139 OBSERVATIONS ON THE CALIFORNIA VINE DISEASE 
hydrogen peroxide), proteid matter (eosin, safranin) ; that they 
contain .also a gum, somewhat similar to wound-gum, would appear 
from their reaction to orcein and rosaniline. 
The deposits in the cells are, then, complex in composition, but 
it is not improbable that a decomposition product of starch forms 
their predominant ingredient. The following facts strengthen 
this view: | 
The cell walls and cytoplasm of healthy tissues stain blue in 
Bóhmer's haematoxylin, but pathognomonic tissues, on the other 
hand, do not. If Bóhmer's haematoxylin and safranin are used 
together the healthy tissues will appear as we have just described 
them, but the cell walls and cytoplasm, when diseased, are red or 
reddish. This would prove that a decided change takes place in 
the cell walls of diseased tissues simultaneously with the change 
in the cytoplasm, but as this change is visible only upon colora- 
tion, it is not likely that the cell walls contribute materially to the 
formation of the deposits in the lumen. The deposits must, 
therefore, arise as decomposition products from one or more of 
the cell contents. I believe that starch is the most important 
contributing substance of them all, and circumstantial evidence 
favorable to this view is not lacking. 
We have seen that the more rapid the death of the leaf the 
denser the occluding deposit. We have seen also that when the 
tissues die suddenly without apparent prenecrotic coloration, 
they are, nevertheless, as free from starch as tissues that become 
previously colored. We have seen also that reddened tissues аге 
replete with starch and practically free from deposit, but 10 
sooner die than occlusion becomes noticeable and starch, to all 
intents and purposes, absent ; we have seen further that the longe! 
the period of chlorosis before death, the freer the cell lumen 
from deposit, and starch, it is well known, is quite scarce in 
chlorotic tissues. It appears to me, therefore, that the substance | 
(of the nature of wound-gum) found in the cell lumen, as атап | 
lar or homogeneous matter, is largely derived from starch. 
The deposits we find in diseased cells do not appear to contain 
pectic substances in any considerable amount, for Victoria blue ` 
and chloriodide of zinc do not give the typical reactions, and the 
