lunkhart Corse in the Yarrow, and probably also Jeffrey's Corse, a hill beyond Bowbeat in the Moorfoot range, are to be attributed to the Anglo-Saxon cross, for which corse is common. Gar occurs more than once, and means shank, also heron, probably from its long shank. It appears in Garlet and Garlavin, hills at the head of the Cymric Talla. Gar or gair also means fort, as in Kittlegairy. Lavin is probably the Cornish lawan, birds. Coomb is not uncommon, probably from the root cum, a cup-shaped depression in the hills, hence a shelter or place between hills, and thence probably applied to the hill itself. We have Coomlees in the original sense. White Coomb is on the borders of Peeblesshire and Dumfries, and Kingledoors Coomb is in Tweedsmuir. is found in districts where the Cymric element is strong, and seems to have been adopted by the Anglo-Saxons, and changed by them into coomb or combe, as is common in the south of England. But Professor Leo of Halle combe or coomb is due to a

Cwm

maintains that the Saxon Saxon root, cimban, to join.2 Lin, a pool, is common, as in linn, a waterfall, or pool at the foot of it. It is also in the Water of Lyne, and in Linton. Lin is the Cornish form of the root; Welsh, llyn; Gaelic, linne. Man, a place or district, appears in Caersman. This is the same word as in the Isle of Man, and in Slamannan. Pen, a head, hence a hill, is a true Cymric test-word. It occurs frequently in Tweeddale. It will fall to be noticed more particularly in connection

1 It is possible that gar, for fort, may be Scandinavian, from gar, to hedge; hence garth, an enclosure or fence.

2 Nomenclature of the Anglo-Saxons, 83.

with the hill-names of the district. Ros, a moor, Cornish, is to be distinguished from the Gaelic ros, a headland. The former is supposed to appear in Melrose and Roslin; but the Gaelic meaning, projection, would suit both places well. It is very common in Cornwall.

Tor, a prominence or projecting rock, appears in Tortye, Torbank, Torwood, Torwoodlee. Torsonce, Torquhan, Torcraik, are in a line northwards from Torwoodlee. Tushielaw is originally Tor-shiel-law,-showing successive strata of different tongues-viz., Cymric, Scandinavian, Anglo-Saxon. Tor is very common in Derby, Devon, and Cornwall. The tar in Tarcriesh, parish of Stobo, is probably the same prefix. This form in Gaelic is torr. The suffix in Tortye, on the Tweed near Dawyck, is the Welsh ty, a town or settlement. Tortye is thus wholly Cymric. We have tor in Tarbolton, or Torbolton, in Ayrshire, part of Strathclyde.

Tre, trev, treva, tra, and dre, are all forms of the same root, meaning a dwelling-place or home. These are Cymric, in fact mainly Cornish. Akin are the Irish treabh and the Gaelic treubh. The root tre is said to occur ninety-six times in the names of villages in Cornwall, and more than twenty times in the same use in Wales. It is frequent in Brittany, in France, and in Spain.2 It occurs in Trahenna, Traquair, Trebetha, Dreva, and in Trabroun and other places in the eastern counties. Dreva is simply the softened form of Treva, and is analogous to Trêves, Trieste, Trient. The same race who left their names in the plains of Central Europe and on the sunny shores of the Adriatic, pitched their 2 Ibid., 240. G

1 Taylor, Words and Places, 237.

VOL. I.

dwellings on the green sloping hills of the Tweed, and all that is left of them there or nearer home is the symbolic word. Traquair, in a document of the year 1116 is Treverquyrd, and in one shortly after the year 1143 is Trauequair. Trauefquer is also found in the 1200. These are the oldest spellings of the name we know. Mr Skene, than whom there is no higher authority on such a point, regards Traquair and Trabroun as derived from the Gaelic treabher, meaning "a naked side "—hence traver. The old form, Treverquyrd, rather supports this etymology. Trauequer and Trefquer seem to point, on the other hand, to the Cymric trev and trevow. There is at least no ground for supposing Trahenna, Trebetha, or Dreva to be other than the simple and obvious Cymric forms. And if Traquair be Gaelic, it is certainly among the rarer names of the district.

To these we ought perhaps to add the Cymric lan, llan, an enclosure, hence a church. For though this form is not found in the valley of the Tweed, it is found in Llanerch (Lanark), enclosure in the wood, and this was one of the principal towns in the kingdom of Strathclyde, of which Tweeddale was the south-east portion. Llan is said to occur ninety-seven times in names of villages in Wales, and thirteen times in Cornwall. It is probably to be found much more frequently in those districts. It is also very common in Brittany.1 Mr Skene tells us that llan and tor are both Gaelic and Cymric, and that these are therefore not proper test-words between the dialects. No doubt he is right in the first assertion, though lann in Gaelic is given by Williams as an old and obsolete form. Taylor, Words and Places, 241.

1

But the very frequent and preponderating occurrence of these and other roots common to both dialects in Cymric districts, as opposed to Gaelic, points clearly to a certain continuity of race in the Cymri, since the words were first used by the common Celtic stock. And this continuity during a long lapse of time is sufficient to give a distinctive character alike to a branch race and a branch language. The subdivision of a people which keeps and uses certain terms which another subdivision abandons, wholly or in part, during the same period, makes for itself a distinctive dialect. We are thus entitled to regard a district in which these retained words are now found, though the Cymri may have disappeared from it, as once closer in race and language to the one subdivision, the Cymri, than to the other, the Gael. I think, also, that where there are two forms of the name of a place, one of which appears in Gaelic and another in the Cymric of the present day, we are entitled to regard an old local name which is the same with, or similar to the latter, as properly Cymric, having been given by the people whose language the Cymri have retained without material change. Those words in which consonants occur that are susceptible of certain uniform phonetic changes are no doubt the most certain tests, but they can hardly be regarded as the only ones.

If we look at the natural objects indicated by the Cymric names, we shall find among these the principal streams. Thus, Tweed is in Cymric Tywi, from the root tuy, and probably means what limits, checks, or bounds. Twyad in Welsh is a hemming in. Robertson gives the Gaelic Tuath-aid, "the river flowing to the north side,"

the th being mute. The oldest spelling is in favour of the other derivation. Bede writes Tuid. In 1185 it is written Tuede.1 Teviot is apparently from Teifi, a name common in the early Cymric poems. It is retained in the "Teiit," the vernacular pronunciation. The root is tyw, spreading round, and Teifi is supposed to mean the spreading stream. Both Twyi and Teifi are frequent in the oldest Welsh and Strathclyde literature. Fruid is in Cymric frwyd, and means the impulsive stream. The root frw, or frou, impulse, occurs in a great many Welsh words, and is finely imitative of the sounding rush of water. Talla is from tal, that tops or fronts. Tal Ard occurs in the poems of Merlin; and Taliessin, the early Cymric bard of Strathclyde, was "the brightbrowed." No one will doubt the appropriateness of the name of the stream who has seen the Talla gleam in its line of foam, amid the mists of Lochcraig Head and Moll's Cleuch Dod, downwards through its precipitous glen, and then pass swiftly amid the glacier moraines, a thousand feet below its source, to plunge headlong over its linns. It fronts the eye grandly by its high foaming flow, and not less grandly does it surround and possess the ear with its continuously falling sound.

This

Lyne is from the root lyn, a pool, or slow-flowing body of water, and the term is specially descriptive of the Tweeddale Lyne. Manor was originally spelt Maineure. was in 1186. Then Menare and Menar occur in 1401 and 1555. It is always pronounced Mæner in the vernacular. The root is, no doubt, man, a stone, the same form as in Manchester. And as the stream flows brightly

1 Acts of Scots Parliament, temp. William the Lion.