I take first some observations communicated by Mr. H. L. Jenkins to Mr. Romanes, since, though they raise a point which we have already shortly considered, they form a transition from unconscious to perceptual inferences. Speaking of the intelligence of the elephant, Mr. Jenkins says,[GV] "What I particularly wish to observe is that there are good grounds for supposing that elephants possess abstract ideas; for instance, I think it is impossible to doubt that they acquire, through their own experience, notions of hardness and weight." He then details observations which show that elephants at first hand up things of all kinds to their mahouts with considerable force, but that after a time the soft articles are handed up rapidly and forcibly as before, but that hard and heavy things are handed up gently. "I have purposely," he says, "given elephants things to lift which they could never have seen before, and they were all handled in such a manner as to convince me that they recognized such qualities as hardness, sharpness, and weight." Now, the question I wish here to ask is--Do the observations of Mr. Jenkins, the nature of which I have indicated, afford good or sufficient reasons for supposing that these animals possess abstract ideas? And I reply--That depends upon what is meant by abstract ideas. If it is implied that the abstract ideas are _isolates_; that is, qualities considered quite apart from the objects of which they are characteristic, I think not. But if Mr. Jenkins means that elephants, in a practical way, "recognize such qualities as hardness, sharpness, and weight" as _predominant_ elements in the constructs they form, I am quite ready to agree with him. I much question, however, whether there is any conscious inference in the matter. The elephant sees a new object, and unconsciously and instinctively builds the element hardness or weight into the construct that he forms. And he shows his great intelligence by dealing in an appropriate manner with the object thus recognized. But I do not think any reasoning is required; that is to say, any process involving an analysis of the phenomena with subsequent synthesis, any introduction of the conceptual element.