CROSSINGS IN MKLANIUM-VIOLETS 285 



conlimioiis Miriatioii in nalurc ;is well :is in /•':.. wliith tlic exception 

 of the colonr cluiraeters; it is tlieretore impossible to put tlie inter- 

 mediates in any class. The impossibility of usin},' such characters 

 in systematical classification is evident. The possibility of making 

 a tolerably useful classification of the Linnean species \'. tricolor se(;ms 

 therefore an unsolvable task. The systematist may be obliged to confine 

 himself to gixc descriptions, as detailed as possible, of the variation 

 in the species. 



All the characters with the e\ce|)lion of the corolla and the dcNc- 

 loi)ment of the labellum combined freely in F:-, so far as 1 could 

 find. If the statements of Morgan and Federley as to the localiza- 

 tion of the genes in the chromosomes are correct coupling might pro- 

 bably occur. Cr.AUSEN states the very interesting and important fact that 

 the number of chromosomes varies in V. tricolor and V. aruensis. He 

 found 13 chromosomes in V. tricolor and lô or 17 in V. aruensis. This 

 discovery may be of great importanc(; for the theory of the localiza- 

 tion of the factors in the chromosomes. 



The origin of the variation in V. tricolor has to some degree 

 occupied the evohitionist. Jordan (1848) cultivated plants of this 

 (and other) species and found that they bred true. He interpreted 

 the results of these experiments as proofs of the fallacy of any evolu- 

 tion theory. In his book »Arten imd Varietäten und ihre Entstehung 

 durch Mutation* de Vries (1916) states the »elementary species» of 

 V'. tricolor as examples of forms originated by mutation. Lotsy has 

 presented a theory of evolution by means of hybridization, as is well 

 known. In one of his papers (1915) he also mentions the elementary 

 species of V. tricolor. Against the assumption of de Vries as to their 

 origin by mutations he thinks that they are products of segregations 

 as a result of crossings. The same should be the case in the other 

 classical example on elementary species mentioned by de Vries, viz. 

 Erophila verna (see Rosen 1911). 



A preliminary note of mine (1914) led Lotsy to make this state- 

 ment as to Viola and therefore a few words shall be said as to the 

 appearance in F, of types resembling the elementary species of Witt- 

 rock. In the Fo-generation of one cross between V. tricolor and V . or- 

 vensis (2 X 10) at least three types appeared which showed such a 

 resemblance. Some red flowering plants with almost closed pollen- 

 magazine were of the same appearance as V. tricolor ummotropa. A 

 light-yellow and large flowering plant resembled V. aruensis curtise- 

 pala and another had the macH/a^a-character. In another cross bet- 



