68 NATURAL HISTORY OF PLANTS. 



Adanson and admitted by many authors after him, closely approach the 

 BosacecB with concave receptacle and single ovary. The Celastraceoe^ 

 on the contrary, at least those which had then been studied, pre- 

 sented a receptacle, either convex or much less concave, and their 

 insertion was nearly that of a great number of hypogynous types. 

 Again, it was thought that the micropyle, exterior in the ascending 

 ovule of the Celastracece^ became constantly interior in that of the 

 Rhamnacece when it had the same direction. But the study of a 

 large number of more recently discovered types conclusively proves 

 that these two great differential characters between the two groups 

 are not at all constant. In Ferrottetia (notably in Caryospermum)^ 

 Frauenhofera and in many other genera of the Celastracece^ especially 

 in certain Mortomas, the concavity of the receptacle and the mode 

 of insertion of the perianth and androecium become evidently what 

 they are in a great number of the Rhamnacece ^ and it has been truly 

 said that apart from the situation of the stamens, the flowers of these 

 Celasiracece were altogether those of Rhamnacece. It might be 

 added that the habit, the foliage, the inflorescence, the fruit, the 

 direction of the seed, might be in one point or another identical. 

 Under these circumstances, to separate the two families, there remains 

 only the oppositipetalous character of the stamens in the Rhamnacece^ 

 invariably alternipetalous in the Celasiracece^ a character which we 

 admit to be sufficient, although it would not be so in other natural 

 groups, since we retain the two families as distinct ; but we main- 

 tain i this consequence of what has just been established : ^' that the 

 Rhamnacece^ oftener perigynous or epigynous than the Celasiracece^ 

 but not constantly, might strictly and justly be considered a series 

 with oppositipetalous stamens." This character suffices to dis- 

 tinguish them from a great number of other families, especially from 

 those which constituted the Buckthorn Order of Jussieu. Ad. 

 Bbongniart^ has completely differentiated them. The Ilicinece^ whose 

 coroUa is most often gamopetalous and which, in their descending 

 seed with micropyle interior and superior, have an abundant 

 albumen, with a small apical embryo, are neighbours of the 

 Ehenaceoe and Sapotacew.^ The Staphylece^ studied in the family 

 of Sapindacece^"^ have neither the androecium nor the ovarian cells 



^ Adansonia, xi. 273. 3 dq. Theor. Elem (ed. 1), 217. 



2 Ithamy,. 11. * Nat. But. of I'lants, v, 342, 392. 



