86 = Self-Incompatibility in Hermaphrodite Plants 
descent. There are all grades in the degree to which both self-compati- 
bility and cross-compatibility may appear, and even the reciprocal 
matings of two hermaphrodites may give opposite results. 
It is clear, as it has been since the time of Darwin at least, that 
within the species compatibility and fertility, both self and cross, is the 
rule, and represents-the primitive condition, and that incompatibilities 
are the special cases. Cross-sterility within a species is a relatively 
rare phenomenon, often accompanied by morphological modifications 
in the sex organs such as dimorphism. Self-incompatibility is more 
common, and has evidently a selection or hereditary value which leads 
to its continuation and intensification whenever it appears. 
The very wide distribution of homomorphic species Showing self- 
incompatibilities (and also cross-incompatibilities to some degree) among 
the families of flowering plants (Gramineae, Liliaceae, Papaveraceae, 
Cruciferae, Rosaceae, Pomaceae, Amygdalaceae, Solanaceae, Scrophu- 
lariaceae, Cichoriaceae, and Compositae, for example) suggests that the 
condition has arisen many times independently, and may be one of 
fundamental significance. The generally accepted view that inbreeding 
and continued self-fertilization are of themselves injurious is certainly 
- favoured by the evidence that self-incompatibility is so obviously a pro- 
gressive character. The application of this doctrine in these cases is 
to some extent subject to experimental test, for it would seem that a 
species whose members exhibit wide variation in self-compatibility affords 
opportunity to test whether continued selfing is of itself injurious. 
Furthermore the intensive study of the fertility of individuals that - 
are feebly self-compatible gives positive evidence as to how the condition 
of self-sterility has arisen in species originally self-fertile. The marked 
fluctuation of the character in practically all self-sterile species is 
especially illuminating on this point. 
thn continuing my experimental studies (1916, 1917, 1918) of physio- 
logical incompatibilities further evidence has been obtained in support 
of the following general conclusions. 
L In all so-called self-incompatible species investigated, with the 
exception of Raphanus sativus and the dimorphic species Linum grandt- 
florum (which, however, have not been so extensively studied), self- 
incompatibility and self-compatibility are highly variable in their be- 
haviour ; although the variations are generally indiscriminate, there are 
ee 6 marked correlation with conditions of age or vegetative vigour 
operating during the period of bloom, giving such results as “end- 
season ” and “ mid-season ” self-fertility. 
