ts failed to report the deaths. From the fact that a
comparatively large percentage of these were reported as defective, we
should expect a higher death-rate than among the unbiased genealogical
cases.

Dr. Bemiss found a very high death-rate among the children of
consanguineous marriage, due partly to the fact that his cases were
reported by physicians. He reports that of the offspring of marriages
between first cousins and nearer relatives, 23 per cent "died young;"
of the offspring of more remote consanguineous marriages, 16 per cent;
and of non-related marriages 16 per cent. There is, therefore, a
strong indication of lowered vitality as a result of consanguineous
marriage.

A determination of even the approximate percentage of degenerate
offspring resulting from marriages of consanguinity by direct inquiry
is exceedingly difficult. The average human mind is so constituted as
to exaggerate unconsciously the unusual in its experience. Herein lies
the fallacy in the work of Dr. Bemiss. His material was "furnished
exclusively by reputable _physicians_ in various states," and of the
3942 children of consanguineous marriages in the cases thus furnished
him, 1134 or 28.8 per cent were in some way "defective." Of these, 145
were deaf and dumb, 85 blind, 308 idiotic, 38 insane, 60 epileptic,
300 scrofulous and 98 deformed. It is evident that a physician in
reporting such data to a physician would naturally give cases in which
something pathological existed. Even if there were no conscious bias,
such cases would be the ones with which a physician would be most
likely to come in contact. Dr. Bemiss himself recognized the
possibility of this bias. To quote him:

It is, natural for contributors to overlook many of the more
fortunate results of family intermarriage, and furnish those
followed by defective offspring and sterility. The mere
existence of either of these conditions would prompt inquiry,
while the favorable cases might pass unnoticed. Contributors
have been particularly requested to furnish without prejudice
or selection all instances of the marriage of consanguinity
within their various circles of observation, whatever their
results.[51]

[Footnote 51: Bemiss. see _Trans. of Am. Med. Asso._, vol. xi, 1858,
p. 323.]

Yet he does not seem to believe that this bias seriously affects his
conclusions.

In order as far as possible to avoid this bias, I sent my own
circulars to