GYNCECIUM IN ANGIOSPERMS. 267 



detected, while certain plants of the same families, of otherwise 

 identical structure, retain the dissepiments even in the fruit. 



500. But a similar condition may equally arise 

 from a modification of parietal placentation, namely, 

 with the margins of the leaves ovuliferous only at 

 bottom, and the placentae there conspicuously devel- 

 oped and completely united. The basal placenta- 

 tion of Dionrea is unavoidably so explained, its 

 nearest relative, Drosera (Fig. 553), having parietal 

 placentae. And this leads to a probable explanation 

 of the case in Primulacea?, where a large free central 

 placenta fills the centre of the cell, and no trace of 

 dissepiments can be detected. 1 



501. The idea maintained in former editions is 

 still adhered to ; namely, that placentae belong to 

 carpels and not to the cauline axis, in other words, 



that ovules arc productions of and. borne upon leaves, usually 

 upon their margins, not very rarely upon other portions of their 

 upper surface, rarely over the whole of it. 2 



502. Ovules cover the whole internal face of the carpels in 

 Butomus and its relatives, also of the Water-Lilies (both ?~ym- 

 phaea and Nuphar, Fig. 551) excepting the inner angle, to which 

 the}* are usually restricted in other plants. And in the allied 

 Brasenia and Cabomba, where the ovules are reduced to two or 

 three, one or more of them is on the midrib, but none on the 



1 The placenta in this and like cases is rather to be regarde;! as an out- 

 growth from the base of the carpellary leaves, combined over the floral 

 axis. Upon this interpretation, a central portion of the column may be 

 (and sometimes must be) of axile nature, yet the ovules be borne upon 

 foliar parts. See Van Tieghem, in Ann. Sci. Nat. ser. 5, xii. 329 (1869) ; 

 Celakowsky, Vergleichende Darstellung der Placenten, &c. (1876) ; Warming, 

 in Ann. Sci. Nat. ser. 6, v. 192. 



2 This view was first maintained as a general theory, and on critical 

 grounds, by Brown, in Plantae Javanica? Rariores, 107-112. ' Schkiden, End- 

 licher, and others took the opposite view, i. e., that ovules are productions of 

 the axis, even in parietal placentation, an exceedingly far-fetched suppo- 

 sition. In later days, the commoner view has regarded ovules as of both 

 origins, as productions of the carpels in parietal, of the axis in at least 

 some free central or basilar placentation. But at present the theory of 

 foliar origin without exception, revindicated by Van Tieghem, and espe- 

 cially by Celakowsky and Warming, again prevails. For the bibliography 

 and an abstract of the various views, see Eichler, Bliithendiagramme, espe- 

 cially the note in the preface to the second part (where he gives his entire 

 adhesion to this conclusion) ; also Warming's memoir, De 1'Ovule, in Ann. 

 Sci. Nat. ser. 6, v. 1877-78. 



FIG. 549. Vertical section through the compound tricarpellary ovary of Spergularia 

 rubra, showing the free central placenta. 550. Transverse section of the same. 



