518 JOHIST E. MUELIN 



In the modified form of apparatus devised by Voit(d) for experiments 

 on small animals the suction pumps were replaced by a large meter driven 

 by a water wheel which served at once to aspirate the air through the cham- 

 ber and to measure its volume. The chamber devised by Voit was of small 

 capacity containing only 64 liters. Larger chambers, however, were used 

 as, for example, the chamber in the accompanying figure which had a ca- 

 pacity of 340 liters. 



The construction of the small suction pumps also was somewhat modi- 

 fied in the Voit construction and a very useful type of valve with mercury 

 seal known as the Voit valve was employed to give direction to the air 

 sample. (See figure 1.) With this type of apparatus in five control ex- 

 periments in which pure olein was burned in the form of a candle or tal- 

 low dip, an average error of 1.75 per cent was found for the CO 2 , and for 

 the absorption of water an error which varied from 1.4 to 5.5 per cent. 



Wolpert(a) working under the direction of Rubner also made some 

 improvements on the Pettenkofer type of apparatus. His chamber 

 measured 1.5 x 2.5 x 2 meters with a cubic capacity of 7.5 cubic 

 meters. The measuring drum was driven by means of a water motor. 

 The apparatus differed otherwise in only minor details from the Voit 

 construction, but Rubner(_/) succeeded in measuring the water vapor 

 with a much greater degree of accuracy. 



b. The Apparatus of Sonden and Tigerstedt. This apparatus erected 

 at Stockholm and first described in 1895 was so constructed as to accommo- 

 date a number of individuals employed as subjects at the same time. The 

 chamber consisted of a room measuring 5x5x4 meters and had a total 

 capacity of approximately 100 cubic meters. The walls were sealed with 

 sheet metal carefully soldered together and the room was ventilated through 

 a zinc pipe measuring 14 cm. in diameter which was carried up above 

 the roof of the room and capped with a ventilator containing a valve to 

 guard against aspiration of air from the room by action of the wind. The 

 room was heated by steam and the air was kept stirred by means of an 

 electric fan. Ventilation was accomplished by means of pumps gauged 

 to three different speeds which could be adapted to the number of indi- 

 viduals serving as subjects. Samples of air were withdrawn from the 

 exit tube near its mouth and were analyzed by means of the Sonden-Petter- 

 son apparatus. Check experiments with burning candles or petroleum 

 gave an average error of 1.16 per cent on the CO 2 * In other series of ex- 

 periments performed later by Rosenberg the error was reduced to 1 pex 

 cent. This apparatus and a later one on the same principle at TIelsirigfors 

 ( Tigerstedt (g)) have been used especially for the study of metabolism in 

 school children. 



c. The Apparatus of Atwaier and Rosa. This apparatus constructed 

 with the aid of the U. S. government in the chemical laboratory at Wesley- 

 an University, Middletown, Conn., was first described in 1897. It con- 



