FOUMER LAND COXXECTIONS. 327 



Again, on the characters of the tentacle, the genera can be 

 divided into three groups, each containing both primitive and 

 degraded members. One result of tiiis is to emphasize a special 

 resemblance between Africa and America, for the largest section, 

 comprising 10 genera, has no representative whatever in Asia, 

 though it extends to the Seychelles. The second section of six 

 genera is also represented both in America, Africa and the 

 Seychelles, and in addition has Indian members. The smallest 

 section of three genera is confined to the Old World, ranging 

 from Africa to S.E. Asia. Thus we see that Africa behaves like 

 a centre of dispei-sal for the whole group, and the relationship 

 between the American and Asiatic elements is through the African 

 fauna. 



Some Evidexce Relating to Connections Between the 

 Southern Continents Through Antarctica. 



Certain faunal resemblances have long been known between 

 the extreme south of Africa and other portions of the southern 

 hemisphere. These also ixiight be regarded as last remnants 

 of a primitive cosmopolitan fauna, which, owing to competition 

 with more finished products of the same stock, have been driven 

 southwards to the ends of the earth : or, another view would 

 explain them in terms of a Gondwanaland theory. But, at any 

 rate, it cannot be claimed for them, as some authorities may 

 do for Peripatus and the Coecilians, that they left the northern 

 hemisphere and persisted only in the tropics because the northern 

 climate became too cold. Some advocates of the view first 

 mentioned lay much stress on the occurrence of fossil leaves, 

 supposed to belong to the family Proteaceae in the Miocene 

 shales of Florissant in Colorado, this family at the present day 

 being specially well developed in South Africa and occurring alsu 

 in South America, Madagascar and Australia, but not at all 

 in the northern hemisphere. Several eminent botanists have 

 rejected these identifications as entirely untrustworthy, for leaf 

 form alone is useless as a family character. Yet, Prof. T. D. A. 

 Cockerel!'" does not hesitate to identify them thus, and tells 

 us by way of verification that he took a certain lady up to some 

 young plants of Grevillea robusta in a greenhouse, and asked 

 without explanation, " Where have you seen that? " The reply 

 came instantly: " In the shale." She did not l:)ioic irhy I 

 asked, nor what the plants were: the impression made by the 

 cut of the leaves was naive ajid. immediate. (Italics mine). If 

 such is the evidence that Proteaceae were formerly northern, it 

 can only be regarded as speculative. According to Dr. 

 Schonland, the best authorities have treated all the identifica- 

 tions of southern plants in the Tertiaries of the northern hemi- 

 sphere as worthless when based only on leaf characters. 



A few instances of these southern resemblances are men- 

 tioned here. 



(1) The freshwater fishes of the genus Galaxias occur in 

 "South Africa (restricted to Western Cape Province), the southern 



