metres--of the birthplace of the Rio Novo, the head-waters of the
River Arinos, of the Rio Verde (Green River), and of the several sources
of the Rio S. Manoel or das Tres Barras, or Paranatinga; and not distant
from the sources of the great Xingu River.

The Serra Azul, extending from west to east, was interesting
geographically, not only because it marked the northern terminus of the
highest terrace of the great central plateau, but also because from it or
near it rose two of the greatest rivers of Central Brazil--the Xingu and
the Arinos (Tapajoz), the latter the most central and important river of
Brazil, crossing the entire Republic from south to north, as far as the
Amazon.

On June 21st we crossed the Piraputangas (elev. 750 ft. above the sea
level), where, owing to the steep banks, we had much difficulty in taking
mules and baggage to the opposite side. We then proceeded across another
large plain, skirting the spurs of the Serra Azul. Nine kilometres from
camp we came to a stream 80 metres wide, which flowed from north-east to
south-west. It had an average depth of 11/2 ft. It was, I think, the Cuyaba
Grande.

It was not easy to identify those rivers, as the existing maps of that
country were absolutely worthless, most of them being filled in with
fancy mountains and rivers, which either did not exist at all or were
sometimes hundreds of kilometres out of their position. There were
frequently mistakes of two, three, and more degrees in the latitudes and
longitudes even of important places. As for the tributary rivers, of
which merely the mouths were known and named, they had supplied good
material for the imagination of more or less artistic cartographers in
order to fill in the rest of their course. Even the German map and the
American maps of the International Bureau of American Republics, which
were the two best, were extremely inaccurate in their representation of
that region. For instance, the latter map--and nearly all the other
maps--placed the Serra Azul some 180 or 200 kil. south of its actual
position. The German map was some 70 kil. out. The Serra Azul could be
seen from a great distance, and had been marked approximately and not by
actual observations on the spot. Nor, of course, had the tributaries of
the Cuyaba been explored or even seen except at their mouths; hence their
imaginary courses.

[Illustration: A Street of Diamantino.]

[Illustration: The Dogs of the Expedition.]

Considering h