Non-Helletiic Portion of the Latin Language. 545 



Deva (Chester), the station of the twentieth legion, would to this 

 day almost serve as a boundary between the English and Welsh 

 counties. Add to this, that the inscriptions of Roman workman- 

 ship to the west of this line are too trifling to allow us to suppose 

 that any long occupation of the country could have taken place. 

 Were it otherwise it would be difficult to account for the station- 

 ary position of two out of the three legions by which Britain was 

 garrisoned, the one at Caerleon, the other at Chester. The Ro- 

 man legions in the provinces, liable to hostile attacks, had their 

 head-quarters almost invariably on the frontiers. It is, therefore, 

 almost impossible to account for the selection of two such points 

 as these, without the supposition that the legions were placed 

 there to guard against the attacks of a race which, if vanquished, 

 was nevertheless unbroken in spirit. Even the very grammar of 

 the language and the traditions of better times long anterior to 

 the Roman invasion, are proofs that as far as the western portions 

 of Great Britain are concerned, the amalgamation with Rome was 

 never completed. Still, however, the long residence of the Ro- 

 mans in the island, with the known influence always produced by 

 such a state of things, renders every statement grounded on the 

 similarity alone of the languages of the two races, the conquered 

 and the conquerors, liable to suspicion. I have, therefore, been 

 compelled to enter upon an exceedingly difficult investigation, 

 which, if successful, must prove the radical identity of the Latin 

 and Cumrian tongues. The proof is this — 



If there are derivative words in the Latin, of which we must 

 seek the primitives in the Cumrian, and if these primitives be 

 shewn to furnish an explanation of many words before inexpli- 

 cable on etymological principles. For example, if the verb " to 

 tread," under various forms, be found with the meaning, "to 

 trample with the feet," in most of the western languages of Eu- 

 rope, and have no noun to base itself upon in these languages, 

 and yet the noun, " traed the feet" be found in one of them, the 

 inference is irresistible that the verb in all its forms was derived 



