64 MB HIEEN, ON EBENACE^E. 



rous, and the section Trichanthera of MABA approaches closely the section Rospidios of 

 DIOSPYROS. The geographical distribution of the two genera is nearly identical and co- 

 extensive. MABA however is more frequent in New Caledonia and other islands in the 

 South Pacific Ocean as well as in Australia, and one species (M. buxifolia, Pers.) seems 

 to have the widest range of any species of the family. 



The best character to distinguish MABA is its 3- or 6-celled ovary, but some spe- 

 cies both of EUCLEA and DIOSPYBOS occasionally, but not I think normally, possess this 

 peculiarity ; I have therefore made this the fundamental character of the genus, and feel 

 no doubt, notwithstanding certain cases of perplexing variability, that it is convenient 

 for the practical classification of the family to maintain the genus MABA. I have on the 

 other hand merged with it Alphonse De Candolle's genus Macreightia, Dalzell's genus 

 Holochilus (the only species however of which I have not seen), and Hasskarl's genus 

 Rhipidostigma, none of which three genera by such incorporation weakens the main cha- 

 racters of MABA. 



DIOSPYROS being the largest genus, and indeed including the majority of the species 

 of the family, exhibits the greatest amount of variation, and possesses points of contact 

 with all the other genera. I have united with it the following genera, Cargillia R. Br., 

 Leucoxylum Blum., Noltia Schum., Gunisanthus Alph. DC., and Rospidios Alph. DC., inas- 

 much as I fail to find any good or even plausible ground for maintaining any separation 

 amongst them. Cargillia with two species (C. laxa and C. australis) was made a genus by 

 Robert Brown principally on account of having 2 ovules in each cell of a 4-celled ovary. 

 The former species often and so far as my observation goes always has an 8-celled ovary 

 with a solitary ovule in each cell, and it is identical with Diospyros maritima Blum. ; and 

 the second species (C. australis) has sometimes at least an imperfect septum, partially dividing 

 each cell of the 4-celled ovary. 



Again in certain other species of Diospyros, the septa of the ovary when 8-celled are 

 alternately thinner, and are therefore difficult to discern ; so much so that Dr Solander in 

 his manuscript notes now in the British Museum described the immature fruit of D. chloro- 

 xylon as 4-celled with 2 seeds in each cell, although the ovary is really 8-celled. 



In some cases in the lower part of the ovary the alternate septa are imperfect and do 

 not reach the axis, and therefore a transverse section across this portion would give a 

 4-celled ovary ; but in the middle part of the ovary the same septa are joined with the 

 axis, and a transverse section there would shew an 8-celled ovary. 



Failing also to detect any peculiarity of habit to distinguish the genus Cargillia, I am 

 obliged notwithstanding the eminence and reputation of its inventor to treat it as a mere 

 section of DIOSPYROS. For a similar reason I consider Macreightia to be a section of 

 MABA. 



The genus ROSPIDIOS was founded on the combination of a 3-celled ovary associated 

 with a tetramerous symmetry in the flower; but this observation was made on a cultivated 

 specimen, while the wild specimens since observed reveal an 8-celled ovary : the genus 

 therefore lapses into DIOSPYKOS ; the name however I have retained for a section of the latter 

 genus. 



