Page:Through the Brazilian Wilderness by Theodore Roosevelt (1925).djvu/113

 with its fore legs off the ground. Their long ears were very prominent. The dogs raced at them. I had always supposed that armadillos merely shuffled along, and curled up for protection when menaced; and I was almost as surprised as if I had seen a turtle gallop when these two armadillos bounded off at a run, going as fast as rabbits. One headed back for the nearest patch of jungle, which it reached. The other ran at full speed—and ran really fast, too—until it nearly reached the other patch, a hundred yards distant, the dogs in full cry immediately behind it. Then it suddenly changed its mind, wheeled in its tracks, and came back like a bullet right through the pack. Dog after dog tried to seize it or stop it and turned to pursue it; but its wedge-shaped snout and armored body, joined to the speed at which it was galloping, enabled it to drive straight ahead through its pursuers, not one of which could halt it or grasp it, and it reached in safety its thorny haven of refuge. It had run at speed about a hundred and fifty yards. I was much impressed by this unexpected exhibition; evidently this species of armadillo only curls up as a last resort, and ordinarily trusts to its speed, and to the protection its build and its armor give it while running, in order to reach its burrow or other place of safety. Twice, while laying railway tracks near São Paulo, Kermit had accidentally dug up armadillos with a steam-shovel.

There were big ant-hills, some of them of huge dimensions, scattered through the country. Sometimes they were built against the stems of trees. We did not here come across any of the poisonous or biting ants which, when sufficiently numerous, render certain dis-