'228 SPICILEGIA FLOE^ SINENSIS. 



23. CEnanthestolonifera,'DC., var. (= Dasyloma japojiicnm, Miq.) 

 In incultis circa Cantonem, a. 1864, leg. Sampson. 



24. G^nanthc stolonifera, DC, var. (= Dasijloma hipinnatum, 

 Miq.) Secus fl. Lien-cliau, prov. Cantonensis, m. Octobri, 1876, 

 coll. Rev. J. C. Nevin. 



25. Geophila reni/ormis, Don. In fissuris lapidnm circa monas- 

 terium biulcUiisticum Fi-loi-tsz, d. 18, Sept., 1866 : in silvis ad 

 Ting-ii-shan, secus fl. West River, m. Julio, 1870, coll. Sampson 

 et Hance. 



Extensively distributed within the tropics, but not before 

 detected in China. 



26. Paederia tomentosa, Bl. ? In ins. Hai-nan, juxta Hoi-hau, 

 aest. 1877, coll. Bullock. If Mr. Kurz is correct in reducing 

 (' Journ. As. Soc. Bengal,' xlvi., 139) to the above-named species 

 Miquel's P. harhulata and P. densiflora, the Hai-nan plant unques- 

 tionably belongs here. It is, however, almost quite glabrous, the 

 inflorescence only being strigillose. I distinguish specifically the 

 ovoid-berried plant from India (of which beautiful specimens were 

 distributed in Hooker and Thomson's Khasia collections) from the 

 round-fruited one referred by Bentham (Fl. Hongk., 162) to 

 P. fcctida, Linn. This name probably belongs to the first-men- 

 tioned plant ; for Linnseus ascribes to it a "bacca ovata" (' Mant. 

 Plant,' i., 62) ; and he appears, moreover, to have founded his 

 species particularly on Rumphius' " Convolvulus foeiidus," which 

 that author describes as having the fruit " sensim oblonga," and 

 figures it as ovoid ('Herb, amboyu.,' v., 436. t. 160.) I would 

 propose the name of P. chinensis for the Chinese and Ja^^anese 

 plant, which is in reality much more nearly allied to P. fietida than 

 to P. tomentosa, to which latter Mr. Kurz refers it with a mark of 

 interrogation. Gaertner describes the fruit as "globosa 1. ovata," 

 but figures it as globose ('De fruct.,' iii., 85, t. 195), after a 

 specimen in the Banksian herbarium, the native country of which 

 he does not state. 



27. EnjKitoriwn stccchadosiniim, Hance. In prov. Cantonensi, 

 ad fauces Tsing-yiin, fl. North River, necnon ad fauces Yeung-shui, 

 fl. Lien-chau, copiose crescentem, ofiendit m. Octobri, 1876, Rev. 

 J. C. Nevin. The first time this has been found wild. The serra- 

 tures of the leaves are frequently deeper than in the cultivated 

 specimens, and the achene is destitute of glands. 



28. Artemisia parviflora, Roxb. var. ? In arena litorea profunda 

 ad Pak-sha, extremitate australi prov. Cantonensis, ipse legi, 

 d. 20, Nov., 1866. I believe the specimens gathered, which look 

 very different from the Indian plant, and approach somewhat to 

 A. scoparia, W. & K., are referable to Roxburgh's species, but the 

 leaves of the flowering stems are very fleshy, smooth, with nari'ow 

 linear divaricate segments, those of the young sterile shoots cano- 

 tomentose, with broader cuneate trifid lobes. I suspect this varia- 

 tion from the normal form is due to the locahty where it grows, 

 exposed to heavy gales, which blow the loose sand up in enormous 

 hillocks all along the coast, and carry it in clouds inland. In 

 A. japonica, Thunb., which, contrary to Miquel and Maximowicz's 



