14 SIR. D. OLIVER ON ATTRANTIACEjE. 



Atalantia, as DeCandolle himself observed, differing solely from 

 the other species of this latter genus in its ternary outer whorls. 

 I have not had good specimens to examine myself, hut M. De- 

 caisne states the ovary to be bilocular, as it is in several Ata- 

 lantias. T. glauca, an Australian shrub described by Dr. Lindley 

 in Mitchell's 'Tropical Australia' (p. 353), is a doubtful congener 

 of T. trifoliata. For the present I leave it in the genus. Before 

 determining where finally to dispose it, the examination of more 

 numerous flowers is desirable. I have thought that perhaps it 

 might be associated with Atalantia (Limonia) missionis, an Indian 

 species, rather isolated in its general fades as well as by precise 

 characters. T. sarmentosa, Bl., is quite unknown to me, excepting 

 from the description. Wight and Arnott think it probably not 

 a Triphasia. I have not re- written the generic description of Tri- 

 phasia for this paper. 



LraoNiA. 



The solitary Limonia originally described by Linnrrus in the 

 'Species Plantarum' (ed. 2), is a very distinct plant, Xanthoxy- 

 leoua in its general features, but with the flowers and fruit of a 

 true Aurantiacea. The figure of it in Eheede's « Hort. Malab.' 

 (iv. 14), quoted by Linnams, is very characteristic ; as to Eumpf's 

 figure, also quoted by him, I am not clear if it belong to the same 

 species. This genus has been most unmercifully hampered by very 

 diverse species which have been allotted to it, especially by the 

 older botanists. Of the eleven species described in the ' Prodro- 

 mes' (three as doubtful Limonias), 1 is a Murraya?, 1 Skim- 

 mia, 2 Glycosmis, 1 Triphasia, 1 Micromelum, while 3 are refer- 

 able to L. acidissima. The remaining 2 may also be Limonias : 

 they are described by Lamarck from Madagascar and Mauritius 

 specimens, and would seem to be allied to L. alata. I have not 

 re-written the description of the genus, as I do not feel that the 

 material to which I have had access would enable me to offer a 

 satisfactory one ; especially failing Lamarck's plants, and seeing, 

 further, that the two species which I still retain in Limonia differ 

 from each other considerably in habit and some floral characters. 

 It has occurred to me that L. alata may be related to Luvunga 

 more nearly than has been suspected. I find the geminate ovules 

 of this species to be sometimes both obliquely superimposed and 

 collateral in cells of the same ovary. Limonia missionis I have 

 removed to Atalantia, with the species of which it appears to me 

 to have closer affinity than with Limonia proper. 



