D— ZOOLOGY 99 



ecologically — i.e. by the kind of country inhabited : desert, savannah, 

 forest. In fact, the peculiarities of a race are best understood if it is 

 considered as part of the environment. 



Not all geographical races amalgamate when they come together. 

 Many of them have become so different that they can live side by side, 

 each being an independent community not interbreeding with the other. 

 As instructive examples I will mention some Swallowtail butterflies : 

 Papilio thoas L. occurs, split up into many races, in South and Central 

 America, its range reaching into U.S.A. A very closely related species, and 

 evidently originally its northern race, P. cresphontes Cram, flies in U.S.A. 

 and extends far south into Central America, the two common insects keeping 

 perfectly distinct in all characteristics, no hybrids being known. In the 

 Oriental region Papilio eurypylus L. ranges in various subspecies from the 

 Bismarck Islands westward to India, and P. doson Feld., which we only 

 recognised as a separate sister-species after a more careful study of the male 

 genital organs, occurs from Ceylon and South India eastwards to the 

 Philippines and the Lesser Sunda Islands. The two species, therefore, 

 are found together over a large area, but the most western districts are 

 inhabited only by P. doson and the most eastern by P. etirypylus ; originally 

 they were the Western and the Eastern forms of one species. It is evident 

 that these butterflies represent a further step in the evolution of species 

 than the species-pairs which still amalgamate in the area common to them. 



Sometimes we find both amalgamation and specific distinctness among 

 the forms divided from a parent stock, as is the case in the sister-species 

 Cat-flea and Dog-flea. The home of the genus Ctenocephalides to which 

 both belong is Africa. Tropical and South Africa are inhabited by a 

 subspecies with short head, and the Nile countries by one with a long 

 head, the two intergrading in the Sudan and Uganda. From India to 

 the Papuan countries, with the exclusion of Australia, a third race occurs, 

 and in Europe and Central and North Asia the cat-fleas were represented 

 by the flea occurring on dogs and wolves. When the Egyptian house- 

 cat came to Europe, it brought with it the long-headed form of Ctenoce- 

 phalides felis Bouche, which thereby came into contact with the Palsearctic 

 shortheaded dog-flea. One might have expected that they would hybridise 

 and amalgamate, but they did not. The morphological differences are 

 but slight, but a physiological barrier had arisen which kept and keep the 

 cat- and dog-fleas as species, although they may occur together on the 

 same individual of the host. When my brother pointed out the specific 

 distinctness of the two fleas, he encountered a good deal of criticism 

 before his opinion was generally accepted as correct. 



Before leaving this subject I will mention a type of local form which 

 stands apart from the usual kind of geographical race. In gregarious 

 mammals, such as the African buffalo, one herd seems frequently to 

 differ from another herd, and as herds keep to their particular district, 

 the difference has all the appearance of being geographical and having 

 originated in the same way as the geographical distinctness of which 

 I have spoken. But we know that family likenesses are inheritable, and it 

 appears to me that the herd distinctions are really family characteristics 

 impressed on the herd by the dominant bull. The point requires further 



