— 217 — 



does V. diffusa agree with 26 diploid (probably 13 haploid). This 

 number is the same as that found in tricolor. Accordingly, the 

 Melanium section of Viola is not the only one in which the 

 chromosome numbers are not plainly proportionate to each other. 



In this connection can be mentioned that Rosenberg (1918) 

 has dealt with the numbers of chromosomes in the Crepis-species, 

 and he has shown that the numbers in this genus are not plainly 

 proportionate to each other. He found the following haploid num- 

 bers in this genus: 3, 4, 5, 9, 20 and the number 8 occurs also. 



Further Ishikawa (1916) has found in Lactuca the dis- 

 agreeing numbers 5, 8, 9, 12, 24 and in Senecio 5, 10, 19. 



A species with 13 chromosomes haploid might come into 

 existence from one with 12 chromosomes in the following way: 

 Two 12-chromosome forms cross with each other. For one reason 

 or another only 11 of the chromosomes are able to unite in pairs. 

 Therefore the diakinesis of F 1 may exhibit 11 double chromo- 

 somes and 2 single ones, which all split up, 13 chromosomes going 

 to each of the poles. After self-fertilization or mutual fertiliza- 

 tion in F x there might be a chance to obtain an F 2 with 26 chromo- 

 somes diploid, and it might happen to be constant in the future. 

 A behaviour of chromosomes in this way Federley (1913) has 

 detected in some Pygaera-hjbrids. — From 13 chromosomes we 

 might come to 17 in a similar way. 



But in nature tricolor and arvensis hybridize and perform 

 fertile offspring; 13 x 17 chromosomes give an F x with 30 chro- 

 mosomes and in this way we might have a chance to obtain a 

 Viola with 15 chromosomes. In 1919 I crossed 13 chromosome 

 plants with 17, 13 with 15, 15 with 17 and vice versa. In 1920 

 circumstances did not allow me to continue my Fto/a-experiments, 

 but in this summer (1921) I have about 150 F 1 plants in a very 

 good condition from those crossings, and it is my intention to 

 examine the reduction division in those hybrids in order to deter- 

 mine the behaviour of chromosomes after hybridization and to 

 compare it with the segregation in F 2 . A mixed mendelian segre- 

 gation can be expected according to the behaviour of the un- 

 paired chromosomes, whether they distribute after the Pygaera- 

 schedule (Federley 1913) or after the .Drosera-schedule (Rosen- 

 berg 1909 a), and according to whether the genes carried by the 

 unpaired chromosomes are different from those in the paired ones 

 or not. 



