ANGIOSPERMS. 
5^5 
surmount the ovary, which may therefore be considered to be composed of a single 
carpel which rises up from the floral axis as an annular zone. In Piperaceae 
however the stigma, which is sessile on the apex of the ovary, is often placed 
obliquely or divided into several lobes ; and this, like the two or four styles 
which surmount the ovary of Naias^, indicates that the ovary is not composed of 
one but of several carpels, which first make their appearance, like the leaf-sheaths 
of Equiseium^ as an unbroken ring, which only at a later period becomes resolved 
at its upper margin into teeth. This hypothesis appears the more admissible 
since, in other Angiosperms where a comparison with nearly allied forms justifies 
Fig. 391. — Longitudinal section of the flower of Rheian 
undiilaHcin ; s leaf of the outer, / of the inner perianth-whorl ; 
a aa three of the nine anthers, ovary, n stigma, k k nucellus 
of the ovary, dr glandular tissue at the base of the filaments 
forming tlie nectaries. 
Fjg. Tß-z. — A^iagaliis UTVoisis ; A longitudinal section of 
a young flower-bud, / sepals, c corolla, a anthers ; A' carpel ; 
J> apex of the floral axis ; B the gynaeceum further developed, 
the stigma n being now formed, and the ovules on the central 
placenta S ; C the gynagceum ready for fertilisation, / pollen- 
grains on the stigma n, gr style, 5 central placenta, ^A'ovules ; 
D unripe fruit, the placenta 5 has become fleshy and swollen 
so as to fill up the spaces between the ovules. 
us in inferring a number of coherent carpels, these carpels originate as an un- 
divided annular zone which developes into the ovary, style, and stigma ; as, for 
instance, in Primulaceae (Fig. 392) (free central placentation). In Polygonaceae, on 
the other hand, where the ovary also forms eventually a closed cavity containing the 
central ovule (Fig. 391), the cohesion of two or three carpels to form the ovary may 
not only be recognised from the corresponding number of the styles and stigmas ; 
but separate carpels appear at first distinct on the floral axis, and only amalgamate 
in the course of their growth, their zone of insertion becoming elevated as a ring. 
Since the wall of the ovary does not in any of these cases form placentas from 
^ I am unable to understand why Magnus calls the wall of the ovary 'perianth.' 
