168 H. B. GUPPY^ M.B.;, F.R.S.E., ON PLANT-DISTRIBUTION 



single birthplace. Eavenala, a genus of the Musaceae, offers us 

 very singular instance of disconnected distribution. It contains 

 only two known species, of which one (It. viadagascariensis) is 

 the Traveller's-tree confined to Madagascar, whilst the other 

 {R. guiancnsis) is restricted to tropical South America. Then 

 there is the genus Adansonia, to which the familiar Baobab-tree 

 belongs. Of its four species, two are African, one belongs to 

 Madagascar, and the fourth is Australian. Then we have the 

 genus Mesembryanthemum, which, though mainly African, 

 possesses a few Australian and South American species. Again, 

 the breadth of an ocean lies in each case between the South 

 American, Australian, and African species of Podocarpus. 

 These examples have been selected because they raise the same 

 questions that are suggested by the disconnected distribution of 

 animals like the marsupials and the tapirs. Evidently we are 

 not here concerned with capacities for dispersal. 



The testimony of the rocks only adds to our difficulties in the 

 search of the home of a genus. What are we to say, for 

 instance, when many living genera of trees, both tropical and 

 temperate, such as Eucalyptus, Ficus, liriodendron, Myrsine, 

 Quercus, etc., present themselves in association and without 

 w^arning in the Cretaceous deposits of North America ? How is 

 it possible, again, to speculate on the home of Eucaljnptus, 

 when we know that it existed in IMesozoic times both in Europe 

 and in ^^orth America ? As far as concerns their former wide 

 dispersal, the marsupials and the gum-trees behave in a similar 

 fashion. \Yhere, it may be asked, ought we to look for the 

 home of Liriodendron ? Found fossil in the Cretaceous and 

 early Tertiary deposits of North America, Greenland, and 

 Europe, its once numerous species are now only represented 

 by a solitary species growing in North America and China. It 

 would seem, indeed, with this evidence l)efore us, that it is not 

 legitimate to raise the question of a home at all. 



But the difficulties are not restricted to the disconnected 

 distribution of genera. The distribution of families presents 

 almost insuperal^le difficulties when viewed from the standpoint 

 of disj)ersion from a centre. It would indeed appear that the 

 farther we trace them back in geological time, the wider is their 

 range. Where, for instance, should we look for the home of the 

 I)alnis at present ffourisliing throughout the tropics but extending 

 far north into temperate latitudes (hiring Eocene times ? 



With some of the families that are well represented in the 

 geological record we cannot even detect the commencement 

 of the differentiation of their tribes. With the Taxacene, ibr 



