228 Mr. Bentham's Observations on some Genera of Plants 



authors, however, appear to have much studied the generic character, which 

 they have taken more or less from some of the above-mentioned botanists, 

 Don observing only " Genus fortfe iterum dividenduni." 



Since the above I am not aware of any modification in the character or spe- 

 cies of Symplocos, until the publication of the last volume of G. Don's General 

 System of Gardening and Botany, where all the hitherto published species are 

 collected, those merely named in Wallich's Catalogue are described, the genus 

 is retained as established by L'H^ritier and Bonpland, but raised to the rank 

 of a natural order, and divided into three sections: Alstonia, containing all 

 the American species said to be distinguished by an 8 — 10-parted corolla, the 

 segments in a double row, stamens in 3 or 4 series, and a half inferior drupe ; 

 Lodlira, consisting of 17 Asiatic species, to which are attributed a 5-parted 

 corolla, stamens inserted without order, and an inferior drupe ; and Palura, 

 described as having the same corolla, with stamens in a triple series, and an 

 inferior ovary. Under this section are enumerated two remaining Asiatic 

 species. 



These characters, however, by no means correspond with the specific cha- 

 i-acters given in the same work to several of the species, and will be found on 

 examination still more at variance in many instances with the plants them- 

 selves. Thus in the section Alstonia, three species at least have a 3-parted 

 corolla, the segments in many of them are not in a double row ; the stamens 

 of S. tinctoria are arranged as in the Asiatic species, and the calyx is as ad- 

 herent to the drupe in Alstonia as in Lodhra and Palura. In the latter respect 

 I cannot see any difference between the S. sinica and the several plants referred 

 to Lodhra ; and if there is any greater regularity in the arrangement of the 

 stamina in S. crata^goides than in Lodhra, it is that they are more decidedly 

 pentadelphous and not biseriate. 



Amidst all these conflicting opinions, after a careful examination of a con- 

 siderable number of both American and Asiatic species, it appears to me that 

 there do exist three distinct groups, which it might be adviseable to consider 

 as so many genera. In the true Symplocos of Linnseus the stamina are erect, 

 the filaments are flat, monadelphous at the base, free in the upper part, where 

 they are distinctly imbricated in three or four rows, and suddenly attenuated 

 below the anther; the corolla is erect and adherent to the staminal tube, 



