ULMACEAE. — ULMUS 247 



the length of the stipe being very variable, as are also the dentation and shape of 

 the leaves. Daveau, in his key, says that the typical U. parvifolia has 16-20 

 pairs of lateral nerves and the perigones are more or less reddish or red, while in 

 U. Shirasawana the number of the lateral nerves is only 8-10, the perigones being 

 more or less greenish or yellowish. I have failed to see a specimen with more than 

 16 pairs of lateral nerves, and I do not beheve that the color of the perigone is of 

 any importance in distinguishing even a form. U. Sieboldii Daveau may represent 

 a variety if the large fruits and the different bark prove to be constant; otherwise 

 I cannot accept these new species. 



At low altitudes this species is common in western Hupeh, especially in the neigh- 

 borhood of the city of Ichang. On rich soil and near water-courses it grows fully 

 25 m. tall and has a trunk from 1.5 to 2.5 m. in girth and clear of branches for half 

 its height. The bark is gray and smooth, and on old trees peels off in roundish 

 patches showing the brown inner bark; on young trees it is merely scaly. The 

 branches are numerous and slender and form a rather broad, rounded crown. In 

 open country it is a low broad-topped tree with a trunk seldom more than 1 m. in 

 girth. It also grows on the cliffs of the glens and gorges in western Hupeh and in 

 eastern Szech'uan, where it is usually a bush. It flowers in September and the fruit 

 ripens in October. The leaves are retained late into the autumn and often assume 

 rich shades of red and purple; in localities where the climate is warm and moist 

 the leaves on individual trees sometimes remain quite green through the winter 

 and fall as the young leaves unfold in the spring. 



Japanese botanists consider this Elm indigenous in Japan, but I saw there only 

 planted trees and not very many of them. At Meguro, near Tokyo, I saw the tree 

 from which came the material on which U. Shirasawana Daveau is based, but I 

 could detect no valid difference between it and the typical U. parvifolia Jacquin, 

 trees of which grew near by. 



Pictures of U. parvifolia Jacquin will be found under Nos. 474, 635 and 654 of 

 the collection of my photographs and also in my Vegetation of Western China, 

 Nos. 494 and 495. E. H. W. 



CONSPECTUS SECTIONUM SPECIERUMQUE ASIAE ORIENTALIS 

 NEC NON HIMALAYAE.i 



Perigonia circiter 14 v. ad medium (rarius ultra medium) incisa, lobis late oblongis 



V. fere rectangulis hand v. vix 2-23^-plo longioribus quam latis. Folia fere 



semper decidua. Flores vere apparentes . . . Sect. I. MADO CARPUS. 



Semina ab emarginatura samarae distincte remota, plus minusve in centre 



samarae sita (confer etiam 14. U. pumilam) . . Subsect. 1. GLABRAE. 



Samarae undique pilosae et ciUatae Ser. a. Wallichianae. 



Pedicellorum - pars superior nuda perigonio et parte inferiore pilosa circiter 

 2-3-plo longior. Samarae tenuiter pilosae et satis sparse ciliatae. 



1. U. Wallichiana. 



' To present the distinguishing characters more clearly the two European species, 

 U. foliacea Gilibert and U. glabra Hudson, which have been constantly confused 

 with certain species of eastern Asia, have been included in the key. 



' The pedicels of the flowers are articulated, the part above the joint being 

 mostly glabrous and more or less gradually passing into the perigone, while the 

 part below the joint is mostly pilose., like the peduncle of the inflorescence. The 

 " perigone " forms that part of the flowers which surrounds the stamens and the 

 ovaries, which are inserted on the top of the pedicel at the very base of the perigone. 



