We know that the larva which develops into a neuter bee, and that again which in time becomes a queen bee, are the same kind of larva to start with; and that if you give one of these larvæ the food and treatment which all its foremothers have been accustomed to, it will turn out with all the structure and instincts of its foremothers—and that it only fails to do this because it has been fed, and otherwise treated, in such a manner as not one of its foremothers was ever yet fed or treated. So far, this is exactly what we should expect, on the view that structure and instinct are alike mainly due to memory, or to medicined memory. Give the larva a fair chance of knowing where it is, and it shows that it remembers by doing exactly what it did before. Give it a different kind of food and house, and it cannot be expected to be anything else than puzzled. It remembers a great deal. It comes out a bee, and nothing but a bee; but it is an aborted bee; it is, in fact, mutilated before birth instead of after—with instinct, as well as growth, correlated to its abortion, as we see happens frequently in the case of animals a good deal higher than bees that have been mutilated at a stage much later than that at which the abortion of neuter bees commences. The larvæ being similar to start with, and being similarly mutilated—i.e., by change of food and dwelling, will naturally exhibit much similarity of instinct and structure on arriving at maturity. When driven from their usual course, they must take _some_ new course or die. There is nothing strange in the fact that similar beings puzzled similarly should take a similar line of action. I grant, however, that it is hard to see how change of food and treatment can puzzle an insect into such “complex growth” as that it should make a cavity in its thigh, grow an invaluable proboscis, and betray a practical knowledge of difficult mathematical problems.