1906] Brainerd, — Hybridism in the Genus Viola 59 



none of the forms of this triad have I observed any sign of impaired 

 fertihty. Some of this intergradation may be due to the imperfect 

 evolution of the three species, the links ordinarily missing being in 

 this instance still extant ; but I have no doubt these species can and do 

 interbreed. 



I will speak of only one other instance of confluence, and I am in- 

 debted for this to the acute observation of Mr. F. F. Forbes of Brook- 

 Une, Mass. He has discovered on "Dedham Island," along the 

 Charles River, a station for V. pedinata, Bicknell (Torreya, iv. 129, 

 Sept., 1904). The type is from Long Island; but Mr. House also, 

 found it, in May, 1904, near Dayton, N. J. (Bull. Torr. Club, xxxii. 

 255, pi. 18), and Dr. C. C. Godfrey, in June, 1905, at Stratford, Conn. 

 The species is, to me, distinguishable only by the configuration of the 

 leaf from V. septemloba, with which it is associated in all four stations.^ 

 At "Dedham Island" jNIr. Forbes found that the two plants passed, 

 by the most gradual stages, one into the other. 1 visited the station 

 last May and again last August, and found the plants abundantly 

 spread over a large territory, and showing no sign of impaired fertility. 

 Plate 69 presents the two extreme leaf-patterns and some of the inter- 

 mediates. Often they closely simulate the leaf of V. cucullata X 

 septemloba (no. 15); but that is a larger plant, with infertile capsules 

 and dark-brown seeds, due to the black seeds of V. cucullata; the seeds 

 of V. septemloba, V. pectinata and their intermediates are of a light 

 straw color. 



The accompanying diagram presents a bird's-eye view of these 

 hybrids and their parentage. Though the forms are perplexingly 

 numerous, only half of the possible sixty-six hybrids have been brought 

 to light. Others will doubtless be discovered. Moreover, still fur- 

 ther complications may be expected; we may find quarter or eighth 

 hybrids, produced by crossing a simple hybrid once, or twice, with 

 one of the parent forms ; or we may find triple or quadruple hybrids, 

 produced by crossing a simple hybrid with an allied species other than 

 the parents, or with another hybrid. Some of the hybrids of V. pal- 

 Tnata X sagittata (no. 20) at New Brunswick are very likely not from 

 pure V. sagittata but from the V^ fimbriatula-sagittata combination, 



1 It will be helpful to our understanding of the relations of V. septemloba to V. pectinata, 

 if we observe the exactly parallel relation of V. pedatifida to V. indivisa, Greene (Pittonia 

 V. 124, pi. xiii.). The variation of V. palmata into V. Bernardi, Greene is a movement in 

 the same direction, though not to the same extent. 



