214 



Rhodora 



[November 



tion ; and when colonies of two species grow closely intermingled, 

 interbreeding is what might be reasonably anticipated. 



In the present paper I shall restrict the discussion to the behavior 

 of the five species of this group that are most abundant in Western 

 Vermont, and that have naturally been studied in the field with most 

 assiduity. I name them with their contrasting characters in the 

 following table : 



V. AFFlNtsl . V. SORO- 

 RIA 



Plants : 

 Leaves: 



glabrous 



cordate- 

 acuminate 



V. SEPTEN- : V.FIMBRI- 

 TRIONAI.IS : ATUl.A 



pubescent 



broadly cordate-ovate 



ovate- 

 oblong 



Si'URRED Petal : with few or no hairs 



bearded 



Cleistogamous 



Peduncles : 



horizontal or decumbent 



Cleistoc.amous ovoid-acum ate 



Flowers: 



Cleistogamous 

 Capsules: 



V. CtTCUL- 

 LArA 



^ A ^ 



glabrous 



^ A 



cordate 



^ A ^ 



beardless 



erect or ascending 



sagittate 



subulate 



_.^ ^ 



subglobose 



oblong 

 usually m ttled with purple 



Cl'st. Sei'als : lanceolate : ovate-lanceolate 



oblong 

 always green 



Auku:lks : 



short, inconspicuous : long, spreading, 

 : hispidulous 



narrowly lanceolate 



, -^ . 



long, not 

 spreading 



The number of possible combinations in pairs of these five species 

 will be found to be ten. In general, if // represent the number of 

 species, and iVthe number of possible pairs, we have the formula, 

 N=hn{n — \). That is, among twelve species of Viola there 

 would be 66 possible ways of hybridizing. If there were twenty 

 species of Crataegus so nearly allied that a cross could be effected 

 between any two of them we might have possibly 190 hybrids. 



1 V. venustula, (Ireene, is found not to be specifically distinct from V. affinis, 

 Le Conte, as understood by Prof. Cireene, or from V. obliqua, as interpreted by 

 Mr. Pollard in the Hritton Manual. 



