554 
Journal of Agricultural Research 
Vol. XI, No. ii 
pens that the least and most damaged rows stood side by side near the 
middle of the field. If location was a factor, adjacent rows should, on the 
average, show a closer agreement in the extent of damage than pairs 
taken at random. On eliminating Phi29, which is the reciprocal of 
Phi 30, the correlation between adjacent rows for the upper ears alone 
is 0.161 ±0.189 and for upper and second ears combined 0.173 ±0.189. 
There is, thus, little or no tendency for adjacent rows to be damaged to a 
similar extent, and position in the field seems not to be an important 
factor in causing the observed differences. 
On the other hand, if the immunity which some of the rows enjoyed is 
due to plant characteristics, there should be some agreement between 
the degree of immunity of upper and second ears of the same row. A 
comparison of the value in columns 2 and 3, Table II, shows such an 
agreement. The row in which the upper ears were most damaged is also 
the row in which the second ears were most damaged. It can also be 
seen that the row with the least damage to the first ears has a very low 
damage in the second ears. Beyond these rather outstanding cases, the 
agreement or lack of agreement is not obvious from inspection. The 
correlation of damage between upper and second ears is believed to be a 
fair measure of this agreement; it was found to be o.7i8±o.o87. This 
correlation alone would seem to establish the fact that the individu¬ 
ality of progenies is an important factor in determining the extent of 
damage. From the 1915 results we may safely oonclude that there is 
something about the plants descended from certain ears which affords 
them an appreciable measure of protection. The next step was to deter¬ 
mine, if possible, whether this protection could be referred to any of the 
recorded plant characters. 
As soon as a detailed analysis of the 1915 data was attempted it became 
evident that there were sources of error that had not been adequately 
guarded against. The ears were not harvested at a uniform stage of 
development, and those left longer were more severely damaged. There 
were also many ears bagged to secure pure seed, and since the bagging 
was in a measure selective, it introduced another source of possible error. 
These disturbing factors made it appear unwise to place confidence in any 
detailed analysis of the 1915 data. These results will therefore be con¬ 
sidered only in connection with the results of the following season. 
EXPERIMENTS IN 1916 
The 1916 plantings were made at Lanham, Md., on May 14 and con¬ 
sisted of 35 rows, as follows: A repetition of the 14 progenies grown in 
1915 (ancestry described on p. 551), 9 progenies from ears secured by hand 
pollinations within the rows of the progenies grown in 1915, 8 progenies 
from ears obtained by crosses between the rows in 1915, and 2 first- 
generation crosses between 1915 progenies and Hopi maize. There were 
