But there is not only sexual dimorphism, there is also dimorphism of larvæ, e.g. green and brown caterpillars in certain species of hawk-moth (Sphinx), and there are sometimes not only two but three or more forms of a species; and in all these cases determinants of the differential parts must be represented twice, thrice, or several times in each germ-plasm, in each fertilized ovum, at least in all cases in which the different forms live together on the same area. In discussing mimicry we spoke of species of butterfly which were everywhere alike or nearly so in the male sex, while the females were not only quite different from the males, but differed greatly in many respects among themselves. Three different forms of females of Papilio merope occur in the same region of Cape Colony, each of these resembling a protected model. All three forms have been obtained from the eggs of one female. In this case the female ids of the germ-plasm must be represented by three different sets, one of which, when it is in the majority in the fertilized ovum, gives rise to the Danais-form, the second to the Niavius-form, and the third to the Echeria-form of the species. Phylogenetically considered, it is probable that each of these three kinds of ids originated by itself, on a more limited area on which the protected model lived in abundance; but with a wider distribution the different female ids mingled together, were united through the males into a single germ-plasm, and now occasionally exhibit all three forms on the same area. I doubt whether there is any other theory that can offer an interpretation of these facts, and I regard them, therefore, as affording further evidence of the real existence of ids.
The polymorphism of social insects must be thought of as similarly based in the germ-plasm.
In bees there are in addition to the males and females the so-called workers, and this can only depend on the existence of special kinds of ids. Those of the workers were originally truly female, but as many of their determinants underwent variations advantageous for the maintenance of the species, they were modified into special 'worker-ids.' I postpone for the present any inquiry into the causes by which these ids come to dominate the ontogeny; obviously it cannot be by the mere fact of being in a majority over the rest of the ids, as I indicated in the case of the butterflies with polymorphic females.
In many ants the division of labour goes further still; there are two kinds of workers in the colony, the ordinary workers and the so-called 'soldiers,' and in this case the worker-id must have developed in two different directions in the course of phylogeny, and have separated into two kinds of ids, so that the germ-plasm of these species must contain four kinds of ids.
I might cite many more cases in regard to which the assumption of two or more kinds of determinants seems imperative, but I believe that what has been said is enough to enable any one to think out other cases for himself.
LECTURE XIX
THE GERM-PLASM THEORY (continued)