450 
BROOKLYN BOTANIC GARDEN MEMOIRS 
and other differences exist; otherwise, I 
speak only of “isolations.” There are cer- 
tainly two of these crown gall strains, and 
probably many. 
(3) Feeble and virulent strains exist in nature, 1. e., there 
1s variation in virulence of colonies from the 
same source—hop, carnation, rose, sugar beet, 
etc. The author believes that the crown 
gall bacteria not only lose virulence on cul- 
ture media but also in the tissue of the gall. 
Examples are: (a) of three colonies selected 
as typical from plates poured from a hop 
tumor in 1910 only one proved infectious; 
(b) of six colonies plated from a witch broom 
on carnation, all of which looked alike and 
typical for crown gall, only one was found 
to be able to cause tumors when inoculated; 
(c) of seven colonies selected as typical from 
a plate poured from a rose gall only three 
proved infectious, and of these two were 
actively infectious, while the other was only 
feebly so; (d) of five colonies selected as 
typical from a plate poured from a tumor on 
Pelargonium none proved infectious, al- 
though in advance we felt quite sure of all 
of these colonies; (e) of thirty colonies 
selected from plates poured from natural 
tumors on sugar beets only five were infec- 
tious and all feebly so (Bul. 213, pp. 192-194 
and Plate XXXVI). 
The first and natural supposition when a 
culture has lost virulence is that some in- 
truder has displaced the right organism; and 
when only certain colonies on a plate are in- 
fectious, that the others are intruders how- 
ever much they may resemble the right 
organism. I cannot say that we have abso- 
lutely excluded this hypothesis, to which I 
held strongly in the beginning, but we are 
gradually coming to believe that it does not 
explain all the facts. 
II. Some other types of plant galls. 
(1) Nonparasitic intumescences. These can be produced in 
several ways: by overwatering and exposure to 
