OF THE FOSSIL PLANTS OF THE COAL-MEASURES.

235

e, springing from the portion of the testa at a', have now increased in size. Kg. 55 isthe third of this descending series of sections. We now see that two of the crenulationsof the testa, at e, e, are assuming the form seen in fig. 53, e, whilst those at <?' have nowwholly detached themselves from the testa, a', with which they were connected in figs.53 & 54. The fourth section (fig. 56) now shows the testa a as a strong well-definedmembrane of uniform thickness, whilst the crenulated outlines, e, of fig. 55 are now yetmore entirely separated from the testa.

The value of this unique, though crushed, fragment lies in the clear proof which itaffords us that the remarkable structure e, e', which I propose to designate the canopyof the lagenostome, originates superiorly in the hard testa of the seed. Fig. 57, whichrepresents a portion of the testa a , with the crenulations of the canopy e intermediatebetween the sections figs. 55 & 56, further demonstrates the origin of some other curiousstructures seen in sections to be described. At a! a cross bar of the testa unites thatstructure with the centre of each of the crenulated curves of the canopy e. Inthe other crenulations this bar is imperfect, only projecting from the testa like atooth and not reaching the concavity of each crenulation. It is now, in fact, a smallprojecting ridge running down the inner surface of the testa, but which has disappearedbefore reaching the section 56, a. Figs. 58, 59, 60, & 61 represent a series of four ver-tical sections of another specimen made in parallel planes. Fig. 58 is made nearlymidway between the most prominent lateral surface of the seed and its centre. Fig. 59is made midway between fig. 58 and fig. 60, which latter has been carried exactly throughthe central axis of the seed ; whilst fig. 61 is from a plane nearly corresponding withfig. 59, but on the opposite side of the seed to the latter section. Fig. 58 exhibits thetesta a , which appears to have been a hard, dense, sharply defined structure, having athickness of from •()! to ‘005, but usually having a mean thickness of '006. I have notbeen able accurately to determine its composition, but there are indications that it con-sists of an assemblage of oblong cells. Immediately within it is a second and muchthinner membrane /, but which obviously has adhered closely to the inner surface of thetesta, as is shown by the numerous irregular shreds of tissue which still link the twotogether. I have designated this the nucular membrane. This term seems convenientlyto represent M. Brongaiart’s term “ membrane nucellaire.” Within this again is thethird membrane g, which is also extremely thin. This I regard as the membrane imme-diately investing the endosperm, the “ membrane perispermique ” of M. Broagaiart.

In fig. 59 all these tissues reappear at the lower part of the seed, as in fig. 58; butat its upper part we find new conditions. The endospermic membrane g is there removedmuch further from the testa than at its opposite end. The nucular membrane/splits intotwo portions at f',f. The outer layer, e, e, ascends towards the upper part of the testa,where it is inflected into four or five crenate curves, the concavities of which are directedtowards the inner surface of the testa, the outermost ones on each side being the largest.Within this structure, which I have designated the canopy, we find another space partlyfilled by a very delicate parenchyma, h, and enclosed superiorly by the innermost, c,of the two layers into which the nucellar membrane splits at/',/'.