AXGIOSPERMS. 



377 



by the cohesion of two or three carpels, both from the corresponding number of 

 the styles and stigmas and from the fact that the carpels are at first separate on 

 the floral axis, and unite only as they develope, their zone of insertion rising at 

 an annular wall. Since in all these cases the wall of the ovary does not form 

 placentas, the number and position of which would indicate the number and position 

 of the carpels, we are driven to direct observation of the early stages of development 

 and the numerical relations of the styles and stigmas. We have to deal moreover 

 in this case with morphological relations, which are not yet made out with sufficient 

 certainty, notwithstanding the many researches that have been made into the develop- 

 ment of the flower. 



Beside the number of carpels which have united to form the ovary, it is im- 

 portant also to know in this division whether in any given case the ovule appears 



FIG. 308. Rheum undulatum. Longitudinal section of the flower; s leaf of the outer,/ leaf of the inner perianth- 

 whorl, a the anthers, three only of the nine being visible, f the ovary, the stigma, kk nucellus of the ovule, dr glan- 

 dular tissue at the base of the filaments forming the nectaries. 



FIG. 309. Anagallis arvensis, A young flower-bud in longitudinal section ; / sepal, c petal, a anther, A" carpel, 

 s the apex'.of the floral axis. older gynaeceum after formation of the stigma and the ovules on the axial placenta .y. 

 C gynaeceum ripe for fertilisation ;/ pollen-grains on the stigma x.^rthe style, 5 the axial placenta bearing the ovules SA". 

 D unripe fruit ;.f?-the style; the placenta 5 has become pulpy and so swollen that it fills the spaces between the seeds .S'A". 



as the terminal structure on the floral axis, or laterally on the axis. It is obvious 

 that the ovule may be a terminal structure of the floral axis, where there is only 

 one ovule springing from the base of the ovary, as in the Piperaceae, Naias, and 

 the Polygonaceae &c., and it has been shown in fact by the researches of Hanstein 

 and Schmitz, Magnus and Payer that not only the ovule as a whole, but the nucellus 

 itself is to be considered as a terminal structure. But it cannot be concluded from 

 this that every ovule which springs from the bottom of the cavity of the ovary 

 necessarily represents the apex of the floral axis, for it is conceivable that the axis 

 may produce an ovule at the side of its apex, though it is not itself advancing, a case 



