DORSIVENTRAL FLOWERS 



129 



position the side of the flower upon which the sterile stamens stand is 

 turned towards the outer side of the zigzag inflorescence. There can be no 

 doubt that this arrangement stands in direct relation with the pollination. 

 The relationships of symmetry of the several flowers to one another and the 

 subordination of the symmetry of the individual flower to that of the 

 whole inflorescence, even in completely asymmetric flowers, are matters 

 requiring consideration, although up till now they have not attracted 

 attention. In Calathea the flowers are in pairs, each flower is so con- 

 structed that it cannot be divided by one plane symmetrically, but the 

 two flowers together form a symmetric whole 1 . The flowers of Valeriana 

 are also quite asymmetric. 



With reference now to dorsiventral flowers, we have to distinguish 

 two cases: either (i) the flowers are laid down radially and become 

 dorsiventral in the course of their further development ; or (2) they are 

 dorsiventral from the beginning the dorsiventrality 

 appearing at the vegetative point of the flower when 

 its parts are laid down -, as happens in Reseda and 

 Leguminosae. 



i. Most dorsiventral flowers are laid down as 

 radial structures and only subsequently become 

 dorsiventral. The period at which the change to 

 dorsiventrality takes place varies, the earliest is 

 observed in cases where during the unfolding of 

 the flower-bud the position of the flower-parts is 

 so altered that a dorsiventral construction appears, 

 and the change is the result of external influences :i . 

 In Epilobium angustifolium and Epiphyllum trun- 

 catum, for example, it is produced by the reaction 



to gravity, and according to Focke the curvature which the style of 

 Lilium auratum and of L. lancifolium exhibits is a heliotropic pheno- 

 menon. The consideration of the details of these movements belongs 

 more to experimental physiology, here I will only point out that to the cate- 

 gory of which I speak belong probably those plants in which radial and 

 dorsiventral flowers both occur in one and the same individual, or in 

 different individuals of the same species. According to H. Muller 4 , 

 Saxifraga stellaris, which has usually radial erect flowers, produces 



FIG. 82. Commelina coelestis. 

 Diagram of an intlorescence. The 

 arrows indicate the plane of sym- 

 metry of the several (lowers ; these 

 all turn the same side outwards, 

 that, namely, on which the sterile 

 stamens, indicated hereby crosses, 

 stand. After Eichler. 



1 Plofmeister has, as in so many other cases, not overlooked this; see Allgem. Morph. p. 581, 

 \Vith regard to the Zingiberaceae see F. Muller, Schiefe Symmetric bei Zingiberaceenblumen, in Ber. 

 der deutsch. bot. Ges. v. p. 99. 



2 For details see the account of the development of the flower in Part II of this book. 



* Dufour, De 1'influence de la gravitation sur les mouvements de qutlques organes floraux, in 

 Arch. d. scienc. phys. et nat., periode 3, xiv. p. 41 3 ; Yochting, Uber Zygomorphie und deren Ursachen, 

 in Pringsh. Jahrb. xvii. 



1 II. Muller, Alpenblumen, p. 535. 



GOEIiEL 



K 



