﻿iqi6] LLOYD— ABSCISSION 223 



show that such autolysis does not occur, and, if this be true, 

 Hannig's inference is not justified. The appearance of collapse 

 or compression is accidental, and that the abscission cells in Mira- 

 bilis should appear so is not to be wondered at, in view of the 

 facts presented in this paper. While some starch grains may 

 indeed persist even until after abscission is completed, the granules 

 are quite obviously very much reduced in size, so that, in a par- 

 ticular plastid, the stroma enveloping the granules may readily be 

 seen between them, and the granules of starch may appear as 



To draw a sure inference from the presence of starch in, or its 

 absence from, separation cells is at best precarious. For example, 

 the separation cells in Hydrangea, in which "maceration" occurs 

 as in Impatiens, retain their starch; while the immediately neigh- 

 boring cells lose it. But, when abscission is complete, the sepa- 

 ration cells, while entirely loosened from each other, are quite alive. 

 In such a case as this the starch may not be used at all, and may 

 be retained because the loosening of the cells prevents its move- 

 ment (this being essentially Hannig's explanation), or it may have 

 been secreted in amount much in excess of that needed. But it is. 

 in any event, doubtful if the loosening of the cells would prevent 

 the movement of starch. 



b) The behavior of the cell wall.— Evidence has already 

 been brought forward to show that, previous to any changes in the 

 dimensions of the cells involved in abscission, there occurs an 

 alteration of the cell walls. The degree of alteration appears not 

 to be the same throughout the whole of the cell wall, judging by 

 its behavior during those steps preceding separation. For the 

 sake of simplicity, I take an ideal case, that of a fairly thick- walled 

 cell of the prosenchyma sheath, in which the physiological activity 

 proceeds in a plane normal to the axis of the cell (pi. fig. 15), it 

 being premised that this plane may lie in any oblique direction 

 (pi. fig. 10) passing through any of the walls. The changes observ- 

 able are presented seriatim. 



(1) The cell wall is softened in the previously described man- 

 ner, most completely, however, in a narrow zone nearer the upper 



