WOODWORTH: GEOLOGICAL EXPEDITION TO BRAZIL AND CHILE. 27 
surmounted the dissected trap plateau at an elevation of some 700 
feet above the valley of the Rio Ponte alto. The broad valley of the 
Rio Cachoeiras which next succeeds exposes a bed of sandstone in its 
banks below which lies a trap sheet over which in turn it flows. About 
ten miles north of the river we camped for the night near a frog pond 
between two of the mule bridges which mark the approach to Coryti- 
banos. 
August 31st.— An hour’s ride from camp brought our cavalcade 
again to Corytibanos, at which point we took the road northwest 
towards Sao Joao en route to Porta da Uniao on the Iguasst. For 
several miles along this route we traversed campos with a scattered 
growth of the the araucarian pine and then of the Butia palm. One 
deeply weathered trappean hill bore numerous lakelets or lagoas 
bordered by tall tufts of grass. Blackened heads of trap rock 
cropped out over the surface of the inosculating ridges which separate 
these solution-basins. At noon we halted beside a small stream whose 
milky waters were without apparent cause since it had not rained. 
On resuming our march we shortly heard the roar of falls in the course 
of the Rio Marombas on the left in the forest where the river tumbles 
over the bedded traps. The river was crossed on a balsa or ferry, a 
sort of raft supported by four or five wooden canoes. (Plate 17). 
A forest fire was burning on the east bank of the river. After travers- 
ing two deep valleys and a broad hill of dry campo we camped on 
another small milky stream in a cattle country. The milkiness of 
these small streams was possibly due to cattle wading in them. The 
day was warm, and beetles, butterflies, and dragon-flies were out 
fluttering over the muddy flats of a stream at noon. 
September 1st.— About an hour and a half after leaving camp this 
morning, we came to the Rio Correntes which was forded on a trap 
bottom. After proceeding for an equal length of time we crossed 
another broad shallow stream with rapids, the Rio dos Patos, also on 
trap. We followed up the right bank of this river in a northwest 
direction until nearly noon when we came to a coboclo hut in the 
forest, where a woman gave us directions. Our halt for breakfast at 
the noon hour was beside another small sluggish milky stream. 
Several of the solution-basin lakelets were passed during the fore- 
noon. As the trail was not distinctly marked from the paths leading 
from one hut to another in this forested region, much of the after- 
noon was spent in seeking directions. A large forest fire at one time 
was raging to the west of our position. Fires in the standing clumps 
of dry dead bamboo burn with great rapidity. At length we entered 
