THE MORPHOLOGY ОҒ THE PRIMULACEA. 291 
In the place of the normal stamens, but wholly detached from the corolla, were 
numerous free hypogynous pistilodie stamens, showing various degrees of transition 
between the conformation of the stamen and that of the carpel (Pl. XLI. fips. 4-7). 
The filamentary portion of these pistillodie stamens was, in some cases, but little changed ; 
in other instances it was more or less dilated and leafy, and often crumpled or spirally 
twisted. In some the filament was dilated into an open carpel, destitute of ovules, but 
prolonged into a long style terminated by a stigma (РІ. XL. fig. 2). The ovarian portion 
was sometimes absent; and then the appearance was that of a style only (Pl. XL. fig. 3). 
In other eases the midrib of the filament or its central vaseular bundle was thickened, 
and bore imperfect ovules (Pl. XX XIX. figs. 33-39). It is especially noteworthy that 
this placenta-like rib, while generally congenitally adherent to the leafy or carpellodic 
filament, was not unfrequently detached from it by chorisis, the significance of which 
fact will be further insisted on in the sequel. 
In another series of these flowers the dilated filament was irregularly and pinnately 
lobed, the terminal lobe being incurved—a general occurrence, indeed, in the flowers 
in question. Тһе uppermost of the lateral lobes very often bore marginal ovules, 
while the lower ones were infolded at the edges and prolonged into a long style and 
stigma (Pl. XL. fig. 1). 
Here, then, we have a branched or divided carpellary leaf with some of its lobes open 
and others closed, both ovule-bearing. We have, in fact, the exact homologue of the 
so-called compound or branched stamen. Тһе bearing of these facts on the nature of 
the free central placenta will be made apparent hereafter. 
In the illustrations just cited it was the filamentary portion which was especially 
involved; but in another set of flowers the filament was unaffected or nearly so, and the 
changes were confined to the anther. 
In the Rev. G. E. Smith's sketch the filaments were all normal, while in Miss Dowson's 
specimens they were usually more or less dilated. Тһе changes observed in the antheral 
portion of the stamen, if stamen it can be called, were similar to those described in the 
case of the filament. Of anther, properly speaking, there was little or no trace, though 
‚ the central placentary rib may be taken as the homologue of the connective, the latter 
prolonged occasionally into a style. In no case was I able to find any pollen at all. 
Тһе ovules, indeed, occupy the situation of the pollen; but it could not be expected that 
there should be any transition between the pollen-cells and the оущез. Nor was I able 
to find, as I had hoped I might do, any trace of pollen in the ovules, as I have before 
seen in Passiflora and Rosa. 
It should further be noted that the pistil in these flowers was quite normal *. 
* Singular as these bisexual phyllomes are, they are relatively by no means unfrequent in certain plants, such as 
Sempervivum tectorum, Cheiranthus cheiri, and other plants mentioned in teratological records (see ‘ Vegetable Tera- 
tology,’ p. 319). Even Conifers occasionally show similar phenomena. Thus, at the time of writing this paper, 
Mr. Syme, of the Elvaston Nurseries, sent me a catkin of Cupressus Lawsoniana in which the lower bracts were 
antheriferous and the upper ovuliferous; and, more than this, one of these bracts (РІ. XLI. figs. 9-11) bore on the 
outer surface, at the lower edge, an anther-cell, while on the inner face was a shortly stipitate ovule. (Ersted has 
recorded similar phenomena. 
