118 MERRILL AND ROLFE. 



Leaves 5- to 10-foliolate. 



Panicle-branches scattered along the common rachis; leaflets 7 to 10, strongly 



caudate-acuminate 1. 8. caudata. 



Panicle-branches fascicled at the ends of the branches, no common rachis 

 present; leaflets 5, slightly acuminate, 3-nerved at the base.. 2. 8. Ctimingii. 

 Leaves 3-foliolate; leaflets caudate-acuminate, not 3-nerved at the base. 



3. 8. trifoHata. 



1. Schefflera caudata (\iilal) Merrill & Rolfe conih. nov. 

 Heptaplcurum camlatum Vidal Plian. Cuming. Pliilip. (1885) 175; Rev. PI. 



Vase. Filip. (188G) 145, excl. Cuming 800. 



Hchefflcra acttminatissima Merr. in Philip. Jouni. iSci. 1 ( 1906) Suppl. 109. 



Lrzox, Province of Albay, Tivi, \'i(lal 1D3, H29a, the former being the type 

 of the species ex descr. !: Province of Bataan, Mount IMariveles, Whitford 112, 

 1222; For. Bur. 3005 Meyer. 



The original diagnosis of lleplapleuium caudal urn Vidal, applies entirely to 

 Vidal 793, and not at all to Cuming 800, although the latter is the first number 

 cited, and as no type was indicated, Cuming 800 would naturally be taken to 

 represent the type of the species, unless the diagnosis was examined carefully 

 and compared with the original specimens. Schefflera acuminatifisima Merr., was 

 described as "quite distinct from Heptaplcurum caudatum Vidal," owing to the 

 fact that the conception of Vidal's species in Manila was based on a specimen 

 of Cuming 800. As a matter of fact, however, the type of this species is identical 

 with Vidal 193, which we consider to be the type of Schefflera caudata. 



2. Schefflera Cumingii (8('om.) Merrill & Rolfe comb. nov. 

 Heptaplcurum Cumingii Seem. Journ. Bot. 3 (1865) 81; Rev. Hederac. (1868) 



45, excl. Cuming 800. 



Seemann describes this species as follows: "Eoliolis 5 cUipticis acuminatis v. 

 ovato-elliptieis longe acuminatis integerrimis 3-plinerviis; paniculis terminalibus 

 pube steliato albido vestitis; drupis obovatis, 5-locularibus. Pliilippino Islands 

 (Cuming! n. 800 et 1293)." 



As was the case with Schefflera caudata (Vidal), Cuming 800 is the first speci- 

 men cited, and would therefore naturally be taken to be the tj^pe of the species. 

 However, in 5 sheets of Cuming 800 that we have examined, including Seemann's 

 type material at the British Museum, all the leaves are 3-foliolate, and although 

 they are long-acuminate, they are not "3-plinerviis." In four specimens of 

 Cuming 1292 examined, the leaves are 5-foliolate, and althougli not long-acuminate, 

 are strongly "3-plinerviis." It is apparent that Seemann drew up his desoriptk)n 

 from both specimens, but mostly from tlie second number cited, the characters 

 of which predominate in his diagnosis, and which we consider to be the type of 

 the species. Cuming 800, while the first specimen cited in the original descrip- 

 tions of both the above species, was really described in neitlier, and is here de- 

 scribed as a new species. Seeman cites the nimilifr 1293, an error for 1292. 



3. Schefflera trifoliata Alcrrill & Rulfc sp. nov. 



Scaiidciis ; i'oliulis o, glabi'is, subiiK'iiihranac-cis, 10 ad 30 cin longis, 

 ()ltloii,i,ns vcl oblongo-ovati.'^, jiilcgris. caudulo-acuiiiiiiatis : pnniciilis icr- 

 iiiinaliltiis, raiiiis elongatis, imilt illoris, riirruraccis : llorilms sii|i('ris i'as- 

 ciuiilalis, j)r()j)c I'aiiinlonini hasiii iimlx'llal is. 



Scandciil. glaliroiis. l)raiuiics liglit-gi'a_y. ljoa\('s altciiiali'. tril'nlidlaU', 



