[Vol. 6 



24 



ANNALS OF THE MISSOURI BOTANICAL GARDEN 



of the error from the mechanical carrying-over of the alkali 

 may be seen in table vi where ''first" and "second" distilla- 

 tions are recorded. This apparatus was only moderately 

 satisfactory. The volume of the receiver flasks varied from 

 130 to 150 cc., whereas a range of 100 to 115 cc. would have 



B 



A 



C 



IB' 







>\*Y\> 



Fig. 1 



rig. 2 



Nitrogen-distilling apparatus. 



been more satisfactory. Moreover, the error incident to redis- 

 tillation detracts from the greater accuracy resulting from 

 the elimination of rubber stoppers. The final apparatus 



adopted is that shown in fig. 2, which is self-explanatory. 



Pyrex glass was used throughout except for condenser jackets. 

 Distillation was carried on slowly for 20 minutes through a 

 cooled condenser and 20 minutes after draining the condenser. 

 It was found by a series of distillations on ammonium sulphate 

 solutions that quantitative recovery was effected by this pro- 

 cedure, and the volume of the receiver flasks as a rule re- 

 mained below 115 cc. at the end of the distillation period. This 



