ECCENTRIC GROWTH. 275 



system of quasi-squares. Since, again, an eccentric system in 

 which growth is unequally distributed on different sides conveys 

 to the eye the subjective appearance of a displacement of the growth- 

 centre towards one side, while the general approximation to a 

 circular outline may not be affected, it may be convenient to make 

 use of this phraseology as a simple way of describing the construc- 

 tion, although it has no causal significance. For example, by 

 selecting the (5 + 5) system of construction curves a figure will be 

 obtained (fig. 96) which is an eccentric homologue of fig. 55, and 

 represents definitely, therefore, the general construction mechanism 

 of a zygomorphic pentamerous flower. In this special case the 

 growth-centre has apparently been displaced towards the upper 

 surface (posterior side of the floral diagram); the converse con- 

 struction is seen on turning the figure the other way round. 



The diagram will now be seen to illustrate several points of 

 interest which will be useful in the interpretation of floral 

 construction : — 



I. The system has lost its radial symmetry and become definitely 

 bilateral, or, as it has been termed, " dorsiventral " ; that is to say, 

 as soon as the growth-centre is displaced, a line may be drawn 

 dividing the construction into two halves as a simple geometrical 

 consequence of the eccentricity. 



II. Notwithstanding this the construction remains definitely 

 (5 -f 5) ; that is to say, all previous deductions based on the 

 corresponding centric type continue to hold. The system is 

 growing, and the growth-centres in each whorl of five are still 

 initiated simultaneously : the fact that they may afterwards grow 

 at different rates is wholly secondary. 



III. But while the strict alternation, the contact-relations, and 

 the simultaneous initiation of five new centres remain unaffected, 

 the appearance of the system at any given moment will always 

 present the subjective effect that the largest members must have 

 started first ! Each cycle of five has in fact been described as a 

 " successive whorl " ; while a construction of the type of fig. 96 

 would be termed " ascending development," and its inverted 

 homologue a case of " descending development." It is at once 

 clear that a " successive whorl " is a contradiction in terms, and 



