SPICILEGIA FLOR^ SINENSIS. 15 



sent specimens of a fourth species, unfortunately in an equally 

 unsatisfactory condition, and also entirely leafless. It is 

 unquestionably distinct from the others, and, judging only from 

 the figures (t. ix., figg. 6, 8, 9-12), m Dr. Eegel's ' Monographia 

 Betulacearum,' pubhshed at Moscow in 1861, I was disposed to 

 refer it to the free he at first considered as a variety sibirica of 

 B. nana, Linn., but which, in his second review of the order for 

 DeCandolle's ' Prodi'omus,' he transferred as a variety rotundifolia 

 to B. glandidosa, Mx., for it agrees well both in the shape of the 

 catkin- scales and in the scarcely margined nucules. In this case it 

 would, of coui-se, be a dwarf species, but I have no information as 

 to the habit of Dr. v. Moellendorff 's tree. It may possibly be new, 

 but this can only be determined when proper materials are obtained. 

 64. Salix tetrasperma, Eoxb. Ad ripas rivuli juxta Shek-mun, 

 ditionis Cantonensis, Nov., 1869, leg. Sampson. 



55. Salix ? popiilifoUa, Anderss. In ripis limosis, Cantone, 

 Jan., 1870, specimina $ florifera detexit Sampson. The leaves of 

 this plant exactly resemble Andersson's figure (' Monogr. Salic' 

 1. 1, f. 5), but the branchlets, instead of being glabrous, are densely 

 fusco-tomentose. Though, in the absence of female flowers, it is 

 impossible to form a trustworthy judgment, I doubt, notwith- 

 standing the very differently shaped leaves, whether this is more 

 than a variety of the preceding. My 8. cantonensis, according 

 to Andersson's classification, falls into the group Pleiandr(B,fragiles, 

 and I feel satisfied that its nearest relationship is with 8. fragiUs, 

 Linn. My diagnosis is so imperfectly copied into the ' Prodromus' 

 as to give no clue to its affinities. 



56. Zingiber Zerumbet, Rose. In silvis ch'ca coenobium ad 

 Ting-ii-shan, prov. Cantonensis satis copiose crescens, invenerunt 

 Sampson et Hance, d. 17 Julii, 1872. Undoubtedly truly wild. 



57. Microtis parvijiora, R. Br. In graminosis ckca Tam-sui, 

 ins. Formos^e, m. Aprili, 1864, legit infaustus Oldham. 



58. Lloydia serotina, Echb. In m. Siao-wu-tai-shan, ChinsB 

 borealis, m. Julio, 1876, coll. Hancock. 



59. 8milax perfoliata, Lour. In ins. Hai-nan, Martio, 1868, 

 coll. b. Swinhoe. On the stout main stem of this rare species the 

 cirrhi are fully developed, and arise from the axil formed by the 

 petiole and the upper margin of the very conspicuous amplexicaul 

 rigid scarious stipule ; in the branchlets, where the stipules are 

 much smaller, the cirrhi are reduced to short spinous processes, 

 occupying, of course, the same position. 



60. Arisoima ringens, Schott. In silvis ad Tam-sui, Formosae 

 septentrionahs, m. Apr., 1864, leg. def. Oldham. Only recorded 

 hitherto, I believe, from Japan. 



61. Amorphophallus campanulatus, Bl. Not particularly uncom- 

 mon, I heheve, in South China, but rarely flowering. Widely 

 spread over the tropical regions of continental and insular India, 

 in the Archipelago, Madagascar, and New Guinea. Very trouble- 

 some and disagreeable to dry, from the size and fleshiness of 

 the inflorescence, and its dreadfully overpowering odour of putre- 

 fying flesh. 



