OP ARTS AND SCIENCES. 205 



bromide was formed as low as 130° in a sealed tube, or at 180° in 

 open vessels, while at its boiling-point the nitrotoluol was almost 

 entirely decomposed by bromine. 



In the work described in this paper, we have tried to give a more 

 definite statement to Keilstein's Law, — which may be stated thus: 

 When toluol (or its derivatives) is treated with chlorine or bromine in 

 the cold, the halogen attaches itself to the benzole ring ; if hot. to the 

 side-chain ; in presence of iodine exclusively to the ring, even when 

 boiling. To do this, we have tried to determine the amount of benzyl- 

 bromide formed from toluol and some of its derivatives, in the first 

 place at different temperatures, and secondly in presence of different 

 amounts of iodine. 



Effect of Differences in Temperature. 



In these experiments a weighed amount of toluol or the substituted 

 toluol was heated in a flask with a return-cooler by a parafline bath, 

 and the calculated quantity of bromine added with a drop-funnel at 

 such a rate that no bromine could be seen iu the fumes which escaped 

 through the cooler. The temperature of the hath and of the vapor 

 within the flask was noted throughout the experiment. To determine 

 how much benzylbromide had been formed, we used the following 

 method, which is very similar to that used by Cannizzaro* for the same 

 purpose. The product of the reaction, after being freed from hydro- 

 bromic acid by washing first with a dilute solution of sodic carbonate 

 and afterwards with distilled water, was heated to 100° with alcoholic 

 ammonia for two hours ; the bromides of ammonium and the three 

 benzylamines thus 'formed were decomposed by warming with a solu- 

 tion of baric hydrate, and the amount of baric bromide determined by 

 Volhardt's f sulphocyanate method of titration after the organic mat- 

 ter had been filtered out. To test the method of estimating tho 

 amount of removable bromine, pure parabrombenzylbromide was 

 treated with alcoholic ammonia and baric hydrate, as described above, 

 and 98 per cent of the calculated amount of bromine obtained. The 

 experiments collected in Table I. were then made. 



* Ann. Chem. Tharm., cxli. 198. t Ibid., cxc. 1. 



