THE FRUIT OF OPUXTIA FULGIDA. 21 



fruits showed it to be sterile. No information is given conceniing the 

 presence of axillary buds on the primary fruit, though it is clear that these 

 must have been present to initiate the secondary flowers. 



The nearest approach to this structure of the wall of the ovary in these 

 CactacesB of which I find record in any other family is seen in the genus 

 Calycanthus. In this form the greatly developed floral axis is depressed at 

 the top to form a cup, from around the edges of which arise the perianth 

 members and the stamens. The numerous distinct carpels, on the contrary, 

 are developed from the bottom of this cup and remain surrounded by it, but 

 do not play any part in roofing it over, as the carpels of the Opuntia flower 

 close above its concave growing-point. The outer wall of the cup of this 

 fruit of Calycanthus bears the scars of many fallen petaloid floral leaves, 

 distributed much as the leaf-scars and areoles are over the fniit of Opuntia 

 fidgida. A study of the development of the flower of Calycanthus shows, 

 however, that there are no axillary buds above its leaves. There is not even 

 the smallest recognizable rudiment of this, and hence no possibility of the 

 development of secondary flowers or shoots from the wall of the flower or 

 fruit. 



The facts of development and structure suggested in the paragraph before 

 the last, together with those detailed earlier in this paper, furnish important 

 if not conclusive evidence regarding the morphological nature of the ovary of 

 this opuntia. To the writer these facts seem to offer very strong evidence 

 for the view that the flower of the opuntias consists of a shorter or longer 

 vegetative joint, into the depressed upper end of which the ovary is com- 

 pletely submerged, and around the margin of which the stamens and 

 perianth members are inserted. This view has been advanced in more or 

 less definite form by various workers on the Cactacea; in the past. (See 

 Schumann, 1894, p. 168; Zuccarini, 1844; Toumey, 1905, p. 235; Harris, 

 1905, etvC.) The evidence for this view, however, has hitherto always 

 seemed somewhat inadequate. Since it is felt that the present study offers 

 the most complete chain of evidence thus far produced, especially from the 

 developmental standpoint, this evidence will be stated here in some detail. 



In the first place the external structure of the ovary at the time of opening 

 of the flower is very like that of a vegetative joint, having prominent mam- 

 milla?, each of them bearing a conical leaf and a bud in the axil of the latter, 

 and (rather rarely) a spine like those of the stem itself. As stated above, 

 the ovaries, and the fruits formed from them, differ markedly in length and 

 breadth. Thus, the basal part of the ovary may sometimes be short, witli 

 relatively few mammilla? and areola^, and so give rise to a nearly globular 

 fruit in which the ovarian cavity reaches nearly to tlie base (figs. 11, 24, 26, 

 27). In other cases the basal portion of the ovary may be far more 

 developed, with 20 or even 30 tubercles, and the ovarian cavity may occupy 

 only the upper third or fifth, or even less, of the portion of the whole floral 

 joint below the perianth (figs, lb, 28). lu these extremely long flowers the 



