INTRODUCTION 5 



sion to refer to plants of this kind in the following 

 chapters, I will here give a few details respecting them. 

 It is familiar to every one that the ray-florets of the 

 Compositae often differ remarkably from the others ; and 

 so it is with the outer flowers of many Umbellifera j , 

 some Craciferae and a few other families. Several 

 species of Hydrangea and Viburnum offer striking 

 instances of the same fact. The Rubiaceous genus. 

 Mussasnda presents a very curious appearance from 

 some of the flowers having the tip of one of the sepals 

 developed into a large petal-like expansion, coloured 

 either white or purple. The outer flowers in several 

 Acanthaceous genera are large and conspicuous but 

 sterile ; the next in order are smaller, open, moderately 

 fertile and capable of cross-fertilisation ; whilst the 

 central ones are cleistogamic, being still smaller, closed 

 and highly fertile ; so that here the inflorescence con- 

 sists of three kinds of flowers.* From what we know 

 in other cases of the use of the corolla, coloured bracteae, 

 &c., and from what H. Miiller has observed f on the 

 frequency of the visits of insects to the flower-heads of 

 the Umbelliferae and Composite being largely deter- 

 mined by their conspicuousness, there can be no doubt 

 that the increased size of the corolla of the outer 

 flowers, the inner ones being in all the above cases 

 small, serves to attract insects. The result is that 

 cross-fertilisation is thus favoured. Most flowers wither 

 soon after being fertilised, but Hildebrand states J that 

 the ray-florets of the Compositae last for a long time, 

 until all those on the disc are impregnated ; and this 

 clearly shows the use of the former. The ray-florets, 



* J. Scott, ' Journal of Botany,' men,' pp. 108, 412. 



London, new scries, vol. i. 1872, J See his interesting memoir, 



pp. 101-164. 'UtberdieGeschlechtsverhnltniss 



t ' Die Befrucbtung dcr Blu- bei ilcn Compositen,' 1869, p. 92. 



