1920] SCHNEIDER, NOTES ON AMERICAN WILLOWS. IX 17 



G. Anderson." How this plant came to England has never been exactly 

 explained. It is not mentioned in later books on the English flora. 



Koch (1828) made S. petiolaris a var. ^ suhglabrata of S. grisea using 

 grisea for S. sericea, Andersson (1858) mentioned it only briefly, while in 

 1867 he distinguished the varieties rigida and suhsericea. The last is now 

 generally regarded as a hybrid between S. petiolaris and S, sericea. Of the 

 first Andersson described two forms: f. latijolia and/, longifolia but he does 

 not cite any specimens. His diagnosis of f. latijolia runs: '' foliis !2-3 poll, 

 longis, basi latiori subrotundatis, infra medium f-1 poll, latis, subito cus- 

 pidatis, argute serratis, rigidissimis, subtus valde intense glaucis. Haec 

 forma non raro modificationibusS. lucidae aut S. discoloris similis est ob folia 

 basi sat lata et subito acuminata, rigida et sublucida; sed facile distinguitur 

 amentislonge alienls." I have not seen a specimen really agreeing with this 

 description, to which Andersson in 1868 adds that the leaves are '* aut valde 

 intense glaucis aut pure et lucide viridibus." He probably confused differ- 

 ent plants. Andersson's f. longifolia is nothing but the typical S. petiolaris, 

 and he himself says in 1868 '* haec fere typica." In 1867, his variety consists 

 of forms Avith leaves lustrous silky beneath " ut *S. discolor,'' and others with 

 glabrous leaves. In the Prodromus he adds a f. hrevijolia with shorter leaves 

 (1-2 inches long, instead of 3-4 inches as in f. longifolia) which are hardly 

 I inch wide. I do not tliink this is a form of any taxonomic value. 



In 18o8 Andersson described a S. gracilis, from specimens collected by 

 Drummond and Richardson In Saskatchewan to which he cites as a syn- 

 onym 5. rosmarinifolia Hooker (1839). The type of this willow is Richard- 

 son's specimen from Cumberland House (Hooker says only Saskatchewan). 

 There is a co-type of it in Herb. Barratt. Andersson also named Richardson's 

 no. 387 (Herb. K.) S, gracilis which is called by Richardson 5. purpurea. 

 Hooker, too, cites S. purpurea from ** Cumberland House Fort, Dr. Richard- 

 son." Besides this Andersson himself named S. gracilis a species of Drum- 

 mond's from Saskatchewan in Herb. K. All these specimens agree well with 

 each other. From typical S. petiolaris they chiefly differ in their more linear 

 leaves (according to Andersson *'latitudine fere decies longioribus ") and 

 in their longer pedicels (about 4 mm. long instead of 3 mm.); the whole as- 

 pect of the plant therefore is a little different. In 1858, Andersson speaks 

 of S. gracilis as '* quasi hybridae S. vagante [S. Bebbiana] et 5. myrtilloide 

 [S, pcdicellaris],'' while in 1867, regarding S. gracilis as a kind of subspecies 

 of *S. petiolaris he more correctly remarks: " Habitu a S. petiolari majori et 

 typica valde recedit sed modificationibus permultis cum ea intime connexa 

 est. Indumentum foliorum idem, ut etiam capsularum, sed forma in hac 

 angustior magisque elongata. Ramis in forma arbusculae humilis erectis, 

 foliis angustis et amentis primum subglobosis habitum S, rosmarinifolia^ 

 nostrae prae se fert, sed ab ea luculenter diffcrt capsulis longissime pedi- 

 cellatis.'' 



Andersson, in 1858, also made a var. rosmarinoides of S. gracilis which 

 differs only in its ** foliis apicem versus serrulatis in gracilis integerrim s 

 marginibus acutiusculis, subrevolutis.'* This sentence apparently should 



