[TINGLE-FERGUSON] NICOTINE IN TOBACCO 29 
hydroxide solution, followed by a few drops of baryta water. This 
last addition is not essential, but tends to clarify the solution when any 
turbidity has developed. The contents of the flask were next made up 
to the 100 cb.c. mark and the precipitate was allowed to settle. The 
supernatent liquid was decanted through a dry filter, and the filtrate 
examined in polarimeter. 
A Schmidt & Haensch Triple Field instrument, graduated to read 
in Ventzke sugar divisions “j light,’ of which divisions 2:56 = 1° 
circular, was used. The source of light was a suitably shaded tung- 
sten incandescent lamp. 
Viewed in a 2 d.m. tube the above solution read — 2-05 sugar 
divisions (—0-7907° circular) as a mean of two observations. We have 
already shown (loc cit.) that a rotation of — 1-0 sugar division corre- 
sponds to 0-2198 gramme of nicotine in 100 cb.c. of solution. There- 
fore the total amount of nicotine found in this sample was 0-4505 
gramme or 2-25%. 
Experiment 2. New Method (modified). A sample of the same 
tobacco, weighing 14 grammes was taken. This was mixed with 
150 cb.c. of water and 15 grammes of barium hydroxide. Dis- 
tillation proceeded exactly as in the foregoing experiment. The 
distillate was not acidified, but was extracted with chloroform four 
times in succession, 30 cb.c. of chloroform (which must be free from 
acid) being used for each extraction. The chloroform extracts were 
collected in another separating funnel and shaken twice with JN. sul- 
phuric acid, using 15 cb.c. for the first, and 10 cb.c. for the second 
operation. The combined acid solutions were run into a 50 cb.c. 
graduated flask and warmed to expel the dissolved chloroform. <A 
few drops of baryta water were added and the solution was cooled. 
Aqueous potassium hydroxide solution was added till present in excess, 
and the contents of the flask were made up to the 50 cb.c. mark and 
filtered. 
The filtrate, examined polarimetrically under the same conditions 
as that in the previous experiment, read — 2-9 sugar divisions 
(—1-1327° circular). This corresponds to 0-6374 gramme of nicotine 
in 100 cb.c. or a total of 0-3187 gramme in the 14-gramme sample 
and therefore 2-27%. 
Experiment 3, Tôth's Method. That we might compare our own 
method with this well known one, we took a sample of the same tobacco 
weighing six grammes and rubbed it up with 10 cb.c. of a 20% 
sodium hydroxide solution. To this was added a _ convenient 
quantity of plaster of Paris. The mixture was then well shaken 
with 60 cb.c. of ether and 60 cb.c of petroleum ether in a stopped 
