546 



SECONDARF CHANGES. 



division they go on in the centripetal order, com}). Figs. 216, 217 ('centripetal-inter- 

 mediate order '). 



2. Division begins centrifugally, and then changes over to the centripetal order. If 

 the latter takes place immediately on the third division, in such a manner that the 

 second cell from the inside then becomes the meristematic cell, which henceforth 

 divides centripetally, Sanio calls the process centrifugal-intermediate ; if the reversal 

 only takes place after later divisions, Sanio calls the order centrifugal-reciprocal. 



In the later stages of growth of those phellogenetic meristems which are active 

 for a long time, the reversal of the mainly centripetal order into the centrifugal, and 

 its immediate return to the centripetal, take place in most cases from time to time 

 without strict regularity. 



iig. 216 



- - Fig. 2r7. F'iJ- 2i3. 



Figs. 216— 218.— Transverse section through the surface of a branch of Sorbus Aucuparia. After Sanio. Origin of 

 the periderm.— Fig. 216. Four epidennal cells, already divided once tangentially ; in the two pairs to the right the lower 

 cell is once more divided into a and />. a meristematic cell, A phelloderm.— Fig. 217. Further stage of development ; 

 a—n meristem, 6—6 phelloderm. The cork -cells corresponding to the outer halves of the epidermal cells have acquired 

 strongly thickened walls.— Fig. 218. Still further developed stage ; n—<7, 6—6, as in the former figure. Outside a— a are 

 three layers of cork-ceIIs,the two outer ones with thickened walls; at >i, « the inner layer of the wall is partially released 

 from the limiting lamella. To the right, at <?, the radial division of a peridermal row is beginning. 



Whatever the course of the divisions may be, it is nearly always only one cell of 

 each radial row% and consequently only one layer of cells in the entire peridermal zone, 

 which remains in the meristematic'condition, so as to carry on the divisions ; all other 

 cells directly become elements of the tissue, after they have originated by division in 

 the meristem. An exception to this rule has only been observed with certainty in the 

 case of Philadelphus coronarius \ and in this case the division in general proceeds 

 centripetally, the innermost layer of cells remaining meristematic, but in each of those 

 products of division which are given off externally, one or two tangential divisions (in 

 the latter case proceeding centrifugally) take place, and only after this does the de- 

 velopment of their products into tissue occur. According to Sanio, a similar process 



^ Compare Sanio, /.c. p. 



