296 Kreider—Quantitative Determination of Perchlorates. 
without any fear of liberating more iodine. An excess of 
decinormal arsenic is then run into the beaker and titrated back 
with iodine. | 
The many little precautions essential to note for the manipu- 
lation are in practice accomplished in a few moments. Seven 
determinations (not counting one which was all but completed, 
when an accident terminated it) from the weighing of the per- 
chlorate to the titration, were completed in one day; and the 
results recorded in Table VII show with what reliability. In 
making the series of experiments recorded in Table VII, it was 
found expeditious to have a partial vacuum always accessible 
instead of waiting each time for the exhaustion. This was 
obtained by connecting a vacuum flask with a two-holed stop- 
per to an ordinary water pump, and having the other perfora- 
tion fitted with a glass stop-cock. The bulb was merely 
attached to the vacuum by a piece of rubber tubing; the stop- 
cock opened and closed immediately, by which means a sufii- 
cient exhaustion was secured. To have the vacuum always in 
readiness a valve, described in a former article of mine,* was 
placed in the rubber leading to the pump, and when lubricated 
with glycerine would hold the vacuum perfectly. The nitric 
oxide employed was supplied by a Kipp generator, in which 
giobules of copper were acted upon by nitric acid mixed with 
an equal volume of water. To purify the gas evolved from 
any possible trace of the higher oxides, it was first passed 
through an acidified solution of potassium iodide in Geissler 
absorption bulbs, the latter one of the three being alkaline. 
This method of generating nitric oxide in a Kipp generator 
(preferably charged with dilute acid and kept warm by immer- 
sion in hot water when large amounts of the gas are to be 
drawn at frequent intervals) was devised by Professor Gooch, 
by whom it has been employed for some time. It is automatic 
and eminently satisfactory. The hydriodic acid was obtained 
from a solution of potassium iodide containing one gram in ten 
cubic centimeters ; thirty cubic centimeters being taken for 
each experiment, and acidified with the required amount of 
hydrochloric acid immediately before using, so as to prevent 
any liberation of iodine by the oxygen of the air. In those 
experiments in which more than this amount of potassium 
iodide was employed a correspondingly stronger solution of the 
latter was used, so that the volume of water was in all cases 
thirty cubic centimeters. 
* This Journal, 1, p. 132. 
