THE GENUS MANETTIA 9 



The general evolutionary aim, then, is the segregation of the 

 flo\vei-s to a position of freedom from the leaves. This aim is attained, 

 to a limited extent in § Pyrrhanthos, by enlargement of the individual 

 flowers, by reduetion of leaves in the neighbourhood of the flowers, 

 and to a lesser degree by aggregation of the floral shoots, Avhich 

 themselves are abbreviated, so that the flowers tend to mutual 

 proximity. In § Heterochloi-a the same result is aimed at by en- 

 largement of the individual flowers and by elongation of the pedicels. 

 In § Lygistum the same goal is aimed at in the progressive mul- 

 tiplication of small flowers and their close aggregation into definite 

 inflorescences; it is the " umbellifloral " tendency in operation. 

 These tendencies are accompanied as a matter of course by the aggre- 

 gation of the inflorescences toward the terminal nodes of the branch- 

 lets ; never, in the case of Manettia, by the elongation of peduncles. 

 Where necessary — e.y. in the case of sessile mnbels — the pedicels are 

 more or less elongated, as in the series M. congest a — JX. conqesioides 

 — M. verticillata. These last-named represent the highest degree of 

 aggregation in the genus, with the exception of the two species with 

 capitulate inflorescences on relatively long stalks — M. capiiata and 

 JSL. Mie)'sia)ia. Between the extremes the aggregation -tendency is 

 observable in a series of species, reflecting the evolution of axillary 

 umbelloid groups of flowers, seen already in M. calycosa, and well- 

 advanced in M. sabiceoides, 31. umhellata, 31. liispida, etc. 



In the way of these inflorescence-tendencies, § Ysginanthus dis- 

 plays various stages of progress. 31. coccinea and its nearer allies 

 are even less advanced than 3£. cordifolia in this regard. Aggre- 

 gation, very similar to that which produced 31. congesta, is observable 

 in 31. dominicensis, 31. tarapotensis, etc. 31. fimhriata bears its 

 flowers many together in more or less elongated racemoid clusters, 

 much like 31. Lygistum. It would appear that inflorescence-develop- 

 inent has taken place within § Ysginanthus independently of, and 

 parallel to that within the other three sections together. 



The tendency of the flowers to free themselves from the leaves is 

 thus in most cases subserved by aggregation into dense inflorescences, 

 culminating in the capitulate forms. But such is not invariably the 

 case even among the Lygistum-species ; for a few tend to secure the 

 same object by the formation of a lax paniculate inflorescence, fi'eed 

 of the leaves by elongation of the inflorescence-branches. The same 

 tendency is accompanied by considerable reduction in the size of the 

 flower. In this case there is no tendency to suppression of the main 

 rachis of the inflorescence, such as is manifest in the umbellate- 

 fasciculate forms ; but rather a general lack of arrest of any of the 

 branches. This tendency is foreshadowed in the small-flowered 

 31. alha, and advances a step in the inany-flowered cymose panicles of 

 31. tenuis. 31. Beyrichiana has very lax panicles and long slender 

 pedicels, the flowers being more or less free of the leaves and situate 

 mostly well beyond them. The very short stout corolla of 31. alba 

 is seen also in 31. paniciilata and M. barbata, both of which reflect 

 the same tendency to a kx paniculate grouping. 



JOUKXAL OF BOTAKY, OcT., 1918. [SuiTLEMEKT.] C 



