MR. G. BE>^THAM ON MTRTACE^E. 133 



distinct. Others, again, have broken up what remains of the genus 

 into eight or nine small ones ; and, as long as only a few species 

 showing remarkable differences in the anthers, ovary, &c. are taken 

 into account, that course would seem justifiable ; but on a repeated 

 examination of about 45 species as we understand them, or of 

 above 60, if all those proposed by Sehauer, Miquel, and others be 

 adopted, the differences have been observed to pass so gradually 

 from the one to the other, that, even as sections, the best groups 

 we have been able to establish are but vaguely defined or purely 

 artificial. Of these sections we have adopted six, for which we 

 have taken the names of Schauer's genera, Hinzia, EiiryomyrtiiSy 

 Scliidiomyrtits^ Harmogia^ and Oxymyrrliine^ and of Lindley*s 

 Bahingtonia^ all of them founded upon single species (except 

 llarmogia, which had three) now placed in the corresponding 

 sections, but with sectional characters necessarily very much 

 modified by the grouping around them of additional species. 



The most constant character we have found to divide the genus 

 into two groups is one which appears to have been overlooked by 

 8chauer. In Itinzia and Euryomyrtus the stamens are never re- 

 duced below ten, and amongst them there is always one opposite 

 to the centre of each petal, usually larger than the others, or with 

 a more flattened filament ; whilst in the other four sections the 



■ 



stamens are much less definite in number, varying from under 5 to 

 above 30, and none are opposite to the centre of the petals, except 

 perhaps in B. polyandra^ where they are very numerous. Some other 

 characters are general, but not constant, in the two divisions. Tlie 

 ovules are reduced to 2 or 3 in each cell in several species of the 

 first, never in the second ; the anthers are uniformly parallel-celled 

 and rimose in the first, heteromorphous in the second. The ovary, 

 as observed by Schauer, is superior, or nearly so, in the original 

 Binzia and one other species, half superior in three others, and 

 almost entirely inferior in B. dimorjpliandra, which cannot other- 

 wise be removed from Rinzia. The section, indeed, is only distin- 

 guishable from JEuryomyrtus in the remarkable dilatation of the 

 filaments, B, diffusa alone in Euryomyrtus showing some approach 

 to it. 



The four sections of the second division are technically, but 

 unfortunately not very definitely, distinguished by the anthers ; 

 the cells are (as in the first division) distinct, parallel, and rimose 

 in Schidiomyrtus, united and porose in Babingfonia, intermediate 

 in Harmogia and Oxymyrrhine. The ovary is also 2-celled in all 

 the species of Schidiomyrtus^ except B. astarteoides, 3-celled in the 



