294. Krewder— Quantitative Determination of Perchlorates. 
out the loss of a particle. if a perforated rubber stopper contain- 
ing a capillary tube, which reached just even with the narrower 
end, was by a slight twist forced tightly into the neck of the 
bottle. In this way a regular funnel shape was obtained and 
the oxygen could be withdrawn without the slightest bubble 
remaining. The other end of the capillary was fitted with a 
short piece of vacuum tubing and screw pinch-cock, which 
worked incomparably better than the glass stop-cocks. The 
larger capacity of the bottle was favorable for the volume of 
oxygen evolved and its shape offered superior facilities for the 
absorption of carbon dioxide. 
For the action of the oxygen on hydriodic acid through the 
medium of nitric oxide, various devices were tested. Passing 
it directly into nitric oxide over a solution of hydriodic acid in 
a Hempel absorption bulb was found to yield low and irregular 
results, due doubtless to the formation of nitric acid wherever 
the nitrogen tri-oxide, or per-oxide as the case might be, met 
water in which the hydriodic acid had been exhausted ; as, for 
instance, along the sides of the bulb. Shaking the bulb as the 
oxygen entered improved the action, but was not sufficient. It 
was evident that for a complete action, the hydriodic acid solu- 
tion must be strong and on the spot where the higher oxide of 
nitrogen is formed; and to avoid excessive use of the iodide 
the volume of water must be kept at a minimum. Letting a 
solution of hydriodicacid saturated with nitric oxide flow slowly 
into the Hempel bulb in which the oxygen was contained over 
water, was so slow in its action that a quantitative test was not 
applied. ‘The plan of mixing the two gases under astrong soln- 
tion of hydriodic acid by means of two capillaries with adjoin- 
ing openings, was more effective and rapid, but it was wasteful 
of nitric oxide, which for complete action would have to flow 
in continual excess, whereas only a small amount of nitric oxide 
would really be necessary for the reaction, since it could be 
used and reused for the transfer of free oxygen to the hydri- 
odic acid. A simple piece of apparatus was then devised to 
meet all these conditions. It consisted of a 100 em* bulb pipette, 
cut off short at either ends with stop-cocks sealed to both 
stubs. The delivery tube of one of the stop-cocks was cut off 
rather short after being tapered and restricted so as to hold a 
rubber connecter tightly, while the other delivery tube was 
left long enough to reach to the bottom of an Erlenmeyer 
beaker. It is. a convenience to have these conducting tubes 3 
or 4™™ in diameter rather than capillaries, since for the various 
connections all air may be expelled from them by displacement 
with water, which is easily accomplished by using a long noz- 
zled wash bottle. By attaching the shorter end to an ordinary 
water pump the air was partially exhausted when the stop-cock 
