131 VEGETABLE ORGANOGRAPHY. 



There may happen two cases which correspond to those 

 mentioned above. 



Sometimes the partitions may originally be in a state 

 of great tenuity, or the carpels, at the time of flowering, 

 may not be longer than the placentae, and then the 

 filament which proceeds from the placenta can reach the 

 base of the style and transmit the fecundating matter 

 to the ovules. This filament is destroyed after fecunda- 

 tion, either by the destruction of the partitions, or by 

 the elongation of the carpels; and then at maturity we 

 do not find it, and can only conceive how the fecundation 

 could take place by having recourse to the anatomy of 

 the ovary at the period of flowering. We see this in 

 all the Caryophylleae with a central placenta; sometimes 

 the filaments, which at the time of fecundation arise 

 from the placenta, are distinct, as in Lychnis dioica, 

 where there are five ; in Stellaria, where we see three 

 (PI. 20, fig. 1) ; sometimes they are all united into one, 

 as in Arenaria (PL 20, fig. 3, 4). An analogous organ- 

 ization is met with in the Portulaceae (PI. 20, fig. 7, 8), 

 where we find three distinct filaments; in Primula, where 

 the placentae are all united into a nearly globular body, 

 the filaments also being united into a point which 

 reaches the base of the style. In all these examples, 

 the filaments are wholly or in part destroyed after fe- 

 cundation, and the placenta seems isolated from the style. 



It may also happen that the branch of this pistillary 

 cord may go along the margin of the carpel which does 

 not retreat, as in fruits with parietal placentae, and thus 

 it arrives at the base of the fruit and to the seeds which 

 are there situated. It is this, probably, that takes place 

 in the fruit of Luzula, for example, where we find the 

 seeds attached to the base of the valve. 



The placenta is usually placed at the inner angle of 

 the more or less retreating part of the carpel, either 



