18 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY. 
this point the road enters a heavily wooded district of moderate 
dissection with flat-topped interstream areas. We camped for the 
night near the road 2.3 miles south of a small settlement called Sepul- 
tura. 
August 18th.— It was so cold at midnight that one of the camara- 
das made coffee with which we were served. At this hour a pan con- 
taining water was coated with a thin sheet of ice. From this camp 
to the Rio Contagem the route traverses a soft sandstone formation. 
Just before halting for breakfast at 10 a. m. we passed a shaded mud- 
puddle covered with a thin film of ice. These occurrences of frost and 
ice I note particularly because they were so contrary to my precon- 
ceived notions, based upon the imperfect accounts in North American . 
geographies of the winter climate of even this elevated region in south 
Brazil. Apparently horizontal beds of deeply weathered sandstone 
continued to form the surface rock as far as camp No. 2 on the head- 
waters of a small stream, the Rio Sao Joao, a north flowing branch 
of the Rio Negro. The gentle descent from the plain to camp lay 
through a forest of Araucaria, tree-ferns, and a graceful bamboo with 
long curling tips. 
August 19th— The night was cold again, with frost on the ground 
and ice in the camp dishes at an early hour this morning. Shortly 
after leaving camp we traversed a plain with large areas of dead brakes, 
scattered palms, and tree-ferns. Araucaria is however the most 
abundant forest tree; there are large tracts of it, and many young trees 
are in evidence. For over a kilometer the path led through a dense 
growth of tree-ferns, taquara or bamboo, Araucaria, and large imbuias 
(Canella imbuia of Brazilian writers) with numerous mud-holes 
crossed-ridged with the pildes made by mules, through which our 
train of animals wallowed with extreme slowness. This jungle 
suddenly gave way to a wagon road on good ground, the sign of 
approach to some German or Polish settlement, of which however we 
saw nothing. Along this route we came upon a fine example of the 
tall Brazilian sassafras tree (Nectandra cymbarum), standing alone 
in the forest. Descending through swampy places of the wet-wood 
type we entered another forest of Araucaria. Tree-ferns partially 
hidden on the edge of these forests or standing in the more or less open 
growth of tall trees were in sight most of the day, but no rock ex- 
posures were seen and no pebbles bestrewed the trail which seems to 
have lain over decomposed shales. Here and there in the valley 
bottoms two or three huts were encountered, around which were pigs, 
a few sheep, cattle, and horses. We camped beside a stream on the 
soft spongy ground of tree-fern growth in the forest. 
