ORIGINATION OF FAUNAS. 15 



ral of the antelopes have related, and apparently ancestral, forms 

 in the Miocene deposits ot Greece (Pikermi), which also contain a 

 foim not very far removed from the giraffe (Helladotherium), and 

 a species of tnie giraffe itself (Camelopardalis Attica), so that 

 possibly a contingent of the African fauna may have been derived 

 from this region. Whether the southern or Ethiopian portion of 

 the continent was at one time since the introduction of the placental 

 Mammalia completely severed from the northern part or not there 

 are as yet no means for detei-mining. That Madagascar at one time 

 formed part of the continent is indisputably proved by the character 

 of its fauna ; but that its subsequent isolation is of very ancient 

 date is conclusively shown by the complete absence of all the more 

 distinctive Ethiopian placental mammals. 



The few examples that have been cited in illustration of the 

 appearance and disappearance of faunas are sufficient to show the 

 character of the investigation that is open to the zoogeographer. 

 While from the data that we now possess much can be done towards 

 shaping our suppositions, it must be confessed that our knowledge 

 is still much too limited to permit of very satisfactory conclusions 

 being drawn therefrom. The principal danger that besets any in- 

 vestigation in the direction here outlined arises from the very 

 natural assumption that the greater antiquity in any one region over 

 another of a given type of animal indicates its prior appearance 

 there, and migration thence to one or more secondary regions. 

 This assumption might be well founded if we were only half con- 

 versant with the past paleontological histories of the regions under 

 consideration ; but where at best our knowledge is still very imper- 

 fect, as it is in the case of Africa, Asia, and South America, it would 

 be, to say the least, highly injudicious. Eor what evidence have 

 we that animal types not yet found, or dating back only to a com- 

 paratively recent period, might not some day be turned up in abun- 

 dance, and in deposits of such age as to completely overthrow any 

 deductions that may have been based upon their supposed non- 

 occurrence ? A single illustration of this kind will suffice. Pale- 

 ontologists are in the habit of considering the camels a New World 

 family, which by migration finally occupied the region which it 

 now inhabits. This conclusion is based upon the circumstance that 

 numerous cameloid forms (Pliauchenia, Procamelus, Protolabis, 

 Poebrotherium) carry this line of animals back in the North Ameri- 



