Taking all factors, anthropological and archæological, geographical and linguistic, into consideration, and in spite of the difference in opinion expressed by Dr. Giles, whose authority to pronounce on the linguistic data all must acknowledge, I am venturing to identify the nomad steppe-folk with the primitive Wiros, while admitting the possibility that the beginnings of their language may date back to Magdalenian and Azilian times, when they may have been living surrounded by the Carpathian ring. /Chapter XIII/ P’S AND Q’S We have seen that with one notable exception, little attempt has been made to explain the early history of the Wiros since 1889, and the position of the Aryan hypothesis has remained stationary.[460] It is true that fresh evidences of such languages have been discovered in the uplands of Asia, and a new group, known as Tocharian,[461] have been identified. Certain affinities to the group have also been noted in the Hittite language, which has been claimed by some writers to be a true Wiro tongue.[462] But this view has not received general acceptance. Little use, however, has been made of this fresh evidence towards solving the problem of the Aryan cradle. But early in 1891 an important communication was made to the Philological Society by Professor, afterwards Sir, John Rhys.[463] This paper raised a storm of hostile criticism, especially in Germany,[464] and its conclusions have not found favour in philological circles. As, however, some of Sir John’s conclusions coincide in certain particulars with the reconstruction offered in the previous pages, based on other evidence, the thesis demands reconsideration. To summarise briefly, Rhys pointed out that the Celtic languages, now confined to the north-western fringe of Europe, fell naturally into two well-defined groups. One of these, the Gaelic, or as he preferred to call it the Goidelic, was spoken in Ireland, North-West Scotland and the Isle of Man. The other, formerly called Cymric, but by Rhys styled Brythonic, was spoken in Wales, Cornwall and Brittany. There are several marked differences between these two groups of languages, the most important being that the C in the Goidelic, which represents an earlier Q or Qu, is replaced in Brythonic by a P or perhaps a B. Thus the Celtic languages fall into two well-defined groups which may be called the Q and P dialects. Rhys pointed out, too, that in the Italian peninsula the same phenomenon appeared. In Latin, and the dialects most closely allied to it, Q or Qu was found, while in the Umbrian forms of speech, used over the greater part of the peninsula this sound was replaced by P. Thus there were Q and P dialects in Italy also.