64 THE PLANT SPECIES 



can be explained on the hypothesis that G. tenuiflora, G. latiflora, 

 and G. cana went through a stage of barely completed primary 

 speciation which was followed by extensive hybridization in 

 various combinations (Fig. 2). The hybridization has gone far 

 enough to prevent the original extreme populations from persist- 

 ing as clear-cut species, but has not fully completed the process 

 of secondary speciation. The major constituents thus possess some 

 of the characteristics of subspecies in so far as they replace one 



Fig. 2. The syngameon of Gilia tenuiflora-latiflora, consisting of four 

 semispecies linked by natural hybridization in nearly every possible com- 

 bination. The abbreviations stand for Gilia tenuiflora, G. cana, G. latiflora, 

 and G. leptantha. 



another geographically, intergrade in the zones of contact, and 

 retain a certain degree of interfertility; they also possess some 

 characteristics of species, as manifested in marginal sympatric 

 contacts and in a degree of morphological and genetic divergence 

 which is comparable to that of species elsewhere in the phylad. 

 The situation just described is not confined to Gilia. It is 

 known also in Finns, Quercus, and Aquilegia, to cite other exam- 

 ples from among those listed in Table I, and in numerous other 

 genera. 



