that he "divorced them on the ground of consanguinity."

In the absence of direct investigation by the Census Bureau, or other
public records of consanguineous marriages, perhaps the most promising
field for research is in the genealogical records of American
families. Several thousand volumes of such material have been
published within the last half-century, and a large number of these
are very carefully and scientifically prepared. The material gathered
from such sources is very accurate in regard to the number of births,
youthful deathrate etc., but mental or physical defects are rarely
mentioned. The greatest objection to the utilization of this material,
however, is the amount of labor necessary in order to glean the
desired facts from the mass of irrelevant data. For example, in order
to find one case of first cousin marriage it is necessary on an
average, to examine the records of nearly two hundred other marriages.

The collection of data from personal sources is likewise open to grave
objections. Not only is the informant likely to be biassed, but the
cases which he will remember will be those in which something unusual
has occurred. Herein lay the fallacy in the conclusions of Dr. Bemiss.
I have endeavored to overcome this bias by restricting my requests
for information to genealogists and others who would more naturally
appeal to records, but my efforts have been only partially successful.

The number of cases of consanguineous marriage, embracing all degrees
of consanguinity, which I have collected from these two sources,
genealogies and correspondence, is 723, a number too small in itself
to establish any definite conclusions; but by using this material in
connection with other related data, I trust I may be able to add
something to the comparatively small amount of real knowledge which