M. Treviranus on the Structure of the Fruit of the Cruciferse. 369 



the cord of the umbilical vessels is wanting on the depressed lower 

 side of the much-inflated silicle, the other alone being present, 

 running over the elevated side. The second occurs, for instance, 

 in Neslia paniculata, Rapistrum rugosum, Crambe orientalis ; and 

 indeed in the two last-named the silicles are only two-celled in 

 the upper joints, from the presence there of a perpendicular sep- 

 tum which is wanting in the lower. In Myagrum perfoliatum 

 and Erucaria aleppica no seed is developed in either of the two 

 chambers which occupy the uppermost part of the silicle. This 

 absence of the septum in the lower part of the fruit has its ana- 

 logue in the slit or opening which exists in the septum in that 

 situation in Farsetia and other genera, in the whole central part 

 in Tetrapoma, and it indicates the necessity of an uninterrupted 

 union between it and the style or the stigma ; while on the other 

 hand an interruption at the other end, namely the lower, appears 

 to be attended with no detriment to the formation of the fruit. 



If this be the correct view of the septum, it will then always, at 

 least originally, have a perpendicular position in the silicle. In 

 Bunias orientate indeed, where the ripe fruit is two-celled and 

 two- seeded, one cell with its seed lies over the other with an 

 almost horizontal septum *, but this is a consequence of develop- 

 ment. For in the earliest state of the fruit, when the stamens 

 have scarcely fallen off, the septum here is vertical as usual, and 

 the cells therefore originally lie side by side and not one above 

 the other. In Bunias Erucago, DeCandolle finds the silicles two- 

 celled in the young state, in a certain measure four-celled in the 

 fully developed, as each cell is frequently again divided into two 

 by a transverse septum f- But if a perpendicular section is made 

 through the somewhat oblique silicle, at the period when the 

 calyx, corolla and stamens fall off, in such a manner that the sec- 

 tion crosses that oblique position, the four cavities are already 

 visible. A membranous septum descends obliquely from above 

 downwards, which by projecting in and out forms two curves, from 

 each of which a fleshy transverse process runs to the outer wall. 

 The septa of both kinds are present therefore when the fertiliza- 

 tion is yet scarcely complete, so that we certainly cannot consider 

 the production of any of them an after-growth. 



In the genera Cakile, Crambe and Rapistrum,, the silicle is 

 divided into two joints by an apparent articulation in the middle, 

 the upper being commonly externally of a different structure 

 from the lower. By this means its cavity is indeed divided into 

 an upper and lower cell, but the two cells are never perfectly se- 

 parated, for there always remains an actual, though very narrow, 

 communication between them, as the apparent articulation con- 



* Gartn. d. Fruct. ii. t. 1 12. f Syst. Natural, ii. f>70. 



