70 PLANT RESPIRATION 



follows. The thick-walled rubber tubing d of the flask is con- 

 nected with the tube h of the gas pipette (Fig. 9) and the tube // 

 is connected with the tube ^ by a suitable turn of the three-way 

 stopcock r. At the same time the bulb / must be filled with 

 mercury up to the stopcock r. The bent end of the tube g is 

 connected by thick-walled rubber tubing with the apparatus 

 for the production of nitrogen and the air in the tubes g and // 

 and the research flask is replaced with pure nitrogen. The stop- 

 cock ;- is then so turned that the bulb / communicates with the 

 tube g, and the tube h together with the research flask is sepa- 

 rated. The rubber tube is now taken from the end of the tube g. 

 this tube is filled with mercury from the bulb /, the end of the tube 

 is immersed in mercury, and then the bulb / is connected with 

 the tube // and the research flask. A test portion of gas is 

 withdrawn from the flask in the same manner as was described 

 for the experiments on oxygen respiration. This gas is passed 

 into an analysis tube, the upper part being prepared with the 

 oxygen excluded^ and containing almost colorless alkaline pyro- 

 gallol solution. If the solution remains colorless, the displace- 

 ment of oxygen is complete. If the solution turns brown or 

 dark-brown, more nitrogen must be passed through the flask. 

 If the internal atmosphere of the experiment flask is free of 

 oxygen the flask can be closed and the experiment begun. 

 Through the removal of the control portion the gas pressure in 

 the flask has already been lowered and the mercury has risen 

 in the manometer. The bulb / is now connected with the tube 

 //, and the rubber tubing d and the outer side of the entrance 

 tube a on the experiment flask are filled with mercury. If the 

 screw pinchcock on the rubber tubing d is then closed, the flask 

 is left filled with pure nitrogen and sealed air-tight. Fig. 8 

 shows that the inner atmosphere of the flask is entirely sepa- 

 rated from the air by glass and mercury. 



Of course the flask can be filled as above not only with nitro- 

 gen but also with hydrogen or another gas. Either nitrogen 

 or hydrogen may be used as an inert gas. Moeller- has been 



1 This is accomplished as follows. Into an analysis tube full of mercury are introduced 

 under the mercury with the aid of two pipettes with bent outlet tubes, first a couple of cubic 

 centimeters of concentrated caustic potash, then some pyrogallol solution. 

 2 Moeller. Ber. d. bot. Ges. 2: 306. 1884. 



