396 ZOO-GEOGRAPHICAL NOTES. 



tions actually take place among the fishes for spawning pur- 

 poses, I was able to convince myself at the rapids of the river 

 Rol, near Mvolo, where twice a year a movement of immense 

 shoals of fish goes on up and down stream, an occurrence which 

 is well known to the natives, and much desired by them, for 

 it supplies the country far and wide with fish. I observed the 

 fish moving down to the north at the end of October, while 

 the contrary movement takes place, according to the statement 

 of those who dwell there, in June. 



Let us now turn our attention to another series of pheno- 

 mena, which are the more interesting because they exhibit 

 facts which seem to be opposed to the universal rules of the 

 distribution of animals ; I mean the appearance of certain 

 species in localities where they are separated from their nearest 

 relations by very long tracts, sometimes by whole continents. 

 Such cases are to be found in Europe in the presence of 

 Cyanopica cyanea, Pall., in Siberia, whilst their nearest rela- 

 tions (Cyanopica, Cooki) reside in Spain, of Myogale moschata 

 in Russia, and of If. pyrenaica in the Pyrenees, while no 

 species even allied to them exists in the intervening countries. 

 An entirely similar instance is afforded by the appearance 

 of the Angola magpie {Pitta angolensis) in a limited district of 

 West Africa, while the genus Pitta most essentially belongs 

 to Wallace's oriental region, between which two haunts beside 

 the ocean, lies the whole breadth of the African continent. 

 The same holds good of Atherura (tufted-tailed porcupine) 

 and the West African representatives of the Tragulidse ; both 

 genera are properly oriental. 



Now, as it is so difficult to account for such very exceptional 

 cases, they must challenge us to an exact study of the interven- 

 ing regions ; and that surprising disclosures may often be found 

 there is shown by the following facts : — According to my 

 experience, Atherura is by no means rare in Central Africa, 

 and is fairly equally distributed ; in Monbuttu I succeeded in 

 collecting within a short time a number of living specimens. 

 In contradistinction to nonda, the common variety of porcu- 

 pine {Hystrix) which is found beside it, it is called holla, 

 and its region may at present, as far as my experience goes, be 

 defined as extending northwards to about 3° 40' N. lat., and 



