962 
Journal of Agricultural Research 
Vol. XXIV. No. II 
and it was found that while Rasmussen’s method is somewhat similar 
in principle, it is entirely different in technic from the one described 
in this paper. As no determinations by the former method have been 
made here, its accuracy is not known. Certain features of it, however, 
are mentioned for comparison with the writer’s procedure. 
The extraction of nicotine from the tobacco to which alcoholic soda 
has been previously added is carried on in a flask containing a mixture 
of equal parts of ethyl ether and petroleum ether, by frequent shaking for 
5 hours. The solvent is then filtered from the tobacco through a folded 
filter covered with a glass plate. An aliquot is taken for acid extraction. 
With extracts, 3 or 4 gm. of the sample are placed in a 100 cc. flask together 
with 5 cc. concentrated caustic soda, 5 cc. water, and 50 cc. of the above 
mixed ethers. The mixture is allowed to stand 4 or 5 hours with frequent 
shaking, then 25 cc. of the solvent is drawn off and treated as before. In 
both cases the nicotine is precipitated with silicotungstic acid. Doubts 
might be raised as to the complete extraction of the nicotine, especially 
from extracts handled in the manner indicated, as compared with a 
Soxhlet extraction of the sample as prepared for analysis here. Further¬ 
more, errors may arise due to evaporation when the solvents are filtered 
and aliquots wi^drawn in the manner described. Finally, no provision 
is made for the prevention of emulsions often obtained in the acid washing 
of the ether extract of some samples and without which no method of 
this character is practical. 
EXPERIMENTAL WORK 
The writer, having occasion to determine the nicotine content of a 
large number of various grades and varieties of tobacco grown in Ken¬ 
tucky, found that it would be very convenient to have a method which 
would be more rapid than either of the official methods and equally 
accurate. As several determinations were to be made at the same time 
and facilities for making simultaneous steam distillations were not at 
hand, it was desirable to have a method that would, if possible, eliminate 
such distillations. A steam distillation of nicotine is objectionable inas¬ 
much as it is often a long, tedious procedure that requires careful atten¬ 
tion in order that the boiling liquid shall be maintained at a low volume 
to remove all nicotine. 
The writer’s previous experience with the Kissling method has shown 
that there are two principal errors to be avoided: First, the presence of 
ammonia in the nicotine distillate, due to its incomplete separation from 
the ether extract of the sample, and, second, the possible loss of nicotine 
in the evaporation of this ether extract previous to its distillation. The 
first error will, of course, give high results, due to the ammonia being 
titrated and calculated as nicotine, whereas the other causes low results; 
consequently, satisfactory figures obtained by the method may be some¬ 
times the result of a balancing of errors. On the other hand, experience 
with the silicotungstic acid method in this work has shown that while it 
is decidedly preferable to the Kissling, nevertheless certain points in its 
use have to be carefully supervised in order to obtain satisfactory re¬ 
sults. The writer’s experiments show that this method as outlined by 
Chapin gives more satisfactory results on nicotine extracts than on to¬ 
bacco. The reasons for this will be mentioned later. 
Experience here has shown that the distillation as prescribed in the 
silicotungstic acid method, even though carried on under proper condi- 
