£31 THE FLOWER. 



and a single style. A pistil of two carpels may be two-celled, with 

 two placentfe, two styles, or two stigmas, &c.* 



* Tliere are, however, some exceptions which qualify these statements : — 



1. Each placenta being a double organ (556), it occasionally happens that 

 the two ]iortions arc separated more or less, as in Orobancliaccous plants, wliere 

 a dicarpellary ovary appears on this account to have four parietal phiceiitie ; 

 either approximate in pairs (as in our Cancer-root, Conopholis), or equidis- 

 tant (as in Apbyllon). 



2. Analogous to tliis is the case where the two constituent elements of the 

 stigma (the only essential part of tlio style) separate into two half-stigmas, as is 

 partially seen in Fig. 494, 495. The stigma, no less than the placenta, belongs 



to the margins of the infohled leaf (545), these margins being 

 ovaliferous in tlie ovary and sligmatiftrous in the style ; as Mr. 

 Brown, the most profound botanist of this or any age, has 

 clearly shown. These two constituent portions of the style or 

 stigma occasionally separate, either entirely or in part, as in 

 Euphorbiaceous plants, in Grasses, and especially in Drosera 

 (Fig. 510), where there are consequently twice as many nearly 

 distinct styles as there arc parietal placenta in the compound 

 ovary If the two comi)oncnt parts of the style of eacli carpel 

 were reunited into one, in the usual manner, tlieir number 

 <^r'''*^ would equal the placentse, and their position would ha alter- 



\ f nate with the latter. But since, in parietal placentation, each 



^"' h(i1f-placenta is confluent (not with its fellow of the same 



cai-pel, but) with the contiguous half-pktcenia of the adjacent carpel, it were surely 

 no greater anomaly for the elements of such half-siigmas as those of Drosera to 

 follow the same course. This is precisely what takes place in Parnassia, and in 

 other cases wliere the stigmas are opposite the parietal placentaj ; — cases which 

 were thouo-ht to be very anomalous, merely on account of the adoption of a 

 false principle (that of the necessary alternation of the stigmas and ijlacenta), 

 but which are really no more extraordinary than parietal placentation itself 



3. Furthermore, the production of ovules is not always restricted to what 

 answers to the margins of the cai-pellary leaves. In the Poppy, the whole sur- 

 face of the long, imperfect partitions is covered with ovules ; in Butomus, they 

 are borne over the whole internal face of each carpel, and in Watcr-Lilies over 

 the whole surface, except the inner angle of each cell, where alone they normally 

 belong. Reduced to two in the allied Water Shield (Brasenia, Fig. 684), the 

 ovules grow from the dorsal suture, or the midrib of the caqjellary leaf alone ! 

 And in the allied Cabomba itself we usually find its three ovules, one on tlie 

 dorsal and one on the ventral suture, and the third on some variable part of the 

 face of the cell in the vicinity of either suture. In Obolaria, Bartonia (Centau- 

 rella, Michx), and in several species of Gentian, a compound one-celled ovary is 

 ovuliferous over the whole face of the cell ! 



All placentation is very differently explained by those who adopt the hypoth- 



riG. 510. Pistil of Drosera filiformis, with three deeply two-parted styles : the ovary cut 

 across, showing three parietal placentae. 



