W. F. R. Weldon 
303 
of young shells, and the mean value of each of these dimensions, associated with 
a columella!- length identical with the mean found for adults in every half- 
revolution, has been determined. The two sets of values are given in Table II, 
from which it will be seen that the values in both series a.re very close together, 
and although the differences between the corresponding values of tiie perpen- 
diculars are considerable in the lower whorls, they are so irregular that it is 
TABLE II. 
Mean Peripheral Radius and Mean Perpendicular corresponding to identical 
Values of the Golumellar Radius. 
Columellar 
Eadius 
Mean Peripheral Radius 
Mean Perpendicular 
Adult 
Young 
Adult 
Young 
2-2201 
1-9428 
1-9783 
0-5832 
0-5919 
2-7807 
2-4470 
2-4383 
0-6886 
0-6927 
3-4528 
3-0427 
3-0020 
0-8191 
0-8134 
4-1334 
3-6747 
3-6527 
0-9707 
0-9569 
5-0040 
4-4703 
4-4711 
1-1512 
1-1358 
6-0519 
5-4216 
5-4268 
1-3272 
1-3.555 
7-0755 
6-4632 
6-4951 
1-5326 
1-5432 
difficult to regard them as significant. We may therefore say that in young shells, 
from 7 to 8 mm. in length, the mean length of the peripheral radius, or that of the 
perpendicular from the peripheral spiral on the columella, corresponding to a 
given length of columellar radius, is sensibly identical with the mean value of the 
corresponding chai-acter of an adult shell, — or more shortly, the mean spiral is 
sensibly identical in young and in adults. There is clearly no room here for the 
suggestion that selection, or any other process, is changing the mean character of 
the spiral at a rate which produces any sensible effect between one generation and 
the next. Such an identity between the mean character of young and of adults 
was demonstrated for the peripheral radii of C. laminata, and the speedy establish- 
ment of such an identity is to be expected in every local race, if the Law of 
Ancestral Inheritance be well founded. The only selective process, which remains 
to be looked for, is what Pearson has called " periodic selection," by which the 
variability of the race is reduced in every generation during growth, the mean 
character remaining unchanged in this, though not in all such cases. 
The standard deviations of the groups of measurements included in every half- 
revolution studied are given in Table I. ; but these are obviously affected by the 
planes of the sections, and cannot be used in comparing variability. The measure 
of relative variability, which seems most reliable, is that given by a comparison 
between the standard deviations of arrays of one dimension in young and adult 
respectively, corresponding to the same fixed type of either or both of the others. 
