SPICILEGIA FLOR.E SINENSIS. 



229 



opinion, I agree with Mr. Bentham in thinking probably distinct, 

 the segments of the leaves of the sterile shoots are narrower than 

 those of the flowering stems. 



29. SciEVOLA (Crossotoma) hainanensis, sp. nov. Frntescens, 

 caulibus diffusis cortice glabro snberoso cinereo obductis, ramulis 

 novellis hii'tulis, foliis saepius fasciculatis spathulato-linearibus 

 obtnsis V. obtusiusculis glaberrimis 6-7 hn. longis 1-1^- lin. latis 

 axillis cinereo-lanatis, floribus axillaribus solitariis sessilibus v. 

 breviter x^edunculatis bracteolis binis foliis homomorphis calycem 

 duplo siiperantibus suffultis, calyce glaberrimo l^- lineali lobis latis 

 rotundatis tubo triplo brevioribus, corolla extus glabra 4-5 lin. 

 longa, indusio ciliato. 



Cu'ca Hoi-han, ins. Hai-nan, aest. 1877, coll. Bullock. (Herb, 

 propr. n 20252.) 



An interesting i^lant, very closely allied to S. spinescens, E. Br. ! 

 from which, however, it seems truly distinct by its woolly leaf-axils, 

 sessile or sub-sessile flowers, foliar bractlets and developed calyx- 

 tube. It is the only extra- Australian representative of the small 

 section to which it belongs. 



30. Jasminum trinerve, Vahl. Li fruticetis prope pagum Sai-nam, 

 secus fl. North Eiver, prov. Cantonensis, d. 15, Sept., 1866, legg. 

 Sampson et Hance ; in ins. Hai-nan invenit Bullock. 



Recorded only from Upper India and Java. 



31. Styrax serndatum, Hoxb. In silvis densis circa monasterium 

 buddhicumFiloi-tsz, secus fl. North River, prov. Cantonensis, d. 18, 

 Sept., 1866, frf. coll. Sampson et Hance : in fruticetis paulo infra 

 verticem montium Pak-wan, florif. invenit Sampson, sub fine 

 Marth, 1868. 



32. Wrightia puhescens, R. Br. Secus viam a portu Hoi-hau ad 

 Kieng-chan, metropolin ins. Hai-nan ducentem, legi Novemb. 1866. 



A shrub or small tree of about twelve feet, the foliage much like 

 that of W. tomentosa, but only hairy beneath, and with small glan- 

 dular guttules beneath, and a very delicate reticulation of deep 

 purplish vemlets between the hahs ; whilst the leaves of IV. tommtosa 

 are quite pannose on the under surface, so that the network cannot 

 be seen. The branchlets, too, of the Chinese plant are quite 

 smooth. The follicles green and not quite ripe when Pgathered, 

 are eight inches long, quite soldered into one, drying black, and 

 marked throughout their whole length with innumerable elevated 

 striae and interjacent fiuTows, the seeds 5 lines long, with a coma 

 about an inch in length. 



33. Marsdenia / iirceulata, J) one. In ins. Hai-nan, juxta Hoi-hau, 

 a. 1877, leg. Bullock. I have seen no authentic specimens of this 

 species, which had only been gathered in Cochin-china, but the 

 present ones agree very weU indeed with the diagnosis, except that 

 the 4-5 flowered cymes have a very short common peduncle, 

 instead of the flowers being quite sessile. 



34. Limnanthcmum cristaUim, Griseb. In stagnis ditionis 

 Amoyensis, legi Oct., 1857 : prope Li-pau, secus fl. North River, 

 prov. Cantonensis, m. Sept., 1866, coll. Sampson et Hance. 



35. Hydrolea zeylanica, Vahl. In uliginosis extra Cantonem, 



