468 WILSON EXPEDITION TO CHINA 



The Himalayan and Yunnan forms are very much alike, and this species seems 

 to differ from B. cylindrostachya Spach in its naore glossy and more or ■wholly 

 glabrous, longer and narrower leaves wliich are more distinctly acuminate and 

 also differ somev/hat in the character of their serration. It may be possible to dis- 

 tinguish several forms, as Regel did, but I should like to see much more material 

 from India before making such an attempt. The fruiting catldns seem to be always 

 more slender than those of B, cylindrostachya Spach. 



4. Betula Imninifera Winkler. See p. 455. 



5. Betula Baeumkeri Winkler in Engler, Pflanzenr. IV.-61, 91, fig. 22 D-F 

 (1904). — Schneider, III Handh. Lauhholzk. 11. 882, fig. 552 c, 553 i-k (1912). 



CHINA. Yunnan: without locality, J. Delavay (type ex Winkler). 



Having seen so many specimens of B. luminifera Winkler, showing the varia- 

 bility of this species, 1 suspect that B. Baeumkeri^ of which I have been unable to 

 see any material, may prove to be only a form of that species. 



Sect. 2. COSTATAE Regel in Bull. Soc. Nat. Mosc. XXXVIIl. pt. 2, 412 

 (1865). — Prantl in Engler & Prantl, Nat. Pflanzenfam. III. abt. 1, 45 (1887). — 

 Koehne, Deutsche Dendr. 107 (1893). —Schneider, III. Handh. Lauhholzk. I. 98 

 (1904). 



Betula^ sect. Euhetula Regel, subsect. Costatae Regel in De Candolle, Prodr. 

 XVI. pt. 2, 175 (1808). — Dippel, Handh. Lauhholzk. II. 183 (1892).— 

 Winkler in Engler, Pflanzenr. IV .-61, 57 (1904). 



This section is rather an artificial one. It includes very different forms like B. 

 Schmidiii Regel and B. corylifolia Regel & Maximowicz, which may represent dis- 

 tinct groups. The latter species and B. grossa Siebold & Zuccarini (probably also 

 B. insignis .Franchet and B. costata Trautvetter) are characterized in a hving state 

 by the same peculiar odor of the inner bark of young branchlets found in that of 

 the American B. lenta Linnaeus and B. lutea Michaux, but it seems to me an un- 

 natural classification to bring together these species in one group on account of 

 this peculiarity which unfortunately cannot be detected in dried specimens with 

 certainty. I distinguish the following subsections to which we have to add sub- 

 Ecct. NiGRAE, nov. subsect. (including only the American B. nigra Linnaeus) and 

 subsect. Lentae, nov. subsect. (Sect. Lentae Regel in Bull. Soc. Mosc. XXXVIIL 

 pt. 2, 417 [1865]. — Sect. Euhetula, subsect. Lentae Regel, in De Candolle, Prodr. 

 XVI. 2, 179 [1868]), including only American species. 1 cannot accept the different 

 subgenera proposed by Nakai (in Tokyo Bat. Mag. XXIX. 40 [1915]). 



Subsect. a. Ermanianae Schneider, n. subsect. 



Betula, subg. Ermani Nakai in Tokyo Bot. Mag. XXIX. 41 (pro parte) (1915). 



The species united in this group seem to be closely related. They have a s°i*JP^^ 

 whitish or reddish or creamy bark peeling off in transverse flakes or strips. J-he 

 following species are well separated geographically, but it is not always ^^^ ^^ 

 distinguish them without sufficient judiciously collected material. What is neeaed 

 to establish a good classification of a difficult genus like Betula are observations 

 of living plants in the field. 



5. Betula Ennanli Chamisso in lAnnaea, VL 537, t. 6, fig. D, a-e (1831)- — 

 Erman, Reise um die Erde, Naturh. Atlas, 56, t. 17, fig. 2 (1835). —Spach m Ann. 

 Sci. Nat. ser. 2, XV. 190 (1841). — Ledebour, Fl. Ross. HI. pt. 2, 653 (1850). 

 Trautvetter m Mem. Sav. Sir. Acad. St. Petershourg, IX. (Maximowicz, Pr^^- 

 Fl. Amur.) 252 (1859). — Regel in Nouv. Mem. Soc. Nat. Mosc. XIIL P^-/' V" 

 (Monog. Betulac. 62), t. 6, fig. 35, 37, 38, t. 12, fig. 13-28 (excl. var. y) (ISOl), 



