THE INFLORESCENCE. 



77 



sarily a true dichotomy, but may apparently be so owing to the 

 abortion of the terminal bud. 



Examples of the indefinite form are seen in the Ouciferse, especially 

 the Wallflower, where a few flowers at first appear in a tuft, while the 

 seed-vessels are afterwards wide apart on an elongated raceme, the upper- 

 most being the youngest. In the Foxglove and similar plants we may 

 produce a very long development of the indefinite structure by picking 

 off the low r er flowers as they wither, when, as no seed is formed, the in- 

 definite terminal bud retains its energy, and continues to lengthen until 

 the plant is exhausted. On the other hand we observe, in the Sweet- 

 William, the Elder, and the Hydrangea, the centre flower of a tuft opens 

 first, and the definite inflorescence becomes wider and wider, but never 

 elongates or Arrows out in the centre. 



Fig. 132. 



Fig. 134. 



Fig. 135. 



Fig. 133. 



Fig. 132. Spike of Verbena afflcinalis. 

 Fig. 133. Spadix and spathe of Calla. 



Fig. 134. Compound spadix and spathe of a Pa]m. 

 Fig. 135. Compound spike, with spikelets, ofJLolium, 



When an indefinite inflorescence is elongated the lowermost 

 flowers open first, while if it be of a flat-topped or crowded character, 

 the outermost flowers open first and the central ones last, as in 

 the capitula of the Compositse. Hence the indefinite forms of 

 inflorescence are sometimes called centripetal or progressive, and the 

 definite centrifugal or regressive. 



There is an exception to the ordinary regularity in the capitula of 

 Dipsacus (Teazel), where the florets open first halfway up ; and then 

 proceed both centiipetally and centrifugally. 



