38 BOTANY OF SOCOTRA. 



Distrib. Endemic. 



A small tree or small shrub in general appearance not unlike G. salvifolia, 

 Heyne, but distinguished readily by its oppositifoliar inflorescence and large 

 turbinate fruits. In well-developed fruits there are four stones, and each 

 is usually multilocular (up to 6), but frequently, as is common in the genus, 

 all the stones are not equally developed, and then the smaller ones have only 

 one loculus. Neither Schweinfurth nor our party obtained flowers of this 

 species. The inhabitants give to it the same name as to the next one. 



5. G. bilocularis, Balf. fil. in Proc. Roy. Soc. Edin. xi. (1882), 504. 



Arborescens glabra ; foliis breviter petiolatis magnis ovatis acutis basi cordatis aBquilateralibus ; 

 cymis bitloris fere oppositifoliis ; pedimculis petiolo suba?quilougis ; drupis 4-pyrenis 

 4-lubatis subcubicalibus glabrescentibus nitidis ; pyrenis bilocularibufi. 

 Arbor parva ramulis glabris. Folia magna 4^-7 poll, longa 2£- 3-^ poll, lata ovata acuta basi 

 5-nervia cordata, crenato-serrata subcoriacea pilis pancis adpressis dispersis subtus vestita 

 cseteroquin glabra ; petiolus §-f poll, longus subpilosus sub lamina expansus. Cymce 

 biflor-se v. abortu 1-floree suboppositifoliae rbacbi pedunculisque subiucanis petiolo subsequi- 

 lougis. Flores ignoti. Drupce 4-lobatse §-f poll. diam. subeubicales aurese nitidse siccoe 

 subrugosas glabrescentes basin filamentis staminum marcidis cinctas ; pyrense biloeulares, 

 loculis superpositis monospermis. 

 Nom. Veen. Eireit (B.C.S.). 



Socotra. Near Adona, at an elevation over 3000 feet. B.C.S. nn. 498, 

 516. 



Distrib. Endemic. 



Species of this genus are in the dried condition notably hard to identify, and, 

 as Masters has pointed out, the synonymy is most complex and the whole genus 

 wants revision. Nevertheless this plant and the preceding are so very marked, 

 and their characters are so unconformable with those of any described species, 

 that I have, although our specimens are incomplete, named them as novelties. 



The present species finds its nearest ally in G. bracteata., Roth. (Nov. Spec. 

 243), a plant of the Carnatic and Ceylon, from which its nearly glabrous 

 leaves and fruits on almost oppositifoliar axes separate it. It is a rare plant in 

 Socotra ; we have it from only one locality. 



This plant frequently sends out adventitious twigs (n. 516) bearing very 

 small leaves, often hardly \ inch in diameter, somewhat cordate, deltoid in form 

 with prominent dentations, the teeth and margin being coloured dark purple, the 

 rest of the leaf green, and the whole being very hairy. Thus the plant is 

 heterophyllous, Schweinfurth sends specimens with this character, which he 

 refers to his n, 475, which is G. turhinata, but I doubt if this is correct, and 

 think his specimens are really like ours from G. bilocularis. 



