PHANEROGAMIA : FLOWERS. 429 



become torn from the cells beneath them, and then surround the seed as 

 a loose sac (e. g. in Drosera and Parnassia\ or, transformed in a pe- 

 culiar manner into an elastic tissue, tear open and discharge the seed (in 

 Oxalis). The rest of the tissue of the integument, beneath the epi- 

 dermis just described, is very variously developed. Sometimes next the 

 epidermis is a layer of looser cells, with intercellular passages or spaces 

 (e. g. in Leguminosce\ into which, in the only cases, Canna and Ne- 

 lumbium, where the epidermis exhibits stomates, these latter lead. We 

 usually find next to the epidermis and firmly attached to it, a thin layer 

 of parenchyma (the whole outer integument), and then, distinct from 

 this, a special coat consisting of an extremely delicate layer of cells (the 

 inner integument alone, or with the nuclear coat) : this is the case in 

 most LiliacecB. 



A different formation is not unusually met with, where two integuments 

 exist, and the inner is not formed by a mere fold of the epithelium. Here 

 the epithelium of the inner integument is generally exactly in similar con- 

 ditions to some of those generally described above, while the outer in- 

 tegument becomes gradually suppressed, and falls off in shreds (e. g. in 

 Euphorbiacece), or remains as a thinner envelope (e. g. in Cistacece, 

 Thymelacce, Laurctcece). Elegant spiral thickenings also occur here in 

 the epidermis of the inner integument (Lauracece, Sparrmannia afri- 

 cana (?) and others). 



The occurrence of spiral, reticulated, and porous deposit-layers in 

 the epidermis of the seed is so usual that it is not worth while to 

 enumerate the cases now. A very rich variety of forms is exhibited, 

 for example, in the Scrophulariacece, especially the Verbascece and An- 

 tirrhinece, but almost all the Solanacece, particularly those with a berry- 

 fruit, exhibit sometimes true spiral fibres (e. g. Solanum\ sometimes 

 reticulated thickening (e. g. Datura]. It is most remarkable, however, 

 that this structure of the epidermis occurs extremely rarely in the seed- 

 buds of the Monocotyledons, which almost always have two integuments, 

 and is presented in the Dicotyledons, particularly in the Monopetalce, 

 which usually possess only one integument. 



In the formation of new vascular bundles in the integuments of the 

 seed, I have hitherto found the law almost unexceptionably confirmed, 

 that the vessels are never distributed in the nucleus, or inner integument, 

 but only in the outer, or the simple integument. Treviranus had pre- 

 viously propounded the law opposed to this, that vessels are only formed 

 in the inner integument, because, starting from ripe seeds, he confounded 

 the very hard and thick epidermis of many seeds with their outer inte- 

 gument, and the parenchyma of the same .with their inner integument. 

 Link * has the same incorrect assertion, here doubly wrong, since he dis- 

 tinctly refers the shell of the seed (testa) to the outer integument, and 

 the inner coat (membrana internet) to the inner integument of the seed- 

 bud. 



That the operculum of the radicle (fig. 265. c.) may be formed from 

 very different parts, follows from what has been said in the paragraphs. 

 Mirbel first demonstrated its peculiar development in Commelinacece and 

 Marantacece ; I myself in Canna. 



172. Very important alterations take place in the funiculus 

 also during the development of the embryo. It has been remarked 



* Elem. Phil. Bot. ed. 2. vol. i. p. 285. 



