1641 
CIC,H,COC,H, + C,H,CHOHC,H, and 
C,H,COC,H, + Cl. C,H,CHOHC,H, 
one and the same monochlorobenzopinacone ought to have been 
formed. 
If, on the other hand, the reactions were as stated above, very 
different produets ought to be obtained in each ease. 
Indeed there was obtained in the first case a chlorine containing 
pinacone, the chlorine content of which certainly was not so high 
as that calculated for a dichlorobenzopinacone [this, however, 
is self-evident because the benzophenone obtained in the reaction 
from the benzohydrol must be converted together with some unchanged 
benzhydrol into the ordinary benzopinacone) but still much higher 
than that calculated for the monochloro compound. 
In the second case a pinacone mixture containing but little chlorine 
was obtained; this also is in full agreement with what was to 
be expected, as first the chlorobenzophenone formed is at the 
start present in small quantities only and, secondly, it is converted 
into the correlated pinacone slower than the benzophenone itself. 
From the progressive change of the reaction, as proved by different 
ways, it thus follows that both hydrogen atoms of the alcohol are 
rendered active. 
A fortunate incident subsequently enabled us to further elucidate 
the course of the reaction. 
During the illumination of solutions of terephthalophenone and of 
isophthalophenone in different alcohols a very distinct yellow colo- 
ration was observed. In the beginning this colour disappeared on 
shaking the tube. When the coloration persisted on shaking, the 
liquid could still be decolorized by allowing the air to enter into 
the reaction mixture. Hence, it was obvious to attribute the decolo- 
rization to oxidation; that in this case the oxygen played indeed 
an important role could be shown by placing the mixture (freed 
from oxygen by boiling) in a space filled with oxygen-free nitrogen; 
on exposure to ordinary day-light the contents of the tube turned 
immediately yellow; if now a trace of air was admitted, the liquid 
got again decolorized. 
In order to show the oxidation still more distinetly, the saturated 
solution of terephthalophenone with a slight excess of the ketone 
was placed in a vessel furnished with a doubly bent tube which 
dipped into mercury ; the space above the liquid was filled with air. 
On illumination, the mereury gradually rose and the small excess 
of terephthalophenone did not perceptibly dissolve, This only took 
