detail, with special reference to the names mentioned;

and let us begin with that of Twrch Trwyth: the word twrch means the male of a beast of the swine kind, and twrch coed, 'a wood pig,' is a wild boar, while twrch daear, an earth pig,' is the word in North Wales for a mole. In the next place we can practically equate Twrch Trwyth with a name at the head of one of the articles in Cormac's Irish Glossary. There the exact form is Orc tréith, and the following is the first part of the article itself as given in O'Donovan's translation edited by Stokes:-'Orc Tréith, i. e. nomen for a king's son, triath enim rex vocatur, unde dixit poeta Oinach n-uirc tréith “fair of a king's son," i. e. food and precious raiment, down and quilts, ale and flesh-meat, chessmen and chessboards, horses and chariots, greyhounds and playthings besides.' In this extract the word orc occurs in the genitive as uirc, and it means a 'pig' or 'boar'; in fact it is, with the usual Celtic loss of the consonant p, the exact Goidelic equivalent of the Latin porcus, genitive porci. From another article in Cormac's Glossary, we learn that Treith is the genitive of Triath, which has been explained to mean a king. Thus, Orc Tréith means Triath's Orc, Triath's Boar, or the King's Boar; so we take Twrch Trwyth in the same way to mean 'Trwyth's Boar.' But we have here a discrepancy, which the reader will have noticed, for twrch is not the same word as Irish orc, the nearest form to be expected in Welsh being Wrch, not Twrch; but such a word as Wrch does not, so far as I know, exist. Now did the Welsh render orc by a different word unrelated to the Goidelic one which they heard? I think not; for it is remarkable that Irish has besides orc a word torc, meaning a 'boar,' and torc is exactly the Welsh twrch. So there seems to be no objection to our supposing that what Cormac calls Orc Tréith was

known in the Goidelic of Wales as Torc Tréith, which had the alliteration to recommend it to popular favour. In that case one could say that the Goidelic name Torc Tréith appears in Welsh with a minimum of change as Twrch Trwyth, and also with the stamp of popular favour more especially in the retention of the Goidelic th, just as in the name of an ancient camp or fortification on the Withy Bush Estate in Pembrokeshire: it is called the Rath, or the Rath Ring. Here rath is identical with the Irish word ráth, 'a fortification or earthworks,' and we seem to have it also in Cil Râth Fawr, the name of a farm in the neighbourhood of Narberth. Now the Goidelic word tréith appears to have come into Welsh as trēth-i, the long vowel of which must in Welsh have become oi or ui by about the end of the sixth century; and if the th had been treated on etymological principles its proper equivalent in the Welsh of that time would have been d or t. The retention of the th is a proof, therefore, of oral transmission; that is to say, the Goidelic word passed bodily into Brythonic, to submit afterwards to the phonological rules of that language.

A little scrutiny of the tale will, I think, convince the reader that one of the objects of the original story-teller was to account for certain place-names. Thus Grugyn was meant to account for the name of Garth Grugyn, where Grugyn was killed; Gwys, to account similarly for that of Gwys, a tributary of the Twrch, which gives its name to a station on the line of railway between Ystalyfera and Bryn Amman; and Twrch ILawin to account for the name of the river Twrch, which receives the Gwys, and falls into the Tawe some distance below Ystrad Gynlais, between the counties of Brecknock and Glamorgan.

Besides Grugyn and Twrch ILawin, there was a third

brother to whom the story gives a special name, to wit, ILwydawc Gouynnyat, and this was, I take it, meant also to account for a place-name, which, however, is not given it should have been somewhere in Ystrad Yw, in the county of Brecknock. Still greater interest attaches to the swine that have not been favoured with names of their own, those referred to simply as banw, 'a young boar,' and benwic, 'a young sow.' Now banw has its equivalent in Irish in the word banbh, which O'Reilly explains as meaning a 'sucking pig,' and that is the meaning also of the Manx bannoo; but formerly the word may have had a somewhat wider meaning. The Welsh appellative is introduced twice into the story of Twrch Trwyth; once to account, as I take it, for the name Mynyd Amanw, 'Amman Mountain,' and once for Dyffryn Amanw, 'Amman Valley.' In both instances Amanw was meant, as I think, to be accounted for by the banw killed at each of the places in question. But how, you will ask, does the word banw account for Amanw, or throw any light on it at all? Very simply, if you will just suppose the name to have been Goidelic; for then you have only to provide it with the definite article and it makes in banbh, 'the pig or the boar,' and that could not in Welsh yield anything but ymmanw or ammanw1, which with the accent shifted backwards, became Ammanw and Amman or Aman.

Having premised these explanations let us, before we proceed further, see to what our evidence exactly Here, then, we have a mention of seven

amounts.

1 In some native Welsh words we have an option between a prefix ym and am, an option arising out of the fact that originally it was neither ym nor am, but m, for an earlier mbi, of the same origin as Latin ambi and Greek ȧupi, 'around, about.' The article, its meaning in the combination in banbh being forgotten, would fall under the influence of the analogy of the prefix, now am or ym, so far as the pronunciation was concerned.

swine, but as two of them, a banw and a benwic, are killed at one and the same place, our figure is practically reduced to six 1. The question then is, in how many of these six cases the story of the hunt accounts for the names of the places of the deaths respectively, that is to say, accounts for them in the ordinary way with which one is familiar in other Welsh stories. They may be enumerated as follows:

1. A banw is killed at Mynyd Amanw.

2. A twrch is killed in the same neighbourhood, where there is a river Twrch.

3. A swine called Gwys is killed in the same neighbourhood still, where there is a river called Gwys, falling into the Twrch.

4. A banw and a benwic are killed in Dyffryn Amanw. 5. Grugyn is killed at a place called Garth Grugyn. 6. A swine called ILwydawc is killed at a spot, not named, in Ystrad Yw or not far off 2.

Thus in five cases out of the six, the story accounts for the place-name, and the question now is, can that be a mere accident? Just think what the probabilities of the case would be if you put them into numbers: South Wales, from St. David's to the Vale of the Usk, would supply hundreds of place-names as deserving of mention, to say the least, as those in this story; is it likely then that out of a given six among them no less than five should be accounted for or alluded to by any mere accident in the course of a story of the brevity of that of

1

1 Possibly the benwic was thrown in to correct the reckoning when the redactor discovered, as he thought, that he had one too many to account for it has been pointed out that he had forgotten that one had been killed in Ireland.

2 It is just possible, however, that in an older version it was named, and that the place was no other than the rock just above Ystrad Yw, called Craig Lwyd or, as it is said to be pronounced, Craig ILwyd. If so, wyd would seem to have been substituted for the dissyllable Lwydog: compare the same person called ILwyt and ILwydeu in the Mabinogion, pp. 57, 110, 136.

Twrch Trwyth. To my thinking such an accident is inconceivable, and I am forced, therefore, to suppose that the narrative was originally so designed as to account for them. I said 'originally so designed,' for the scribe of the Red Book, or let us say the last redactor of the story as it stands in the Red Book, shows no signs of having noticed any such design. Had he detected the play on the names of the places introduced, he would probably have been more inclined to develop that feature of the story than to efface it.

What I mean may best be illustrated by another swine story, namely, that which has already been referred to as occurring in the Mabinogi of Math. There we find Pryderi, king of Dyfed, holding his court at Rhudlan on the Teifi, but though he had become the proud possessor of a new race of animals, given him as a present by his friend Arawn, king of Annwn, he had made a solemn promise to his people, that he should give none of them away until they had doubled their number in Dyfed: these animals were the hobeu or pigs to which reference was made at p. 69 above. Now Gwydion, having heard of them, visited Pryderi's court, and by magic and enchantment deceived the king. Successful in his quest, he sets out for Gwyned with his hobeu, and this is how his journey is described in the Mabinogi: 'And that evening they journeyed as far as the upper end of Keredigion, to a place which is still called, for that reason, Mochdref, "Swine-town or Pigs' stead." On the morrow they went their way, and came across the Elenyd mountains, and that night they spent between Kerry and Arwystli, in the stead which is also called for that reason Mochdref. Thence they proceeded, and came the same evening as far as a commot in Powys, which is for that reason called