424 



FOURTH GROUP. SEED-PLANTS. 



FIG. 351. 



If the relative positions of the parts are first considered, it is obvious that they 

 can never be symmetrically distributed in flowers with a purely spiral structure, but 

 that in hemicyclic flowers the members at least that have a cyclical arrangement 

 may possibly be symmetrically distributed. If on the other hand the parts of the 

 flower are all arranged in whorls, they are also usually monosymmetrically or poly- 

 symmetrically distributed on the torus ; thus the diagram in Fig. 340 may be divided 



symmetrically and regularly by three planes, 

 that in Fig. 341 by four, that in Fig. 342 by 

 five, while Fig. 344 can only be symmetrically 

 halved by one plane, which is at the same time 

 the median plane. The diagram in Fig. 345 

 may be divided by the median section into two 

 symmetrical parts, which are different from 

 those produced by the lateral section; the 

 diagram is zygomorphous like those in Fig. 343 

 J5, C and Fig. 344, but the latter are singly, the 

 former is doubly symmetrical. 



The symmetry of the mature expanded 

 flower is usually genetically connected with 

 the relations of symmetry of the diagram 

 which represents only the number and position 

 of the parts, as is seen by comparing Figs. 352 

 and 354 with Fig. 344 A ; but inasmuch as the 

 general form of the mature flower is essentially 

 determined by the outlines, dimensions, torsions 

 and curvatures of the separate parts, these are 

 the things which chiefly affect the relations of 

 symmetry in the expanded flower, and they may 



affect them to such a degree that even flowers with their foliar structures arranged 

 spirally may be monosymmetrically zygomorphous in reference to their general form, as is 

 the case to a great extent in Aconitum and Delphinium. It must however be observed 

 that in these instances the zygomorphous form is produced chiefly or exclusively by the 

 calyx and corolla, in which the spiral arrangement may perhaps still be disputed, but 

 which in any case are inserted on so narrow a zone of the torus that their position may 

 be considered as equivalent to a cyclical (verticillate) one. If on the other hand the 

 floral axis is sufficiently elongated for the ascending spiral arrangement to be distinctly 

 shown, as in the perianth and androecium of the Nymphaeaceae and in the androecium 

 and gynaeceum of the Magnoliaceae, the full development of the organs does not seem 

 to produce a zygomorphous or any other really symmetrical general form. 



On the other hand the zygomorphous and monosymmetrical form is frequently found 

 in flowers, whose parts are arranged in whorls. Very distinct zygomorphism is often 

 combined with partial or entire abortion of certain members, as in Columnea (Fig. 352) 

 and other Gesneraceae, in which the posterior stamen is changed into a small nectary, 

 while in Labiatae it is entirely wanting ; in Orchideae, where of the six typical stamens 

 only the median anterior outer stamen or two obliquely anterior inner stamens are 

 developed, the abortion is still greater. Sometimes the subsequent monosymmetrical 

 general form is to some extent provided for by the order of succession of the parts of 

 the flower in the first rudiments, so far as these are not formed simultaneously in a 

 whorl and do not follow one another in the circle with a definite angle of divergence, 

 but their development begins with an anterior or posterior member and then proceeds 

 simultaneously right and left of the median line to the opposite side of the whorl, in 

 the manner already described in the Papilionaceae in the one case and in the 

 Resedaceae in the other. 

 The diagram of the zygomorphous flowers of the Fumariaceae (Fig. 345), as has been 



Flower of Heracleum pubescens with zygo- 

 morphous corolla. 



