Let me take, as an illustration, a well-known passage, in which the “Chronicle” tells how the West Saxons, in 577, “took three ceasters, Gleawan ceaster and Ciren ceaster, and Bathan ceaster.” Now, each of these cities would be severally known as a “ceaster,” and, in due course, as _the_ “ceaster”--just as “forum” was developed by the Romans, and as “market” has been by ourselves (_vide ante_, v. 250)--but when, as here, mentioned together, they would have to be distinguished from one another. The _hams_ and _tuns_ which covered the land were so distinguished by prefixing to them the names of their English owners. This could not be done with the _ceasters_, which did not become the homesteads of English owners. The distinctive prefix was, therefore, sought in some existing (although, to the invaders), meaningless name, either that of the river on which it stood, as “Exan ceaster” (the Chester on the Exe), or that of the place itself, “Glewan ceaster,” a form which may be paralleled in the “Fort Chipewyan,” “Fort Winebago,” &c., of their descendants in North America.[47] It is often, of course, most difficult to say whether the prefix is derived directly from the river, or indirectly, through the original place-name. But, in any case, we must dismiss the hypothesis that the prefix was “the Romanised title” of the town, for the termination “an” (“Ex_an_,” “Gleaw_an_”) is found in cases where the Roman forms differed so widely as “Isc_a_” and “Glev_um_.” We must guard against the idea that such prefix was ever the “Roman name” itself, used in apposition to “ceaster.” To resume, then, we have seen that there is no ground for supposing “_castrum_” to have ever formed part of the “Roman names” of those cities whose modern names end in “Chester,” &c. From this it has followed that the terminal in question is the result and badge of the English invasion, representing the English word “ceaster,” the invaders’ term for a walled town, and not the equivalent of a mere _castrum_, though etymologically derived from it in the first instance. We have also seen that the terminal in question was not a mere “addition” to the “Roman name,” but was itself the new name imposed by the conquering English, to which, when and where necessary, a prefix was in time permanently added, for the sake of distinction. It is not, surely, too much to say that if these conclusions were satisfactorily established, they must gravely modify, if not revolutionise, the view which has hitherto universally prevailed, and which is based, I think, on a too hasty induction from the resemblance between the English and the Latin words.