which had taken a name before the division of Teuton and Italian. Some
of the members of that family may have joined the band of which came the
Goths, while other members joined the band of which came the Romans.
There is no difference but the length of time to distinguish such a
supposed case from the case of an English family, one branch of which
settled in the seventeenth century at Boston in Massachusetts, while
another branch stayed behind at Boston in Holland. Mr. Sayce says truly
that the use of a kindred language does not prove that the Englishman
and the Hindoo are really akin in race; for, as he adds, many Hindoos
are men of non-Aryan race who have simply learned to speak tongues of
Sanscrit origin. He might have gone on to say, with equal truth, that
there is no positive certainty that there was any community in blood
among the original Aryan group itself, and that if we admit such
community of blood in the original Aryan group, it does not follow that
there is any further special kindred between Hindoo and Hindoo or
between Englishman and Englishman. The original group may not have been
a family, but an artificial union. And if it was a family, those of its
members who marched together east or west or north or south may have had
no tie of kindred beyond the common cousinship of all.

Now the tendency of this kind of argument is to lead to something a good
deal more startling than the doctrine that language is no certain test
of race. Its tendency is to go on further, and to show that race is no
certain test of community of blood. And this comes pretty nearly to
saying that there is no such thing as race at all. For our whole
conception of race starts from the idea of community of blood. If the
word "race" does not mean community of blood, it is hard to see what it
does mean. Yet it is certain that there can be no positive proof of real
community of blood, even among those groups of mankind which we
instinctively speak of as families and races. It is not merely that the