E— GEOGRAPHY 123 



5000 B.C., for all the major language families had differentiated before 

 that period. 



The cradleland is represented as sending forth successive flows of lava 

 from a centre of eruption. These form concentric zones about south 

 central Asia, and each flow pushes its predecessor to the margin. The 

 eff"ect of one flow on its neighbour — involving some contact and assimila- 

 tion — is also rather usefully indicated. The flows reach the four ' pen- 

 insulas ' of Asia (i.e. Europe, Africa, America and Australasia) according 

 to the relative advantages of the connecting corridors. I have elaborated 

 this concept in several essays already published (Griffith Taylor 1936c). 



G. Ecological Notes on the Aryan Problem. 



We may use the stage-diagram to correlate our scanty knowledge as to 

 the early wave-fronts of the Aryan languages. There are three fairly 

 definite subdivisions of Aryan : (i) the early Kentum or' K' speeches like 

 Gaelic and Latin, (2) the Intermediate ' P ' languages like Welsh (with 

 which we may associate Teutonic and Greek for convenience), (3) the 

 later Satem languages like Slav and Indian. 



Turning to Fig. 13, some idea of our knowledge of the language dis- 

 tribution in Sumerian times is given in the lowest map of the series. 

 At this time Hamitic languages were used by the Pharaohs in Egypt, akin 

 to those still spoken by the Berbers in the Atlas Mountains. Semitic 

 languages characterised Arabia and Syria, as they still do. Sumerian 

 itself has some resemblances to the Altaic, though its affinities are not yet 

 clearly understood. In Europe at this early date there were racial allies 

 of the present-day Hamitic-speakers — all of Mediterranean race — living 

 in the western regions, who probably spoke Hamitic according to Rhys 

 and Jones (Griffith Taylor 1936a). Central Europe was occupied by early 

 migration of Alpine ' Brakephs ' (broadheads), of whose language we 

 know nothing. It was almost certainly not Aryan, and something akin 

 to ' Basque ' seems most likely. This problem is taken up later. In 

 view of the important corridor linking Turkestan with China by way of 

 the Tarim Basin, I have ventured to suggest that a linguistic kinship 

 between early Chinese (Sinitic) and Sumerian or early Aryan is only to 

 be expected. 



In the second map (Fig. 13 at B), for the period around 1200 B.C., we 

 are on surer ground. Vast migrations of ' Satem '-speakers had poured 

 into India from Turkestan. The Hittites, who seem to have spoken an 

 Aryan tongue somewhat akin to the Kentum Group, were in control of 

 Anatolia. Semitic was now the chief language of Egypt and Mesopotamia. 



In central Europe (if we adopt the suggestions of H. Peake) Kentum 

 languages were spoken in the regions east of the Alps, while Brythonic 

 (one of the Intermediate ' P ' type) was that used by the Cimmerians of 

 the Ukraine and Caucasus areas. It seems logical to assume that many 

 Satem-speakers still remained in Turkestan, and were perhaps allied to 

 the Sarmatian tribes. 



In the next map (for 400 B.C.) we see the first great Aryan conquest in 

 the Near East, that of the Persians. They spoke a Satem language, and 

 it is probable that their Sarmatian kin were occupying the European 



