Gooch and Ensign — Determination of Bromine. 151 



not less than 600 cm 3 — the amount lying midway between the 

 proportions with which divergences begin to be noted. 



Other test determinations were made along the same lines, 

 excepting that, instead of generating the nitrous fumes outside 

 the licpjid under treatment, pure sodium nitrite was introduced 

 directly into the solution. The nitrite was prepared free from 

 halogens by adding to its solution a little silver nitrate, acidu- 

 lating distinctly with nitric acid, and filtering off the little 

 silver chloride thus precipitated, together with a small amount 

 of silver nitrite thrown down at the same time. The strength 

 of the solution thus prepared was determined by acting upon 

 a definite portion with ferrous sulphate in excess and titrating 

 the residual ferrous salt with potassium permanganate. The 

 details of these experiments are given in Table XIII. 



Table XIII. 







■c 



rH 





3 



pH 





6s 



.4 





tt 



o 





*% 



w 



^ 



££ 



ffi 



W 



• r 



cm 3 



grm. 



grm. 



3 



0-5 



0-35 



3 



0-5 



0-35 



3 



0-5 



0-35 



3 



05 



0-35 



3 



0-5 



0-35 



3 



05 



1-75 



3 



0-5 



1-75 



KBr=HBr 

 taken. 



grm. , grm. 



0-55080-3745 



0-55130-3747 

 0-5513,0-3747 

 0-3005 0-2042 

 0-2759 0-1875 

 0-5513:0-3747 

 0-5510 0-3746 



3 a 



S 3 



650 



650 

 650 

 C50 

 650 

 650 

 650 



-3 

 <& 3 



I! 



cm" 



never 



below 



500 cm* 





About 

 30 



AgBr= 



= HBr 



3 



found. 



t. 





o C 





Zm 





Hffi 



grm. 



grm. 



grm. 



0-8689 0-3744 0001 — 



08694 0-3746 0001 — 

 0-8699 0-37481 0-0001 + 

 0-4746 0-2045, 0-0003 + 

 04358 0-1878 0-0003 + 

 0-8705 0-3750 0-0003 + 

 0-8707 0-3751 ; 0-0005 + 



The mean error of these seven determinations is 0'0002 

 grm. + , lying between the extremes - 0005 grm. + and 0-0001 

 grm. — . In the first five of these determinations enough 

 nitrite was employed to break up one and a half times the 

 amount of iodide taken, if the action is supposed to go to the 

 point of setting NO free. In the last two experiments, eight 

 times the quantity of the nitrite theoretically thus called for was 

 taken with no apparent change in the effect. 



That iodine may be removed with reasonable accuracy from 

 mixtures of iodides and bromides without disturbing the bro- 

 mine is evidently established ; and, inasmuch as the proportion 

 in which the reagents are taken in the corresponding process 

 for the separation of chlorine from iodine lie far within the 

 limits found applicable to the bromine separation, it would be 

 natural to suppose that in the presence of a chloride associated 

 with the bromide the sum of the hydrobromic and hydrochloric 

 acids would be given with exactness under the conditions suit- 



