278 Transactions of the Society. 



flattened, which would obviously favour dispersal by the wind. 

 Where the pods are narrow, as in Nasturtium sylvestre and Brassica 

 Sinapistrum (Charlock), there is only one row of seeds; where the 

 pod is broader, as in Nasturtium amphibium and Brassica (Dip/o- 

 taxis) tenuifolia, there are two. 



In many species the seeds are carried away as food by animals, 

 and being no doubt often dropped, are thus dispersed. In others 

 the seeds are much flattened, and no doubt carried by wind. 



In some species of Cardaminc and Dentaria the valves of the 

 pod open elastically at maturity and scatter the seeds. 



In some species of Brassica the pods terminate in a kind of 

 beak which often contains one or two seeds. It seems possible 

 that they may thus escape being eaten by birds. 



Lepidium sativum, the common Cress, is remarkable for its tri- 

 partite cotyledons. This character is perhaps due to a longitu- 

 dinal folding in ages long gone by, so as to enable the embryo to 

 fill the seed. 



A Brazilian species of Cardamine, C. chenopodifolia, produces 

 underground pods as well as others of the common aerial type. 

 These underground pods differ in being shorter and containing 

 fewer seeds. We shall find one or two similar cases in other 

 orders, and the reason I think is that if there were a number of 

 seeds they would interfere with one another, and all but one or 

 two would perish. 



Eesedace^e. — In the genus Reseda the seeds are contained in a 

 capsule as in some preceding genera, but it is unique in the fact 

 that the cup is open long before maturity. It contains numerous 

 seeds arranged along a number of placentas ecpual to that of the 

 styles. 



The seeds are rugose, but so finely that they appear smooth to 

 the naked eye, and are black with a lustrous sparkle. 



Those of R. lutea are much larger than those of R. luteola. 



Violarie^e. — In the Violariese, again, the fruit may be an 

 indehiscent berry, or a capsule opening elastically by as many 

 valves (3) as there are placentas. This is the case with our only 

 indigenous genus, Viola. The species, however, fall into two 

 groups. In one (V. hirta, V. odorata, etc.), fig. 68, the capsules 

 nestle on the ground, and are even said (as, for instance, by 

 Vaucher) to bury themselves. They are, at any rate, pushed among 

 moss, decaying leaves, etc., close to the ground. In other species 

 (V. canina, fig. 69) the capsules when open resemble an inverted 

 tripod. Each valve contains a row of from three to five brown, 

 shining, pear-shaped seeds, slightly flattened at the upper (free) end. 

 As the capsule dries the sides approach one another (figs. 70, 71), 

 and grip the smooth seeds more and more tightly, till at length 

 the attachment is ruptured and the seeds are thrown several feet. 

 I have suggested elsewhere that we get a clue to the existence of 



