SUPPLEMENT 121 



particles wander through the protoplasmic bridges from cell to cell, and so 

 might be the means of conveying the initials of characters. That nuclei are 

 of no significance in the transference of characters in graft- hybrids, is shown 

 also by the histological researches of STRASBURGER (1905 a) on Cytisus Adami, 

 and by NOLL (1905) on Crataegus x Mespilus : in both cases the hybrids do 

 not possess the double number of chromosomes possessed by their parents, as 

 might have been expected from a nuclear fusion (comp. NOLL, 1905, p. 19). 



We may still expect a far-reaching insight into the problems of heredity 

 from a searching study of graft-hybrids, more especially if the experimental 

 treatment of the subject be successful. For the moment the known data are 

 too few to enable us to come to any definite conclusion. 



Still it may be noted, in conclusion, that it is quite possible that to employ 

 them, as we have done, in opposition to the ' chromosome theory of heredity ' 

 is not quite free from possible criticism. As STRASBURGER (1907) correctly 

 points out, there are many phenomena which appear at first to stand in opposi- 

 tion to this hypothesis, but which on closer study may really be brought to 

 support it, though superficially. The same may be true of graft -hybrids. 



383, title of lecture, delete ADAPTATION and note. 



11. 18-24, f or Two questions . . . par excellence, read Doubtless, however, 

 hybridization is not the only mode of origin of new species ; it must now be 

 our task to investigate other possible factors. 



We are thus led to consider a problem 



384, 11. 48-9, after generations read by multiplying vegetatively or by self- 

 fertilization. 



386, 11. 22-4, for At the same time . . . obviously two read We must obviously 

 recognize two 



387, 11. 1-3, for three (DE VRIES . . . fourth type, read two, fluctuating varia- 

 tions and mutations. To these may be added a third type, 



1. 36 P. 388, 1. 17, for It should be noted . . . KLEBS, 1903.] read Although 

 the curve takes this form very frequently, still there are other forms of varia- 

 tion curves also. Thus, for instance, a character may vary only in one direc- 

 tion, and hence half-Galtonian curves present themselves; or, again, curves may 

 appear with two or more apices. The latter phenomenon may be associated 

 with the fact that a certain number occurs more readily than those close to it, 

 but it may be due to the material under investigation not being all of one kind 

 to the mixing of two forms whose averages are different, but whose rarer 

 numerical values merge into each other (transgressive curves). We cannot 

 deduce from this that a variation curve with a single apex is an indication of 

 the uniformity of the material, for a mixture of many forms may indeed pre- 

 sent a Galtonian curve with one apex. 



The reason that the characters of a species vary round a certain average 

 is that, even with the most careful cultivation, irregularities in the growth 

 conditions (' nutrition ' in the widest sense of the word) must occur. 



1. 38 P. 393, 1. 2, for If individual variations ... is formed, read That 

 the variations from the mean follow, however, certain laws is associated with 

 corresponding variations in the external conditions round a certain mean 

 (comp. KLEBS, 1903). There is no reason for regarding the variations in 

 quantitative characters, which are demonstrable statistically, in any way 

 differently from the variations induced in the course of experiment which we 

 have studied in Lectures XXIII and XXIV. Variations are induced in the 

 present instance by the sum of the factors which we have termed collectively 

 ' nutrition ', as they were in that case by one individual factor. Since, 



