I have now no doubt that he can distinguish between different words. For instance, when he is hungry he will bring a "food" card time after time, until he has had enough, and then he lies down quietly for a nap. Again, when I am going for a walk, and invite him to come, he gladly responds by picking up the "out" card, and running triumphantly with it before me to the front door. In the same way he knows the "bone" card quite well. As regards water (which I spell phonetically, so as not to confuse him unnecessarily), I keep a card always on the floor in my dressing-room, and whenever he is thirsty he goes off there, without any suggestion from me, and brings the card with perfect gravity. At the same time he is fond of a game, and if he is playful or excited will occasionally run about with any card. If through inadvertence he brings a card for something he does not want, when the corresponding object is shown him, he seizes the card, takes it back again, and fetches the right one. No one who has seen him look along a row of cards, and select the right one, can, I think, doubt that in bringing a card he feels that he is making a request, and that he can not only perfectly distinguish between one word and another, but also associates the word and the object. I do not for a moment say that Van thus shows more intelligence than has been recorded in the case of other dogs; that is not my point, but it does seem to me that this method of instruction opens out a means by which dogs and other animals may be enabled to communicate with us more satisfactorily than hitherto. I am still continuing my observations, and am now considering the best mode of testing him in very simple arithmetic, but I wish I could induce others to co-operate, for I feel satisfied that the system would well repay more time and attention than I am myself able to give. JOHN LUBBOCK. INSTINCT OF LOCALITY IN DOGS. [_March 4, 1893._]