But among insects this is not so. All the sex characters of the insect are the direct expressions of the action of the chromosomes and owe nothing to the physiological activity of the sex-gland. The differentiation of each cell is pursued under the control of its own set of chromosonie-borne factors : in mammals, though each cell of the female body is chromosomally, genetically dif ferent from each cell of the male body, this chromosomal dif ference does not bring about the whole differentiation of sex characters locally, within the individual cell. but, once the sex glands have become differentiated, the testes and the ovaries take charge.
Goldschmidt it was who made the first serious attempt to dem onstrate the method by which the sex-determining factors in their action lead to the production of the sex-characters of the adult. He started from the fact, long known to entomologists, that when species or even geographical races of moths are crossed, sexual abnormalities are commonly found among the hybrid off spring. For his material he chose the gypsy moth, Lyma)itria dispar, which has a wide distribution and is very variable. If European specimens of this forest pest are bred among them selves the offspring are unremarkable in every way. The same is true of the Japanese variety, Lymantria japonica. But if a Japanese male is mated with a European female, normal male off spring and females which show a number of modifications in the direction of the male type are produced. When such female intersexes are mated with their brothers, of the females of the generation thus produced half are normal, half are intersexual. The reciprocal cross, European male Japanese female, produces normal females and males in the first hybrid generation, but if the individuals of this are then interbred they produce a certain proportion of males with female characters.
Further investigation demonstrated that there were many dif ferent sub-races of European and of Japanese gypsy moths that were quite distinct in respect of intersexuality, in that the degree or grade of intersexuality was definite and typical for a particular mating. Goldschmidt classified strains as relatively "strong" or "weak." For example, a "strong" male mated to a "weak" female gives 5o per cent. normal males and 5o per cent. intersexual females. A "very strong" male mated to a "weak" female would give offspring all male. A mated to B gave a low grade of inter sexuality, C X D a high grade, E X F a grade intermediate between these, and so on. If strong race A gave moderate sexuality with weak race P, whilst with race Q it gave strong intersexuality, and if strong race B gave moderate intersexuality with Q, then it could be predicted that B with P would give only a slight grade. Similar males from one culture mated with females from different cultures gave intersexes that could be arranged in a series accord ing to the degree of their abnormality and so on. It was possible, by calling on experience, to produce every stage from an almost complete male to an almost complete female intersex at will by making the appropriate mating. In fact, it was as possible to turn the "determined" females into fully equipped males as to ensure the regular production of normal males and females.