THE STELA R THEORY. 141 



add a new region to those already distinguished. He 

 finds a zone situated at the periphery of the pith, i.e., just 

 internal to the ring of bundles, corresponding exactly to 

 the pericycle external to the ring, as well characterised 

 histologically as the pericycle itself, and indeed resembling 

 the latter very closely in structure and role. This zone, 

 the perimedullary zone, is according to Flot (and his 

 figures entirely support this) separate in development from 

 the pith proper, or internal conjunctive, and belongs rather 

 to the hollow cylinder of tissue (the "thickening ring " of the 

 older German anatomists) giving rise to the bundles and 

 the conjunctive immediately surrounding them {external 

 conjunctive). It is impossible sharply to separate the peri- 

 medullary zone on the one side, just as Morot found it 

 impossible to separate the pericycle on the other, from the 

 ray tissue, and we should rather regard the contrast of 

 the pith with the external conjunctive tissue, as of greater 

 importance than the division of the latter into pericycle, 

 rays and perimedullary zone, which are in the main 

 topographical regions marked out by the limits of the 

 bundles. In many adult stems it is however impossible to 

 fix the limits of external and internal conjunctive, just as 

 it is often impossible to fix the limits between external con- 

 junctive and cortex. Flot is of opinion that this is owing 

 to a growth in breadth of the cells of the external conjunc- 

 tive continued longer than in the pith, the whole of the tissue 

 of the cylinder thus becoming approximated in size and 

 shape. This same cause, together with a masking of the 

 endodermal thickenings (in cases where these are originally 

 present) by a general thickening of the walls of all the 

 parenchyma cells may very conceivably account for the 

 frequent absence of the obvious limit between cortex and 

 cylinder, though we are not aware that such an occurrence 

 has been either established or suggested. 1 Further in- 

 vestigation on this point, as well as on the separation of 

 the regions in root cylinders with a well-developed con- 



X I now find that Sanio (24, pp. 371-2) states that this is practically 

 what occurs in the stem of Ranunculus acris. 



