4 JOURNAL OF THE ARNOLD ARBORETUM [vol. h 



" Closely rcseni])les the last [S. discolor]; but the uuicnts are more compact 

 and silky, and the scales rounder." He does not mention the pubescence, 

 and cites as synonyms " S. prinoidcs Pursh? S. crassa, Barratt." 



Andersson (18.5S) said very little of S. discolor making it a quasi form of 

 S. pJnjlici folia. In 1867, he dealt thorout^hly with it. His figure 41 on t. 5 

 represents rather short-petioled leaves loosely pubescent beneath, and a 

 glabrous female flower which rather agrees with that of S. cordata but un- 

 fortunately a good many of Andersson 's figures are very inaccurately drawn. 

 He divides S. discolor in the three forms: rigidior, latifulia and viiescens. 

 The last is identical with Muhlenberg's type in Herb. Willdcnow, and re- 

 presents a form with " foliis angustius lanceolatis." Andersson does not 

 mention it agriin in the Prodromus (18(58). Here he also changed the two 

 other forms of 1867. The name rigidior is not used again but is apparenth 



foha. This form, however, is, in my oi)ini<)ii, the same 

 as var. prinoidcs of recent authors. The name rigidior seems the oldest, 

 tJierefore this variety has to be called var. rigidior (see later). 



Andersson's var. latijolia (1867) is retained in 1868, but in the Prodro- 



mus lie excludes the hairy forms keeping them sejiarate as " 3 hirsida " 

 of which the lyj^e Is said to be in " II. C. Vindob." Unfortunately I have 

 not been able to find it in the Herbarium at Vienna. 



In 1867 Andersson also mentions two subspecies : * S. eriocepJmla and * S. 

 prinoidcs. In 1858 (Ofv. Svensk. Vet.-Akad. Forh. xv. 117) he said of S. 

 criuccphala: " E Salicibus americanis haec mihi valde obscura " placing 

 it next to his irrorata. He dealt extensively with specimens which he had re- 

 cei\ed as S. erioccphala, prinoidcs, and conifcra but his statements are of 

 little value. In 1867 he made two varieties of his subspee. erioccphala: 



ifolia and nif 



if< 



" Ad fiunicn Saskatchavan." Without having soon it, I cannot decide 

 whelher this really is a form of S. discolor or of S. Bchbiana. Andersson 

 himself says: " facie S. vagantis rostratae." Tlie var. rufescens came from 

 " America occidentalis," and it belongs to S. Scoulcriana. Andersson adds 

 to his description the following remark: "Haec forma S. braclnjstachyac- 

 Scoulcriauac similis est, ab ca attamen amentis multo longioribus magisque 

 laxifloris capsuHs demum longius pccidellatis differt." See also 5. Scoidcri- 

 ana. In 1868, Andersson does not mention again these two forms. He 

 makes S. erioccphala a var.? erioccphala of S. discolor and says: " In herbariis 

 variis fornias indubie identicas nunc S. discolor nunc S. erioccphala pro- 

 miscuc signatas vidi, nee e scrii)tis auctorum differentias nlla-, extricare po- 

 tui. nisi j.nn allatas, e forma squamarum dcsumtas sed viles." 



I1ie subsjiccies S. prinoidcs Andersson (1867) is founded on Pursh's 

 S. prinoidcs. Andersson says: De hac sj;ecies iterum incertus sum; e des- 

 criptione a pracccdcntibus et omnibus S. cinerascentibus differret "stylo 

 elongalo " et forsitan melius S. phylicifoliae afiinis cemieretur; specimina ta- 

 mcn quae plurima \idi, etiam in eo S. discolorcm magis revocaberunt, quod 

 stylum ullum sub])roductum vix observare posscm. Pursh ipse de hac 

 dicit "resembles ,?. discolor." Nuttall cam cum .b'. discolore cum .? con- 



