840 MR JOHN M'LEAN THOMPSON ON 



three are small and abaxial. Occasionally the leaf-trace, on its way through the 

 cortex, consists of only three bundles. Not uncommonly a thin commissure locally 

 unites any two of the small abaxial bundles, or forms a short bridge between 

 an abaxial and an adaxial strand. No such commissures link the two large adaxial 

 bundles. 



Here and there a root springs from the base of a leaf-gap (figs. 31, 40), and, 

 pursuing its course on the outer side of the leaf-trace from the same gap, passes 

 quickly to the surface of the axis. Many of the roots are not, however, directly 

 connected with leaf-gaps, but arise as offshoots of the small abaxial leaf-trace bundles, 

 either close to the points of insertion of the leaf-trace bundles on the margins of the 

 leaf-gaps, or at quite considerable distances from the points of insertion on the gaps 

 (text-fig. xxiv and fig. 40). Other roots arise independently on the outer surfaces of 

 the strap-like meristeles. 



The axis is by no means of uniform thickness throughout, and while the vascular 

 system of the more robust portions in the specimens investigated shows fairly 

 advanced dictyostely, a much less divided condition of the stele is found in other 

 parts (fig. 31). Those portions are more slender, and were formed, no doubt, during 

 less favourable periods when the apical region and leaf production were small. In 

 those more slender parts of the axis leaf-traces seldom spring from the lower surface 

 of the stele, which is then locally devoid of leaf-gaps, and may take the form of a 

 thin, curved strap of conductive tissue quite uninterrupted for some distance, or 

 pierced only by small circular or oval perforations. 



Some support is lent to the above interpretation of the fluctuations in the general 

 condition of the dictyostele by the following facts. In an old plant — about half the 

 stele of which is represented in the reconstruction (fig. 40) — -the axis became gradually 

 reduced in bulk towards the apex, and here bore only a few leaves inserted laterally 

 and on the upper surface. The series of sections into which this apex was cut are 

 represented in text-figs, ix-xxvi. 



Accompanying the reduction of axis and leaf was a corresponding progressive 

 reduction in the dictyostele. At some distance from the apex transverse sections 

 showed the stele to be a fairly advanced dictyostele composed of some half-dozen 

 strap-shaped meristeles (text-fig. xxvi). But on passing forward the number was 

 reduced to four and then to three, this being effected by the union of the meristeles 

 towards the lower surface of the axis into a long narrow strap from which not a 

 single leaf arose (text-figs, xxv-xxii). Still further forward this strap was irregularly 

 pierced by small circular and oblong perforations (text-figs, xxi-xiv), while close to 

 the apex an almost solenostelic condition was established, and the diameter of the 

 stele was greatly reduced (text-figs, xiii-ix). Further, it was found that while in 

 the more robust portion near to the apex of the axis the leaf-trace bundles traversed 

 in the sections were relatively large and strong and usually in groups of four and five, 

 leaf-traces consisting of three or even two bundles, and usually fairly small, were found 



