IN a 2-l. round-bottom flask, fitted with a stopper holding a dropping
funnel and a mechanical stirrer, is placed a mixture of 275 cc.
of concentrated nitric acid (sp. gr. 1.42) and 275 cc.
of concentrated sulfuric acid (sp. gr. 1.84). This is cooled to 10'0
in a freezing mixture, and 100 g. of benzyl cyanide (free from alcohol
and water) are run in slowly, at such a rate that the temperature
remains at about 10'0 and does not exceed 20'0. After all the benzyl
cyanide has been added (about one hour), the ice bath is removed,
the mixture is stirred for an hour and then poured on to 1200 g.
of crushed ice. A pasty mass slowly separates; more than half
of this mass is _p_-nitrobenzyl cyanide, the other constituents
being _o_-nitrobenzyl cyanide, and a variable amount of an oil
which resists hydrolysis; apparently no dinitro compounds are formed.
The mass is filtered on a porcelain funnel with suction, pressed well
to remove as much oil as possible, and dissolved in 500 cc.
of boiling alcohol (95 per cent). On cooling, _p_-nitrobenzyl
cyanide crystallizes; the mother liquor, on distillation,
gives an impure alcohol which can be used for the next run.
Recrystallization from 550 cc. of 80 per cent alcohol (sp. gr.
0.86 to 0.87) yields 70 to 75 g. (50-54 per cent) of a product
which melts at 115-116'0.

This product is satisfactory for most purposes, and incidentally
for the preparation of _p_-nitrophenylacetic acid. Occasionally
it must be free even from traces of the ortho compound,
and in this case should be crystallized again from 80 per cent alcohol;
it then melts at 116-117'0.