340 ON LYEI.LIA, LEPTOSTOMUM,
in those species having quadrangular capsules they nearly reach the parietes opposite to the external angles, towards which they are directed.
As the columella of Polytrichum retains its regular form in the ripe capsule, its real structure may even then be in a great measure determined. In this stage I find its wings, or compressed sides, to consist of a double membrane with an intermediate spongy substance, in which there is no appearance of granules ; and the internal denser substance of the axis is equally free from granular matter. But as there is no indication of lateral rupture, the sides in several species remaining perfectly smooth, nor of any central cavity, this structure affords a powerful argument in refutation of those hypotheses which assume the existence of two kinds of granules in the capsules of mosses ; the one produced in the cavity formed by the internal membrane, the other in the substance or supposed cavity of the columella itself; the latter being considered as seeds in one of the hypotheses, 1 and in the other as pollen. 2
This argument, however, is not here advanced to disprove the existence of two kinds of granules in the capsules of mosses, but merely against their production in the distinct cavities assigned to them in the hypotheses referred to.
In the greater number of Polytricha as well as in Lyellia and Dawsonia the seeds are extremely minute ; a fact with which the increased surface for their production is probably connected : for in P. undulatum, where the seeds are larger than in most other species of the genus, this increased surface does not exist : and in P. Jcevigatum, where they are of still greater size, the plicae of the inner membrane are probably also wanting.
5711 Although there is but little resemblance in the structure of the peristomium among the different genera of Poly- trichoidecB, they may still be said essentially to agree in the function of this part : for in all of them the complete separation of the seeds is ensured by the smallness of the apertures for their discharge. It may be remarked, that the
1 Palisot Beauvois, ^Etheog. p. 5. 2 Keith, Physiol. Bot. ii. p. 346.
�� �