932 Journal of Agricultural Research vol. vi.no. 24 



under investigation; the butter flask B. F., of about i liter capacity, for 

 retaining the sample after its passage through part of the system; the 

 moisture flask M. F., of 200 c. c. capacity, for retaining the greater part 

 of the moisture liberated from the sample; and a small evacuation flask 

 or globe, E. F., of about the same capacity. These detachable parts are 

 connected with the upper part of the system by means of various mercury 

 seals, S, S. 



The operation is begun late in the afternoon of the day before the actual 

 determination by putting the entire division 1 under vacuum and allowing 

 it to stand in this condition overnight. This is done in the following 

 manner : 



With the exception of the butter tube B. T., the apparatus is connected 

 as illustrated: The mercury-seal stopcocks S^ S 2 , and S 4 are closed and 

 stopcocks S 3 and S 5 are so turned as to open the system from E. F. through 

 a, B. F., M. F., b } and c to the Topler pump T. P. The Topler pump is 

 now given one stroke, which serves somewhat to exhaust the air confined 

 in the system; a small beaker, x, filled with concentrated sulphuric acid 

 is brought under stopcock S 2 so that the tube projecting downward from 

 the stopcock is plunged well beneath the surface of the acid, and the 

 beaker is supported in this position. Stopcock S 2 is now cautiously 

 opened until the acid rises to form a long level of drying agent covering 

 the bottom of tube c, when the influx of acid is stopped. The pump is 

 now worked to its limit of exhaustion (about 0.3 mm. on the McLeod 

 gauge). A turning back and forth of stopcock S 5 accompanied with 

 successive strokes of the pump will evacuate tube d to stopcock S 4 , and 

 this is followed by turning stopcock S 3 and working the pump to exhaust 

 the tube e. The entire division 1 is now under exhaust and is allowed 

 to remain so for a considerable length of time, preferably overnight. 

 The next morning, if the gauge indicates that no leakage of air into the 

 system has occurred, the actual determination is made as follows: 



The moisture flask M. F. is covered by a beaker which is then packed 

 with cracked ice and salt (sodium chlorid). The butter tube B. T. is 

 connected at S 4 by means of a mercury seal. Funnel B is closed at stop- 

 cock H, and filled with a three-fourths saturated sodium-chlorid solution 

 at a temperature of 50 C. A little of this brine solution is allowed to 

 trickle from, a pipette into the small side tube of the butter tube B. T. 

 until the latter is completely filled, whereupon it is connected with the 

 funnel tube below H by means of a piece of tight-fitting rubber vacuum 

 tubing. A large glass jar (not shown in the illustration) is now used to 

 cover the butter tube B. T., the base of which rests upon a large rubber 

 stopper with its center removed. The mercury-seal stopcock S 5 is now 

 turned to connect the evacuation flask E. F. with d, and a turn of S 4 toward 

 d followed by a closing of the same serves to evacuate the tube from S 4 to 

 the glass stopcock K of the butter tube B. T. S 3 is opened to connect e 



