1919] SCHNEIDER— AMERICAN WILLOWS 311 



Virginia to Alabama and Florida. The center of its development 

 is from California to Washington, Montana, and Texas. 



Among the American willows the Longifoliae occupy an 

 isolated position, and of the willows of the Old World it is difficult 

 to say which can be taken for the nearest relatives of this group. 

 I shall discuss this point later, and I can now only repeat that 

 probably the forms of the sect. Albae Borr. might be regarded as 

 rather closely related genetically to the Longifoliae. 



In the following key it is recognized' that there are two rather 

 well marked types in the group based on the form of the stigmas. 

 In one, represented by 6". taxifolia and S. sessilifolia, the lobes of 

 the stigma are narrow and elongated, and in the older flowers 

 mostly more or less revolute; while in the other group, the types 

 of which are S. exigiia and S. longifolia, the lobes are shorter and 

 broader, not linear-lanceolate, the whole stigma often being quasi 

 capitate. In some forms of S. longifolia, especially of var. Wheeleri 

 from the northeast, the shape of the stigmas is rather intermediate. 

 In the first group S. taxifolia is well distinguished from 5. sessilifolia 

 and its relatives by the short small aments, the small more or less 

 globose anthers, and the small hnear leaves; while S. sessilifolia 

 and its varieties and 5. fltiviatilis have long cyhndric aments, 

 oblong-ellipsoid anthers, and longer, broader leaves. In the second 

 group it is more difficult to separate the species because the main 

 characters, glabrousness or pubescence of the ovaries and leaves, 

 are more liable to variation. S. melanopsis .with var. Bolanderiana 

 represents a rather well marked t^-pe with glabrous ovaries, but in 

 S. exigua as well as in S. longifolia we meet with forms of which the 

 ovaries vary from densely pubescent to entirely glabrous. The 



' It seems to be of interest to quote Bebb's opinion as to the possibility of a taxo- 

 nomic arrangement of the forms of this section (Box. Gaz. 16:104. 1891): "Clearly 

 marked as are the outer limits of the group it presents no Imes of cleavage within by 

 which it can be satisfactorily divided. No natural characters are found to coincide 

 with such assumed distinctions, for instance, the 'Hnear lobes of the stigma,' made 

 promment in the attempt to separate S. sessilifolia. Each portion after subdivision 

 remains as heterogeneous as was before the aggregate group. It may be possible, by 

 emphasizing first one character and then another, as these are found to predominate 

 in the different forms, to designate a number of subspecies and varieties; but so 

 bewildering and intangible is the reticulated intergradmg that the difficulty of segre- 

 gation seems only to be heightened by every fresh acquisition of the material." 



