146 
Journal of Agricultural Research 
Vol. XXI, No. 3 
infected, while some beds where no disease occurred in 1918 showed the 
highest percentage of infection in 1919. These observations, however, 
are complicated by the fact that different varieties, the susceptibility 
of which is not known, were planted the second year. From present 
knowledge it seems that the disease must be carried over on the root¬ 
stocks, especially since the trouble begins to develop before the plants 
are taken from the hothouse. 
Overwatering of the foliage appears to be a large factor in the develop¬ 
ment of the disease, since rootstocks which were taken from heavily infected 
clumps and kept under favorable conditions during the winter and care¬ 
fully watered when potted in early spring gave not a single case during 
the entire summer, while rootstocks from the same source without special 
care showed a large percentage of infection before setting-out time. 
GEOGRAPHICAL DISTRIBUTION 
The disease has thus far been observed only in the District of Columbia 
and in Illinois. Typically infected plants were found at Urbana, Ill., in 
the summer of 1920. 
ISOLATIONS 
When sections were made of diseased tissues, motile bacteria were 
invariably found in great numbers swarming out on the slide. Plates 
were poured on peptone-beef agar from leaf spots, pedicels, and from 
petioles near the ground level. In every case practically pure cultures 
of a white bacterial organism were obtained. Judged from plate colo¬ 
nies, the organism was the same in every case, and comparative cultural 
studies of several isolations have corroborated this judgment. 
INOCULATIONS 
Inoculations were made in 1918 on potted cannas in the hothouse. 
These plants were not in good condition but were the only ones available 
at the time. Young leaves just unfolded were inoculated by placing 
drops from a young agar slant culture on the surface and making deli¬ 
cate pricks in the blade through these. Drops were also poured into the 
tips of tightly rolled leaves without wounding. Subcultures of isola¬ 
tions from leaf blades and from petioles were used. Some plants were 
kept moist for 36 hours by spraying in cages with sterile water; others 
were left in the open hothouse. 
In most cases no infections appeared. One inoculated leaf showed 
on the fourth day several water-soaked streaks 1 to 10 mm. long run¬ 
ning from needle pricks. These turned yellow then brown, but did not 
spread further. Plates were poured from the edge of the longest streak, 
and colonies were obtained which appeared to be right. Transfers from 
these conformed to the original isolation in subsequent cultural tests. 
Attempts were made to infect the slow-growing cannas with the re¬ 
isolation recorded above, but all of these failed. Further inoculation 
