"Now, on one occasion, in a flower belonging to an individual of _Salvia verticillata_, and only on the left stamen, I observed a perfectly developed and pollinigerous lower cell, perfectly homologous with that which is normally developed in _Salvia officinalis_. This case of atavism is truly singular. According to the theory of Pangenesis, it is necessary to assume that all the gemmules of this anomalous formation, and therefore the mother-gemmule of the cell, and the daughter-gemmules of the special epidermic tissue, and of the very singular subjacent tissue of the endothecium, have been perpetuated, and transmitted from parent to offspring in a dormant state, and through a number of generations, such as startles the imagination, and leads it to refuse its consent to the theory of Pangenesis, however seductive it may be." This seems a strong confirmation of what has been here advanced. The main objection raised against Mr. Darwin's hypothesis is that it (Pangenesis) requires so many subordinate hypotheses for its support, and that some of these are not tenable. Professor Delpino considers[224] that as many as eight of these subordinate hypotheses are required, namely, that-- "1. The emission of the gemmules takes place, or may take place in all states of the cell. "2. The quantity of gemmules emitted from every cell is very great. "3. The minuteness of the gemmules is extreme. {214} "4. The gemmules possess two sorts of affinity, one of which might be called _propagative_, and the other _germinative_ affinity. "5. By means of the propagative affinity all the gemmules emitted by all the cells of the individual flow together and become condensed in the cells which compose the sexual organs, whether male or female (embryonal vesicle, cells of the embryo, pollen grains, fovilla, antherozoids, spermatozoids), and likewise flow together and become condensed in the cells which constitute the organs of a sexual or agamic reproduction (buds, spores, bulbilli, portions of the body separated by scission, &c.). "6. By means of the germinative affinity, every gemmule (except in cases of anomalies or monstrosities) can be developed only in cells homologous with the mother-cells of the cell from which they originated. In other words, the gemmules from any cell can only be developed in unison with the cell preceding it in due order of succession, and whilst in a nascent state. "7. Of each kind of gernmule a great number perishes; a great number remains in a dormant state through many generations in the bodies of descendants; the remainder germinate and reproduce the mother-cell. "8. Every gemmule may multiply itself by a process of scission into any number of equivalent gemmules."