III. Inheritance m Shirley Poppu 
405 
Thus a fairly stringent selection of margins took place between the Parkstone 
general population and those members of it who actually provided germinating 
seed, and so daughter-plants at Crewe. But this is not the whole selection. 
Parkstone has for stigmatic bands a low variability, as compared with what we 
know of Shirleys in general. It is extremely unlikely that it has not also a 
reduced variability in extent of margin. How much, however, it is impossible to 
say, for we cannot compare Mr Blanchard's estimate of what is classed as " slight " 
in margin with estimates on other crops by other observers. We have no know- 
ledge of the relative personal equations. Thus the above stringency of selection 
can only be considei"ed as a minimum, /u, in reality is probably considerably lower. 
(c) Number of Petals. If we consider number of petals we find : 
Parkstone General Parkstone Mothers of Kidderminster II. 
Population Crewe Crop General Population 
4 Petcals 758 333 798 
4 Petals and some"! 
petaloid stamens J 
More than 4 Petals ] 
with and without > 125 41 122 
89 21 47 
petaloid stamens ) 
972 395 967 
Or, reduced to percentages : 
4 Petals 78-0 84-3 82-5 
4 Petals and some] 
petaloid stamens J 
More than 4 Petals J 
with and without V 12-8 10-4 12-6 
petaloid stamens ) 
Taking d as the range from normal four petals to the end of the range of 
normal four petals accompanied by petaloid stamens we have, if ap, a', and cr^ 
be the respective standard deviations : 
d = -SeUap = ■2GUa' = •2090(7^.. 
We see that in this case the Parkstone maternal j^lants are an "inverse" 
selection of the general population, being more variable than the general Parkstone 
crop, but they are considerably less variable than a typical crop like the Kidder- 
minster. Thus on the general population of Shirleys there is a stringency of 
selection in the maternal plants of the Crewe crop at least equal to fj, = S and 
possibly more intense. At Streatham even the percentages being 8G'l, 2-3, and 
11-6 respectively, we have d = -llOSa-s, and the stringency of selection of the 
Crewe maternal plants would give /j, as low as -53. But there are all degrees 
of petaloid stamens, and all observers may not have noted equally every slight 
deviation from normality in the stamens of their flowers. 
