1020] SCHNEIDER, NOTES ON AMERICAN WILLOWS. IX 23 



The third variety is var. angustifolia with " foUis 1-2 poll, longis vix 

 I poll, latioribus, supra opaco-cinerascentibus subtus pallidis tomentosis nee 

 dcnudatis," while in 1868 the sentence runs " foliis 1-2 poll, longis vix ^ poll, 

 latis subtus glaiicescentibus vel tenuiter tomentosis." In 1868 he puts un- 

 derthis variety f. " recurvata (Psh) : capsulisrecurvatis,"andf. "opaca: cap- 

 sulis brevissime vel vix pedicellatis (Wright, Coll. Nov. Mexic. n. 1878)." 

 In 1868 he says: " Hue forsan etiam pertinent: S. recurvata Pursh . . . et 

 forma opaca . . ." Of S. recurvata Pursh I have already spoken. Wright's 

 no. 1878 Is a form with glabrous (not only glabrescent) capsules which does 



not belong to S. humilis at all. 



S. Hariueyii Bentham which is regarded by Andersson as a subspecies 

 (1867) or doubtfully as a variety (1863) of S. humilis has nothing whatever 

 to do with this species. See my note in Bot. Gaz. lxv. 28 (1918). 



To S. humilis seems to belong also Andersson's S. tristis glabra (1867 and 

 1868) of which he himself says: " Ad S. humilem aperte accedens." I have 

 not seen a type. It is extremely difiBcult to distinguish well marked varieties 

 of S. humilis until full series of the variations of the different regions and 

 localities have been collected. The first question that arises is : can S. tristis 

 be regarded as a good species, or only as a variety of S. humilis, as asserted 

 by Griggs (see later under S. tristis). Almost all authors agree that S. tristis 

 is, as Ball (1910) says, "very similar to S. humilis but smaller in every 

 way." It differs from humilis chiefly in its thinner branches, shorter petioles 

 (hardly more than 3 mm. long), smaller leaves (scarcely up to 5 cm. long and 

 1-1.3 cm. wide), the absence of stipules, present and linear only on vigorous 

 shoots and in its smaller aments. Probably it is nothing but an ecological 



subspecies. 



I am not able to discuss certain forms of S. humilis from the northeast 



r 



which partly resembles S. discolor. There are others, too, in the middle west 

 that need careful observation in the field. I have seen material of S. humilis 

 sensu lato from the following states and counties: Vermont (Caledonia 

 County), Massachusetts (Middlesex, Hampshire, Essex and Suffolk Coun- 

 ties), Connecticut (Fairfield County), New York (Tompkins, Warren, 

 Wayne Counties and Long Island), Rhode Island (Providence County), 

 New Jersey (Bergen and Ocean Counties), Maryland (Montgomery County), 



District of Columbia, Virginia (Bedford County), North Carolina (Bun- 

 combe County), West Virginia (Randolph County), Georgia (Sumter County, 



uncertain form), Florida (Leon County), Pennsylvania (Chester and York 

 Counties), Kentucky (Jefferson County), Mississippi (Alcorn County) 

 Arkansas (Polk and Clay Counties), Texas (Pottawatomie and Atchinson 

 Counties), Nebraska (Nuckolls, Cass, Howard, Thomas, Holt and Brown 

 Counties), Iowa (Lyon, Cerro Gordo, Floyd, Humboldt, Delaware, Story, 

 Decatur and Poweshiek Counties), Missouri (St. Louis, Pike, Wayne, 

 Howell, Carter, Newton, Jasper, Scott, Webster, Shannon, Jepson, Macon, 

 Putnam and Clark Counties), North Dakota (Richland County), Minnesota 

 (Ramsay County), Wisconsin (Sauk County), Michigan (Ingham, Wayne 



and St. Claire Counties), Ohio (FrankUn s^nd Scioto Counties), Illinois (St, 



