EVPHOBBIAOE^. 155 



In the present state of our knowledge, there is only one character 

 common to all the Euphorhiacece : theii' descendent ovules, whose 

 micropyle is turned upwards and outwards. The number is always 

 defluite, but there is sometimes one and sometimes two in each cell. 

 This is the character we employed in the first place, and we still 

 think it is the only one practicable ; we have divided all the 

 Euphorbiacece into uniovulate and biovulate. Other characters, 

 formerly considered to be constant in this family, are now found 

 to be only so in the majority of cases, but are wanting in certain 

 exceptions. First there was the presence of albumen around the 

 embryo ; but this sometimes disappears or more often is reduced to 

 a membrane in certain species whose thickened cotyledons become 

 l^lano-convex, without which, after our manner of limiting the genera, 

 we might exclude other species which have all the other common 

 characters, but where the albumen is thickened in consequence of the 

 foliaceous conformation of the cotyledons. The existence of a 

 placentary projection called by us obturator, is also a nearly constant 

 character, and this orgun often attains svich a development, that it 

 much exceeds the size of the ovules iaserted below it. Evidently 

 we could not, for such differences, place two plants in two distinct 

 genera or even in two distinct families. The Euphorbiacece are all 

 provided with diclinous flowers, according to most authors ; it has 

 been seen that we only consider this character as very frequent, but 

 not as constant. The divisions of the family in Frodromus ^ are 

 based on the form of the cotyledons, certain Euphorhiacccp Iiaving 

 them much larger than the radicle {Flatylobece'), whilst others liave 

 them tliick, semi-cylindrical, or nearly so, and of about the same 

 size as the radicle (^Stcnolobea) ; on the prsetioration of the calyx, 

 sometimes valvate, sometimes imbricate ; on the presence or absence 

 of petals ; on the mode of insertion of the androceum, sometimes in 

 the centre of the receptacle, sometimes under the base of a central 

 body (usually a riidimentary gynseceum -) ; on the form of the 

 stamens, ■«'hose anthers have the cells longitudinally adnate to the 

 connective, or free and only attached by an extremity, and whose 

 filament is upright from the bud or incurvate at tliat period in such 

 a manner as to bear outwardly the fi'ont of the anther, which is 



' See the table in this volume, p. 189. ' Frequently described as a central disk. 



