110 WILSON EXPEDITION TO CHINA 
29. Salix hondoensis Koidzumi in Tokyo Bot. Mag. XXVII. 88 (1913). 
Salix dolichostyla, var. hirosakensis Léveillé & Vaniot in Fedde, Rep. Spec. 
Nov. ITL. 22 (1906). 
Salix hirosakensis Koidzumi in Tokyo Bot. Mag. XXVII. 264 (1913). 
JAPAN. Hondo: prov. Mutsu, Hirosaki, “secus aquas," May 1905, U. 
Faurie (No. 6602, type of var. hirosakensis, Q); Aomori, in streets, May 1905, U. 
Faurie (No. 6622; 9); “ basi montis Iwagisan in pago Dake,” May 1905, U. Faurie 
(No. 6601; 9); Hokkaido: prov. Ishikari, * in planitie Sapporo," June 14, 1908, 
U. Faurie (No. 256; 9). 
Koidzumi founded S. hondoensis on a specimen from Sapporo without mention- 
ing a precise type. Taking up the name hirosakensis he says: floribus femineis 
saepe biglandulosis. In Faurie's type the 9 flowers have only one gland, but 
otherwise the specimen looks very much like S. eriocarpa Franchet & Savatier. 
The g^ flowers in No. 6601 have rather small glands sometimes appearing like 
a single ventral gland almost surrounding the pedicel. Our species certainly is 
very nearly related to S. eriocarpa and S. jessoensis Seemen. Whether they are 
hybrids or varieties of one species has not yet been determined. 
30. Salix Makinoana Seemen in Fedde, Rep. Spec. Nov. 1. 173 (1905). 
Saliz gymnolepis Léveillé in Fedde, Rep. Spec. Nov. 111. 22 (1906). 
Salix purpurea, subspec. gymnolepis Koidzumi in Tokyo Bot. Mag. 
267 (1913). 
JAPAN. Hondo: prov. Mutsu, “in sylvis Aomori,” May 1904, U. Faure 
(No. 5761, type, 9, ex Seemen), “in plateis Aomori,” May 1904, U. Faurie (No. 
2769; co-type, 9, ex Seemen); Hirosaki, “ secus rivos," May 1905, U. Faure 
(No. 6615, type of S. gymnolepis; 9). 
This species has been very differently interpreted by Koidzumi, l. c., but unfortu- 
nately I have not seen one of von Seemen’s types. Koidzumi says, that No. 5761 
partly belongs to S. stipularis Smith (in which he includes S. opaca Andersson and 
S. sachalinensis Schmidt, see p. 143), and partly to S. Miyabeana Seemen, see 
p. 166. No. 2769 is referred by Koidzumi to his subspec. gymnolepis. Léveillé& 
type agrees very well with von Seemen’s good description of S. M. akinoana. Other- 
wise S. gymnolepis Léveillé much resembles S. hondoensis Koidzumi (S. dolichostyla, 
var. hirosakensis Koidzumi). The bracts are not perfectly glabrous, as Lé 
says, but hairy at the base. The ovaries have very short pedicels, as indicated 
by von Seemen for S. Makinoana, and von Seemen's whole description of 
9 flowers and fruits fits exactly the specimen of S. gymnolepis Léveillé before 
me. I believe that S. Makinoana Seemen and S. gymnolepis Léveillé are n 
related to S. hondoensis Koidzumi and have no connection with S. purpurea 
Linnaeus. 
31. Salix jessoensis Seemen, Salic. Jap. 31, t. 3, fig. F-L (1903). 
JAPAN. Hokkaido: prov. Ishikari, Sapporo, Park, August 23, 1914, E. H. 
Wilson (No. 7414; tree 15-17 m. tall, girth 2.4-3.9 m., bark grayish, shallo 
fissured; sterile); same locality, May 18, 1891, Y. Tokubuchi (type, ex Herb. Sap- 
poro Coll., 9 and 3); same locality, May 1890, Y. Tokubuchi; same locality, May 
11, 1892, Y. Tokubuchi (d); same locality, July 1, 1891, Y. Tokubuchi (sterile). 
The sterile specimen of Tokubuchi has very hairy yellowish brown branches 
(suckers?), and the leaves are up to 13.5 cm. in length and 3 cm. in width a 
grayish and loosely silky beneath; the petioles are about 1 em. long, and the stip- 
ules are broadly ovate at the base, very acuminate at the apex and 34 the pnt 
of the petioles. This species differs from S. eriocarpa Franchet & Savatier 
sca na Aratus atento 
