486 
Journal of Agricultural Research 
Vol. XXV, No. n 
Infected plants continued to develop new infections, but there seemed 
to be little or no spread of the infection to other plants not in direct 
contact with infected leaves. 
All of the strains isolated at Washington produced infections not only 
on the original host species but also on other species, and in the numerous 
cross inoculations no difference was observed in the degree of patho¬ 
genicity on the various species. Strains from either red, white, alsike, 
or other species produced the characteristic lesions on any of the species. 
The following clovers were artificially infected: Trifolium pratense L., 
T. pratense var. perenne Host., T. repens L., T. repens var. latum Mc¬ 
Carthy, T. medium L., T. hybridum L., T. incarnatum L., and T. alex - 
andrinum L. 
Lima bean ( Phaseolus lunatus) and velvet bean {Stizolobium deeringia- 
num) were also infected by the Virginia strains of the clover bacteria. 
On lima beans the infection was slight. The lesions were small, reddish, 
with a surrounding white zone and without any translucent tissue. On 
velvet bean the infection was moderate, lesions dark, almost black, cir¬ 
cular to angular, and with translucent borders. 
At Washington many attempts were made to infect soybeans (i Glycine 
hispida). Three varieties—I to San, Wilson Fine, and Black Eyebrow— 
were used in the tests. These were inoculated at various ages and with 
various strains of the clover bacterium, but no infections were ever se¬ 
cured. Similar plants used as controls were readily and typically infected 
with Bact. sojae and Bact. glycineum. The number and the thoroughness 
of these experiments on soybeans give evidence of definite specific dif¬ 
ferences between the clover and the soybean bacteria. 
Two attempts to infect alfalfa gave negative results. 
Further inoculation experiments will be necessary to determine how 
to interpret the differences in pathogenicity observed in the Mississippi 
Valley and the two seaboard stations. It may be that we are dealing with 
specialized races, the one restricted to the red clovers, the other having 
wider host range. It will be recalled that the Wisconsin strain 1920-III 
which was pathogenic to the red clovers in Wisconsin in the 1920 trial 
was found in 1921 in North Carolina to be culturally indistinguishable 
from the strains isolated in that State. It had, however, then lost its 
virulence even for red clover, indicating that pathogenicity is not a fixed 
and constant character with this species. 
The results at all three stations are, however, in full accord so far as 
concerns the nonpathogenicity on soybeans. These failures to infect 
soybean with bacterial leafspot of clover furthermore substantiate the 
cultural studies in showing that the clover organism and the one from 
soybean are specifically distinct. 
In the Washington laboratory comparative cultural tests, made with 
the Wisconsin and the Virginia clover strains, indicate that they are not 
identical in cultural characters. The tests were repeated several times 
and always gave the same results. These variations may be inherent or 
merely due to differing ages of culture. The chief cultural differences ob¬ 
served in parallel tests at Washington are summarized below. 
