214 INTERCELLULAR SPACES. 



numerous hollow stems of Equisetum, Grasses, Umbelliferae, Labiata^, Compositae, 

 &c., of hollow leaves and petioles (Allium, Asphodelus, Umbelliferae, &c.), also the 

 axile cavity of the internodes of Nelumbium '. 



In many cases the origin of the chambers has not been carefully investigated, 

 but it may be recognised with some certainty from the structure of their mature 

 walls (which will be described below), in the same way as in the instances marked (?) 

 above. INIore accurate detail on this subject has but slight interest. On the one 

 hand the two modes of origin are fundamentally different in extreme cases, usually 

 also certain differences in structure of the walls of the chambers are peculiar to 

 each, and lastly they are distributed as a rule in different systematic groups. But on 

 the other hand all sharp limits are here again obliterated by all sorts of intermediate 

 cases, while forms differing in origin and structure may be substituted one for the 

 other, in the same position, in closely allied plants. For instance, the peripheral air- 

 passages which alternate with the vascular bundles in the stems of the Equiseta are of 

 intermediate structure and origin : they originate schizogenetically; finally, some of the 

 separating cells are broken up and their membranes remain attached to the wall of 

 the passage -. The same holds for the large air-passages opposite the two shorter 

 sides of the quadrangular transverse section of the stem of the Eucallitrichae, while 

 the smaller ones there and in Pseudocallitriche are all schizogenetic ^ In the 

 petioles of the jMarantacese the arms of the very loose stellate-lacunar bands are often 

 finally torn asunder by extension of the surrounding tissue, so as to form continuous 

 air-passages, while their torn, often pointed and thick-walled ends rise free into the 

 cavity. In the halms of Scirpus lacustris* the meristem differentiates first into 

 prismatic bands of stellate-lacunar tissue, and plates of dense parenchyma, usually one 

 layer thick, which separate these, and appear in transverse section as a net with 

 angular meshes. The stellate cells follow the surface-growth of the latter, their arms 

 elongating greatly, but finally they are for the most part broken up, and only dried 

 remnants are left behind in the prismatic spaces. For further examples comp. infra. 

 * As examples of the substitution of one form for another in a similar position 

 in allied plants the above-mentioned halm of Scirpus lacustris and that of Papyrus 

 may be cited. The latter has air-passages with very similar distribution, but of ap- 

 parently purely schizogenetic origin ; I have however at hand no direct observations 

 on their development. Carex arenaria has in the inner cortex of the rhizome a circle 

 of large air-passages, separated by many-layered, radial lamellae of parenchyma; 

 these, as is the rule in the Cyperacese, are lysigenetic : C. disticha has in the same 

 place 7-10 circles of narrow schizogenetic passages, separated by simple layers 

 of cells. 



As regards the /brffi of the larger air-cavities it has already been indicated that 

 they are either short polyhedral chambers, e. g. in the leaves of Pistia, the swellings 

 of the petiole of Trapa, and in the Lemnas, or elongated prismatic canals or passages, 

 and this is usually the case in elongated parts, such as stems, petioles, or narrow 

 foliage leaves. The latter but rarely traverse the whole of the elongated part 



' Trecul, i.e. p. 166. ^ Compare Frank, /. c, 



' Compare Hegelmaier, Monogr. v. Callitriche, p. 24, Taf. I. 



* Frank, /. ^. p. 147. 



