STRUCTURE OF COLLATERAL BUNDLES. 335 



and arrangement as though it had proceeded from the longitudinal division of 

 one of the wide elements. So far as investigations extend, the narrow elements 

 are here sieve-tubes, which may be accompanied by narrow cambiform-cells ; 

 the wide ones are cells, whether they be called cambiform or simply parenchyma. 

 More minute and extended investigations of these elements are wanting ; they are 

 usually curtly disposed of under the name cambiform or soft bast. In the thick 

 bundles, often mentioned above, of Dicotyledonous leaf-stalks and ribs, the paren- 

 chymatous rows, resembling medullary rays, are continued directly outwards from the 

 xylem into the phloem, and in the latter pass between bands of tissue, appearing 

 irregularly narrow-meshed in cross-section, which no doubt contain the sieve-tubes. 



3. In the primary bundles of the Coniferae, especially of the leaves, where there 

 is no disturbance from the secondary growth which quickly comes on in the stem, 

 and also in the large bundles of the leaf of Welwitsphia, the cross-section of the 

 phloem shows regular rows of similar elements with soft membranes, which are in a 

 very high degree capable of swelling. They are either present alone (Figs. 63, 157), 

 or a single row consisting of relatively wide parenchymatous cells runs here and there 

 between them, e. g. in the leaf of Dammara alba. The general form of the regularly 

 serial elements first mentioned is elongated prismatic. How far they are sieve-tubes 

 or cambiform-cells is undecided. 



The development of the elements of the phloem of collateral bundles begins at 

 the external edge and proceeds towards the xylem, and thus in the opposite order to 

 that in the latter, i. e. centripetally in the phloem, if the centrifugal direction is 

 maintained in the xylem. In the middle of the bundle, on the border between the 

 two parts, active meiistematic division may still be going on when the elements of 

 the external and internal edge are completely differentiated. 



The outermost, first-developed elements of the phloem (Russow's Protophloem) 

 are often distinguished from those which follow by their smaller width, and thicker, 

 apparently gelatinous walls ; as regards their nature, however, they are in the cases 

 now in question, which admit of more accurate investigation, partly sieve-tubes, 

 partly cambiform-cells. In the thicker bundles they not uncommonly become com-' 

 pressed in the radial direction, owing to the expansion of the surrounding tissue, 

 while their walls apparently swell up to the obliteration of their lamina — a pheno- 

 menon which ensues to a much greater extent in the old sieve-tubes and cambiform- 

 cells of the secondary bast. (Cf. Chap. XV.) 



The boundary between phloem and xylem is in most cases on the whole sharply 

 marked, owing to the contrast between the delicate and non-lignified membranes on 

 the one hand, and the characteristically thickened and lignified membranes on the 

 other. The elements which form the boundary on the side of the phloem, no doubt 

 always have the characteristics of cambiform-cells ; it is not known that a sieve-tube 

 ever borders directly on a trachea. In the thick bundles of the great majority of 

 Dicotyledons the cells on this border-surface long remain capable of division, and 

 when the differentiation of tissues has once begun at the periphery, their divisions 

 take place chiefly in the tangential direction, parallel to the outer edge. Owing to 

 the arrangement of the cells in radial or tangential rows thus determined in the zone 

 indicated, the boundary is pretty sharply defined, even in cases where the actual 

 divisions soon come to an end. In the stems of Dicotyledons and Gymnosperms 



