SPACES CONTAINING AIR AND WATER. 215 



continuously — e.g. the petiole of Nuphar luteum, according to Frank; they are 

 usually divided into chambers by numerous transverse plates, or diaphragms, which 

 will be described below, or, in the case of stems, are interrupted, at least at the nodes, 

 by plates of parenchyma. 



The structure of the walls of large air-cavities shows many notable phenomena. 

 Stem, leaves, and also roots of many water and marsh plants are traversed by nume- 

 rous passages or chambers, and these, especially the schizogenetic ones, are separated 

 in most cases only by plates of parenchyma one layer of cells thick: it is only where vas- 

 cular or fibrous bundles traverse them that these septa are several or many layers in 

 thickness. Meyen^ has drawn attention to the fact that the cells of the one-layered 

 lateral septa are often quite uninterruptedly connected over large areas, a condition 

 which seems in fact to be frequent, e. g. in the stems of species of Potamogeton and 

 Myriophyllum, the stems and leaves of Papyrus, Scirpus lacustris, the petioles of the 

 Nymphaeacese, and of Pontederia crassipes; according to investigations already made, 

 I cannot join in this statement with full certainty : in thg first place, because of the 

 great difficulty of proving beyond doubt the absence of very small interstitial spaces, 

 and in the second, because the passages in the petioles of Ceratopteris and Villarsia 

 parnassifolia communicate one with another by very narrow spaces, and those in 

 the internodes of species of Marsilia even by rather large ones. An indirect com- 

 munication of the passages occurs also where the sides are thus completely closed, 

 through the interstitial spaces in the leaf-lamina, the nodes of the stem, and in 

 certain cases through special diaphragms to be described below. 



Schizogenetic spaces are limited by the smooth membranes of the cells which 

 form their wall. In many large passages, e.g. Nuphar-, or even in cavities, 

 e.g. Rhizome of Aspidium Filix mas, these are covered by a delicate cuticle. 

 It remains for further investigation to decide how far this phenomenon is of genera] 

 occurrence. 



The wall of the lysigenetic spaces is as a rule more or less covered by the 

 remains of those disorganised cells, at the expense of which the space originated, or 

 the latter is here and there loosely filled with remnants of cells. Various special 

 lorms originate in connection with this, according as the transitory cells are 

 mechanically torn asunder, or are broken down, or dried up, or according as 

 these several phenomena are combined. 



In the pith of most internodes of stems, which become hollow, i.e. perforated by a 

 large, axile, lysigenetic air-passage (Grasses, Unibelliferac, Compositae, Equiseta, &c.), the 

 formation of the hollow begins by those transitory roundish or polyhedral cells, which do 

 not follow the growth of the tissue surrounding them, becoming first separated from one 

 another, so as to form schizogenetic cavities which gradually increase in size. The cells 

 of the tissue thus broken up then gradually lose their protoplasm, dry up, and coalesce to 

 flaky or membranous masses, which are attached to the wall of the cavity. The whole 

 process begins in an internode, either simultaneously at many points in the pith (e.g. 

 Phragmites), or in the middle line, spreading centrifugally outwards from it (e. g. Cicuta 

 virosa, peduncle of Taraxacum-^). 



Essentially the same process, on a smaller scale, occurs in the small air-passages which 

 alternate with the vascular bundles in the internodes of Equisetum : also, according to 



' Physiol. I. p. 295. 2 Frank, I.e. p. 155. ^ Frank, I.c p. 145. 



