The alternation of generations as it is found amongst the Entoparasitic Trematodes and most Cestodes, is to be explained in a slightly different way.
It appears that in these parasitic forms a complicated metamorphosis first arose from the parasite having to accommodate itself to the different hosts it was compelled to inhabit, owing to the liability of its primitive and subsequent hosts to be devoured[9]. A capacity for asexual multiplication—obviously of immense advantage to a parasite—appears to have been acquired in some of the stages of this metamorphosis, and an alternation of generations thus established.
A nearly parallel series to that exhibiting alternations of sexual generations with generations which produce by budding is supplied by the cases where sexual generations alternate with parthenogenetic ones, or in some instances even with larvæ which reproduce sexually or else parthenogenetically.
The best known examples of this form of alternations of generations are found amongst the Insecta[10]. A simple case is that of the Aphides. The ova deposited by impregnated females give rise to forms differently organised to the parents but provided with an ovary[11]. The eggs from the ovary develop parthenogenetically within the oviduct, and so long as there is plenty of food and warmth the generations produced are always parthenogenetic forms. The failure of warmth and nutriment causes the production of true males and females, and so the cycle is completed. We must suppose that the capacity possessed by so many female insects of producing eggs capable of developing without the influence of the male element, has been, so to speak, taken hold of by natural selection, and has led to the production of viviparous parthenogenetic forms, by which, so long as food is abundant, a clear economy in reproduction is effected. The continuance of the species during winter is secured by the production of males and females, the females laying eggs in autumn which are hatched in the spring.
In Chermes there is less modification of the primitive condition in that the parthenogenetic generations lay their eggs like the impregnated females. In the gall-flies (Cynipidæ), there is frequently an alternation of generations of the same kind as in Chermes; there being no viviparous forms. The individuals of the different generations differ from each other to some extent in all these cases.
A second type of alternations of parthenogenetic and sexual generations is exemplified by the cases of Chironomus and Cecidomyia, where the larvæ which develop from the eggs of the fertilized female produce parthenogenetically, by means of true ova, forms which eventually after several generations (Cecidomyia) of larval reproduction give rise to sexual forms. The explanation is here practically the same as in the case of Aphis, and is paralleled in the gemmiparous series by the production of buds in the larval forms of Trematodes, etc. A very similar occurrence takes place in Ascaris nigrovenosa (vide chapter on Nematoidea), except that larval forms, which carry on reproduction and then perish without developing farther, do so by a true sexual process. Thus there is an alternation of generations of adult and larval sexual forms. The Axolotl is an intermittent example of the same phenomenon.
As might be anticipated from the mode in which alternations of generations have become established, incomplete approximations to it are not uncommon. Such approximations are especially found in the Arthropoda, where alternations of sexual and parthenogenetic generations frequently take place, in which the individuals of different generations are similarly organised (Psychidæ, Apus, &c.). Another approximation is afforded by the parthenogenetic winter eggs of Leptodora amongst the Phyllopods, which give rise to Nauplius larvæ, while the young hatched from the summer eggs do not pass through a metamorphosis. Numerous transitional cases are also found amongst the forms in which there is an alternation of sexual and gemmiparous generations.
The whole of the cases to which allusion has been made in this section may be conveniently classed under the term alternations of generations, but the cases of alternation of two sexual generations, and of sexual and parthenogenetic generations, are classified by Leuckart, Claus, etc. as cases of heterogeny, which they oppose to the other form of alternation of generations. If special terms are to be adopted for the two kinds of alternation of generations, it would be perhaps convenient to classify the cases of alternations of sexual and gemmiparous generations under the term metagenesis, and to employ the term heterogamy for the cases of alternation of sexual and parthenogenetic generations.
The term Nurse (German Amme), employed for the asexual generations in metagenesis, may advantageously be dropped altogether.
[1] To this general statement Wolff forms a remarkable exception, for though without any clear knowledge of what we call cells he had very distinct notions on the relations of growth and development.