Main Botanic Garden, Russian Academy of Sciences, 127276 Moscow, Russia

Abstract. Salix hainanica (sect. Humboldtianae Pax, Salicaceae), apparently endemic to Hainan Island, China, is described on the basis of collections at A and LE dating back to 1934.

Among the eastern Asian collections housed at A and LE, I discovered two sheets of Salix---duplicates of a single collection made in 1934 on the 10th Hainan Expedition of Lignan University---apparently representing a new, as-yet-undescribed species.

At first glance the plant looks like a small-leafed specimen of S. babylonica L., and so it was determined to be by E. D. Merrill. On the other hand, the sheet in A has a note by F. P. Metcalf expressing doubt about that determination but not proposing a different one. According to Flora Hainanica (Chun et al., 1965: 340), the only species of Salix recorded from Hainan is S. tetrasperma Roxb.---the most widespread tropical willow of Southeast Asia. But it is difficult to imagine that S. hainanica could have been mistaken for S. tetrasperma. Although they belong to the same section, in their general outlook these two are strikingly different.

The placement of S. hainanica in section Humboldtianae is mainly based on two features: (1) the very characteristic small, triangular, sharp-tipped buds with the outer bud scales having free, adaxially overlapping margins; and (2) the shortly pubescent bracts, well-developed pedicels, ovoid capsules with minimal styles and stigmas, nectaries very short but broad and sometimes half-embracing the pedicel of the female catkins.

Section Humboldtianae contains 15 to 20 species distributed mainly in the tropics and subtropics of America, Africa, the Near East, and Southeast Asia. Some of the species, such as S. chilensis Molina in America, S. mucronata Thunb. s.l. in Africa, and S. tetrasperma Roxb. s.l. in Southeast Asia, occupy very extensive areas, whereas others, such as S. thorelii Dode in Laos and S. kusanoi (Hayata) C. K. Schneider in Taiwan, have quite limited distributions. Salix hainanica seems to belong to the latter group of taxa: in several major herbaria I have had access to (A, GH, MO, LE, W), I was unable to find additional specimens besides the above-described type, although to the collector, S. K. Lau, the species appeared to be locally "fairly common."

Some authors (including Hao, 1936, and Chou et al., 1984) split sect. Humboldtianae, referring the eastern Asian species to sects. Tetraspermae (Anderss.) C. K. Schneider and Glandulosae Kimura (= Wilsonianae Hao). If such splitting is accepted (which I prefer not to do), S. hainanica should be referred to sect. Tetraspermae, and within this group I would place it close to the Near Eastern S. acmophylla Boiss., which also has narrow, acutely attenuate leaves and rather small (as compared with S. tetrasperma or S. mucronata) capsules. Still, the very small, densely sericeous leaves with peculiar microscopically sharp serrations and the small number of ovules per capsule are distinguishing features of S. hainanica.

I am deeply grateful to Dr. Peter H. Raven, director of the Missouri Botanical Garden, for providing support for the continuation of my studies on eastern Asian willows. I also express my sincere thanks for the friendly assistance I received from the herbaria A, GH, LE, MO, and W. My special gratitude is due Mr. Boris Sosnovski for providing the photograph, and Mrs. Raisa Trokhinskaya for typing the manuscript.