380 METAMORPHOSIS 



therefore, the vegetative-cell evidently contains the same initials as the germ- 

 cell. Now there are, on the one hand, animals which are like typical plants in 

 possessing extensive powers of regeneration, and there are also plants which 

 are comparable with the higher animals in their slight powers of regeneration. 

 How are such plants to be regarded ? Are the root initials absent in a twig which 

 cannot strike root, and does a leaf which forms roots but not shoots, possess 

 only root initials ? We can give no final answer to this inquiry ; but when we 

 reflect how much regeneration depends on external conditions (e. g. high tempera- 

 ture) we shall be inclined to believe that only a favourable environment is 

 wanting and not initials. Moreover, initials originally present may easily be 

 lost or so very much attenuated in efficiency — sit venia verbo — by the growth 

 of the cell that they become inoperative. 



Regeneration can but rarely proceed, as in the case of Marcliantia, from small 

 fractions of the body of the plant ; isolated cells of fully-developed tissue are 

 mostly only just capable of growth (Haberlandt, 1902), or on the application of 

 stimuli only produce a few divisions (Winkler, 1902), but they never regenerate 

 a complete organism. The possibility is not excluded that suitable nutriment 

 is wanting in these cells such as would render possible an increase of the small 

 quantity of idioplasm which they contain. If regeneration taking place in 

 a tissue proceeds from a single cell {Begonia (Hansen, 1881), Torenia (Winkler, 

 1903)) these cells are always seen first to divide actively without growing ; Winkler 

 supposes that these divisions indicate an increase of idioplasm. Also in callus 

 the formation of organs is always preceded by an ample division of cells, which 

 Winkler regards in a similar way, but nothing more definite is known. In any 

 case the botanist may, in opposition to Weismann, maintain that all cells have 

 the same initials, and that there is no qualitative division of hereditary characters. 

 The germ-cells are distinguished from the vegetative-cells only by the fact that 

 in them the idioplasm is predominant, while in the latter, in accordance with 

 their function, trophoplasm prevails. 



But how is this to be reconciled with the segregation of characters which 

 appears in certain hybrids and which seems to be characteristic of the germ-ceU ? 

 Is not this segregation in itself a new argument for the chromosome hypothesis ? 

 If every chromosome were the bearer of one quality, then the reduction division 

 must be an excellent means of effecting qualitative division of hereditary charac- 

 ters, thus leading to segregation. The small number of chromosomes, however, 

 renders it impossible to allot one quality only to each of them, and the theories 

 which allot to the chromosome the hereditary substance are obliged to assume 

 that each chromatin-grannle is the bearer of several qualities. The reduction 

 division as understood by the zoologist does not explain the segregation of the 

 hybrids ; on the other hand, Correns (1902) has formulated a view which 

 renders it possible to understand this process on the grounds of the chromosome 

 theory. We will not pursue further these purely hypothetical questions, but 

 confine ourselves to the facts, regarding it as established that the supposed 

 distinction between somatic- and germ-cells does not in reality exist at all. 



We have already referred to segregations which take place in the vegetative 

 region, but one of the most instructive examples is furnished by Cytisus adami, 

 a hybrid of C. lahurnum and C. purpureus. This hybrid is approximately 

 intermediate between its parents so far as its organs of vegetation and repro- 

 duction are concerned ; thus, for example, it produced flesh-coloured flowers. 

 But reversions to the parental forms also occur, for on individual branches we 

 frequently find the yellow blossom of C. lahurnum, and also, but more rarely, the 

 red flowers of C. purpureus. If a branch has assumed the character of one of 

 the parents it never returns to the intermediate form. The segregation need 

 not involve the whole branch ; it can also be accomplished in the longitudinal 

 half of a bud, so that the branch arising from it has on one side the character of 

 C. lahurnum, and on the other that of C. adami ; and in this case the line sepa- 



