When we were ready to continue our journey on the second morning, we discovered that the Nhambiquaras in departing had taken two of the dogs with them. Colonel Rondon spent some hours hunting for the Indians, but their start was too long and he could not come up with them. I regretted heartily that they had not taken all of the dogs, as they were a mongrel, worthless lot; they were of no assistance in hunting, nor did they watch camp. On the contrary, they brought fleas and ticks into the tents, insisted on eating and drinking out of our dishes, and consumed quantities of food that might have been used to better advantage later. The country beyond the Juruena is somewhat rolling, but there is no appreciable change in the vegetation. We rode twenty kilometres the first day, camping on the banks of the Rio da Fomiga (February 10). Next day we travelled but twelve kilometres, reaching the Jurina, a shallow though rapid stream six hundred feet wide; the crossing was slow and laborious, as there was only one very small _balsa_ or ferry. Camp was pitched on the banks of a small stream a league beyond. Near by were several deserted thatched huts and the comparatively new graves where two Brazilian soldiers and one army officer had been buried. The Indians had killed them, and interred them in an upright position with the head and shoulders protruding out of the ground. The following night, on the Rio Primavera we saw two other graves; the men who were buried there had been slain while asleep in their hammocks. This was the most dangerous part of the whole Nhambiquara country. When we reached a place called Mutúm Cavallo in the afternoon of the 15th, the mules Kermit and I had ridden were so tired that we decided to give them a day’s rest; that meant walking to the next camping-site, and rather than undertake the long journey during the hot hours of the next day we planned to start immediately after supper. There was still some time to spend, however, so we went about our work as usual. An army of ants was foraging near the tents; they had discovered a large, hairy caterpillar, but the half-inch long “bristles” with which it was covered protected its body from the onslaught of the marauding host. The ants, however, were not to be deterred from their purpose; they made repeated rushes at the caterpillar, clipping off a bit of hair each time they struck. After continuing these tactics for twenty minutes, a small patch of the plump insect’s body had been cleared of hair, and one ant got a good hold with its vise-like mandible. The caterpillar, upon feeling the pain, promptly began to wriggle, thus exposing its unprotected under-side, and the ants immediately rushed at that vulnerable part and soon succeeded in overwhelming their victim.