232 BOTANY OF THE LIVING PLANT 



flowers, and may be held as secondary modifications of their free condition, 

 as seen in the primitive state. 



This view is borne out by the study of development. For where the parts 

 are fused in the mature state, they still originate as separate papillae of tissue 

 from the growing point of the flower, just as the foliage leaves usually do. 

 It is later that the growth extends from the individual bases of these papillae 

 into the region between them. Consequently when mature they appear as 

 though borne up on a common base. This is well shown in the flower of the 

 Compositae, in which there is cohesion of the petals to form the tube of the 

 corolla, and adhesion of the stamens to the inner surface of that tube. 

 Fig. 1 88, p. 239, (v) and (vi) show how that adhesion arises. In (v) the stamens 

 and petals are independently inserted on the hollowed receptacle ; but the line 

 where basal growth will take place is indicated. In (vi) the result is seen ; 

 for they are there borne up on a common base, which has been the result of 

 that growth. 



(iv) Pleiomery. By this is meant that the number of parts of one 

 category is greater than the fundamental number for the whole flower. 

 It is most frequently seen in the androecium, so that the stamens 

 are in excess of the other parts. It may be a question exactly how 

 this comes about in each individual example. Branching or fission 

 of original y single parts may account for some cases ; interpolation 

 of additional parts, where there is room for them on the receptacle, 

 may explain others. The distinction between these is not always 

 clear; it turns upon comparison, and the obser- 

 vation of details of development. The essential 

 feature is, however, that more parts of one category 

 are produced than the other whorls of the flower would 

 give reason to expect. 



Fission is most easily recognised where two or more 

 stamens stand side by side in the place normally 

 occupied by one ; it then sometimes happens that they 

 arise from a common stalk. The case of Vellozia 

 (Fig. 181) gives a good example. Interpolation of extra 

 parts may give very similar results. Sometimes it is 

 FlG - lSl - individual parts, sometimes additional whorls that are 



t, thr takfng m ?hl add^- All these methods appear exemplified in the 

 place of one stamen in the Rosaceae. The following diagrams may be quoted as 

 5 illustrating the pleiomeric variations within that family, 



though without suggesting any actual line of Descent 



(Fig. 182). In Sibbaldia, the pentamerous flower has five stamens. Quittaija 

 has two whorls of five (diplostemonous), and this probably represents the 

 fundamental type for the Rosaceae, as it corresponds to that of related families 

 with the formula (S 5 , P 5 , A 5+5 , G 5 ). But in the Rosaceae the matter does 

 not stop there. Further steps are taken till an indefinite number of stamens 



