THE ANGIOSPERMAE 1147 



which is in turn derivable from the type of three-sided apical cell that is 

 prevalent among the Pteridophyta. We have already shown reason for 

 doubting whether the axis in the Spermatophyta can be derived directly 

 from that of the Pteridophyta (see Volume I, Chapter XXI), and of course, 

 Gregoire's theory of the unique nature of the flower axis, if substantiated, 

 would also tell against Salisbury's view. 



The flowers of the Ranunculaceae, where the number of parts is large 

 and fluctuating, afford very suitable material for the investigation of the 

 conditions governing flower-building, and several intensive studies have 

 been devoted to them. One of the most painstaking of such studies, by 

 Burkill, does not support Salisbury's view of the primitiveness of trimery, 

 but on the other hand, reveals a strong tendency to isomerism, that is to 

 say towards stabilization with an equal number of parts in all four zones of 

 the flower, calyx, corolla, stamens and carpels. This is attributed to the 

 action of hormones, determining the character of each primordium on the 

 receptacle, which are supposed to be equally distributed around the axis 

 and to act most strongly along the lines of closest contact of the primordia, 

 that is to say the parastichies. 



If a trimerous, whorled flower is more primitive than a spirally arranged 

 flower, then a shift of primordia must have taken place to bring about the 

 change. Salisbury suggests that this occurred by the fusion of the last 

 member of the outer whorl with the first member of the inner whorl, 

 thus giving a single series of five members arranged serially, instead of two 

 whorls of three. Some change of the plastochrones must also be involved, 

 spacing out the newly arranged members in wider order on the axis. The 

 perianth parts in Ranunculaceae, Rosaceae and many other orders with 

 multipartite flowers, although apparently whorled, are, in fact, inserted in 

 a very closely set series, which needs only a little elongation to produce a 

 true spiral. 



In most of the advanced families the perianth parts are, however, 

 strictly whorled, that is to say their primordia appear simultaneously on the 

 young receptacle, without any trace of precedence among them. The 

 change from a spiral order, even a compressed spiral, to this ringed arrange- 

 ment, is analogous to the change on the vegetative shoot between a spiral 

 phyllotaxy and opposite or whorled leaves and has presumably come about 

 in a similar way. It is no mere matter of shortening of internodes or of 

 rearrangement of parts by suppression or fusion; a perfectly definite change 

 in the growth pattern of the apical meristem is involved and of its inner 

 physiological meaning we are still ignorant, though its consequences in 

 flower-building have been obvious and profound. At one step it hmits 

 the number of parts of each kind, it crystallizes the tendency to isomerism 

 and it opens up possibilities of constructive fusion on which the higher 

 flower-forms depend. 



The arrangement of the perianth parts in relation to each other is called 

 the aestivation of the flower. There are two fundamental types. When the 

 parts meet edge to edge they are said to be valvate, and this might be 



