378 METAMORPHOSIS 



number of chromosomes as the nucleus of the ovum is regarded as a proof that 

 the chromosomes are really the transporters of hereditary characters, for in 

 general the offspring inherit in equal degree from father and mother. 



We have hinted at the basis on which certain modern theories of heredity 

 are founded, and more especially Weismann's germ plasm theory (1892 and 

 1902), although their characteristics have by no means been treated exhaustively. 

 Weismann's theory is at least the first theory of heredity to be worked out in 

 detail, and leads to results of the greatest consequence. It is for this reason 

 worthy of consideration and its scientific importance is greatly enhanced when 

 we reflect how many specific researches it has given rise to. If we do not deal 

 with it here in extenso it is not merely from considerations of space, but also on 

 its merits, for botanists must, in our opinion, reject this theory. Detailed reasons 

 for rejecting Weismann's theory we cannot give here, all we need say is that to 

 attribute to the chromosomes, or indeed to the nucleus at all, the exclusive 

 possession of the initials is a view which has in no sense been justified. 



Let us first of all consider the question of the individuality of the chromosomes . 

 [The principal supporter of the individuality of the chromosomes of late years 

 has been Boveri (1904), for Strasburger, who also held that view in the past, 

 appears to have abandoned it (1905). The parts of the chromosomes, the chro- 

 matin -granules or ' ids ', are still to be considered as organs which may increase 

 in number by dividing, but which cannot be created afresh.] In the majority of 

 nuclear divisions an unprejudiced observer will conclude that the chromosomes 

 are formed in the so-called prophase stage of the division and will disappear in the 

 anaphase. The constant return to the same number of chromosomes results from 

 the fact that before each division the mass of chromatin is approximately of the 

 same amount. There are statements enough available which suggest that the con- 

 stancy of the number of chromosomes is often more a pious wish on the part of the 

 observer than an actual scientific fact. Very frequently the number of chromo- 

 somes cannot be accurately counted at all, and the observer contents himself 

 with ascertaining whether they correspond approximately to the normal number, 

 or to its half, or its double. Guignard (1891) adduces an example of a very 

 remarkable anomaly in the number of the chromosomes. The primary embryo-sac 

 nucleus of Lilium has in its first division twelve chromosomes ; one of the resulting 

 daughter-cells always exhibits in the two following divisions twelve chromo- 

 somes, the other and lower one, however, shows, in the first division, sixteen or 

 more, and in the second twenty to twenty-four. This, at all events, proves that 

 the chromosomes can reproduce themselves otherwise than by longitudinal 

 division. Dixon (1894) has made similar observations in the prothallium of 

 Pinus, where the large nuclei of the archegonium-wall cells possess more than 

 double the number of chromosomes found in the nuclei of the first prothallial 

 cells. Finally, we may draw attention to the increase in the number of chromo- 

 somes in apogamous ferns (compare p. 368). In spore formation also, as above 

 explained, a diminution in the number of the chromosomes arises, not, as one 

 might assume, because, as in the case of animals, the longitudinal division fails to 

 appear, but because the chromatin of the nucleus breaks up into only half the 

 number of chromosomes. We conclude from all this that the chromosomes are 

 not definite organs of the cell as are the chromatophores and the nucleus ; they 

 are reformed at each division, and hence the chief basis for believing them 

 to be transmitters of hereditary characters disappears. 



Starting from the fact that the internodal cells of the Characeae are 

 incapable of regeneration, it has been claimed that only by normal nuclear 

 division can two cells with equal hereditary properties arise. The above-named 

 internodal cells show later, as a matter of fact, the so-called direct cell division, 

 in which no chromosomes and consequently no longitudinal division of the 

 chromosomes can be found. When first produced, however, the nucleus divides in 

 a thoroughly typical manner, and the sister cells at the nodes possess fully the 



