18 



August Krogh and Marie Krogh. 



and was driven by the water wheel. This was fed from a small constant 

 level reservoir. The meter was protected against direct sunlight but 

 could not of course be kept at a constant temperature. On account 

 of the large quantity of water in it, the changes were, however, com- 

 paratively slow. From the pipe between the mixing vessel and the 

 meter a piece of very narrow lead-tubing led to a sampling vessel. 

 The sampling apparatus consisted of two sampling vessels of 100 cc. 

 capacity each, mounted on a stand (fig. 8). The vessels could be inserted 

 in the short pieces of rubber tubing (1) and filled from the receiver 

 (2) through the three-way tap (3). Through the second boring of this 

 tap connection was established with a narrow rubber tube ending 

 in a finely pointed glass tube through which the mercury could flow 

 out. The resistance in the point of the tube was regulated in such 

 a way that the 100 cc. of mercury would take a suitable time (2- — 6 

 hours) to flow out. In order to maintain a constant rate of flow the 

 glass tube was inserted in the cork of a bottle in which the outflowing 

 mercury was coflected and each bottle was suspended by means of 

 a spring. The weight of the mercury would lengthen the spring to 

 such an extent that the vertical distance between the level of mercury 

 in the sampling vessel and the bottle would remain practically unal- 

 tered. When the bottle was so far lowered that the sampling vessel 

 was nearly empty it would make an electric contact and ring a bell. 

 By this signal the operator was called and the corresponding period 

 of the respiration experiment was then brought to a close. The sam- 

 pling vessel was closed and the lead tube transferred to that on the 

 other side. The thermometers in the chamber and the meter were 

 read. The meter itself was read and the exact time noted. Samples 



of atmospheric air and samples 

 of the air in the chamber were 

 taken immediately afterwards. 

 These samples were taken in dry 

 medicine bottles of 30 and 50 cc. 

 capacity. By means of the arrange- 

 ment shown in fig. 9 a current of 

 air was passed through them. 

 The glass tube was narrowed to 

 such an extent that a con- 

 siderable pressure was obtained 

 in the blower. The bottle was accordingly washed out by a uniform 

 but comparatively slow current of the air to be sampled and a cork 

 soaked in paraffin wax inserted immediately on withdrawal of the 

 tube. It is essential when taking samples in this manner, of air, dif- 

 fering in composition from the surrounding atmosphere, that the 

 current at the mouth of the tube is slow and uniform. Otherwise it 

 will on withdrawal act as a jet, and the sample in the bottle will be- 



Fis: 9. 



