1907] CHR YSLER—PO TA MOGETONA CEA E 167 



(3) They may (with the exception of the central bundle of each stipule, 

 which joins a lateral leaf trace) disappear soon after entering the 

 cortex, leaving the stem without cortical bundles^ e. g., P. crispus. 



(4) They may enter the central cylinder immediately, like the leaf 

 traces, leaving the stem devoid of cortical bundles, e, g., P. jnikher. 

 To this should be added the peculiar case of P. NuttaUii {fig. 5), in 

 the upper leaves of which the median bundle of each stipule fuses with 

 the neighboring lateral leaf trace bundle, and from the fusion is given 

 off a strand which descends into the cortex. The cortical bundles of 

 this species arc not numerous (six to eight), and form a complex 

 anastomosis with the central cylinder at the level of the lower part of 

 the node. In general it may be stated that when a species possesses 

 cortical bundles they arc derived from the small bundles of the petiole 



r 



and stipules. This view is confirmed by the fact that such bundles 

 are universally present in the petioles and stipules, even when they are 

 absent from the stem. As to composition, the cortical bundles range 

 from complete collateral bundles to mere groups of a few fibers, and 

 moreover the character of a bundle may change greatly in its course, 

 being composed mostly or altogether of phloem at one level and only 

 of fibers at a slightly different level, Schenck (28, pp. 48, 54) is of 

 the opinion, and my observations confirm his view, that all these 

 strands represent real fibrovascular bundles, which share in the 

 reduction shown by the vascular tissues of the central cylinder. 



The extent to which mechanical tissue is present is, according to 

 ScHWTENDENER (30), dependent on whether the plant grows in still or 

 in running water. But different species occupying a similar habitat 

 may differ widely in respect to the presence of mechanical tissues; 

 for example, P. natans is abundantly provided with cortical strands, 

 while P. pulcher entirely lacks them, yet both are inhabitants of ponds, 

 and both possess floating leaves. Judging from over forty species for 

 which data are available, there is no relation between the presence of 

 cortical bundles and a totally submersed or a partly floating habit. 

 Neither can a relation be established between size of leaf or stem and 

 presence or absence of cortical bundles- That these strands are not 

 primitive structures, but have arisen in connection with the mechanical 

 necessities of the leaves and stipules is indicated by three facts, (i) 

 In the creeping stem, which has rudimentar}' sheathing leaves devoid 



