58 NATURAL HI/^TORY OF PLANTS. 



distinguislied at a glance from the Malvaccce and Geraniacem by 

 tlieir entire leaves, and from the Caryophyllcw by theii* alternate leaves 

 (•with the excejition of two species)." These diiferential characters 

 are however still too arbitrary, and they can only be artificially 

 separated. "We can only imagine," we have said,' "that in the 

 Linacem the union of the carpels among themselves, according to the 

 axis of the gynseceum, is carried further than in the Geraniacece. 

 These are the only ones among which we observe the types with free 

 ovaries, as in Biebersteinia and Floerkea {Limnanthes)^ and the only 

 ones consequently in which the style can become more or less 

 completely gynobasic." By the series where the genera are formed 

 of plants with woody or frutescent stems, the Linacem are connected 

 with certain other families, among which we find types of inde- 

 pendent or nearly independent ovaries ; these are first the Mal- 

 pighiacece^i then the Euphorhiacece and Ternstroemiacece. These latter 

 have sometimes been placed with I.ronaufhcs, which may Lave as 

 many as twenty staniens ; but we know that these are pcrigynous, 

 and if we could hesitate between a woody species of Linacece and 

 an oligandrous Tcrnstroemiacew ., we should recall the fact that in 

 these the stamens, otherwise free, are united with the base of the 

 corolla, while in the Linacea) it is not with the petals they unite, 

 but, when they are not independent, which is the rule^, it is among 

 themselves that they form a short tube or a sort of lu'ceolum from 

 which the free part of the filament afterwards separates. It is so 

 in the Hiigoniece and Erythroxiileiv. In doubtful cases there remains 

 the dii'ection of the ovular regions, the micropyle being su2)erior 

 and interior in the Ternstroemiacece^ bvit constantly exterior and 

 capped by an obturator in the Linacece. In the Malpighiacece ., the 

 general floral organisation is nearly the same as in the woody 

 Linacece ; but they have opposite leaves, which is only the case here 

 with AneulopJms ; the sepals are generally furnished with one or 

 two exterior glands ; the seed is totally destitute of albumen ; the 

 embryo is not generally straight, and the ovary cells are uuiovulate. 

 This is never the case here except in ErutJtroxylon, cliaractcrised by 

 its alternate leaves, axillary stipidos, monadelphous stamens, and 

 ovary cells generally sterile, except one. The Euphorhiacece whose 



' tn Adidisonin, x. 3G0. - Except in tho Moitmineie. 



