— 392 — 



All the plants have been mac and faintly labellata. Regarding 

 the colour of petals, segregation has taken place, but only in 

 tricolor-colours: pall and lut. Apparently some of its characters 

 caused it to be referred to tricolor and some to arvensis. A cyto- 

 logical investigation showed that it usually had 13 chromosomes 

 (hapl.) as tricolor (fig. 9 b). But in some of the anthers the divi- 

 sions were very irregular, and most of the pollen from these 

 anthers degenerated. A rather regular interkinesis of this sort is 

 shown in fig. 9 a. 2 chromosomes here are lost in the first division. 

 Possibly all the irregular pollen degenerates and therefore the 

 type can be constant. According to its chromosome number I 

 regard it as a tricolor, but it was certainly established through 

 a segregation after hybridization and it has inherited some of 

 tricolor's and some of arvensis' characters. 



One of the bright yellow tricolors (grand, lut, n.mac), a well 

 established type that only segregate the two types/?«//, mac and lut, 

 n.mac shows similar irregularities in the reduction division. The 

 chromosome number is 13 haploid, but in a few cases two chro- 

 mosomes are expelled in the heterotypic division. Fig. 9 c shows 

 the homotypic metaphasis in such a dyad. Each of the nuclear 

 plates has 12 chromosomes, and between them are seen the two 

 expelled chromosomes. Apparently there is some instability in 

 the chromosome stock. The cause of this instability is probably 

 a preceding cross. I assume that all the bright yellow tricolors 

 descend from crosses between tricolor and arvensis. 



V. Are all the genotypically conditioned Combinations 

 (Isoreagents) to be regarded as "Species"? 



In Viola there is a multitude of possible types. When crosses 

 are possible and give fertile offspring, and when the variation is 

 caused in this way, we may expect from time to time to find 

 all the combinations of the characters realized. With the clas- 

 sification here adopted there is 2. 2. 2. 3. 3. 6. 2. 2. 2. 2. 2. 3. 

 2.2.2.2.2.2.2 = 5,308,416 possible combinations. But as 

 there are more characters varying than the former (for instance 

 broad petals and narrow petals, but difficult to distinguish from 

 each other), there are still more combinations. The number of 

 possibilities will increase very rapidly, when new varying char- 

 acters are found, and in the same proportion the likelihood of 



