by compass, with no trail whatever, we found
ourselves entangled in a swampy valley with tall reeds, from which we had
some difficulty in extricating ourselves. We eventually had to retrace
our steps for six kilometres in order to find an easier way for our
animals. After an examination of the country with my telescope from a
high spot, I decided to go westward across a flat swampy plain of ashes,
sand and water--most troublesome for the mules and horses. They sank deep
into the soft ground and frequently rolled over, damaging saddles and
baggage. One or two of my men had involuntary baths when the animals'
knees gave way under them.

As soon as we had emerged from that wearisome marsh the animals and men
were so tired--although we had only gone 22 kil. from our last camp,
without counting the deviation (28 kil. with deviation)--that I had to
encamp on the bank of the streamlet Fascina, coming from the west. There
we had the laborious task of spreading to dry all the articles that had
got wet--including my bedding, tent, and a quantity of my clothing, which
was not packed like all the rest in air- and water-tight cases.

The stream Fascina flowed into the Rio Furnas and eventually into the Rio
Manso to the north--the latter a tributary of the Cuyaba River. That
region had been rich in Mangabeira (the _Hancornia speciosa_ M.)--a wild
lactiferous plant of much value, producing a fruit called the _mangaba_.

June 16th. Minimum temperature 54 deg. Fahr.; elevation 1,940 ft. On leaving
camp, after a good deal of trouble in recovering our animals in the
morning, as they had strayed in all directions, we found ourselves
travelling along the edge of a large grassy basin (elev. 2,000 ft.)
extending from south-east to north-west, with a wonderful growth of
_burity_ palms; then upon a second basin (elev. 2,100 ft.) with deep
deposits of ashes. We climbed higher, to 2,150 ft., where we found a
third oval _cuvette_ with a surface layer of ashes--merely a continuation
of the preceding _cuvette_. We here resumed our northerly course, going
through what the Brazilians call _chapada_, or high land scantily wooded.

To the south-west we had a high plateau with round natural towers of red
rock, resembling the walls of a fortress. Those red cylindrical towers
stood all along the summit of the range--with immense square blocks of
grey rock above them in horizontal strata. In the centre of that long
range could be perceived a double-tiere