- 33 - 
Jones (221) in May 1937 published a modified procedure for the crys- 
tallization of rotenone- carbon- tetrachloride solvate from extracts of dcrris 
and cube roots. Accurate results by this crystallization method were obtained 
only when the rotenone present was equivalent to at least 4 percent of the 
root, or in lieu of this, when sufficient rotenone ".'as added, or a sufficiently 
large sample taken, to bring the mount present during crystallization above 
this value. A determination of the precision of replicate results on one 
sanplc of derris with about this rotenone content showed a standard deviation 
of +0,05 percent. In a study of the accuracy eight extracts, containing 
known amounts of rotenone in the range cf most accurate results, gave average 
values which in view of the precision were not significantly different from 
the actual rotenone content. In other words, results for rotenone in the re- 
gion of 4 percent should only be quoted to about 0.1 percent. 
It is probably that the nonrotenone portion of the extract only exerts 
a retarding action on the crystallization and has little or no actual solvent 
effect on the rotenone. The so-called Sunatratype derris extracts show no 
unusual retarding action and no apparent solvent effect, indicating the "hidden" 
rotenone of Cahn and Bean to be a result of retarded crystallization similar 
to that found in other extracts of lev; rotenone content. 
The exact procedure cf the improved method is as follows: 
"The extract from a 25-grar. sample of root in a 125-cc. Erlcn- 
meyer flask is evaporated on the steam bath in a current of air until 
free of solvent. The residue is treated with exactly 25 cc. of carbon 
tetrachloride, the flask is loosely stoppered, and complete solution 
is effected by gentle warning. The extract is next cooled in an ice 
bath for several ninutes end then seeded with a few crystals of ro- 
tenone-carbon tetrachloride solvate. The flask is then tightly stop- 
pered and swirled until crystallization is apparent. If at this stage 
only a small amount of crystalline material separates, an accurately 
weighed quantity of pure rotenone should be added, the mixture warned 
to effect complete solution, and crystallization again induced. Suf- 
ficient rotenone must be present so that the result, expressed as 
"pure" rotenone, is at least 1 gram. At the same time a solution of 
rotenone in carbon tetrachloride containing 0.27 gram per 100 cc,' of 
solution (the solubility at 0° C.) is prepared for use in washing. 
The extract and the washing solution are then placed in an ice bath 
capable'' of maintaining a temperature of 0° C. overnight, 
"In the morning (the method is based 'on a crystallization time 
of 17 to 18 hours } the extract is filtered rapidly through a weighed 
G-ooch crucible in the bottom of which has been placed a disk of filter 
paper, the flask being removed, from the ice bath only long enough for 
pouring a small portion of extract into the crucible. The separated 
crystalline materia -1 is rinsed from the flask and washed under suction 
with sufficient of the ice-cold saturated solution (usually 6 to .10 cc.) 
to remove the excess mother liquor. Suction is applied for about 5 
minutes, and then the material is dried to constant weight at 40° C. , 
which usually requires about 1 hour. This is the weight of crude ro- 
tenone- carbon tetrachloride solvate. 
