the worshippers[21], previous to their offering sacrifices, by obliging them to squeeze themselves through this narrow passage, the inconvenience of which was increased by the awful character attributed to this cromlech. Even in Christian and Mohammedan countries religious ordeals of a similar kind are not unknown; and a very analogous custom in the celebrated mosque of Kairowán may well have its origin in the older pagan practice of the aborigines. However this may be, the religious character of the whole structure can scarcely be doubtful, from the nature of the flat stone, the channel in which was certainly intended to carry off the blood of the victim.[22] It must strike the observer, in regarding these ruins, that while they are so rude in principle, their style of execution evidently bears traces of art; and I think it not improbable that the art may be ascribed to Roman influence. We shall further on see another specimen of these curious pillars combined with the ground-plan of an almost regular Roman temple. But from whatever quarter this artistical influence may have proceeded, there cannot be the least doubt that the character of the structure is, on the whole, not Roman, but indicates quite another race; and if we take into regard what I have just said about the influence of art visible in this structure, and that such influence could scarcely proceed from any other quarter than that of the Carthaginians or the Romans, we must attribute these remains to the Berber race, who, during the historical period, were the exclusive possessors of these inland regions. Analogous structures have been found, however, not only in England and Ireland on the one side, and in several parts of India, principally in the Nilgherries, on the other, but also in Circassia, Southern Russia, on the South Arabian coast, and in the Somali country. This analogy might certainly be explained by a similarity of principle in the simple religious rites of rude people; but there may be also in these curious remains a confirmation of the theories of Sir Henry Rawlinson respecting the wide extension of the Scythians. But while, with regard to other tribes, from the Dravidian group in Southern India to the Celtic in Ireland, such a connexion of origin seems to be confirmed by analogy of language, there exist but very few points of analogy between the Berber and the Central Asiatic languages, except by means of the Coptic. In every respect, however, it may be better to call such remains by the general name of Scythian than by that of Druidical, which certainly can be justified only with regard to the north-west of Europe.