12 
IOWA ACADEMY OF SCIENCES. 
of silver without filtering off the silver halides it seems 
best to escape the disadvantage of titrating a colored solu- 
tion by weighing the silver halide direct. In the following 
experiments the hydriodic acid was precipitated as silver 
iodide which was weighed. 
(1) 20 c.c. iodate solution gave .4726 grams Agl, cor- 
responding to .2554 grams iodine. 
(2) 20 c.c. iodate solution gave .4764 grams Agl, cor- 
responding to .2566 grams iodine. 
(3) 20 c.c. iodate solution gave .4764 grams Agl, cor- 
responding to .2574 grams iodine. 
The theory requires .2539 grams iodine for 20 c.c. of the 
solution. 
Determination of Bromic Acid . — The potassium bromate 
used in the following determinations was made by Merk. 
It contained no bromide. It was dried to constant weight 
at 100^. A deci-normal solution was made up and its 
strength was determined by digestion with hydrochloric 
acid and potassium iodide. 20 c.c. of this solution 
gave on digestion an amount of iodine which required 
120.08 c.c. of a standard solution of sodium thiosulphate, 
deci-normal, while the theory for 20 c.c. of a normal solu- 
tion of the bromate would require 120 c.c. of deci-normal 
thiosulphate. 
The reduction with iron was carried out as already 
described under iodic acid, but it was found better to keep 
the solution at room temperature. In this case the com- 
plete reduction required about two hours. Usually no 
color of bromine was visible above the bulb of the flask, 
and care was taken to prevent the escape of any bromine 
by the same means as in the case of iodine. The amount 
of iron dissolved made it even more difficult than in the 
case of hydriodic acid, to use the Volhard method. Three 
determinations by this method gave results decidedly too 
low; and in fact, in three experiments in each of which 20 
c.c. of bromate were used the volumes of silver used were 
39.11, 39.35 and 39.07 c.c. 
