160 CROWN-GALL OF PLANTS. 



takes place in crown-galls is like that seen in the regeneration tissue 

 of wounds, but that growth is governed by a physiological need and 

 ceases with the repair of the wound, whereas the gall tissue prolif- 

 erates indefinitely, passes beyond the control of the plant, and 

 becomes a wasting disease. So far as the tissues themselves are 

 concerned, the chief difference appears to lie in the different distri- 

 bution of the various elements, the overplus of parenchyma, the 

 weakening of tlie conductive tissues, the persistent prevalence of 

 meristematic (embrj'-onic) cells, and of immature forms generally, 

 e. g., defective vascular bundles. Crown-galls vary greatly not only 

 in virulence, but in their structure from species to species and also 

 from individual to individual within the species, depending, appa- 

 rently, on where the tumor takes its origin, i. e., whether it begins 

 in pith or bark or wood or on the lamina of a leaf. 



Sometimes the tumors are very woody and hard, their structure 

 consisting chiefly of twisted and contorted lignified vascular bundles 

 and woody fibers mingled with more or less parenchyma. At other 

 times, and very often, the structure consists mostly of rapidly pro- 

 liferating nests of parenchymatous tissue of a round-celled or spindle- 

 celled type (PI. XXXII) , intermixed v/ith which are vascular bundles 

 (conductive tissues), more or less lignified, but twisted out of their 

 normal shape, with walls abnormally thin, the total mass of the 

 conductive tissue being less by far than that encountered in normal 

 tissue, i. e., there is an enormous excess of the rapidly proliferating 

 parenchyma and a corresponding reduction of conductive tissues. 



The cells of the hyperplasia are often much smaller than the cells 

 of the tissue in which it originates, e. g., inoculated tumors in cortical 

 parenchyma of tobacco stem (PI. XXIX) . There is never any enor- 

 mous stretching of individual cells such as we find a common phe- 

 nomenon in the galls containing Plasmodiophora and in those 

 formed by the nitrogen tubercle organism of Leguminosae. On the 

 contrary, the stimulus to division is so active that the cells do not 

 have time to attain their normal size. The mechanism of division 

 is a subject for further research. In young, rapidly growing daisy 

 tumors fixed for that purpose we did not find many karyokinetic 

 figures, but in a rose gall numerous double nuclei were observed 

 lying close together in undivided cells. Tourney observed this in 

 almond. In our pure-culture inoculations where several needle 

 punctures have been made close together sometimes only one gave 

 rise to a tumor; sometimes all or nearly all of them developed inde- 

 pendent tumors which fused into one mass as growth continued. 



A study of sections of the earliest stage of tumor development 

 might lead to interesting results respecting the cells first infected. 

 This we propose to undertake. The tremendous proliferation prob- 



213 



