HALDANE GAS-ANALYSIS APPARATUS. 75 



sampler and then drawing air from the sampler through the connec- 

 tions. The drawing of the sample by the washing method is carried 

 out as follows: An additional three-way stopcock is attached to the 

 stopcock, C, above the burette. One of the branches is attached to 

 the sampler. Air is then drawn through the tube from the sampler 

 into the burette, A, in portions of about 15 c.c., and rejected through 

 the free opening of the extra three-way stopcock. The amount of 

 washing depends in part upon the amount of gas available, but the 

 process should be carried out two or three times at least. When the 

 amount of gas is small, it is necessary to use the other method, that is, 

 by filling with mercury the space between the stopcock, C, attached to 

 the burette and the sampler and then drawing the mercury up through 

 the tube into the burette, A. The former method has ordinarily been 

 used in this laboratory, as in practically all cases the sample to be 

 analyzed was of such size that a considerable amount could be rejected 

 in the washing method. In all washing and sampling arrangements 

 the gas must always be under pressure, so that if any of the connections 

 are not tight, the leak would be outward rather than inward, as a leak 

 inward would produce a change in the composition of the gas. 



After the final washing is completed, the amount required is drawn 

 into the burette, A . The stopcock, C, is then reversed and the leveling 

 performed by means of the leveling bulb, D, and the device at the bot- 

 tom of the burette. The burette should contain sufficient water to 

 saturate the gas thoroughly before the setting is made and the actual 

 reading is taken. The water in the water-jacket, B, should be stirred 

 by forcing in a little air through the tube, T. The two right-angle 

 stopcocks, F and G, should then be turned in such a way that the gas is in 

 connection with the pipette, E. A reading is then taken. The gas is 

 passed back and forth several times, care being taken not to force the 

 mercury up into the stopcock, C, at the top of the burette. A reading 

 is then taken, the levels being set again as before. The difference 

 between the two readings gives the amount of carbon dioxide absorbed 

 from the sample. In order to make sure that all of the carbon dioxide 

 is absorbed, it may be passed again into the pipette and a reading taken. 



After the carbon dioxide is absorbed, the air is then passed into the 

 potassium pyrogallate pipette to absorb the oxygen. The routine 

 which has been carried out in this laboratory is as follows : After the air 

 is sent back and forth into the pyrogallate five times, it is left in the 

 pyrogallate pipette for a few minutes, then drawn out and passed back 

 and forth in the potash pipette five times. It is next drawn from the 

 potash pipette and forced back and forth in the potassium pyrogallate 

 pipette five times, and again sent into the potash pipette, when the 

 first reading is taken. After the air has been sent back and forth into 

 the potash once and into the pyrogallate five times readings are again 

 taken. This routine is repeated until the last two readings are constant 



