CHARACTER OF THE TUMOR. 105 



successful inoculations does not suffice to produce this quasi immunity, 

 but several are required. Even then it is possible b}" inoculating with 

 a more virulent strain to induce tumors on these plants, although as 

 far as our observation goes, the tumors are slower to appear and gen- 

 erall}^ smaller and less vigorous in their growth than on check plants 

 (p. 177). Spontaneous recovery from the disease is quite frequent. 



A fourth likeness is the tendency" of the disease to appear in callous 

 or scar tissue, e. g., on pruned roots or at the junction of stock and 

 graft. This appears to be rather more than mere presence of wounds 

 in these places. The wounds must probably exist, but the softer, 

 modified character of the new tissue appears to invite both wounds 

 and infections, just as it also invites secondary fungous and bacterial 

 infections. 



A fifth resemblance consists in the marked tendency of the galls to 

 return after excision. 



Sixth, the fact frequently observed 1)}" us that on agar poured 

 plates made from tumors, especially those of some age, certain 

 colonies which look like those of the right organism and which behave 

 properly when transferred to peptonized beef bouillon either do not 

 produce the overgrowth when inoculated into susceptible plants, or 

 yield only very slow-growing, feeble, soon stationary Iiyperplasias, 

 requires explanation and may be mentioned here. At first these 

 results were interpreted by us as meaning accidental presence in the 

 tumors of organisms resembling Bacterium tumefaciens on agar, etc., 

 but unlike it in other respects. More recently we have come to the 

 conclusion, or rather formed the working hypothesis, that these per- 

 plexing colonies, or at least some of them, must be nonvirulent 

 strains of the gall-producing organisms, not other species. YVe do not 

 know what constitutes virulence, but we do know that on culture 

 media many organisms gradually lose this property. Bacterium tume- 

 faciens being one of them. The question then arises: Why should 

 not virulence often disappear from organisms buried inside the 

 tissue of tumors ? And is not the fact that the tumor has ceased to 

 be active and the host has gained the ascendency evidence of this ? 

 It is certain!}^ conceivable that either by the juices of the host or 

 through their own by-products the bacteria might be so acted upon as 

 to lose power to infect other plants when cultivated out, and this 

 without losing their common cultural characters. The same phe- 

 nomenon is believed to occur in cultures of the organism causing root 

 tubercles of Leguminosae. 



It is believed by us that we have here the beginning of a solution 

 of the cancer problem in men and animals, or at least a most instruc- 

 tive border-line field. The chief objection raised by animal pathol- 

 ogists with whom we have talked to considering these tumors in the 



213 



