DERIVATION OF THE FLORA AND FAUNA 215 



water. lie mentions Conipositae-Labiatiflorae, Cal}xeraceae, \'alerianaceae, Cacta- 

 ceae, Orchidaceae and Dioscoreaceae as good examples. In other cases he blames 

 the diaspores as in Nolanaceae, Leguminosae, Violaceae, Fagaceae, Amaryllidaceae 

 and Liliaceae. 



Also with reference to life-forms the island flora makes an impression of being 

 unbalanced, a fact already discussed at some length in 2ji. 825-830; I shall 

 only repeat that woody species are in overwhelming majority and annuals almost 

 absent, whereas they are very numerous, also pro[)()rtionately, on the mainland. 

 Already in 19 14 [22'/) I argued that climatic differences alone do not offer an 

 explanation and that historical causes must be taken into account. 



Geographical elements. 



In his survey of the flora p. 229-232 JoilOW discussed its composition and 

 distinguished various elements. The endemic genera and a few peculiar endemic 

 species of non-endemic genera form his first (and most ancient) group; the second 

 group contains the remainder of endemic species: (a) markedly distinct, (b) nearly 

 related to continental species. Both groups make up "Continjente A". The third 

 group, "Continjente B", contains the species found elsewhere, all occurring in South 

 America except Ilalorrhagis ''alata". America is claimed as the source of the en- 

 demic element as well: "tanto las especies del primero como las del segundo grupo 

 pertenecen, con la unica excepcion del Santalum, a jeneros representados, si no 

 en Chile, a lo menos en alguna parte de la costa occidental de Sud-America" — 

 this did not, however, apply to the isolated Compositae, to Lacioris etc., which 

 JoilOW regarded as originated from ancestors to be looked for in the Tertiary 

 flora of Chile. Halorrhagis, Santalum and Coprosma, mentioned later, were sup- 

 posed to have immigrated from western Pacific. 



It serves no purpose to go into detail; when JOHOW wrote his book our 

 knowledge of the flora was too defective to allow him to arrive at anything like 

 safe conclusions. The same may be said of my 1914 paper [221), even if my short 

 visit to the islands revealed the existence of an up till then unknown element. 

 Four main groups were distinguished: I. Old Pacific (Engler's Altoceanisches 

 Element), comprising genera or species supposed to have a long history behind 

 them within the precincts of the Pacific and lacking near relatives; subdivision 

 A, P2ndemic genera, with 11 sp., and Non-endemic genera with 10 sp.; subdivision 

 B with allied species in Hawaii, Polynesia, Australia and New Zealand, 16 en- 

 demic species^the genus Dendroseris was also placed here — and i non-endemic 

 [Halorrhagis). Group II, called Neotropical, contained 6 endemic species; Cnminia 

 and Juania were included here. Group III, called Chilean, was the largest and was 

 divided into 3 lots: A, very distinct species, 18 ((9c//^^rtz/?Vz placed here); B, less 

 well-marked species, 1 1 ; C, also found in Chile, 28, but among them were 6 not 

 now regarded as native. So far the main difference between this arrangement and 

 Joiiow's lies in the greater number of species with supposed west Pacific connec- 

 tions. Finally Group IV, Subantarctic-Magellanian, 4 species, was added. 



The study of our 1916-1917 collection added many species not known before 

 and gave rise to a fresh analysis of the vascular plants {2jg), of which a 



