40C) ZOOLOGICAL GEOGRAPHY. [part hi. 



110 actual land connection is required for birds which can cross 

 consideraLle arms of the sea. 



Reptiles again seem to offer no more support to the view than 

 do mammalia or liirds. Among snakes there are no families in 

 common that have not a very wide distribution. Among lizards 

 the Gymnopthalmidse are the only family that favour the 

 notion, since they are found in Australia and South America, 

 but not in the Oriental region. Yet they occur in both the 

 Pahearctic and Ethiopian regions, and their distril)utioii is alto- 

 gether too erratic to be of any value in a case of this kind ; 

 and the same remarks apply to the tortoises of the family 

 Chelydidse. 



The Amphibia, however, furnish us with some more decided 

 facts. We have first the family of tree-frogs, Pelodryade, con- 

 fined to the two regions ; Litoria, a genus of the family Hylida: 

 peculiar to Australia, but with one species in Paraguay ; and in 

 llie family Discoglossidse, the Australian genus Chiroleidcs- has 

 its nearest ally in the Chilian genus Cahj'ptoccphalvs. 



Fresh-water fishes give yet clearer evidence. Three groups are 

 exclusively found in these two regions ; Aiihritis, a fresh-water 

 genus of Trachinidffi, has one species in Tasmania and two 

 others in Patagonia ; the Haplochitonidic inhabit only Terra del 

 Puego, the Falkland Islands and South Australia; while the 

 genus Galaxias (forming the family Galaxidte) is confined to 

 South Temperate America, Australia, and New Zealand. We 

 have also the genus Ostcoglosswm confined to the tropical 

 rivers of Eastern South America, the lndo-]\Ialay Islands and 

 Australia. 



It is important here to notice that the heat-loving lieptilia 

 afford hardly any indications of close affinity between the two 

 regions, while the cold-enduring amphibia and fresh-water 

 fish, olfer them in abundance. Taking this fact in con- 

 nection with the absence of all indications of close affinity 

 among the mammalia and terrestrial birds, the conclusion seems 

 inevitable that there has been no land-connection between the 

 two regions within the period of existing species, genera, or 

 families. Yet some interchange of amphil>ia and fresh-water 



