Packard informs us[[415]] that in Bombus the larva, after it is full fed, passes into the pupa state (Fig. 331, A, B) by a series of transformations accompanied by moultings of the skin. Packard's statements have been confirmed by others, but details have not been fully given, so that the number of the moults, their intervals and other particulars, are still unknown. We have remarked that the pupal instar is very like the perfect instar, except that it is colourless and soft, and that each of the members is wrapped in a very delicate skin; the colour appears gradually. This metamorphosis exhibits important differences from that of the Lepidoptera. Packard calls the Insect, during the stages of transformation from the full-fed larva to the pupa, the semi-pupa; the later stages of the pupa, when the colouring has appeared, he terms the subimago. Altogether he considers there is a series of at least ten moultings of the skin. His ideas were apparently derived from examination of a series of specimens after death rather than from observation of the development in living individuals. The parasitic forms of Hymenoptera have apparently extraordinary metamorphoses of very varied kinds.
Parthenogenesis.—One of the most remarkable facts connected with this Order is the prevalence of parthenogenesis in a considerable number of widely separated species. In many of these Hymenoptera it is not a mere occasional occurrence, but plays an important part in the continuity of the species; indeed, it is believed that in some members of the Order the reproduction is entirely parthenogenetic. We shall give particulars as to some of these cases in subsequent chapters, and will here make some remarks on the different forms of parthenogenesis existing in the Order. The three forms of parthenogenesis mentioned on p. [141] all occur in Hymenoptera. In the gall-making Cynipidae parthenogenesis is frequently accompanied with alternation of generations, a generation consisting of the two sexes being followed by another consisting entirely of females, which in its turn gives origin to a bisexual generation. In this case deuterotokous parthenogenesis is established as a part of the normal economy of the species. This same form of parthenogenesis also occurs in other species of Cynipidae unaccompanied by alternation of generations. Thus in Rhodites rosae the generations resemble one another, and the male is very rare, but is still occasionally produced,[[416]] and the same condition exists in other Cynipidae. According to the observations of Adler, we may assume that the male, in the latter cases, is useless, the continuation of the species being effected by virgin females although males exist. Deuterotokous parthenogenesis also occurs in the sawflies, but as a comparatively rare phenomenon.[[417]]
Thelyotokous parthenogenesis is common in sawflies, and it also occurs in some Cynipidae. There are several species of this latter family in which no males have ever been found.[[418]] The phenomena in Rhodites rosae we have mentioned, give rise to the idea that in that species deuterotokous parthenogenesis occurs as an exception, the species being usually thelyotokous. A most remarkable case of thelyotokous parthenogenesis is said to exist in the case of the parasitic ant Tomognathus. This species is said to be monomorphic, only the female existing, and reproducing by uninterrupted parthenogenesis.
Arrhenotokous parthenogenesis—i.e. parthenogenesis in which the progeny is entirely of the male sex—occurs in several species of sawflies. We find it also occurring in the case of the social Hymenoptera; the workers of ants, bees, and wasps occasionally produce eggs parthenogenetically, and the progeny in these cases is always of the male sex. In the honey-bee the queen sometimes produces eggs before she has been fertilised, and the parthenogenetic young are then always of the male sex.
Some species of Hymenoptera exhibit two forms of parthenogenesis. In Nematus curtispina the parthenogenetic generation is generally of the male sex, but a female is occasionally produced;[[419]] while in Hemichroa rufa parthenogenesis may result in either deuterotokous or thelyotokous progeny. No case is yet known of a species exhibiting the three forms of parthenogenesis. From this review we may conclude that parthenogenesis does not favour the formation of one sex more than another; but it is clear that it decidedly favours the production of a brood that is entirely of one sex, but which sex that is differs according to other circumstances.
Production of Sex.—It is believed that a very peculiar form of parthenogenesis exists in the honey-bee, and it is confidently stated that the drones, or males, of that species are always produced from unfertilised eggs. These views are commonly called the Dzierzon theory, and are widely accepted. They assume that the eggs are male till fertilised, and then become female. After the queen-bee is fertilised most of the spermatozoa soon find their way into a small chamber, the spermatheca, near the posterior orifice of the body; it is believed that each egg may be fertilised as it passes the door of this chamber, and that the eggs that produce females (i.e. workers or queens) are so fertilised, but that the eggs that produce drones are not fertilised. Hence it is supposed that the sex is determined by this act of fertilisation, and Cheshire has described what he calls an apparatus for differentiating the sexes. It is also confidently stated that no male honey-bee ever has a father.
The facts we have stated as to the sexes resulting from parthenogenetic reproduction in Hymenoptera generally, are extremely opposed to the Dzierzon theory, in so far as this relates to the production of sex. There have always been entomologists[[420]] who have considered this view unsatisfactory, and the observations of several recent French naturalists[[421]] are unfavourable to the idea that the sex of an egg is determined by its fertilisation.
There can be no doubt that the queen honey-bee frequently produces males parthenogenetically, and the error of the views we are alluding to consists in taking the parthenogenesis to be the cause of the sex of the individual. It must be recollected that the laying of an unfertilised egg by a fertilised female may be different physiologically from the laying of an egg by an unfertilised female; for, though both have as result an unfertilised egg, it is possible that the fertilisation of the female may initiate processes that modify the sex of the eggs produced by the ovaries, so that though these may produce previous to fertilisation only male eggs, yet after fertilisation they may produce eggs of the opposite sex or of both sexes. In other words, the act of fertilisation may initiate a different condition of nutrition of the ovaries, and this may determine the sex of the eggs produced.
Polymorphism, or Castes.—The question of the causes of the modified individuals forming the various castes of the social Hymenoptera has been much discussed. These individuals are many of them very different in size and structure from either of their parents, and are also different in their habits and instincts. This difficult subject is far from being completely elucidated. In the case of the honey-bee it is well established that an egg of the female sex can, after deposition, be made either into a queen or a worker-bee by the mode of nutrition—using that word in the largest sense. On the other hand, Dewitz thought that in the case of the ant Formica rufa, the caste—whether worker or winged female—is already determined in the Insect before leaving the egg.[[422]] Weismann and others associate the caste with some hypothetic rudiments they consider to exist at the very earliest stage of the embryonic, or oogenetic process.
Herbert Spencer says:[[423]] "Among these social Insects the sex is determined by degree of nutrition while the egg is being formed," and "after an egg, predetermined as a female, has been laid, the character of the produced Insect as a perfect female or imperfect female is determined by the nutrition of the larva. That is, one set of differences in structure and instincts is determined by nutrition before the egg is laid, and a further set of differences in structures and instincts is determined by nutrition after the egg is laid."