from the Chlorides of Sodium and Potassium. 107 



chlorides are greatly in excess, the tendency to obey the laws 

 in accordance with which they crystallize asserts itself and we 

 have a more or less coarsely granular deposit. This is espe- 

 cially true in the case of sodium chloride. These facts have a 

 bearing on the question of completeness of separation. 



Consider the amounts of magnesia found in the alkalies. 

 They bear no relation to the amount of magnesium chlorides 

 involved. It seems probable that the magnesia there found 

 has been carried mechanically, and that it is not there as the 

 result of decomposition. How can we better explain the 

 finding of magnesia in the insoluble chlorides of experiment 

 (1) where but 9 mgrm. of magnesia were involved. Here it is 

 essentially the result of inclusion. 



The desirability of adding water to the amyl alcohol solution 

 of magnesium chloride before completing the evaporation is to 

 be emphasized. By so doing it is possible to expel the alco- 

 hol so completely that even those experiments where there was 

 the maximum amount of magnesia gave no trouble from car- 

 bonization. 



In adding sulphuric acid a great excess is to be avoided. In 

 the above experiments about twice the amount demanded by 

 theory was used. The result, in the end, is economy of time 

 and less danger of loss. It is better to add sulphuric acid a 

 second time and repeat the evaporation, ignition and weighing. 

 In the experiments involving an amount of magnesium equiva- 

 lent to 0*16 grm. of the oxide, the ignited sulphate was 

 found in a few instances to increase its weight after a second 

 treatment with sulphuric acid. The lesser amounts were 

 almost without exception wholly converted into the sulphate 

 by a single treatment. 



In the course of the work it was found that amyl alcohol 

 and concentrated hydrochloric acid mix in all proportions. 

 Furthermore if amyl alcohol containing but a small quantity 

 of the acid be boiled down to but a small fraction of its origi- 

 nal volume the residue contains chlorine in some form. 30 cc 

 of the alcohol, to which 0*2 CC of concentrated hydrochloric had 

 been added, having been reduced to 2 CC gave a decided test for 

 chlorine. This suggested the possibility that the use of benzyl 

 chloride was unnecessary. For, if amyl alcohol mixes with the 

 concentrated acid and on being evaporated retains more or less 

 of it to the last, the conditions which would prevent the de- 

 composition of magnesium chloride are maintained. The 

 easy solubility of the anhydrous magnesium chloride is beyond 

 question. Possibly then, in the above experiments, the addi- 

 tion of benzyl chloride was uncalled for as in every case 

 hydrochloric acid had been previously added to the salt solu- 

 tions. To decide the question a few separations were made 



