114 Self-Incompatibility in Hermaphrodite Plants 
gives the number of flowers per head expressed as a denominator and the 
number of seed obtained as the numerator ofa fraction. (Such data show 
that the variations in number of seed per head are quite independent of 
the number of flowers per head.) Data for No. 33 give comparative results 
of hand and autonomous self-pollination, which are typical of the evi- 
dence that careful self-pollination by hand does not increase self- 
fertility over autonomous selfing. 
-The data for No. 50 include for comparison the results of the controlled 
cross-pollinations of 10 heads, and are quite typical of the abundant 
evidence obtained which shows that feebly self-compatible or strongly 
self-incompatible plants are highly productive of seed to compatible 
crosses made on the same dates of bloom. 
Series R 12-11-16-— . All of the 103 seed obtained from the self- 
pollinations of the parent were planted. Sixty-nine germinated, but 
possibly as many as 20 embryos had been injured at the time seed was 
examined to determine if embryos were present. Of the 60 plants 
tested for self-fertility, 16 were self-incompatible and 44 were self-com- 
patible. 
: The data given for No. 37 in Table III include results of self- 
pollinations by hand for 10 heads, and for all heads the numbers of 
flowers and seeds are expressed in the form of a fraction. 
General summary. Table IV presents a grand summary of the 
various crops of the variety red-leaved Treviso with the pedigree 
indicated and the distribution of the individual fertilities given. As 
there shown the variety was kept in culture for the first three years 
(1914, 1915, 1916) by crossing self-incompatible plants. The first plant 
found to be self-fertile was one of the 1915 crop but it was very feebly 
self-fertile and the two offspring -grown from its seed were self-incom- 
patible. Among the 1916 crop grown from self-incompatible parentage, 
there were 11 plants which were self-fertile to some degree and of these 
4 were highly self-fertile. ve 3 
The subsequent generations descended from self-fertile plants, and 
the study of fertilities in them constitutes a test for the heredity of self- 
compatibility that arose sporadically after three generations of self- 
incompatible parentage. Three lines were grown in the J,, and two 
families were continued into the J,. 
The proportion of plants that were self-fertile was decidedly increased 
in the J, and was still larger in the J,. In both families of the J, there 
was also an extension in the range of self-fertilities. 
_ The two families grown into the J, exhibit some differences in self- 
