st have once existed, joining the
promontory on which I stood to the great mass of prismatic red volcanic
rock to the west of us, and ending in a flat triangle with a wide base.
The surface soil on the height of the peninsula was of spattered lava and
black broiled rock and pellets.

The bottom of the abyss formed two sweeping undulations--the second from
the centre much higher than the first--seemingly a great wave of lava
vomited by the crater, by which probably the destruction of the wall
joining the peninsula had been caused.

To the S.S.E. in the distance stood a high mountain range--or rather a
great flat-topped plateau of delicious cobalt blue shades, almost losing
itself in the sky. To the east, completing the circle, were two other
great spurs of red-baked rock, with precipitous, almost vertical, sides
and with much-striated buttresses that ended in conical mounds--eroded
into that shape by the action of water and wind.

To the south, beyond, a sloping table-land with a pronounced dip eastward
extended from east to west. It towered over everything, and was shaped
like a trapezium. In front of this sloping table-land was another long
flat-topped range, stretching from E.S.E. to W.N.W. Again in front of
this, could be seen an interesting series of prismatic mounds--like
parallel barriers. To the S.S.W. rose a large mountainous mass--another
plateau. Then came a second range, cut into clear pyramids with
rectangular bases, and, beyond, a great expanse of lovely green with some
large mounds of a similar shape to those already described. Two more
pyramids were also to be observed far, far in the distance, while others
of a slightly less angular shape were noticeable upon the great flat
stretch due west.

Right under us, at the bottom of the precipice, was thick forest
covering, zigzag fashion, the two depressions, roughly in a general
direction of south-east to north-west. Those two depressions drained that
immense basin. It was there that the streamlet Caxoeirinha had its birth.
The Caxoeirinha flowed north-west and fell into the Ponte de Pedra River,
which flowed south. Those two streams, with a number of others, formed
the head-waters of the great S. Lourenco River, a formidable tributary of
the Rio Paraguay or Parana.

An extraordinary effect of clouds could be seen that day, and a similar
occurrence I saw on many other occasions upon the table-lands of Matto
Grosso. The clouds reproduced--upside-down--the