io 4 The Botanical Gazette. [April, 



In keeping with this view it finds its fullest development and 

 greatest variation in form and stature on the Pacific slope. 

 Eastward it declines in vigor and variability until on the At- 

 lantic coast it is of rare occurrence from New Brunswick to the 

 Potomac. Clearly marked as are the outer limits of the group 

 it presents no lines of cleavage within by which it can be sat- 

 isfactorily divided. No natural characters are found to co- 

 incide with such assumed distinctions, for instance, as the 

 "linear lobes of the stigma," made prominent in the attempt 

 to separate 5. sessilifolia. Each portion after sub-division 

 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 dif- 

 ferent forms, to designate a number of sub-species and varie- 

 ties; but so bewildering and intangible is the reticulated ' 

 intergradmg that the difficulty of segregation seems only to be 

 heightened by every fresh acquisition of material. 



I. b. cordata Muhl— No American willow has a wider 

 distribution than this, from the Gulf States to California and 

 northward in the interior to the Mackenzie River, and per- 

 haps none other— not even excepting aggregate S. longi/oUa- 

 presents more the appearance of a "congeries of species in the 

 making. It differs from S. longi folia however in being, of all 

 our willows, the one which hybridizes most freely with others 

 and this implies that even where actual hybridity can not be 

 proven it is more or less affected by association with other 

 willows ,n different portions of its wide area of distribution. 



«. S Lasiole pis Benth. -Ten years ago this species was 

 known to the writer only from California,! specimens Even at 

 that time three dominant lines of development were recog- 



2S2; v e since been found to lead out to such 



disSt dlVe ^ ent extr V mcs W w ould certainly be admitted as 



ttZl ^ eC, !\i WCre Lt n0t for the ^tergrading. The most 



best de \ h f eSe ' C f iblting the ™**BigJieH in its far- 

 ornL if* H OI V hC typkal iasi0l ^ s oi southern Cali- 



W hin^ton I me K 7 Mn C - V - Pi P e '- from »™ battle, 



wit eaVCS 0b ; )Vate ' ° blon ^ or oblanceolate, coarsely 



and irregularly repand-serrate, 2 inches wide by Ll inches 



hole ITThV^ - ld A C ° piOUsl >' Si % with long hairs as 



overlaid S?^ , A * if S ' ****** ™* "°t already 



overloaded with aberrant forms we have to mention still an- 

 other, provisionally referred only, found by Prof Greene on 



