determination of nitrogen by means of a solution of sulphuric acid of 

 known strength. A second fifth is then distilled. In this there is 

 generally no more ammonia, and the quantity found in the first por- 

 tion is usually correct. 



The normal solution of sulphuric acid for the determination of the 

 ammonia is so composed that 5 cub. centims. of this fluid shall be sat- 

 urated by 0-0106 of ammonia. As the alkaline fluid which serves for 

 the testing of the solution is so for diluted that 33 cub. centims. of it 

 saturate the 5 cub. centims. of dilute acid, 1 cub. centim. of the alka- 

 line fluid represents 0-00033 of ammonia; and as the burette is divid- 

 ed into tenths of a cubic centimetre one division represents 0-000032 

 of ammonia. 



As regards the accuracy of the determination, it is necessary in this 

 process to take care that the normal acid is first added to the fluid to 

 be tested for ammonia, and that the alkaline fluid intended for the sat- 

 uration of the acid is then poured in. Any inaccuracy must therefore 

 arise from an error of two divisions of the burette. The determination 

 of the ammonia can therefore only be brought within 0-06 milligrm. ; 

 but as the operation can always be performed twice, even if the errors 



24 



