VMBELLIFEBM. 



171 



the importance of its study in this family. Sprengel ^ and Hoff- 

 mann ^ especially had recourse to it in establishing their classification 

 of the UmheUiferce ; so also had Koch/ ten years later. Then A. P. 

 DE Candolle published his " Observations on the Family of the 

 Umbelliferae," ^ and a description of the family in his ''Prodromus." ^ 

 He there admits 157 genera, and preserves the Araliacece^ as a distinct 

 family with only 13 genera. The UmheUiferce proper he divided 

 into seventeen tribes, distributed in three sub-orders, now inadmissible, 

 Ortlwspermce, Campylospermce and Goelospermce, About ten years later 

 Endlicher,'' adopting the same divisions, with the same number of 

 genera of Araliacece,^ enumerated 195 genera of UmheUiferce. Ten 

 years since Bentham and Hooker ^ retained only 153 genera of 

 UmheUiferce and 38 of AraUacece,^^ whilst B. Seemann ^^ admitted 43 

 in the latter, though he did not include the plants with imbricate 

 petals nor those with a unilocular ovary. We now reduce the total 

 number of genera to 113 of which 25 belong to the Araliece.^^ We 

 have arrived at this result by uniting numerous types of UmheUiferce 

 which, in our opinion, are separated by no generic difference ; the 

 same among the AraUece. In the latter only we have established 

 some new genera, as Apiopetalmn, Pseiidosciadium and Eremopanax; ^^ 



J Plant. Uinbe.lL Prodr. (1813). 



2 Syllab. PL JJmhelL ; Gen. (1814). 



3 Mem. Acad. Nat. Cur. (1824). 



4 ColL Mem. (1829). A complete history of 

 this family will be found in this memoir. 



«iii. (1830) 55, Ord. 92. 



6 Log. cit^2ol, Ord. 93. 



7 Gen. 762, Ord. 162. 



8 Loc.cit. 793, Ord. 168. 



9 Gen. 859, 1008, Ord. 80. 



io Loc. cit. 931, 1009, Ord. 81. 



1^ Revis. Hederac.Journ. ofBot. ii.-vi. (1864- 

 1868). 



12 Without counting those which are too im- 

 perfectly known to be retained or classed defi- 

 nitely : 



1. Chamcele (Miq. Ann. Mus. Lugd.-Bat. iii. 

 59). This genus, represented by a single 

 Japanese species (C. tenera), is considered by 

 the author as belonging to the SeseUnecB. The 

 unripe fruit we have examined was compressed 

 perpendicular to the partitioji ; which would 

 ally the plant to Carum. 



2. Asciadium (Griseb. Cat. PL Cub. 118). 



3. Galbanum (Don, Trans. Linn. Soc. xvi. 

 603 ; — Endl. Gen. n. 4486). A genus created 

 for a Persian plant which produced the galba- 

 num, whose fruit, alone known, having meri- 

 carps compressed parallel to the partition and 

 having at the same time primary and secondary 

 ridges, could not be assimilated to any of the 

 genera we have studied. 



4. Opoidia (Lindl, Bot. Mag. xxv. [1839] 64). 

 A genus likewise established for a galbanum 

 plant, considered by Bentham and Hooker 

 {Gen. 920) very like a Pcucedanum., left by 

 Boissier {Fl. Or. ii. 1089) among those of un- 

 certain place. However, after remarking that 

 the mericarps have three obtusely angular pri- 

 mary ridges each containing a thin vitta, and 

 flat furrows to each of which corresponds a 

 wide and deep vitta, with four commissural 

 vittse, the author is inclined to think that this 

 Persian herb probably belongs to a group of 

 iimyrnieoe. 



5. Platyraphe (Miq. Ann. Mus. Lugd.-Bat. 

 iii. 56) " Ajnminea." 



^'"^ AdaHsonia,xu. 130, 133, 158 (1878). 



