624 SECONDARY CHANGES. 



radial rows of meristem at the periphery also develope into parenchymatous cells, a layer 

 immediately adjoining the tabular cells remains meristematic, and acts as cambium 

 throughout life. This layer of cambium, surrounds the whole axile bundle, with the 

 exception . of the apex, and the place where it is simultaneously perforated by the 

 insertions of bundles from the leaf and root. It appears in the strict sense to be a 

 single initial layer of cells, but this is not made out with certainty. The cells are of 

 similar form to those of the tabular layer, but are on the average shorter in a radial 

 direction,, and are also similar in all points to those of the cambium of .Monocoty- 

 ledonous stems. Like these they show further successive radial extens'on, and 

 reciprocal, tangential, longitudinal divisions. The products of the latter pass over on 

 the side :next;the bundle in centrifugal succession, and on the side, next the cortex 

 in centripetal succession, into a definitive condition of tissue. In the relative extent 

 of the thickening, however, in the two directions, there is this difference from all other 

 known instances, that the axile bundle always remains thin and narrow in comparison 

 with the cortex, and only increases in thickness by a few layers, while the secondary 

 cortex grows in the course of years hundreds of times as much. The elements 

 added in centrifugal succession to the axile bundle have the form, of the tabular cells 

 often mentioned. They retain the original radial arrangement and close connection 

 one with another, and, at least in I. lacustris, are without intercellular spaces. Oblique 

 divisions occur, deviating from the tangential longitudinal direction, and require more 

 exact study. The cells retain for the most part in I. lacustris the above-described 

 brilliant walls and transparent contents ; the latter also require more exact investi- 

 gation. Between the transparent cells there are isolated cells or longitudinal layers 

 of them, which contain numerous starch grains. In terrestrial species, as I. hystrix, and 

 Durieui, as Hegelmaier found, there is a regular alternation between transparent and 

 starchy concentric layers, so that one layer of starchy cells lies between each series of 

 3-5 layers of transparent ones. Tracheides are entirely absent in most cases in the 

 secondary thickening of the bundle; but, they have been observed in single specimens : 

 of I. lacustris and Durieui, situated singly or in groups between the tabular elements, 

 and resembling these in form and arrangement, but similar in structure to the original 

 tracheides of the bundle. Finally, attention may be called to the fact that it is 

 uncertain whether the above-mentioned- first layer of tabular elements is the first 

 product of the cambium, or belongs to the primary bundle. 



The cambiogenetic secondary cortex consists exclusively of parenchyma, the 

 cells of which contain according to the species large quantities of starch, or of 

 starch and oil. According to their origin they remain arranged, in radial and con- 

 centric rows; as they develope, wide air-containing spaces appear between the 

 rounded corners, especially in species with an aqueous habit. Like the primary 

 cortex the secondary thickenmg is also much more considerable between the furrows 

 than opposite them. It is known that in each period of vegetation there is a large 

 production of secondary cortex, which pushes the older parts of the cortex outwards, 

 together with the old leaf-bases and roots which adhere to it, and with the vascular 

 bundles belonging to these, which lie internally, and are stretched and finally torn 

 away by the thickening; also that the old layers of cortex successively die and rot 

 oif, while their empty cell-walls turn brown ; but they are not thrown off by a 

 formation of periderm. 



