BACTERIAL WILT OF CUCURBITS. 27 
and the resulting fraction reduced to a decimal. In making up the 
relative-virulence graphs these fractions were used in the abscissae 
to express varying degrees of relative virulence between possible 
instant death (unity) and no infection at all (zero). For example, 
one isolation gave (1) incipient wilt in 4 days, (2) wilt of 
the inoculated leaf in 5 days, and (3) wilting of the whole plant 
in 11 days after inoculation. Added together, these three figures 
equal 20, the number used as the basis of comparison. The expres- 
sion 1 over 20 reduced to a decimal equals 0.05, which is the figure 
taken as the final index expressing the degree of relative virulence 
for this particular isolation of Bacillus traclieipJiilus. This happens 
to be one of the most virulent isolations found thus far in our study. 
The relation in number of days between 1 and 2 and between 2 
and 3 tended to be similar in the different isolations tested. That is, 
if the first signs of wilt appeared late a longer time usually elapsed be- 
fore complete wilting than when the first signs appeared early. How- 
ever, this rule was not strictly true in all cases, so it was felt that a 
truer comparison could be made by adding together the three cardinal 
points above mentioned rather than by taking any one of them alone 
as the basis for comparison. 
In the graphs the ordinates show the number of isolations found by 
numerous inoculation tests to fall at any particular degree of relative 
virulence. In plotting the curves the average virulence index for all 
tests with any one isolation was in each case taken. 
One hundred and three different isolations were tested as to rel- 
ative virulence in Chicago Pickling cucumber, and this list included 
isolations from cucumber, squash, and cantaloupe. Referring to the 
graph (fig. 6-B), it will be seen that most of the isolations are extremely 
virulent. However, a few isolations worked very slowly, and one very 
weak isolation never caused anything more than incipient though 
undoubtedly true bacterial wilt. It will be noted that most of the 
isolations (84 in number) fall above the virulence index 0.033, and it 
may be added that all the isolations from squash are within this group. 
As a supplementary test, 62 of these isolations from cucumber, 
squash, and cantaloupe were further tested as to relative virulence 
when inoculated into White Bush Scallop squash, the method used 
being the same as in the cucumber tests above described. Reference 
to the graph (fig. 6-A) shows the opposite kind of a curve from that 
obtained with cucumber as indicator. Only 12 out of the 62 isolations 
tested caused wilting of entire squash plants, and a lower degree of 
relative virulence was shown here than when these same isolations 
were tested by inoculation into cucumber. Of these 12 highly viru- 
lent isolations it happens that 6 were originally obtained from cucum- 
ber and 6 from squash. Eighteen other isolations tested on squash 
exhibited various shades of virulence from incipient wilt only up to 
