INTRODUCTION lix 



introduced form, allied to Raillardella, if not actually of that genus. These Composites 

 are common and generally distributed, though Dubantia apparently is not known on 

 Hawaii, except from the most ancient part of that island, that part in fact which 

 originally formed the whole island 



Next come the three genera Tetramalopiuni, Lipochaeta and Canipylotlieca only 

 "on the borderland of generic distinctness," the two former reputed to have species 

 existing in California and Mexico, and the latter practically equivalent to Coreopsis. 

 Finally we have Artemisia with two endemic species, each occupying a very different 

 station. 



The era of arborescent Lobelias is supposed to exhibit much the same phenomena 

 as the Compositae. Dr Guppy says that some are in their prime, others on the point 

 of extinction, and that the distribution of genera and species within the islands shows 

 that the original immigrants were not contemporaneous. They are only found in the 

 Eastern Pacific Islands, chiefly in Hawaii, with a few species in Eastern Polynesia. 

 They are absent from the Western Pacific Islands for the same reasons, it is said, as 

 are the Compositae. 



However, on an examination of the Hawaiian endemic Lobelias, we find that 

 there are five genera, all well represented with from half-a-dozen to about thirty 

 described species, excepting the monotypic Brighamia, of peculiar habits and occupying 

 stations impossible for most of the others to thrive or even exist in. But there is this 

 great difference between the endemic Hawaiian Composites and the tree-lobelias; of 

 the latter the genera are not only all endemic, but they are remarkably peculiar, so 

 much so that Hillebrand did not even hint at their affinities. The seven endemic 

 genera of Hawaii and Tahiti are placed by Engler "in a group by themselves"... 

 " and he does not approve of the endeavours of some botanists to isolate one of them 

 from the rest and to connect it {Brighamia) with the Australian Isotoma " (Guppy). 



Now it seems much more natural to suppose that these allied genera of Hawaiian 

 Lobelias have been developed in the islands from some very anciently immigrant 

 species, than that all the genera (or their ancestors) were originally produced elsewhere 

 and have all become extinct in their original homes. Or if they have become in the 

 islands so greatly modified that their American ancestors can no longer be recognized, 

 the time sufficient to produce such modification would have been also ample to have 

 allowed the development of the large number of species (many being still undescribed) 

 and of most of the genera, even from a single anciently introduced form. One may 

 judge of the near relationship of these Hawaiian genera, when one studies the synonymy 

 of the species and notes how competent systematic botanists have constantly referred 

 identical species to different genera ! 



While the endemic Compositae may aptly be compared with the endemic Carabidae 

 in the beetles, the Lobeliaceae are comparable rather with the Prognathogrylline crickets, 

 though the evolution of species in the latter has been much less profuse. 



h2 



