280 WILSON EXPEDITION TO CHINA 



bluish black, about 8 mm. thick, and the stones much resemble those of C. yunna- 

 nensis. The leaves are ovate with a more or less distinctly long-pointed apex or 

 somewhat rhomboid-ovate, and they measure up to 11 cm. in length and 5.8 cm. 

 in width. I do not know the color of the ripe fruits of C. yunnanensis, and I am 

 not quite sure if my specimen represents a form of it or a new species. There is also 

 some resemblance to C. Salvatiana (see p. 283), of which the fruits and the mature 

 leaves are still unknown. The shape of the leaves seems to be rather variable in 

 some species, those of my number are somewhat intermediate in shape between C. 

 amphibola and C. yunnanensis, but the stones of these two species look very different. 



9. Celtis nervosa Hemsley in Jour. Linn. Soc. XXVI. 450 (1894). — Henry 

 in Trans. As. Soc. Jap. XXIV. Suppl. 85 (List PI. Formosa) (1896). — Matsumura 

 & Hayata in Jour. Coll. Sci. Tokyo, XXII. 370 (Enum. PI. Formos.) (1906).— 

 Matsumura, Ind. PI. Jap. II. pt. 2, 32 (1912). 



CHINA. Formosa: Ape's Hill, Playfair (No. 458, type; ex Hemsley); same 

 locality, summit, A. fl'enr-y (shrub 0.6 m. tall; with ripe fruits) ; Takow, December 

 11, A. Henry (No. 2035, in Herb. New York Bot. Garden; tree 7 m. or more tall; 

 with old leaves, fruits wanting). 



The leaves of No. 2035 are up to 5.5 cm. long and 2.5 cm. wide, and are somewhat 

 rough on both surfaces, the midrib and nerves being not quite so glabrous as in the 

 other specimen. The fruits are described as " glauca," but Henry says nothing 

 about the color. The stones may be described as: ovato-elliptica, hilo paullo pro- 

 ducta vix apiculata et foveolata, lateraliter compressa, facie exigue foveolata v. 

 pro parte maxima levia, tenuiter costulata, circiter 5 mm. longa et 3.5 mm. crassa. 

 The flowers are unknown, and the relationship of this well-marked species re- 

 mains uncertain. 



10. Celtis formosana Hayata in Jour. Coll. Sci. Tokyo, XXX. Art. 1, 272 {Mat. 

 Fl. Formos.) (1911). 



Celtis sinensis Henry in Trans. As. Soc. Jap. XXIV. Suppl. 85 (List PI. Formos.) 



(non Persoon) (1896), quoad No. 1616. 

 Celtis philippinensis Matsumura & Hayata in Jour. Coll. Sci. Tokyo, XXII. 369 



(Enum. PI. Formos.) (non Blanco) (1906). 



CHINA. Formosa : Kelung and Pachina, 1896, T. Makino (type ex Hayata) ; 

 Naibun, February 1907, G. Nakahara (co-type, ex Hayata) ; Bankinsing, A. Henry 

 (No. 1616, in Herb. Arnold Arb.; tree 17 m. tall; with ripe fruits). 



I have not seen a type specimen, but Henry's No. 1616 agrees well with 

 Hayata's description, only the petioles and pedicels are a little longer, measuring 

 about 12-14 mm. each, while according to Hayata the petioles are 8 mm. long and 

 the pedicels 10 mm. long. The dried leaves are whitish beneath and densely papil- 

 lose under the microscope. The species is totally different from C. philippinensis 

 Blanco which belongs to sect. Solenostigma. Hayata says it is " near to C. aus- 

 tralis Linn, and C. tetrandra Roxb.," but the former differs widely in its densely 

 cuspidate-serrate leaves and in other respects, while the latter, as far as I know, 

 has distinctly furrowed and ribbed stones, which are almost smooth in Henry's 

 specimen and may be described as : ovato-oblonga, versus hilum truncatum paullo 

 productum leviter foveolata, ceterum facie levia, vix costulata, lateraliter vix v. 

 paulo compressa, circiter 7 mm. longa et 5 mm. crassa. This species may be most 

 nearly related to some of the hitherto very little known subtropical or tropical 

 species of this section from southeastern Asia. 



11. Celtis Bungeana Blume. See p. 269. 



12. Celtis cerasifera Schneider. See p. 271. 



