III. Inheritance in Shirleij Pop]))/ 
403 
Another suggestive manner of examining the influence of environment is to 
consider the percentage of certain characters observed in the different crops of 
one and the same season. We first place all the characters together without 
endeavouring to analyse how far the results may be due to differences of standard 
in different observers. 
Now examining this table we see that although the seed for these four crops 
was all taken from a single local crop by sampling fertile individuals at random, 
yet the environment influenced widely the results. This is clear, because in 
counting the number of stigmata or noting if margins existed there was no room 
for sensible influence of personal equation. Luxuriant crops like those at Parkstone 
and Kidderminster had more stigmata, petals, and petaloid stamens than a poor 
crop like that at Arncliffe, or a starveling one hke that at Streatham. Further 
we see that nearly the same order is preserved for the percentages without margins, 
or we suspect that poverty in a crop increases the number of individuals in whom 
the pigment does not reach the edge of the petal. With the other characters the 
results are not so congruous. Arnclifl'e and Streatham, which agree in having 
petals well margined, are different in basal patch frequency, Streatham showing 
more flowers without basal patch than any other crop and Arncliffe fewer. In 
wrinkling the results are not very definite, but indicate that the more luxuriant 
crops have less wrinkling ; while reds appear to be favoured by the bleaker north. 
The chief point to be noticed is that there is immense diversity from one crop to a 
second, and that this diversity is hardly greater in characters like colour and base, 
where personal equation may be influential, than it is in stigmata and number of 
petals, where it must be of small or no account. All the above percentages have 
been deduced from the results weighted with the number of sister plants, so as to 
bring out the influence of change of environment on the results we have to work 
with in measuring sibling resemblance. These great differences between local 
crops of the same season forcibly suggest that individual plant-environment within 
the local crop, depending on subtle and inappreciable differences in soil, aspect, and 
amount of moisture, will in the case of material thus susceptible produce very 
sensible differentiations. We believe that a good deal of the irregularity, and 
possibly the lessened intensity of the inheritance coefficients for plant life, is due 
to this marked susceptibility of plants to small differences of environment. But 
until crops can be grown year by year in the same place, and under nearly similar 
conditions, it will be impossible to measure these conditions, or accurately deter- 
mine their influence on the intensity of inheritance. 
(4) The Influence of Selection in Modifying the Intensity of Inheritance. 
If p be the actually observed con-elation, and r the correlation if all members 
of a race produce offspring, then it is well known that any selection of parents, 
natural or artificial, tends to reduce p below the value of r. If the selection be 
stringent, then p may be so much lower than r that it may appear to have no 
relation to it at all, and fallacious conclusions be drawn as to the variability of 
inheritance for different species or for different characters in the same .species, 
51—2 
