he derives these words from 'pinna,' a battlement or projection of a wall ('pena' in Provençal); and adds, that if they had been derived from the Celtic 'pen,' they would certainly have kept the masculine gender. The same objection, if it be sound, applies to the case of La Penne cited above.

The Gaelic form of 'pen' is 'ben ;' and the use of one or other of these words, like that of Aber' and Inver,' enables us to trace the line which separated the Cymric and Gaelic branches of the race in Scotland. In the South of Scotland 'Pen' occurs frequently, but to the North and West the Gaelic term is found in Ben Nevis,' Benlomond,' 'Benledi,'' Bencruachan,' and many other cases. 'Cenn' is another Gaelic form of the same root, and appears in Kenmore, Cantire, Kinnaird, Kinross, and Kenmare, in Ireland: perhaps the county of Kent, and other names, may retain traces of it, as Mr. Taylor suggests; but it must be remembered that the Cymric cefn,' already referred to, would easily pass into ken or kenne.

The old Gaelic word for a height or fortress was 'dun,' represented in Welsh by 'din' or 'dinas.' Dintywi' was 'the castle on the Towey river.' Dunedin, Edinburgh; Lugdunum, Lyons; Sedunum, Sitten or Sion in the Vallais; Ebredunum, Yverdun. The last syllable of London is probably the same word. Lexdon is a mixture of two languages-Legionis-dunum, and Camalodunum has become Maldon. Sometimes, on the Celtic principle of composition, it precedes the qualifying word, as in Dumfries, Dunkeld, and Dumbarton.

'Rhos,' a moor, is another element in Celtic names of places. 'Craig' signifies a crag or rock; and 'tor,' a high summit such as those on Dartmoor or on the Cornish moors. The syllable 'ard,' high, occurs in two hundred Irish names, such as Ardagh; and very frequently in Scotland, as in Ardrossan,'' Ardnamurchan,' and probably in Arran.' It forms the first syllable of the name of the forest' Arden,' of which Drayton says—

Mighty Arden, even in her height of pride

Her one hand touching Trent, the other Severn's side.'

Ardennes, the great forest on the borders of France and Belgium, is the same name. The second syllable of both- den,' signifying a deep wooded valley-is supposed to be a Celtic word adopted by the Saxons; and a third great forest-that of Dean -seems to be named from the same root. It also forms the termination in such names as 'Tenterden.'

The Welsh cwm,' and the Anglo-Saxon 'comb,' occur frequently in proper names, and still exist in the language of the people. Combe, in the West of England, means, we believe,

properly

properly a valley or depression in the hills, without a running stream. 'Tre,' a place or dwelling, pretty nearly equivalent to the Saxon town, does not occur in Gaelic or Erse names: it is especially common in Cornwall, and is of course met with in Wales, and less frequently in the neighbouring counties. In one Herefordshire name, the township of Trevil, the process of accumulation, spoken of above, seems to have taken place by the addition of the translation 'vil,' or 'ville,' to the original Welsh word. Nant,' as in Nant Francon,' the Vale of Beavers in Wales, meets us again in the Alps of Savoy, and testifies to the presence of a Cymric race there. Near Chamounix it is constantly applied to a torrent. Llan again, usually an enclosure for a church, is a Cymric, not a Gaelic word. The principal ecclesiastical words in Welsh are, as might be expected, of Latin origin. Thus we have 'Esgob,' a bishop, from 'Episcopus; Eglwys,' a church, from 'ecclesia;' and 'Plwyf,' a parish, from plebs,' the ordinary middle-age word for a congregation.†

We have not space to go through all the Celtic elements of names given by our author; but the Erse 'Magh,' signifying a plain' or field, which appears in Armagh, Maghera, and perhaps in Magdeburg, is very remarkable. The corresponding Cymric form is 'Maes,' a meadow or mead.

'The chief Cymric roots are found scattered over Spain, Northern Italy, Switzerland, and Southern Germany; but the root "magh," the Erse test word, seems to be confined almost entirely to the district of the Lower Rhine and its tributaries. In Switzerland it does not appear, and in Italy it occurs only in the district peopled by the intrusive Boii. In southern and western France it hardly occurs at all, and it is found only once or twice in Britain. We may, therefore, conclude that while the Cymry came from the region of the Alps, the Gaelic branch of the Celts must have migrated from the valleys of the Rhine and the Moselle. It seems to have been from this district that the earliest historic movement of the Celts took place. Three Celtic

* In the time of Giraldus Cambrensis (about 1180), the beaver was found only in the Teivi. Inter universos Cambriæ, seu etiam Loegriæ fluvios solus hic castores habet.' The two furs which the queen was entitled to by the Welsh laws were the beaver (Llostly dan) and the ermine (Charlwng): see Book xiv. ch. iii. 16, 'Ancient Laws and Institutes of Wales' (1841). In chapter iv. of the same book it is called 'Llostlydan y Befyr,' where the story of its self-mutilation is told. The first of these two names makes it clear that no other animal was meant, for it signifies 'Broadtail.'

As to this sense of plebs, see Savigny, 'Geschichte des Römischen Rechts,' v. i. s. 456; and the curious lines quoted by Raumer (Hohenstaufen, v. vi. s. 167) :— 'Sed et omnis episcopus urbis Plebes vendebat, quas sub se quisque regebat.'

Compare Ducange, in v.

tribes

tribes burst through the Alps; they pillaged Rome, and, after returning to Illyria for a while, they broke in upon Greece, and plundered the treasures at Delphi.* They settled for a time in Thrace, and then, crossing the Bosphorus, took possession of the central parts of Asia Minor, to which they gave the name of Galatia, the land of the Gael, and where they long retained their Celtic speech, and the ethnical peculiarities of their Celtic blood. Here, curiously enough, we again encounter this root "mag," which is found so abundantly in the district from which they emigrated. In the Galatian district we find the names of Magarsa, Magydus, Magabula, Magaba, Mygdale, Magnesia (twice), and the Mygdones. In Thessaly, where these Celts settled for a time, we also find two of these names, Magnesia and the district of Mygdonia, which lay on the banks of the Axius, a Celtic river name. Magaba is on the Halys, which is a Celtic word, meaning "salt river." In Lycia, according to Strabo, there was an enormous rocky summit, steeply scarped on every side, called Kpayos.'-(pp. 245-247.)

But we must pause for a moment here and observe that Mr. Taylor can hardly mean to attribute the name Magnesia, in Thessaly, to the inroad of the Celts after the taking of Delphi. In the Catalogue of the ships (II. B. 756) we have

• Μαγνήτων δ ̓ ἦρχε Πρόθοος Τενθρηδόνος υἱὸς,
οἱ περὶ Πηνειὼν καὶ Πήλιον εινοσίφυλλον

ναίεσκον.

Herodotus (vii. 132) tells us that the Magnetes were among the Greek tribes who gave earth and water to Xerxes; Magnesia, on the Mæander, is also mentioned by the same writer (iii. 122). Mygdonia and the river Axius are spoken of in the march of Xerxes (vii. 123, 124). There is, as our readers well know, a story connected with Croesus passing the river Halys (i. 75). If these, therefore, are Celtic names, they were there long before the march of Brennus. In fact, we should not dream of attributing such an oversight to Mr. Taylor if it were not for the words printed in italics above, which seem to leave us no choice. In like manner we find the names of Magarsus and the Magarsian Minerva recorded by Arrian in connection with the expedition of Alexander (B.C. 332), whereas the defeat of the Gauls on their retreat to the Danube by Sosthenes took place B.C. 277, or fifty-five years afterwards. Again it must be remarked that 'Magaba' was a mountain,, which does not very well suit the meaning

* Sir George Lewis says: "The contemporary accounts of the capture of Rome by the Gauls, confirmed by authentic traditions, place this event upon a solid historical basis; but it is difficult to judge how far the circumstantial narrative is deserving of belief. With respect to the bare fact that Rome was taken by the Gauls, there is for the first time in Roman History the testimony of contemporary Greek writers.'-' Credibility of Roman History,' Vol. II., pp. 354-359.

attributed

attributed to the root 'magh' in the passage which we have just quoted; and the name of 'Magabula' is doubtful altogether.

Another oversight of a different kind occurs at page 255, in which we are told that 'even so late as the time of Henry II. Hereford was considered to be in Wales.' But Offa's Dyke, as Mr. Taylor knows, passed over the ridge of hills between Herefordshire and Radnorshire; and Domesday Book, in speaking of the customs of Archenfield, which was a sort of march district, or debateable land, in Herefordshire, expressly says, 'Similiter emendat qui jussus a vicecomite secum ire in Walis, non pergit' -that is, the man who is summoned by the Sheriff to go with him into Wales, and does not go, pays the same fine.' Again, in speaking of the City of Hereford, it is stated, 'Si vicecomes iret in Wales cum exercitu ibant hi homines cum eo.' Surely these expressions are sufficient to show that long before the time of Henry II., Hereford was not reckoned to be in Wales.

6

Still the general conclusions of our author with regard to England stand uncontested. He says: "Over the whole land almost every river-name is Celtic, most of the Shire-names contain Celtic roots, and a fair sprinkling of names of hills, valleys, and fortresses, bears witness that the Celt was the aboriginal possessor of the soil' (p. 256).

It appears that the Cymry held the lowlands of Scotland as far as the hills of Perthshire, and were probably the people called 'Picts,' whilst the 'Scots' were the Gaelic tribes who immigrated from Ireland. The test-words for establishing the limit between these two races of the same family are 'Pen' and 'Aber' on the one side, and 'Ben' and 'Inver' on the other: Aber' is the Cymric word for the mouth of a river, which meets us constantly in Brittany and in Wales; Inver' is the Gaelic form, and is accordingly common in Ireland and in the Highlands. If,' says Mr. Taylor, we draw a line across the map from a point a little south of Inverary to one a little north of Aberdeen, we shall find that (with very few exceptions) the "invers" lie to the north of the line, and the "abers" to the south of it' (p. 259). Substituting 'north-west' and 'south-east' for 'north' and 'south,' this line, as he observes, coincides very nearly with the present limit of the Gaelic tongue. It appears, however, that the Gaels encroached on the Picts or Cymry, and that a few 'invers' have intruded themselves even on the Forth; as, for instance, Inveresk near Edinburgh, Inverkeithing in Fife, Inverbervie in Kincardine. The following passage is worth quoting, as showing the application of our author's principles on a small scale: :

* Livy, xxxviii., 19. Compare Smith's' Dictionary of Ancient Geography,' in v.

The

The ethnology of the Isle of Man may be very completely illustrated by means of local names. The map of the island contains about 400 names, of which about 20 per cent. are English, 21 per cent. are Norwegian, and 59 per cent. are Gaelic. These Celtic names are all of the most characteristic Erse type. It would appear that not a single colonist from Wales ever reached the island, which, from the mountains of Carnarvon, is seen like a faint blue cloud upon the water. There are ninety-six names beginning with "balla," and the names of more than a dozen of the highest mountains have the prefix "Slieu" answering to the Irish "Slievh or "Sliabh." The Isle of Man has the "Curragh," the "Loughs," and the "Allens" of Ireland faithfully reproduced. It is curious to observe that the names which denote places of Christian worship are all Norwegian; they are an indication of the late date at which heathenism must have prevailed.' -p. 260.

Thus the groundwork of the English local nomenclature is Celtic. The names of the rivers and of many of the hills show the clearest traces of the people who held the land when the Romans landed on it. The Romans occupied the country as conquerors, and their colonization was founded on military principles, as ours was in the Highlands, and will probably now have to be in New Zealand. Ardwick le Street in Yorkshire, Chester le Street in Durham, Stretton, Stratton, Streatham, Streatley, and several places called Stretford or Stratford, all tell us of their proximity to a Roman road. Portway,' which name is applied to nine places in the kingdom, is in like manner connected with these military highways. The name Cold Harbour' is said to occur no less than seventy times in the neighbourhood of the ancient lines of road, and it seems to have signified a ruined house or station, where travellers could find shelter and nothing else.* Many important towns are marked by the suffix 'ford,' because at these spots the fords occurred in the great roads, as Oxford' (in Welsh, Rytychen, Rhedycina) very few comparatively have the termination 'bridge. A certain number

The proper sense of harbour' as it occurs in this name is best illustrated by a quotation from the Morte d'Arthur' (cap. xxv.): At last they came unto a courtelage, and there they asked herborow, but the man of the courtelage wold not lodge them.' The question of the origin of this name Cold Harbour' was started in the last century, and an answer on the subject will be found in the 'Gentleman's Magazine' for 1793, 4th July, part ii., p. 603, in which reference is made to a place called Kalte Herberge,' on the road between Strasburg and Basle. There is also a small town called 'Kalter Herberg' close to the Belgian frontier in Rhenish Prussia, 18 miles s.E. of Aix la Chapelle, near the town of Montjoie. The existence of one such name in another European country is sufficient to relieve us from any mysterious sense of the word cold,' such as is suggested in the letter of Captain Smyth to Sir Henry Ellis (Archæologia,' 1849, p. 125), in which it is considered that it may possibly be derived from coluber,' and retain a trace of the ancient Ophite worship!

Vol. 116.-No. 231.