Robinson and Walkden .— A Critical Study of Crown Gall. 319 
This idea was originally suggested in relation to the conception that the 
secondary tumour in the leaf originated from an ‘ invading destructive 
growth ’ (Smith, 23 , p. 16) derived from the stem. We have shown that the 
radial stem-like growth referred to arises by the active division of leaf- 
tissues on the adaxial side of the vascular bundles in relation to bacteria 
situated in the protoxylem. There can be no question of simply explaining 
the transformation of leaf-tissues into those of a ‘ pseudo-stem ’ by the fact 
that the original inoculation was made into the stem. 
The appearances in question are in some respects similar to those 
described by Winkler ( 36 ) for the petiolar bundles on Torenia when 
adventitious buds arise on the lamina of the leaf. Here the remains of the 
desmogen of the leaf-traces give rise to an almost radial bundle by secondary 
thickening. In crown gall an exactly similar stimulation to this latter 
almost invariably occurs in the petiole or midrib of the leaf of Chrysan¬ 
themum frutescens below the position of the primary gall on the leaf. This, 
like the anatomical change in the petiolar bundle of Torenia , seems to be 
a correlation effect. The facts regarding such structural changes, while of 
great interest in themselves, do not, in our opinion, support any detailed 
comparison of the radial structure of secondary galls with the histological 
results of the active transference of tissues by infiltration or metastasis in 
malignant tumours. 
The origin of leafy or bud-like growths on crown galls apart from pre¬ 
formed buds does not, in our view, afford any support to comparisons with 
malignant tumours. The new growths in plants can only in a very general 
way be regarded as equivalent to animal teratomas. Where such structures 
appear in crown gall they are comparable with the adventitious buds or 
roots that occur in ordinary cases of woody callus, or with the buds that 
arise very commonly on the cut surfaces of internodes of shoots or on 
mutilated leaves of Solanum lycopersicmn , apart altogether from any infection 
by organisms. Smith’s later work has afforded new and excellent examples 
of this phenomenon. This aspect of the question, however, lies outside the 
scope of this paper. 
The most striking growths of crown gall are always obtained when 
regions of the plant which are capable of considerable further growth are 
inoculated. When, as in most of our experiments, the inoculations are made 
into immature organs the subsequent development of these in the control 
plant has to be taken into account in dealing with the causation of the effects 
obtained^in the inoculated plant. The changes followed in the normal 
development of organs behind a growing-point are usually in themselves 
difficult to understand, but this only makes it the more necessary to con¬ 
sider the effects of the bacterial stimulation in the light of the potentiality 
for development of these organs. We have found, for example, in the 
present work that the very different effects produced by the inoculation of 
Y 
