Robinson and Walkden.—A Critical Study of Crown Gall . 309 
the exterior with mercuric chloride, have been referred to by Walkden in 
his note on the isolation of the organism. He has pointed’ out that the 
isolation is much more readily accomplished if the external surface of the 
gall is not first sterilized. In isolating B. tumefaciens he also found that 
more numerous colonies of this organism are obtained when portions of the 
exterior of the gall are used than if the inoculum be taken from the interior 
of the gall after removing the surface layer. Further, when sections of 
living galls on aerial shoots of any age from ten days to two months or 
older are mounted fresh in water, we have found that the external surface 
of the gall invariably shows a mucilaginous film from which very large 
numbers of'bacteria of the form, size, and general appearance of B. tume- 
faciens can be observed diffusing into the mounting water. 
Like Smith and other investigators, we have entirely failed to demon¬ 
strate the bacteria, by staining, within the cells of the gall, nor have we 
obtained any evidence that the bacteria enter the living cells of the host 
plant. The individuals of B. hLmefaciens present in the film of mucilaginous 
material on the exterior of the gall are, on the other hand, readily demon¬ 
strated by the ordinary staining methods. In the early stages of the 
development of galls, after the inoculation of the cut surfaces of Chrysan¬ 
themum shoots, it is possible also to demonstrate the bacteria by staining 
some little distance (for about 2 mm.) along the vessels, and also occa¬ 
sionally in the intercellular spaces of the cortex (e. g. in PL V, Fig. 9, at is.). 
The presence of the bacteria in these situations in the early stages of 
gall-formation is, in part at least, responsible for the shape of the gall which 
results from the localization of the disturbing influence in definite regions 
of the stem, i. e. in the vascular bundles and occasionally in loci in the 
cortex. It must be pointed out, however, that in the Chrysanthemum the 
organisms are not found at any considerable distance from the surface, and 
soon the form of the growing gall becomes such that the majority of the 
organisms producing it are localized on its surface. The further evidence 
for this conclusion which is indicated by the facts outlined above may now 
be given. 
It was repeatedly found that merely dipping the unbroken gall into 
mercuric chloride solution (1 in t,ooo) for ten seconds, prior to breaking it 
up for making platings, was sufficient to reduce the numbers of B. tume¬ 
faciens found from two hundred per plate to one. This result was obtained 
no matter how thoroughly the gall was washed after the treatment with 
mercuric chloride. The suspicion therefore arose that the active organisms 
were situated mainly on the outer surface of the gall. To test this the 
following experiment was carried out: 
A gall was taken direct from the plant growing under the usual green¬ 
house conditions, without any attempt to prevent contamination, and 
dropped into a tube containing 10 c.c. of sterile water for ten minutes. 
