Robinson and IValkden .— A Critical Study of-Crown Gall. 315 
Fig. 19 ,shows tumour-strands in the pith of another similarly inoculated shoot 
of Nicotiana. Here the bacteria in intercellular cavities marked i.s. i, i.s. 2, 
i.s. 3 are forming a centre from which the stimulating influence is radiating 
outwards. In these cases it is again clear that there is no true intrusive growth 
of tumour-tissue : the stimulus merely results in the subdivision of the pre¬ 
existing cells of the pith. This is seen even more clearly in PL VI, Fig. 20, 
which also shows the deeply-staining masso i B .tmnefaciens in the intercellular 
space which forms the centre for the production of the gall. This figure also 
indicates that the invaded cavity may be slightly enlarged by the occasional 
necrosis of cells bounding it, but there are never, in our experience of 
crown gall, the marked necrosis cavities described for the Olive-knot 
disease due to Bacteruim savastoni , Smith. This difference between the 
two diseases, however, would seem to be one of degree rather than of kind. 
We have traced, in serial longitudinal sections, the passage of the 
bacteria along the very large intercellular spaces which are present in the 
stem of the Nicotiana affinis. PL VI, Fig. 21, shows a longitudinal section of 
a part of a shoot at some distance below the surface which was inoculated 
eighteen days previously. The dark centres (i.s.) in the cortex are clearly 
seen, with tumour-tissue arising by the subdivision of cells on either side of the 
bacterial strand. PL VI, Fig. 22, shows, more highly magnified, the zoogloeal 
strand of B. tmnefaciens seen advancing through a large, longitudinal, inter¬ 
cellular space in the pith of Nicotiana. Such zoogloeal strands of the 
bacteria, which are readily demonstrated by staining, are invariably found 
in sections of young secondary tumours on Nicotiana affinis , and, since 
detecting them in this plant, we have frequently observed them also in the 
Chrysanthemum. In the latter plant, however, the intercellular spaces are 
much smaller and less abundant, and we have not observed the zoogloea to 
penetrate to a depth greater than two millimetres from the cut surface of 
the shoot. 
In addition to the definite zoogloea of B. tmnefaciens which have thus 
been repeatedly observed in the intercellular spaces of both plants studied, 
it has been mentioned that the bacteria also enter the protoxylem elements 
of both infected shoots and leaves, and, travelling in these vessels, and also 
passively carried in them by growth extension, serve as centres for the 
development of secondary tumours. The secondary galls which we have 
described and figured (Pis. V and VI, Figs. 11, 12, 33) for Chrysanthemum 
frutescens are of this type. They owe their origin and characteristics to the 
fact that they were derived'by the inoculation of the basal region of leaf- 
rudiments at such an age that the first differentiation of protoxylem of the 
leaf-trace had taken place, but the basal growth of the leaf-rudiment had not 
ceased. PL V, Fig. 4, shows a longitudinal section of an apex of Chrysantlie- 
mum frutesce 7 is,a.x\(\ the leaf (/.) shows the stage in development at the time 
of inoculation for results such as are seen in PL VI, Fig. 12, to be produced. 
