558 ON THE RELATIVE POSITION OF
These are the piiiicipal luodificalions of the coiiipouncl ovarium when forming a simple series ; bnt it is necessary to observe that both smfaces of the infiectecl and included portions of the carpels are not unfrequently equally pro- ductive of ovula, a structure which is manifest in numy Cf/rtandracece, especially Cyriaiulra, although in several other genera of the same family the production is confined to the inner or upper surface of the margin. In other cases the polyspermous ovuliferous portion or placenta is connected with the inner angle of the cell by a single point only, which may proceed either from the apex or base of the cavity. This modification of structure, though in some families hardly of generic importance, seems to me to assist in explaining the apparently anomalous structures of Hydnora, liafflesia, and Brugmaima.
On the subject of the origin and type of Stigma, my first observation is, that the style where present can only be regarded as a mere attenuation, in many cases very gradual, of the whole body of the ovarium. Hence the idea natu- rally suggests itself, that the inner margins of the carpel, which in the lower part are generally ovuliferous, in the upper part perform the different, though in some degree analogous, function of stigma. As the function, however, of this organ implies its being external, and as in different famihes, genera, and even species, it has to adapt itself to nrr various arrangements of parts destined to act upon it, corresponding modifications of form and position become necessary ; hence it is frequently confined to the apex, and very often, es[)ecially in the compound ovarium with united styles, appears to be absolutely terminal.
In such cases, as it must always include and be closely approxinjated to the vascular cord of the axis, it has by some botanists been considered as actually derived from it, which it is, liowever, only in the same manner as the marginal pla- centae are derived from the axis of the carpel. But according to the notion now advanced, each simple pistillum or carpel has necessarily two stigmata, which are to be regarded, not as terminal, but lateral.
That the stigma is always lateral may be inferred from its
�� �
