3l6 INTERCELLULAR SPACES. 



Frank, in those which alternate with the bundles in the leaf-sheath of many grasses, 

 the leaf-sheath and lamina of species of Carex, Luzula maxima, and albida. In the 

 last-named cases however, a further and slightly different condition seems to obtain, 

 which may be observed in the passages in leaves of Liliaceae, and Amaryllidaceae, Pan- 

 danus, and apparently in many other places : the group of cells, which occupies the place 

 of the future passage, first loses its protoplasm, the membranes become apparently 

 thinner, are partially dissolved, and are finally ruptured by extension of the surrounding 

 tissue. Thin very obscure tatters of the ruptured tissue clothe the walls of the mature 

 passage. 



The above-described phenomena are modified in the halms of Cyperacese (Scirpus 

 lacustris, species of Heleocharis, and Eriophorum), Juncus effusus, &c., in the leaves of 

 Iris pseudacorus, Sparganium Typha \ &c., in the following way : the strands of tissue, 

 which originally occupy the cavity of the air-passages, become at first lacunar, with many 

 armed cells, and for a time follow the growth of the surrounding tissue, while the arms 

 of the cells are much elongated : finally they dry up, and are partially broken down. The 

 extreme delicacy which they finally assume indicates that the membranes are partially 

 dissolved. As the result of the phenomena described, groups of distorted, more or less 

 collapsed 'stellate' cells are to be found on the walls of these air-spaces, or even, as in halms 

 of Juncus, the whole cavity is loosely filled with such cells. It is instructive that these 

 cases correspond closely on the one hand to those described above for the Marantacese, 

 where many-armed lacunar bands are partially torn away, though their cells do not die : 

 on the other that the dried-up lacunar pith of Juncus corresponds to the numerous cases, 

 where the pith soon dries up to form a cylinder filled with air, without the formation of 

 large cavities or passages. 



Finally, in a number of Cyperaceae those cells, at the expense of which the air-cavity is 

 formed, retain, at least in part, a firm wall, and these walls approach one another, by 

 reason of the tension caused by the surrounding tissue, till their lumen disappears. 

 These collapsed walls are then stretched in the air-cavity like thin plates or threads, 

 which, as Schwendener says, give quite the idea of a glass tube drawn out in a lamp. 

 Schwendener^ describes such a structure in -the case of the many-armed cells in the halm 

 of Scirpus maritimus. The same occurs in the cylindrical-prismatic cells of the cortex of 

 the root of species of Carex and of Cyperus alternifolius : also, though in a slightly 

 developed form, in the air-spaces of the rhizome of Carex arenaria. In the inner cortex 

 of the root of Carex foUiculata, for instance, numerous radial bands, 1-3 rows of cells 

 in thickness, continue to be composed of cylindrical-prismatic parenchymatous cells con- 

 taining starch throughout the tangential extension of the outer cortex : bands, usually of 

 2-4 rows, alternating with them, are widened out to cavities, in which the membranes 

 of the transversely distorted cells are stretched as almost solid thin plates in a tangential 

 direction. 



As has been already indicated, and will be further stated in later chapters, -layers of 

 parenchyma alone take part in the formation of the wall of the air-passages in question- 

 that is, if we leave out of account those hair-like single fibres, which spring in certain 

 cases from the walk The only known exception occurs in the petiole of Thalia deal- 

 bata^, in which each air-passage is traversed throughout its length by numerous thin 

 bundles of sclerenchyma, which for the most part stand quite free in the passage, and are 

 only connected laterally with other elements where they run quite straight through the 

 two sorts of Diaphragms to be described below. Further, it may easily be recognised in 

 the mature plant that the denudation of the bundles is the result of breaking up of the 

 lacunar band of parenchyma, which or.iginally fills the passage, this not keeping pace in its 

 growth with the extension of all the other parts. 



' Frank, /.c. p. 148. > Mechan. Princip. p. 93, Taf. X. 10. 



^ Duval-Jouve, Diaphragmes vasculif^res des Monocotyledones, Mem. Acad. Montpellier, 1873, 

 168. 



