If a species _A_ is to be transformed into a species _B_, both must be united by fertilisation and the resulting hybrids then be fertilised with the pollen of _B_; then, out of the various offspring resulting, that form would be selected which stood in nearest relation to _B_ and once more be fertilised with _B_ pollen, and so continuously until finally a form is arrived at which is like _B_ and constant in its progeny. By this process the species _A_ would change into the species _B_. Gärtner alone has effected thirty such experiments with plants of genera _Aquilegia_, _Dianthus_, _Geum_, _Lavatera_, _Lychnis_, _Malva_, _Nicotiana_, and _Œnothera_. The period of transformation was not alike for all species. While with some a triple fertilisation sufficed, with others this had to be repeated five or six times, and even in the same species fluctuations were observed in various experiments. Gärtner ascribes this difference to the circumstance that “the specific [_typische_] force by which a species, during reproduction, effects the change and transformation of the maternal type varies considerably in different plants, and that, consequently, the periods within which the one species is changed into the other must also vary, as also the number of generations, so that the transformation in some species is perfected in more, and in others in fewer generations.” Further, the same observer remarks “that in these transformation experiments a good deal depends upon which type and which individual be chosen for further transformation.” If it may be assumed that in these experiments the constitution of the forms resulted in a similar way to that of _Pisum_, the entire process of transformation would find a fairly simple explanation. The hybrid forms as many kinds of egg cells as there are constant combinations possible of the characters conjoined therein, and one of these is always of the same kind as the fertilising pollen cells. Consequently there always exists the possibility with all such experiments that even from the second fertilisation there may result a constant form identical with that of the pollen parent. Whether this really be obtained depends in each separate case upon the number of the experimental plants, as well as upon the number of differentiating characters which are united by the fertilisation. Let us, for instance, assume that the plants selected for experiment differed in three characters, and the species _ABC_ is to be transformed into the other species _abc_ by repeated fertilisation with the pollen of the latter; the hybrids resulting from the first cross form eight different kinds of egg cells, viz.: _ABC_, _ABc_, _AbC_, _aBC_, _Abc_, _aBc_, _abC_, _abc_. These in the second year of experiment are united again with the pollen cells _abc_, and we obtain the series _AaBbCc_ + _AaBbc_ + _AabCc_ + _aBbCc_ + _Aabc_ + _aBbc_ + _abCc_ + _abc_.