256 ANNUAL REPORT SMITHSONIAN INSTITUTION, 1907. 
The reverse phenomenon appears when the plates are placed in the 
benzole tank or mounted with a prism. The change of color thus in- 
duced is toward the more refrangible end of the spectrum, and some- 
times produces the shift of more than half a tone. For instance, the 
red becomes orange red, and orange yellowish. Blue and violet, on 
the other hand, are scarcely modified, or, rather gain, in power and 
purity. 
This well-known phenomenon is based according to the author 
on the giving up of water by the gelatine to the benzole or to the 
Canada balsam, so that naturally the distance between the lamine is 
decreased. In order to get over this difficulty development should be 
rather longer, so that the colors shift toward the red, or, still better, 
the plate should be warmed before exposure, and just before placing 
in the mercury slide, in a drying cupboard at 86° F. 
FALSIFICATIONS OF THE COLOR TONES IN THE DARKER PARTS OF THE PLATE. 
With underexposed plates or in places which correspond to the 
shadows of a colored object, the pitcture shows, instead of the true 
color rendering, another color, and, as a rule, it is the opposite to 
the phenomena observed with overexposed plates, the shift being to- 
ward the more refrangible end of the spectrum. 
Thus the shadows of a head in sunlight are brownish-green or 
greenish-yellow, instead of the delicate rosy tint. An orange which 
is correctly reproduced on the illuminated side shows pure green in 
the shadows (fig. 16). ; 
These and ‘other imperfections of dark or only briefly exposed 
objects can be ascribed, according to the author’s researches, chiefly to 
fixation, the action of hypo or cyanide. Keeping to the example of 
the orange, the plate was, as a matter of fact, affected in the bright 
and dark parts by rays of different intensity, reflections from neigh- 
boring objects being excluded, but in the strongly exposed parts there 
were formed numerous dense lamin, while in the shadows these 
were fine and pale; in many cases there were only formed a small 
series of yellowish grains. 
The cause of this phenomenon, which had already been observed 
by O. Cramer, is that in fixation there is more silver bromide dis- 
solved out in the shadows the weaker the action of light, and there- 
fore thin lamin in the dark parts approach one another during 
drying; while in the brightly lit parts, which are therefore poorer 
in silver bromide, they scarcely alter their relative positions. 
From this fact we may deduce the practical lesson that Lippmann 
plates should not be fixed, because the disappearance of the silver 
bromide causes a general reduction of the intervals and a consequent 
falsification of the colors. 
