persons in that same degree. Particular inquiries were made on the
latter point, but, as already stated, the answers were incomplete.
There is, however, enough information to justify three conclusions of
primary importance to the present inquiry--namely, the _average_
number (1) of brothers of the subject, (2) of brothers of his father,
and (3) of brothers of his mother.

The number of Fellows to whom circulars were addressed was 467. The
number of those who gave useful replies was 207, a little more than
one-half of whom sent complete returns of the numbers of their
brothers and uncles; some few of these had, however, placed a query
here or there, or other sign of hesitation. As the number of
completely available returns scarcely exceeded 100, I have confined
the following tables to that number exactly, taking the best of the
slightly doubtful cases. It would have been possible, by utilizing
partial returns and making due allowances, to have obtained nearly
half as many again, but the gain in numbers did not seem likely to be
compensated by the somewhat inferior quality of the additional data.

The first three lines of Table V. show that there is no significant
difference between the average numbers of brothers and sisters, nor
between those of fathers' brothers and fathers' sisters, nor again
between those of mothers' brothers and mothers' sisters; nor is there
any large difference between those of male and female cousins, but it
is apparently a fact that the group of "brothers" is a trifle smaller
than that of uncles on either side. It seems, therefore, that the
generation of the Subjects contains a somewhat smaller number of
individuals than that of either of their Parents, being to that
extent significant of a lessening population so far as their class is
concerned.