38 BULLETIN 727, U. S. DEPARTMENT OF AGRICULTURE. 
extending into a private cucumber patch. On August 17 the course 
of this drainage channel was found to be marked by new infection 
in at least three rows and by a second old center. There was also 
much scattered new infection in the region of the delta, both in field 
3 and in the private patch as well. 
In field 4 the rows were parallel to the very slight western slope. 
On August 18 a spread of even-aged secondary infection to the 
westward from four of the six centers was noted. Examination of 
the anthracnose centers in field 5 on August 18 showed a downhill 
spread of infection across the rows in three instances. In two cases 
me = ow 
QFE CTION OF 
BLOFLZ 
Fic. 13.—Diagram of field 2, Madison, Wis., August, 1916, showing the relation of surface drainage to the 
distribution of anthracnose. The cucumber rows are indicated by broken lines and the slope by the 
4-foot contour intervals. The four trapezoids represent water traps 1 to 4; the black circles and ellipses, 
original centers of anthracnose; and the shaded areas, new infection noted on August 12. 
this infection extended across two rows downhill (14 feet), and at the 
same time into the adjacent row above. 
_ Among the Madison fields that season, anthracnose became prev- 
alent only where there were slopes. In field 19 on level muck soil, 
there was very little spread from the old centers. In a large water- 
melon field near Quitman, Ga., infection was found. to be wide- 
spread on certain slopes, but still restricted to small areas in the 
more level parts. 
Late in the season of 1916 an excellent opportunity was afforded 
in field 1 to observe the results of the heavy rainfall of the first week 
