BETULACEAE. — CARPINT7S 427 



Franchet's plate, but agrees with Winkler's statements. The outer (dentate) mar- 

 gin of the bracts is not lobed at the base, but there is a basal lobe at the inner (entire) 

 margin. The dilTerences between the Kiangsi plants and the Hupeh variety are 

 apparently very sUght, and more copious material probably will show that both 

 are the same and may represent a good spcf^ies confined to central China. Then 

 the oldest name would be C. Fargesii Franchet. 



Carpinus Turczaninovii Hance, var. ovalifolia Winkler in BoL 

 Jahrh. L. Suppl. 505 (1914). 



Carpinus Turczaninowii Franchet in Jour, de Bot. XIII. 203 (pro parte) 

 (1899). — Burkill in Jour. Linn. Soc. XXVI. 502 (pro parte) (1899).— 

 Diels in Bot. Jahrh. XXIX. 279 (pro parte) (1900). —Winkler in Engler, 

 Pflanzenr. IV.-61, 38 (1904), exclud. var. polyneura. 



Carpinus polyneura Burkill in Jour. lAnn. Soc. XXVI, 501 (pro parte, non 

 Franchet) (1899). — Henry in Elwes & Henry, Trees Gr. Brit. & Irel. III. 

 627, t. 201, fig. 5 (1908). — Bean in Kcw Bull. Misc. Inform. 1911, 327; 

 Trees & Shrubs BHt. Isl 1. 297 (1914). 



Western Szech'uan: Mao-chou, valley of Min River, alt. 2300 

 m., September and October 1910 (No. 4105; tree 7-17 m. tall, girth 

 1.2-2.4 m.; fruiting branchlets); between Lungan Fu and Sungpan 

 Ting, mountains, alt. 2300 m., August 1903 (Veiteh Exped. No. 4489; 

 tree 8 m. tall; almost ripe fruits). Eastern Szech'uan; Wushan 

 Hsien, A. Henry (No. 7020; type of var. ovalifolia; fruiting branches); 

 same locality, A. Henry (No. 7219; co-type, ex Winkler); district of 

 Tchen-keou-tin, Farges (without number, fruiting branchlets; and 

 No. 1273; ex Winkler). Kansu; Lotani road from Min-chou, alt, 

 2600 m., W. Purdorn (No. 787; tree 7 m. tall; fruits). 



The specimens which I doubtfully unite under this variety agree with each 

 other in their fruiting bracts of similar shape whicli differ from those of the typical 

 form in being only toothed and not lobulate on the outer margin. The leaves, too, 

 are somewhat more ovate-oblong and slightly more elongated at the apex than those 

 of the typical C. Turczaninovii Hance. There is no real difference in the pubescence 

 of the upper or lower surfaces of the leaves; this is variable in both the varieties. 

 Certainly var, ovalifolia needs further investigations. From C. polyneura Franchet 

 It differs clearly in the glabrous nutlets, and in the serration of its leaves which is dis- 

 tmctly compound with short teeth. Judging by a specimen cultivated at Kew which 

 is referred by Henry and by Bean to C. polyneura Franchet, the male flowers are 

 dilTerent from those of this species, I describe them as follows: Amenta mascula 

 sessilia, usque 3 cm. longa et circiter 0.7 cm. crassa, densa, rhachi dense sericeo- 

 villosa; braeteae late ovatae, acutae, ima apice fere nigrae, ceterum parte superi- 

 ore ruhro-brunneae, parte inferiore olivaceae, intus glabrae, margine satis ciliatae, 

 extus plus minusve adpresse sericeae; stamina circiter 10, toro piloso; thecae 

 separatae, elHpticae, concolores, apice et etiam antica parte superiore pilis longis 

 sericeis instructae, filamentis subbrevioribus tantum apice partitis sufTultae. 



^ There is a specimen collected by Wilson, "Wushan Hsien, ravine, May 1000 

 f^'^'itch Exped. No. 618; tree 3 m. tall), the male flowers of which are similar, but 

 the bracts are nearly round, more obtuse, and even more ciliate. I am not quite 

 Bure whether this specimen belongs to our variety or to another species. 



