136 RELATIONSHIPS OF SYMMETRY 



from the chief axis, 6", and this is due to the fact that at an early period 

 before the appearance of the flowers the axis of inflorescence is flattened 

 upon the other side which lies closely adpressed to the chief axis. This 

 side bearing no flowers then becomes strongly convex, by which a good 

 protection is provided for the young flower-buds. Subsequently the 

 axis of the inflorescence straightens in a negative geotropic direction, 

 and is orthotropous in spite of its dorsiventrality. Comparison with other 

 Leguminosae shows that this dorsi ventral inflorescence is probably derived 

 from a radial one. 



Speaking of this leads me to mention the dorsiventral circinate in- 

 florescences which are found in many, although not all, Boragineae and 

 Hydrophylleae, in Hyoscyamus, and other plants ; such inflorescences arc 

 phylogenetically derived from scorpioid cymes l . The chief consideration 

 for us here is that they are markedly dorsiventral, bearing flowers upon 

 the upper side and the bracts, where these exist, upon the flanks. The 

 axes of inflorescence are usually obliquely ascending and unfold themselves 

 in such a way that an opening flower always stands upon the highest 

 point of the axis ; this flower is thus not only most visible but is also 

 best illuminated, and it is in the highest degree probable that light plays 

 not merely an indirect but a direct part in pollination. It is for example 

 a common belief of gardeners that artificial pollination is more likely to 

 succeed in sunshine than in dull weather 2 . 



I must refer to my previous publications for further details regarding 

 inflorescences which are dorsiventral from the outset, and I now pass 

 to those inflorescences which are unilateral through torsion of their flower- 

 stalks or their axes of inflorescence whilst they were laid down as radial 

 structures. We find such forms amongst the Leguminosae, for example 

 in Hedysarum sibiricum, and others, but also in other families especially 

 in Labiatae, for example in Horminum pyrenaicum, Scutellaria peregrina, 

 and others, in Scrophularineae, for example in Digitalis purpurea and 

 especially Melampyrum, in Pyrolaceae, and also in many monocotyle- 

 donous plants, such as Gladiolus, and Freesia. This unilaterality of the 

 inflorescence is specially striking when the bracts originally standing in 

 decussate pairs are brought through torsion of the axis of inflorescence 

 to lie in two rows which converge in many cases to that side of the 

 inflorescence which bears no flowers, whilst the flowers themselves bend 



1 For the literature of this subject see Goebel, Uber die Verzweigung dorsiventraler Sprosse, in 

 Arb. d. bot. Inst. in Wiirzburg, ii (1880). Where, as in Borago officinalis, the flowers have long stalks 

 and large corollas the inflorescence is as good as not dorsiventral, and is really a true scorpioid 

 cyme as it occurs in Scrophularia. 



2 This may be a consequence of the secretion from the stigma, and we must remember the amount 

 of transpiration is different in the sun from what it is in diffuse light. The matter deserves careful 



investigation. 



