— 216 — 



place in the development of the gametophyte. No special signs 

 of apogamy are found, and castration experiments have affirmed 

 this result. 



The Number of Chromosomes. 



The numbers of chromosomes are very peculiar indeed. The 

 typical tricolor-ïorms, I have examined, as gr, lab, mac, cl, fc, viol 

 have 13 chromosomes haploid. (Figs. 17 — 22). 



Figs. 15 — 16 shows 17 chromosomes in the homoeotypical 

 anaphasis of a V. arvensis. A typical arvensis-\me parv, nib, nmc, 

 op, nfc, alb from Allindelille Fredskov, Sjælland, also has 17 chro- 

 mosomes haploid, but another arvensis-lme, parv, nib, mac, op, 

 nfc, alb and with large leafy stipules, from one of the indivi- 

 duals in the population C (Tysinge Moor) has only 15 chromo- 

 somes haploid. 



Numbers as 13, 15 and 17 are very rare chromosome numbers. 

 In addition to this, we do not expect difference in numbers of 

 chromosomes among species belonging to the same collective 

 species. Finally, if there is a difference between related species 

 the numbers are usually proportionate to each other. 



The chromosome numbers found by Miyaji in Viola-species 

 are (Isjikawa 1916): 



Viola glabella 6 haploid 



— grypoceras 10 — 



— verecunda 10 — 



— nipponica 10 — 



— Okuboi 12 — 



— Okuboi var. glabra 12 — 



— phalacrocarpa 12 — 



— japonica 12 — 



— Patrini 36? — . 



— Patrini var. chinensis 48 diploid 



— diffusa 26 — 



In addition to this I am allowed to state, that Professor Winge 

 has determined the chromosome number of V. odorata to be 10 

 hapl. (Not published before). 



The numbers 6, 12, 24 exhibit a very fine series. But the 

 four species with 10 hapl. agree badly with the others. Even less 



