294 NATURAL HISTORY OF PLANTS. 



because the periantli of tlie female Howers comes off in a circular 

 piece (figs. 335, 330) to discover the carpels, which it at first covered 

 completely. In this respect Mo/Zi/wc/ia is to Peiniiits what Hedy- 

 carya is to Ilorton'ia. The flowers are dioecious,' and even in the 

 same species the perianth varies in form with the sex. It forms a 

 globular, turbinate, or nearly campanulate sac, usually split up into 

 four lobes of variable length, imbricated and decussate in tlie bud. 

 The two outer lobes are not always similar to the inner pair, and 

 sometimes behave quite differently on the expansion of the flower.* 

 The stamens are most frequently very numerous, from twenty to 

 sixty in number, inserted over the whole surface of the perigonal 

 sac in vertical rows, two or three superposed to each division of the 

 calyx. Each stamen consists of a short filament first inflexed and 

 then erect, and a basifixed anther shaped like a horse-shoe. Its two 

 cells^ surround the edges of an oval connective continuous with the 

 filament, and dehisce each by a longitudinal cleft, the two clefts 

 appearing single when the dehiscence is completed. The perianth 

 of the female flower has also an opening whose edges are split into 

 four imbricated decussate lobes. On the bottom of tlie receptacle 

 formed by the dilatation of the pedicel, we find an indefinite 

 number of carpels crowded together, and resembling those of licdy- 

 carya, each with a suspended ovule, whose micropyle looks upwards 

 and inwards (figs. 330, 330). The drupes and seeds are also the 

 same as in the genus Jledycarya. 



The true Mollinedias are of American origin ; a couple of species 

 come from Mexico, and the twenty-five others belong to South 



* They are said to be monoDcioiis in excep- equal, triangular, and firstercct, afterwards spread- 



tional cases (Hentii., PI. Hartiveq., 250). They ing. The periitnth is campanulate anil quadritid. 

 may be incomjiletely hermaphrodite, either be- ^ Tliey are cither exactly marginal, or slightly 



cause the female flowers present rudimentary introrsc ; more rarely, nearly extrorse. There 



sterile stamens towards the throat of the peri- are really two cells to the anther ; but after de- 



nnth, or because at the very bottom of the recep- hiscence the two clefts, which were at first di«- 



tacle of the male Howers are contained ill- tinct, coalesce at the \\\»}\. In the species with 



developed carpels, with, however, a rudimentary elongated anthers, after dehiscence we see two 



ovule in each ovary. This is very marked in panels, one internal the other external, each 



most of the male flowers of M. elliplica (Af. formedof two half-cells, and separatin;; from the 



nitida TuL. Teirafome elliptica Gauun., other from above downwards ; they then take very 



Uook.Journ., 1H12, 530). difrorent forms and directions. The one remains 



' The two inside are often larger, thinner, flat or nearly so, or else its very thin wlges are 



and with the edges less entire than the two out- retluxed outwards, while the other (usually the 



side, besides iK-ing often more rellexed. In the inner one) is much more niark«tlly imhrieate. 



male flowers of M. liffimlrina Ti'l. {Ann. Sr. The opi'ii anther thus presents a conformation 



Aat., B6r. 4, iii. ft), the four divisions of the that is sometimes very peculiar, and it may up- 



culyx, deeply separateil from one another, are pear unilocular. 



