842 TRANSACTIONS OF SECTION H. 



Upper and Central Italy, are now admitted to have spoken an Aryan language, and 

 I have recently given some reasons for believing that the Latin language is simply 

 the native tongue of the ahorieiual Ligurian population of Latium with some 

 admixtures derived from the Italic tribes of Siculi and Sabines. I have also shown 

 that the ancient Iberians, the next neighbours of the Ligurians, used the same 

 forms of place-names as the latter, and that some of the words plainly exhibit 

 Aryan terminations. Thus we may conclude that with the exception of the 

 Basques, who are probably a non-Aryan spurt from North Africa, the melanochrous 

 populations of Spain, Italy, the Balkan Peninsula, France. Britain, Ireland, and 

 Holland have from the first spoken none but an Aryan language. 



(c) Only one argument is now left to the defenders of the non-Aryan 

 theory. When the study of sociology first sprang up in the last century, it at 

 once became a fundamental doctrine that the Aryans had always been strictly 

 patriarchal, and that polyandry and descent through women was unknown amongst 

 them. Though this view has received many rude shocks in later days, Professor 

 Ziramer argues from it tliat the indigenous people of Britain and Ireland were 

 non-Aryan, 



It is well known from the ancient writers that the Picts were polyandrous and 

 that succession was consequently through females. Again, it is certain, both 

 from the ancient Irish literature and al!=o from statements of external 

 writers, that the Irish were polyandrous, and that they also almost certainly 

 traced descent through women. Accordingly Professor Zimmer infers that the 

 indigenous race was non-Aryan. But McLennan has long since pointed out that 

 descent through women was the ancient law at Athens, and I have just shown 

 that the Athenians and Arcadians, the autochthonous, dark-complexioned people 

 of Greece, never spoke any save an Aryan tongue. Moreover, I have shown else- 

 where that the Ligurians, who are now generally admitted to have spoken always 

 an Arvan language, had descent through women, whilst I have also pointed out 

 that there is good evidence that the ancient Latins, who have generally been taken 

 as typical Aryans, had the same system. Again, it is admi+ted that the ancient 

 Illyrians and dark-complexioned Thracians spoke an Aryan language, which, 

 inasmuch as it differed materially in certain ways from that spoken by their Celtic 

 overlords, must have been aboriginal, whilst I have further given grounds for 

 believing that the ancient Iberians (thous'h not the Basques) were also an Aryan- 

 speaking folic. But there is jrood evidence that the lUvrians, melanochrous 

 Thracians and Iberians all traced descent through women. In view of these facts 

 it is useless to urge that because the Picts of Scotland and the ancient Irish had 

 that system of succession through females these peoples must have been non- 

 Aryan. 



We have now reviewed the three main criteria of race at present used by 

 anthropologists: («) pigmentation of the skin, hair, and eyes; (b) the shape of the 

 skull and other osteological characteristics; and finally (c) their system of tr.acing 

 descent. We have seen that osteolo?ical differences may be but foundations of 

 sand, because it is certain that such variations take place within very short 

 periods, not only in the case of the lower animals, as in the horse family, but in 

 man himself. Pigmentation is no true criterion, for we have found a steady 

 tendency to change in colour in the case of the lower animals from latitude to 

 latitude, whilst in the case of man the steady shading off in colour from dark to 

 blonde may be traced from the equator to the Baltic. Unless then we postulate 

 that man is entirely free from the natural laws which condition the osteology and 

 pigmentation of other animals, we must admit that neither bone nor colour differences 

 can be regarded as crucial criteria. Further, we saw that the test of descent 

 through males or females broke down absolutely in the case of peoples who can be 

 proved historically never to have spoken any but a non-Aryan language. Finally, 

 we are forced to the conclusion that language, now that we realise what are the 

 laws which govern its borrowing by one race from another, is reallv the surest 

 of all the known tests of race when dealt with broadly and over wide areas, and 

 not merely in the way of guesswork etymologies. 



II. Hitherto I have dealt only with the need of a rigid application of 



