BACTERIAL WILT OF CUCURBITS. I 
counted. These figures were in each case multiplied by 10 to give 
the assumed actual number of beetles in the whole field at each date, 
and these totals were used in plotting the curves. Obviously it 
would be impossible to make these figures absolutely accurate, but 
the counts were carefully made, always by the same person, and were 
checked up by general observations over the field; and there is no 
doubt that the figures are comparable for purposes of plotting the 
curve. 
In the case of the wilt records, each plant showing wilt was marked 
by a stake dated when the disease was first observed, and careful 
records were kept by row, hill, and plant, so that a wilted plant was 
counted once and only once. These records were as accurate as it 
was possible to make them. 
During the summer of 1917, similar careful records were kept in the 
experimental field near Tuxedo, Md. The cucumber and cantaloupe 
spray blocks and the cucurbit variety block were all contained in one 
rectangular field of 1 J acres, and the beetle and wilt records cover this 
field as a whole. In the graphs (fig. 4) made from these figures the 
wilt is expressed in actual number of new cases at each date of obser- 
vation, and the beetle prevalence is given in percentages, 100 per cent 
representing the maximum number of beetles for each brood. 
Similar records also were kept (1916) of beetle and wilt prevalence 
in two experimental fields at Giesboro Point, D. C. (fig. 5; XI, XVI), 
with records over a part of the season in several other fields in the same 
locality; and records for two other fields at East Marion, Long Island, 
were made. 
A detailed discussion of the beetle and wilt records follows: 
In Fields I and II (figs. 1 and 2), East Marion, 1915, the striped-beetle curve shows 
that the maximum covers several days during the last of June and first of July, while 
the maxima of the corresponding wilt curves cover the last few days of July, almost 
one month later. In Field III, though planted only three days after Field I, the striped 
beetles were much later in appearing. Here the maximum of the beetle curve is about 
the first of August and the maximum of the wilt curve the last few days of August, 
almost a month later. Fields I and III were less than a quarter of a mile apart; in 
fact, both were a portion of one larger field, so that the meteorological conditions were 
similar, yet the wilt curves in the two cases were approximately a month apart. In 
Fields I and II the maxima of the wilt curves came just before the greatest rainfall of 
the season, while in Field III the reverse was true. No definite relation between the 
wilt and the temperature or rainfall could be detected in any case. Thus in these 
three fields wilt prevalence bore a clear and definite relation to striped-beetle preva- 
lence rather than to weather or to time of planting. 
At East Marion (L. I.), N. Y., in 1916 (figs. 1 and 3) the maximum of the striped- 
Ibeetle curve for the first brood occurred about June 26, while for the corresponding 
Fields, II and Ila, the wilt-curve maximum occurred, respectively, around July 27 
and 29, about one month later. As shown in the graph, downy mildew {Pseudo- 
peronospora cubensis) appeared early in August and during the rest of the month, 
while the summer brood of striped beetles was making its appearance, gradually 
destroyed the vines in our experimental fields. No evidence of another brood was 
