OF ARTS AND SCIENCES. 321 



ter's figure ; but that was intended to represent the whole pistil, and 

 the villous hairs ai-e placed, not upon the base of the style, but around 

 the ovary, — indeed, they belong to the corolla. 



The inflorescence is axillary throughout the genus. This is stated 

 by Blume only ; but De CandoUe specifies the axillary flowers under 

 almost every species. 



The species of our Polynesian collections are evidently three, 

 viz. : — 



Geniostoma ligustrifolium, a. Cunn., of New Zealand, known 

 by its didymous or bicapitate stigma, and its triangular pointed stipules. 



Geniostoma rupestre, Forst. ; a polymorphous species, to which 

 I should refer, not only Blume's G. Hcemospennum, but probably Ben- 

 tham's G. crassifolium, at least the variety glaherrimum, and even 

 a plant which has the calyx assigned to G. Ciimingianum, Benth. 

 Our forms of the species — besides the typical one, which is smooth, 

 and with rather small and pointed leaves, like those of the New Zea- 

 land species — may be arranged under the varieties ellipticum, macro- 

 phyllum, and puherulum. The species may be known from the last by 

 the transversely truncate stipules, and the entire stigma, raised on 

 more or less of a style, never absolutely sessile ; nor is the corolla 

 wholly glabrous within, as in 



Geniostoma astylum {Gray, I. c.) : glaherrimum ; stipulis truncato- 

 bifidis ; foliis ovalibus ; sepalis ovatis ; corolla intus glabro testivatione 

 quincunciali ! ; stigmate integro subgloboso sessili ; fructu immature 

 anguste oblongo. — Tahiti, Society Islands. 



The specimen in the Hookerian herbarium from the Sandwich 

 Islands referred by Bentham to his G. crassifolium, ^3. glaherrimum, if 

 really from those islands, is likely to belong to a species which on the 

 whole is better placed in the genus 



Lahordea. Judging from the plate in the Botany of Freycinet's 

 Voyage, Bentham presumed the estivation of the corolla of this im- 

 perfectly known genus to be valvular, and the fruit to be baccate. 

 The first point would seem to have been verified by M. Bureau, who, 

 according to Mr. Bentham's supplementary note, " has been enabled 

 to dissect three flowers of this plant \_Lahordea fagrceoidea, Gaudich.]. 

 He confirms the presumed valvular aestivation of the corolla, but al- 

 ways finds two cells only to the ovary, and very plausibly suggests 

 that the three-celled one examined by Gaudichaud was accidentally 



vol. IV. 41 



