118 ROYAL SOCIETY OF CANADA 
when recrystallized from benzene, melted at 118°— the melting point 
of inactive mandelic. Its aqueous solution was quite inactive when 
examined in the polarimeter. 
Two other experiments were performed using 2 N. HCl to effect 
the hydrolysis of the amygdalinic acid, but the result was the same. 
Inactive mandelic acid was the product in each case. 
Fearing that the inversion to inactive might have been caused 
by boiling with an excess of baryta, we performed another experiment 
in which the latter was added in small quantities with a final slight 
excess, but the product was again inactive mandelic acid. 
In two other experiments, after precipitating the barium with 
the theoretical quantity of oxalic acid, we hydrolysed the amygdalinic 
acid with concentrated hydrochloric acid and hydrochloric acid of 
specific gravity 1:12 respectively, as in the preparation of the active 
acid from amygdaline, but again obtained the inactive acid. 
In order to test finally in which stage the inversion takes place, 
we dissolved 25 grams of amygdaline in an excess of standard barium 
hydroxide solution and diluted with water to 250 c.c. The solution 
was examined in the polarimeter and gave a reading in the 100 m.m. 
tube of —4-3°. A part was then boiled under a reflux condenser 
for ten minutes and then rapidly cooled. Its rotatory power was 
now —5-55°. A second portion was boiled in the same way for thirty 
minutes and also had a rotation of —5-55°. The hydrolysis is, there- 
fore, one which proceeds very rapidly and is complete after ten 
minutes boiling. The solution was then boiled in an open flask till 
all the ammonia was expelled and again made up to 250 cc. The 
angle of rotation was now —5-75°, showing that the substance had 
not been inverted by boiling with the barium hydroxide, and this 
solution yielded on hydrolysing with concentrated hydrochloric acid, 
the inactive acid. The inversion, therefore, must take place when 
the hydrogen atom is being introduced in place of the glucose radical. 
Wislicenus has shown that some of these so-called physical isomers 
are exceedingly sensitive to light, and that ordinary daylight under 
certain conditions can transform the one into the other. To investi- 
gate whether such was the case in the present instance or not, we 
hydrolysed the amygdalinic acid with normal HCl in darkness, but 
obtained again the inactive acid. 
The same result was obtained when potassium hydroxide was 
substituted for barium hydroxide and the amygdalinic acid hydro- 
lysed with 4 N. HCl. 
It seems, therefore, impossible to hydrolyse active amygdalinic 
acid with hydrochloric acid of any strength at least above $N. so as 
to obtain the corresponding active mandelic acid — the inactive race- 
