200 THE SCOTTISH BOTANICAL REVIEW 



grasses, also the carpels which in many flowers monopolise 

 all the residual part of the growing-point of the flower, and 

 the terminal stamen of Naias^ Callitriche, etc. Ambrosia is, 

 however, a specially well-marked example of this procedure. 

 The development of the female flower need not be described 

 further, except to point out that there is only a mere trace of 

 the corolla, while the stamens and calyx have disappeared 

 without leaving a trace. 



Thus we see that Ambrosia has carried still further in its 

 female inflorescence that reduction already indicated in 

 Xanthiwn. Even in Xanthium the male organs of the 

 capitulum are completely suppressed and the number of 

 female flowers is reduced to two. As new structures there are 

 present the hook-like spines on the outside of the concrescent 

 floral-bracts, which, taking the place of the pappus originally 

 present, facilitate dispersal of the fruit. In Ambrosia these 

 spines dwindle to a few rudiments and the number of flowers 

 diminishes to a single one. 



In the male inflorescences, however, the number of capitula 

 is probably increased considerably in comparison with the 

 type-form with hermaphrodite flowers. With this may also 

 be correlated, that these male capitula have departed from 

 the prevailing scheme of development, and as seen in their 

 " reversal " have progressed along new lines. Thus with 

 no great eff"ort, and with a basis of facts admissible in any 

 new speculation, we are able to trace, in what seems to me a 

 satisfactory manner, the historical evolution of the structure 

 of the inflorescences of Ambrosia. At the same time, the 

 case dealt with here illustrates once again the phenomenon to 

 which the author has frequently made reference : namely, that 

 our phylogenetic series, so far as we can depict them with 

 any degree of probability, all represent a reduction-series. 



The following may be advanced in support of this : — 



1. In a descending series we have a definite starting-point 



(that is, some one of the more completely equipped 

 forms as distinct from the more reduced ones) with 

 which we can compare the less completely equipped 

 members of the series. 



2. In many instances the organs in question may still be 



recognised as rudiments. 



