TOPOGRAPHY OF CORNWALL. 



353 



Vertevia, and since arno as in Duriarno means a river, some 

 place on the river Mole is at once suggested, and in all probability 

 South Molton. 



And here the Ravennat again returns to take up the line 

 of the central route left at Tamaris. His next name is Termonin, 

 and bearing in mind that the e sounds a, and that the local pro- 

 nunciation of Tor is still Tar, Termonin has a very natural 

 rendering in Tor-hill, and we are led to locate it somewhere on 

 the line of the great Dartmoor trackway. Without going into 

 detail, I may say that there is a Tor-hill which shows abundant 

 traces of ancient population, in the neighbourhood of Eippon 

 Tor. Mostevia, which comes next, may be variously placed, 

 but Milidunum there can be no hesitation in identifying 

 with the great earthworks on Milber Down, near Newton 

 Abbot. Apaunaris is indicated beyond the reach of 



reasonable controversy in Hope's Nose, the eastern headland of 

 Torbay, which bore ample traces of ancient settlement. The 

 modern Hope's is a clear corruption of the Apau which the 

 Eavennat has preserved, and his naris is of course a translation 

 of the original Scandinavian ness. Masona is doubtful, and 

 may very well be Berry Head, once the site of an important 

 station ; but whether this be so or not I have no hesitation in 

 placing Alongium on Haldon, completing the route to Exeter on 

 the south. 



It will be gathered from this that I assume the Eavennat 

 to have worked methodically, possiblj^ from a map, proceeding 

 from west to east, with all the regularity possible, and taking 

 traverses from the north to the south to complete his work. The 

 same rule applies, so far as I am able to trace, through the 

 counties of Somerset and Dorset ; and I believe it supplies the 

 key to the interpretation of his record for the kingdom. My 

 belief in the correctness of this solution of his problem does not, 

 it will be seen, depend upon a few happy guesses, or possibly 

 chance coincidences, but upon a whole series of consecutive 

 identifications, each of which decreases in geometrical ratio the 

 chances of error. 



The results at which I have arrived are best shown probably 

 in the accompanying map. 



