EXPERIMENTAL METHODS. 17 



ing the idea adopted by Pallidin and Kostytschew the receiving flasks for the 

 plant material were provided with extensions of the necks in the form of 

 collars about 25 mm. high, so that when the stopper was inserted the flasks 

 could be securely sealed by pouring on enough mercury to cover it (a, fig. 1). 

 The stopper was punctured with three holes. In one a thermometer (d, fig. 1) 

 was thrust, in another a U manometer (c, fig. 1), and in the third the delivery 

 tube (6, fig. 1), which ran well down into the flask and was closed on the out- 

 side with a glass stop-cock. Figure 1 shows the general appearance of the 

 receiver. These flasks proved exceedingly convenient, as they could be car- 

 ried easily from place to place without disturbing the seal, could be darkened 

 by covering with tinfoil and paper, and could be immersed in water if occasion 

 required, without any danger that the material inside would become wet. 

 They could also be set up with such rapidity that it was possible to have four 

 sets of experiments running simultaneously without taxing the manipulator. 



The material was placed in these receivers, and, after being left the desired 

 length of time, samples were drawn off and the capacity was determined in the 

 manner to be described. When the sample was to be taken the outlet tube 

 was connected with what may be called the sampling apparatus, which was 

 designed after the form used by Aubert, but differing in that it was detachable 

 and so could be used for an indefinite number of receivers. There is not only 

 this advantage, but also that which comes from having the receiving flasks 

 readily transportable. This sampling apparatus consisted of a bulb of about 

 25 c.c. (/, fig. 1), capacity provided with three openings. The upper one, a 

 bent narrow-bore tube, was connected, when desired, with the outlet tube of 

 the receiving flask, by means of rubber tubing, properly wired, and each time 

 the connection was made it was also wired (e, fig. 1). The two tubes had 

 ground ends and came in very close contact when brought together. This seal 

 was repeatedly tested by exhaustion and could hold a mercury column of 

 50 or 60 mm. indefinitely. It may be mentioned that the ordinary cotton- 

 covered annunciator wire was found to be excellent for such wiring purposes, 

 as it had no tendency to cut the rubber tube when tightened around it. The 

 second outlet, which was really a branch of the first, led to the small-bored 

 delivery tube (h, fig. 1), so bent as to deliver the gas conveniently when the 

 free end was immersed in the mercury trough (i, fig. 1). This delivery tube 

 was provided with a glass stop-cock. There was also an extension upwards of 

 the first outlet with a three-way cock attached, but this could have been dis- 

 pensed with as well as not. The third opening at the bottom of the bulb led 

 by means of a long rubber tube (g, fig. 1) to a mercury reservoir which could be 

 elevated or depressed, as required to fill or empty the bulb. 



To withdraw any gas the whole system up to the stop-cock in the outlet 

 tube of the receiver was filled with mercury, great care being taken to see that 

 no air was caught in the bulb, tubes, or stop-cocks. Inasmuch as the stop- 

 cock on the receiver had a vent looking downwards, it was possible to run any 

 quantity of mercury through the apparatus. In a similar fashion the delivery 

 tube of the sampling contrivance was filled with mercury. Before proceeding, 

 the whole apparatus was once more inspected for air-bubbles, especially those 

 which are apt to stick in the bore of the stop-cocks, and to see that the joint 

 with the receiver was tight. The stop-cock on the receiver was now opened 

 simultaneously with the depression of the mercury reservoir so that some gas 



