a, Germany,
Portugal, and a few other states taken together--and most of which was
little known or absolutely unknown--it was not easy to produce an exact
census.

During my journey, which crossed that immense country in a zigzag from
one end to the other in its broader width, and covered all the most
important regions of the Republic, I became assured that few indeed were
the pure Indians to be found in Central Brazil. One went hundreds and
hundreds of miles without meeting signs of them; and that in localities
where they were supposed to be swarming. The Bororos--a few dozens of
them, all counted, in two or three different subdivisions--were perhaps
the strongest wild tribe in all the immense State of Matto Grosso.

As I have said, I was greatly impressed, from my first contact with the
Bororos, by the strongly Polynesian appearance of some of them. The more
specimens I saw of them the more I became convinced that they were of the
same race. In fact, more: I began to speculate whether the people of
Australia and Polynesia had migrated here or whether it was just the
other way--which theory might also be plausibly upheld--viz. that the
people of Central South America had migrated to the west, into Polynesia
and Australia. Many theories have been expounded of how races always
follow certain rules in their migrations, but in my own experience I do
not invariably find that those theories are always correct. Again, it
does not do to rely too much on the resemblance of words in establishing
a relationship between two or more races. Nor, indeed, can one trust
absolutely to the resemblance in the rudimentary ornamentation of
articles of use. If you happen to be a student of languages, and have
studied dozens of them, you will soon discover how far words will travel
across entire continents. They can often be traced back to their origin
by the knowledge of intermediate languages through which, with
distortions, those words have passed. In Central Africa I actually heard
words of Mongolian origin, and not only that, but even traced Mongolian
characteristics in the type of the ruling classes of natives, as well as
in the construction of their language.

It is easy to be occasionally misled. I remember on my journey across
Africa how amazed I was at first at hearing some Tonkinese expressions
used by the native cannibals. I really could not get over my amazement
until I learnt that some years previously a number of Tonkinese co