DERIVATION OF THE FLORA AND FAUNA 279 



Tropic-, Arcto-, and Antarcto-Cretaceous, and it is the latter that interests us here. 

 Axp:i,KOI) emphasizes that numerous data allow us to conclude that many families 

 anil i^enera (generally rej^arded as ha\ in<4 evoKetl in the temperate zones orif^inated 

 in the tropics. 



The data seem to support the view that angiosperms have not had an exclusively 

 holarctic source, or a wholly austral center of origin. They were being assembled in 

 both regions by the gradual adaptation of basic troi)ical groups to extratropical conditions 

 during the long j)eriod of Permo-Triassic down to the Cretaceous. On this basis the 

 Arcto-Cretaceous and Antarcto-Cretaceous floras of higher temperate latitude represent 

 vegetation types whose genera were derived largely by the long and continued differen- 

 tiation of successively derivative members of original tropical and border-tropical angio- 

 sperms. From this standpoint, the temperate regions to the north (holarctic) and south 

 (antarctic) are subordinate or secondary centers in early angiosperm evolution. 



The conclusions we can draw from this are (i) that the origin and primary seat 

 of the angiosperms w^as in the tropics before anything like the present map existed; 

 (2) that this tropical flora gradually gave birth to a temperate flora also in the 

 south; (3) that Antarctica became the centre of a varied Antarcto-tertiary flora 

 which spread north and over S. America reached the region where now stands 

 Juan Fernandez. 



II. Pteridophyta. 



Owing to their great age and perhaps also to their greater faculty of dispersal 

 many of the genera and some species have attained a very extensive range making 

 it difficult or impossible to assign them to one of the elements proposed in the sub- 

 division of the angiosperms. Of 23 genera found in Juan Fernandez, 16 are very 

 wide-spread; some of them belong to the largest fern genera, and even if we avail 

 ourselves of the generic concept used by COPELAND [dp), the situation remains 

 practically the same. To CoPEL.\XD 50-75 % of the living ferns are of austral or 

 Antarctic ancestry; we shall see to what extent this is true of the Juan Fernandez flora. 



With regard to the Hymenophyllaceae Copelaxd remarks {dy. 174) that "no 

 other plant family of its size and diversity is quite so conspicuously Antarctic in 

 origin as this one". 



Trichomanes-Vandenboscliia is pantropical; as 2 of the 3 insular species are 

 decidedly neotropical in their affinities, also T. philippianum is referred to the same 

 element. 



Serpyllopsis is subantarctic-American and best attached to an Antarcto-tertiary 

 element. 



Hy))ienoglossum should perhaps be referred to the same element, but its re- 

 lations are just as unknown as those of Serpyllopsis. 



Hymenophyllum-Mecodium is a pantropical group, also represented in Xew Zea- 

 land; cnneatum and caudicidatuDi are most likely of neotropical origin and this may 

 be true also q{ fuciforme. H. pectinatum stands, COPELAND says, apart from the other 

 species, but I can find no good reason to call in Antarctic, while the distribution 

 of Sphaerocionium ferrugineum and its near relatives testifies to their Antarctic origin 

 in spite of their affinity to ciliatum (comp. above p. 219). Hymenopliyllum s. str. 



