192 E. C.Andrews — The Geological History of the 



decided description between the great tropical land blocks, and 

 a connection, moreover, which appears upon first inspection 

 to have ceased at the present time. Indeed, were one to con- 

 sider only the evidence of one great family or order only, such 

 as the Myrtacese, Leguminosse, Solanacese, Rutacese, or Pipera- 

 cea?, it would seem impossible to escape the conclusion that the 

 present tropical lands had been directly connected in the not 

 very remote past. But the distribution of the mammals does 

 not bear this out and, more important still, the facts of 

 structural geology are directly opposed to the idea. In the 

 Cretaceous, or Eocene, before the strong zoning of climate 

 mentioned by Chamberlin, these genera most probably had' a 

 much wider distribution than they have at present, and at that 

 time they may have moved along land bridges which existed 

 much farther to the north than the present tropics, yet without 

 leaving the mild and moist climate apparently so necessary to 

 their existence. Such conclusion is indeed suggested by the 

 existence of outposts of types such as Solanum, Euphorbia, 

 Sophora, and Hibiscus, in the temperate regions. 



The case of the oceanic islands should ever also be kept in 

 mind in this connection. All the facts of geology go to prove 

 the non-existence of direct land connection at any time between 

 islands on the one hand such as St. Helena, the Azores, and 

 the Hawaiian groups, and the neighboring continents, on the 

 other hand. These appear to have derived their flowering 

 plants from the great land blocks through the agency of sea- 

 currents, birds, animals, and man. 



For example, are we to conclude that the several species 

 of Commidendron in St. Helena, the seven species of Tetramo- 

 lobium, the 11 species of Lipochceta, the 12 of Campylotheca, 

 6 of Dubautia, 12 of Raillardia, 2 of Hesperomannia (all in 

 Composite), 20 of Pilea (Rutacese), 16 of Kadua (Rubiacese), 

 11 of Clermontia (Lobeliacese), 29 of Cyanea (Lobeliacese), 

 and so on, in Hawaii, all of endemic genera, were one and all 

 wafted or carried to these islands, and yet exterminated on the 

 main lands. Is it not rather an inescapable conclusion based 

 upon the theory of probabilities that, since these lands must 

 have been supplied by waifs and colonists from the great lands, 

 inasmuch as the geology appears to disprove their former land 

 connection with the continents, then they are the descendants 

 of types brought to these islands in old times and which have 

 since produced either endemic species or genera or both ? 



Among the numerous families which can be considered as 

 illustrating the connection of the tropical lands and Australia 



