314 Robinson and Walkden,—A Critical Study of Crown Gall. 
of the structures in which they appear, and we have found no evidence of 
any invasive growth of tumour-tissue from a distance. In the case shown in PI. 
VI, Figs. 13 and 15, we have direct as well as cultural evidence of the presence 
of the bacteria in the protoxylem region in the centre of the modified midrib, 
and we have traced the organisms back along the protoxylem to the 
primary gall by serial microtome sections. The facts thus briefly referred 
to for secondary galls on Chrysanthemum frutescens were made out only 
after a considerable body of our work had been completed since, as has 
been mentioned, like Smith, we experienced great difficulty in isolating the 
organisms from the secondary galls on the Chrysanthemum. 
Parallel studies, however, which we carried out on Nicotiana affinis 
enabled us successfully to stain and thus directly demonstrate the bacteria 
in the interior of secondary galls on this plant, to isolate the organism in 
large quantities from the interior of such smooth galls, and to demonstrate 
conclusively the infective migration of the bacteria from the original point 
of infection for very considerable distances through the plant. These 
secondary galls on Nicotiana will, therefore, now be described. 
Cut surfaces of the internodes of the young flowering shoots of 
Nicotiana were inoculated in a manner similar to that described for 
Chrysanthemum. As in the latter plant, in some cases, a large rough gall 
arose on the end of the inoculated shoot and, as before, the bacteria were 
found to be present in very great numbers on the rough outer surface of the 
gall. More frequently in Nicotiana , however, the primary gall arising on 
the end of the shoot is much smaller; in such cases it is invariably found 
that a number of smooth galls arise as a series of swellings which extend to 
a considerable distance below the inoculated surface. PI. VI, Fig. 16, shows 
such a series of smooth secondary galls on Nicotiana , and it will be seen that 
these are identical in appearance with similar galls on Nicotiana figured by 
Smith ( 29 , Fig. 24,#). PI. VI, Fig. 16, is quite typical of a large number of 
inoculated shoots which we have obtained showing secondary galls. We suc¬ 
cessfully isolated Bacterium tumefaciens from the secondary gall (s.g. 1) most 
remote from the inoculated surface and we have also stained these bacteria 
in situ within the galls. PI. VI, Fig. 17, shows a transverse section of the shoot 
seen in PI. VI, Fig. 16, through the swelling (s.g. 2) immediately above (. sg . 1). 
In this case there are four centres of disturbance in the cortex giving rise 
to three secondary galls and a tumour-strand. Serial sections at different 
levels of the shoot seen in PI. VI, Fig. 16, have demonstrated the longitudinal 
continuity, throughout the length of the stem shown, of the tumour-strand and 
secondary galls seen in the cross-section in PI. VI, Fig. 17. Around the pro¬ 
toxylem of some of the bundles there are also small disturbances obviously 
resulting from the presence of the bacteria in these vessels. PI. VI, Pig. 18, 
which is taken from the inner face of the vascular ring (at px.) of the section 
seen in PI. VI, Fig. 17, shows one of these small tumour-strands. PI. VI, 
