REFERRED TO HEDERACEJE. 205 
celled ovary referred to the latter, and justly considered an anomaly in 
a group where the bulk of the plants have ovaries from 2- to many- 
celled. These genera are Arthrophyllum, Blum., Bursinopetalum, 
Wight, Pukateria, Raoul, and Cuphocarpus, Dene. et Planch. 
Arthrophyllum, Blum., an East Indian genus of trees, presents, as 
now described, the additional anomaly of having opposite pinnate or 
bipinnate leaves with opposite leaflets, all the other Hederacee having 
alternate ones, with the apparent exception of the Sandwich Islands 
genus, Cheirodendron, Nutt., where those of the terminal flowering 
branches are opposite. The idea of removing Arthrophyllum to Cor- 
nace, in which Order one-celled ovaries are admissible, has probably 
occurred to all who reflected on the subject, but in that Natural Order, 
as now constituted, no plants with bipinnate leaves are to be found. 
The difficulty, however, admits of a solution. If Cornus Suecica, Linn. 
(Fl. Lapon. t. v., Engl. Bot. t. 310), be examined, it will be seen there 
is a subterranean rootstock which throws out ramified branches with 
simple opposite leaves and often terminal flowers. As soon as the fruit 
is ripe, these branches wither and die off. Now, if this subterranean 
rootstock were erect, Cornus Suecica would be exactly like Arthro- 
phyllum, and the annual branches would probably be mistaken for 
leaves, as those of the Indian genus were. What I maintain, then, 
is that Arthrophyllum has annual, deciduous, occasionally dichotomous 
branches, or rather branchlets (like Cicca disticha) with opposite simple 
and entire leaves, the terminal ones of which produce flowers and fruit. 
The real branches (as in Phyllanthus, Cicca disticha, and Cornus Sue- 
cica) are woody, and do not fall off as the branchlets do. In the ab- 
sence of specimens, an inspection of t. 14 of vol. i. part 1 of Miquel’s 
Fl. Nederl. Ind., representing Æ. ovalifolium, Jungh. et De Vries, will 
tend to make this clear. There is therefore no reason why the genus 
Arthrophyllum should not be incorporated with Cornacee. 
Bursinopetalum, Wight, was placed by its founder in Olacinee, by 
Miers in Aguifoliacee, and by Thwaites € Mise. vii 
Araliaceg, in which he was followed by C. Koch (‘ Wochenschrif, 1859, 
p.372). But ina subsequent publication (Enum. Ceylon Plants, p. 
42) Mr. Thwaites, abandoning his opinion, referred it once more to 
Olacinee. More recently Bentham and Hook. fil. (Genera Plant. 
p. 345) assigned to it a place amongst Cornacee, with which arrange- 
ment I concur. 
