I 2 8 RELATIONSHIPS OF SYMMETRY 



the horse-chestnut, the lupin, and many other plants, remain smaller 

 than those turned away from it, and thus the leaf-surface is carried 

 more to the outside ; the smaller leaflets are in these cases the last 

 laid down. 



Geum bulgaricum supplies an interesting case (Fig. 81). Here the 

 leaf looks as if it were changing from an interruptedly pinnate one to 

 a simple one. The terminal leaflet is very large, and its position gives 

 it the appearance of a peltate leaf. The lateral leaflets are almost 

 covered by it and are correspondingly reduced ; their differences in size 

 can however be always recognized. In other species of Geum the 

 difference between the lateral leaflets and the terminal one is not so 

 great. 



V 



RELATIONSHIPS OF SYMMETRY OF FLOWERS AND 



INFLORESCENCES. 



FLOWERS. 



I have already pointed out that both radial and dorsiventral construc- 

 tion is found in flowers. Flowers which are divisible by two or more 

 planes of symmetry and flowers in which this is not the case occur, 

 apparently without rule, upon plants with radial flowers. The former 

 are only possible if the flowers be cyclic, the latter when they are acyclic, 

 but this difference is of no moment in a general consideration of the 

 condition. The expressions 'regular' and 'irregular' are perhaps best 

 avoided. In most dorsiventral flowers the plane of symmetry passes 

 through the median of the bract 1 ; seldomer it is transverse to this, as 

 in Corydalis and in Fumaria (in which one of the two transverse petals 

 is spurred), Wachendorffia thyrsiflora 2 , and others. The flowers of 

 Fumaria and Corydalis are twisted subsequently through an angle 

 of about 90, so that the spur which was originally lateral comes to 

 lie in the median plane. An oblique position of the plane of symmetry 

 is not infrequent in flowers which stand in cymose inflorescences, for 

 example in Commelinaceae, Aesculus, and others ; it is not however the 

 cymose character of -the inflorescence, but the position of the flowers 

 to one another which is critical heie. Thus if w r e examine the diagram 

 of an inflorescence of Commelina coelestis (Fig. 82) we observe that the 

 several flowers are all indeed obliquely dorsiventral, but through this 



1 The so-called ' median-zygomorphy.' 



- See Eichler, in Sitzungsbtr. der C.esellsch. naturforsch. Freunde zu Berlin, 1880. 



