mainland as were islands on which we saw evident marks of their having been, such as decayed houses, fires, the before-mentioned turtle bones, etc. Maybe, at this more moderate time, they make and use such canoes, and when the blustering season comes on, may convert the bark of which they were made to the purposes of covering houses, water-buckets, etc., well knowing that when the next season returns they will not want for a supply of bark to rebuild their vessels. Another reason we have to imagine that such a moderate season exists, and that the winds are [not] then upon the eastern board as we found them is, that whatever Indian houses or sleeping places we saw on these islands were built upon the summit of small hills, if there were any, or if not, in places where no bushes or wood could intercept the course of the wind, and their shelter was always turned to the eastward. On the main, again, their houses were universally built in valleys or under the shelter of trees which might defend them from the very winds, which in the islands they exposed themselves to.
Of their language I can say very little; our acquaintance with them was of so short a duration that none of us attempted to use a single word of it to them, consequently words could be learned in no other manner than by signs, inquiring of them what in their language signified such a thing, a method obnoxious as leading to many mistakes. For instance, a man holds in his hand a stone and asks the name of it, the Indian may return him for answer either the real name of a stone, or one of the properties of it, as hardness, roughness, smoothness, etc., or one of its uses, or the name peculiar to some particular species of stone, which name the inquirer immediately sets down as that of a stone. To avoid, however, as much as possible this inconvenience, myself and two or three others got from them as many words as we could, and having noted down those which we thought from circumstances we were not mistaken in, we compared our lists; those in which all agreed, or rather were contradicted by none, we thought ourselves morally certain not to be mistaken in. They very
