THE FRUIT OF OPUXTIA FULGIDA. 19 



of the ovary. The foi-m of this Layer is not a simple transverse plane, but is 

 that of an inverted cone, usually perforated at the apex. This funnel-shaped 

 layer of tissue is initiated in cells 15 or 20 layers beneath the surface of the 

 cup-like upper end of the ovary that bears the stamens, petals, and sepals 

 (figs. 23, 58, 60, 61). 



In a diametric longitudinal section of the ovary this abscission layer 

 usually starts in at the base of perianth, either outside the sepals or, more 

 rarely, between these and the petals. From here it extends downward, 

 parallel to the surface of the cup at the top of the ovar^^, to a level just a 

 little above the base of the style, where the abscission layer again comes out 

 to the surface of this cup (figs. 23, 24, GO, 68). Less frequently the more or 

 less developed abscission layer may extend across beneath the bottom of the 

 cup at the top of the ovary. In this latter case the style is cut off, along with 

 the stamens and perianth attached to the same complete, shriveled funnel 

 (fig. 61). In the more usual case, first noted, the perianth and stamens only 

 are borne on a shriveled funnel that is perforated at the base, while the style 

 is shed separately, breaking across just above its base (fig. 24). In rarer 

 cases, where the abscission layer starts in at the top within the perianth (fig. 

 23), the i-)etals and sepals must evidently be cut off separately. Whether a 

 real abscission layer is developed in each part or not was not determined. 



The first origin of the abscission layer across the top of the ovary is evi- 

 denced by a swelling, chiefly a radial elongation, of a continuous layer of 

 cells in the midst of the wall of the ovary, in the region stretching between 

 the base of the perianth and the base of the style (figs. 61, 68). Apparently 

 any cells along the line of the abscission layer to be may take part in its 

 formation, except such specialized cells as those of the vascular bundle, the 

 mucilage cells, and the crystal-holding cells. The cells that are to form 

 the abscission layer increase in radial length to about twice their tangential 

 diameter. Then the cells divide tangentially into two nearly isodiametric 

 cells (figs. 67, 68). A further tangential division follows very soon in each 

 of these cells, resulting in the formation of a row of 4 cells, of which the two 

 middle ones have only half the radial thickness of the two outer (fig. 67). 

 It is at this stage, or sometimes after one or two further tangential divisions, 

 that abscission occurs. The details of possible changes in the radial walls of 

 tlie abscission cells have not been studied. (See Lloyd, 1914, p. 70, and 

 1916, pp. 213-230). It is clear, however, that in consequence of the shrink- 

 ing of the perianth, on drying, the delicate radial walls of the thin, tubular, 

 cambium-like cells in the middle of the abscission layer are niptured. The 

 separation occurs first at the base of the perianth and continues do-v\mward 

 until the whole top of the ovary, bearing perianth and stamens, and clasping 

 the style within, curls together and droj)s ofi^ the flower on the second or third 

 day after its opening. The surface left at the top of the ovary after the 

 shedding of the periantli shows cleaidy that the break tiikes place in one of 

 the thin cells in the middle of the abscission layer, and also that it occurs 

 in the radial wall rather than as a split between two tangential walls (fig. 



