﻿BOTANICAL GAZETTE 



demonstrate that these membranes are free by means of untreated 

 sections, although not at all impossible. 



The result of all this is to produce a separation cell, the ends of 

 which are invested by thicker, physically relatively unaltered 

 walls (text fig. i,d) with a transverse zone between them, narrow 

 at first, but becoming quite wide at length, of extremely thin 

 membrane. Viewed en face, the thicker portions of the wall show 

 their shallow pits, well seen in the prosenchyma sheath, and 

 between them and the thin portions the pits are seen to disappear 

 and the membrane itself to show the granulation or flocculation 

 due to the chemical alteration (pi. fig. 1 5) . If, as has been premised, 

 the thinning out has proceeded in irregular, more or less oblique 

 zones, very various topographic conditions (pi. fig. 10) ensue with- 

 out any violation of principle. 



Although the above view of the nature of the thin walls seems 

 to accord entirely with the facts, it remains true nevertheless that 

 they could be accounted for by supposing that, as the old walls 

 become softened and broken down, entirely new thin walls are laid 

 down by the growing protoplasm. This is Tison's view. While 

 it is altogether possible, or even probable, that some new wall 

 material is being laid down, the optical evidence is against the idea 

 that the wall is entirely new. One can, if with some difficulty, 

 resolve a relatively unaltered membrane in contact with the pro- 

 toplast in the portions of the cell which are not elongated, and can 

 determine its continuity with the thin membranes in the zone of 

 elongation. The continuity comes out clearly when the chemical 

 alteration of the rest of the wall is far enough advanced so that 

 the optical differences presented by the primary and secondary 

 membranes on the one hand and the tertiary membrane on the 

 other are obvious (pi. figs. 6, 13; text fig. 1, b,f). 



The process as described is the same for all living elements, 

 for example, cambium cells and wood parenchyma. The xylem 

 vessels are fragmented in one or more transverse planes, according 

 to the number of tiers of cells involved in abscission. Tyloses in 

 various degrees of development and in various numbers are to be 

 seen both above and below the abscission plane, but, as Swart 14 

 has shown experimentally, they cannot effectually hinder the 



14 Swart, X., Die Stoffwanderung in ablebenden Blaettern. Jena. 1914- 



