622 ON PHYSIOLOGICAL CHROMOMERES. 



with the view, here proposed. He has collected the coefficients of 

 the mutations for most of the latter; they do not show any definite 

 kind of linkage as in the two cases, just described. Of course it 

 must be assumed that- the activation of the latent factors for lata 

 cannot have an influence on the factors which may be located in 

 other chromosomes, and so it is in other cases too. High figures 

 may occur from time to time, especially in small cultures, just as 

 in our tables, but it is only the regular recurrence of the phenomenon, 

 which can be considered as a proof for linkage. 



Resuming these considerations we may bring the seven chromo- 

 somes of the haploid nuclei of our evening primrose under the follo- 

 wing heads: 



1. The central one, which is one of the three large rods and con- 

 tains the factors for the Drosophila-Wkt mutations, as far as studied 

 until now, those for the differences between laeta and velutina, 

 and those of the new mutant 0. pulla. 



2. The two large lateral ones, each with two physiological chro- 

 momeres in a mutable condition, viz. 0. lata with albida, and 0. 

 scintillans with oblonga. 



3. The small lateral ones, for each of which only a single mutable 

 group of characters has been ascertained until now. 



Before considering these groups separately it seems desirable to 

 point out the connection of these distinctions with the principles 

 of the theory of mutation. This started from the assumption that 

 the characters of all organisms are built up of sharply distinct units 

 {Die Mutations-Theorie, I, p. 3). These units might be linked together 

 into groups and in allied species the same units and groups recur. 

 This principle has since been largely corroborated by numerous 

 researches and especially by those on Mendelian inheritance and 

 may now be considered to enjoy general acceptation. 



From it three main types of mutation may be derived. In the first 

 place characters may be lost, then lost ones may reappear and finally 

 new ones may be added to the stock. It seems probable that new 

 units are produced from existing ones bij means of some constitu- 

 tional change. The loss of a character constitutes a toss-mutation, 

 which is also often called retrogressive or recessive. It is of very 

 general occurrence and almost all our well-known varieties among 

 agricultural and horticultural species as well as among wild plants 

 belong to this group. The same is true for the larger part of the 

 mutations and specific differences which have been studied gene- 

 tically in the last two decades. The reappearance of lost properties 



