ANOMALOUS THICKENING IN DICOTYLEDONS AND GVMNOSPERMS. 613 



annular zones. They take a sinuously curved course in a vertical direction, thus 

 crossing the girdles and the radial connections. In an apical direction tliey may 

 always be followed as far as the broad base of the foliage leaves. Here they are 

 seated, and in all cases several of them of unequal strength, sometimes on the two 

 shanks which pass out from the girdle into the leaf, sometimes on the first branchings 

 of the latter in the base of the leaf itself, the strongest ones as a rule on the first 

 branching. From this point of insertion they descend through the stem, and after 

 a course of varying length they either unite, sometimes with similar ones from the 

 same leaf-base, sometimes with others descending from above, or they insert them- 

 selves with their lower end, or with a lateral branch, on a girdle or on a radial con- 

 nection. Free blind ends are not present; nor is there any connection with the 

 bundle-system of lateral buds or roots. 



The structure of the cortical bundles is such that their middle is occupied by 

 a narrow parenchymatous prism of pith, and this is successively surrounded by a 

 ring of wood and cambium, and a weak ring of bast. On the part of the cambium 

 there is a permanent though slow increase, so that in old stems the thickness of the 

 bundle rises from 2"i™ to 5-6"i™. Wood and bast are traversed by thin-walled medul- 

 lary rays, which are successively increased as the thickening goes on : between these 

 the elements of the bundles are distributed as in the rings. The tracheides of the 

 wood have usually a scalariform pitted thickening in C. revoluta, but few of them have 

 bordered pits ; spirally thickened ones are absent. At the points of insertion there 

 is an alteration of the structure, inasmuch as the pith disappears ; the elements of 

 one bundle then become contiguous with the similar ones of the other. 



The secondary thickening of \he. Roois of the Cycadece^ which have been investi- 

 gated always corresponds, at first, like the primary structure described on p. 357, with 

 that of the typical roots of Dicotyledons and Gymnosperms, and especially with those 

 which are fleshy and highly parenchymatous. The structure of the secondary ele- 

 ments of the bundle is fundamentally similar to that of the wood of the corresponding 

 stem. For their individual characters and distribution, comp. Mettenius, /. c. For 

 the investigated roots of Encephalartos (E. Caffer, longifolius) nothing of importance 

 need be added to the above ; at least, roots 3cm thick showed no more remarkable 

 phenomenon than very considerable extension of the internal parenchyma and 

 consequent distortions of the strands of wood. Also, as far as the subject could be 

 investigated, the primary outer cortex is in these plants thrown off at an early stage 

 by periderm. While accordingly the roots described belong to category (2) of fleshy 

 roots described on p, 516, that of Cycas revoluta must be placed in category [i)(a). 

 The secondary thickening is in this case weak, the starchy parenchyma is still living 

 in roots as thick as the finger, and is only covered on its outer surface by a layer of 

 periderm. How long it persists is not known. As the root becomes older, according 

 to Mettenius, the first cambial ring loses its activity, which is renewed by a peripheral 

 cambium ' and the further growth proceeds exactly as in the stem.' 



On the bushy dichotomous excrescences of the roots of Cycadeae caused by 

 their penetration by Nostoc, comp. Reinke, I.e. 



' Mettenius, /. c. — van Tiegliein, Keinke, /. c. See p. 356. 



