428 MORPHOLOGY. 



taceum) is the epidermis of the inner integument, closely connected 

 with the parenchyma of the same, and the membranous layer of the 

 nuclear membrane ; in Chelidonium, on the contrary, the brittle layer is 

 the whole external integument clothed with delicate epidermis, and the 

 membranous layer is the inner integument. In Ricinus, therefore, the 

 outer integument would come in as stratum externum evanescens ; in 

 Chelidonium, the epidermis as stratum membranaceum me.dio arete 

 adhcerens. To make the confusion quite perfect comes next the circum- 

 stance, that the different observers, in the analysis of ripe seeds, have 

 determined the number of coats, in dissections made according to vari- 

 ous methods, or in transverse sections under weaker or stronger magni- 

 fying power, according to the variations of the cells quite undistinguish- 

 able in them : so that a seed has often been determined to have a simple 

 seed-coat, which has two or three ; others actually with a simple mem- 

 brane, favoured with two or three seed-roots, on account of differences 

 in the development of the cells. From the small number of observa- 

 tions which have as yet been published by Brongniart, Mirbel, Brown, 

 and myself, it already follows, with perfect certainty, that every deter- 

 mination of the coats of the ripe seed is altogether useless, if their nature 

 has not been demonstrated by the course of development. 



The case mentioned in the beginning of the paragraphs, of the coch- 

 lidiosperms of the species of Veronica, has appeared to me as yet the 

 most difficult subject of investigation, and I was obliged to take the 

 research up several years, one after another, till I had completed it, 

 since, in addition to all the other abnormalities, totally unsymmetrical 

 formation of the seed-bud occurs, which renders the investigations very 

 much more difficult. 



The most general condition is where the epithelium of the external or 

 the simple integument, or the nuclear coat, is developed in a remarkable 

 manner. Thus, in most plants, especially those which have a hard shin- 

 ing seed (e. g. the Leguminosce), it is converted into a tissue, composed 

 of relatively long prismatic cells, placed perpendicularly to the surface 

 of the seed, and usually much thickened, even to the partial obliteration 

 of their cavities. In other plants, particularly such where the seed, 

 when placed in water, becomes coated with gelatinous matter, it consists 

 of cylindrical cells placed in the same manner, but with thin walls, and 

 densely filled with gelatinous matter (Quinces, Plantaginacece), and which 

 frequently contain elegant spiral fibres (many Polemoniacece and Cucur- 

 bitacetz). Here it is often easy to observe the gradual filling of the cell 

 with starch, the solution of this to gum, and the conversion of this into 

 the very hygroscopic gelatine, while the spiral deposits are simultane- 

 ously formed on the wall.* More frequently that gelatine is wanting, 

 and the cells, less cylindrical in form, project in a papillose manner as 

 hairs, or several unite together as little spines, gibbosities, or ridges, &c., 

 rendering the surface of the seed uneven ; or they form a flat surface, 

 while their walls are thickened by a variety of spiral, reticulate, or 

 porous deposits (in Hydrocharis, most Labiatas, Solanacece, Scrophu- 

 lariacecB). In a few cases these cells develope, in exceeding delicacy, to 

 a large size, and become filled with fluid, so that the seed resembles a 

 berry (in Punica granatum, in Ribes (?) ). Those cases are remarkable 

 in which these cells expand so much in the surface that they necessarily 



* See Miiller's Archiv, 1838, p. 152. ; and Schleiden's Beitriige zur But. vol. i. 

 p. 134. 



