1 This was written at the end of 1892, and read to a joint meeting of the Cymmrodorion and Folk-Lore Societies on January 11, 1893. ↑

2 Some account of them was given by me in Folk-Lore for 1892, p. 380; but somehow or other my contribution was printed unrevised, with results more peculiar than edifying. ↑

3 In Folk-Lore for 1893, pp. 58–9. ↑

4 In the neighbourhood I find that the word gwaeldyn in this verse is sometimes explained to mean not a worthless but an ailing person, on the strength of the fact that the adjective gwael is colloquially used both for vile and for ailing. ↑

5 Since writing the above remarks the following paragraph, purporting to be copied from the Liverpool Mercury for November 18, 1896, appeared in the Archæologia Cambrensis for 1899, p. 334:—‘Two new fishes have just been put in the “Sacred Well,” Ffynnon y Sant, at Tyn y Ffynnon, in the village of Nant Peris, Ỻanberis. Invalids in large numbers came, during the last century and the first half of the present century, to this well to drink of its “miraculous waters”; and the oak box, where the contributions of those who visited the spot were kept, is still in its place at the side of the well. There have always been two “sacred fishes” in this well; and there is a tradition in the village to the effect that if one of the Tyn y Ffynnon fishes came out of its hiding-place when an invalid took some of the water for drinking or for bathing purposes, cure was certain; but if the fishes remained in their den, the water would do those who took it no good. Two fishes only are to be put in the well at a time, and they generally live in its waters for about half a century. If one dies before the other, it would be of no use to put in a new fish, for the old fish would not associate with it, and it would die. The experiment has been tried. The last of the two fishes put in the well about fifty years ago died last August. It had been blind for some time previous to its death. When taken out of the water it measured seventeen inches, and was buried in the garden adjoining the well. It is stated in a document of the year 1776 that the parish clerk was to receive the money put in the box of the well by visitors. This money, together with the amount of 6s. 4d., was his annual stipend.’ Tyn y Ffynnon means ‘the Tenement of the Well,’ tyn being a shortened form of tyđyn, ‘a tenement,’as mentioned at p. 33 above; but the mapsters make it into ty’n = ty yn, ‘a house in,’ so that the present instance, Ty’n y Ffynnon, could only mean ‘the House in the Well,’ which, needless to say, it is not. But one would like to know whether the house and land were once held rent-free on condition that the tenant took care of the sacred fish. ↑

6 See Ashton’s Iolo Goch, p. 234, and Lewis’ Top. Dict. ↑

7 See my Hibbert Lectures, p. 229, and the Iolo MSS., pp. 42–3, 420–1. ↑

8 A curious note bearing on this name occurs in the Jesus College MS. 20 (Cymmrodor, viii. p. 86) in reference to the name Morgannwg, ‘Glamorgan’:—O enỽ Morgant vchot y gelwir Morgannỽc. Ereiỻ a dyweit. Mae o enỽ Mochteyrn Predein. ‘It is from the name of the above Morgan that Morgannwg is called. Others say that it is from the name of the mochdeyrn of Pictland.’ The mochteyrn must have been a Pictish king or mórmáer called Morgan. The name occurs in the charters from the Book of Deer in Stokes’ Goidelica. pp. 109, 111, as Morcunt, Morcunn, and Morgunn undeclined, also with Morgainn for genitive; and so in Skene’s Chronicles of the Picts and Scots, pp. 77, 317, where it is printed Morgaind; see also Stokes’ Tigernach, in the Revue Celtique, xvii. 198. Compare Geoffrey’s story, ii. 15, which introduces a northern Marganus to account for the name Margan, now Margam, in Morgannwg. ↑

9 M. Loth’s remarks in point will be found in the Revue Celtique, xiii. 496–7, where he compares with tut the Breton teuz, ‘lutin, génie malfaisant ou bienfaisant’; and for the successive guesses on the subject of the name Morgan tut one should also consult Zimmer’s remarks in Foerster’s Introduction to his Erec, pp. xxvii–xxxi, and my Arthurian Legend, p. 391, to which I should add a reference to the Book of Ballymote, fo. 360^a, where we have o na bantuathaib, which O’Curry has rendered ‘on the part of their Witches’ in his Manners and Customs of the Ancient Irish, iii. 526–7. Compare dá bhantuathaigh, ‘two female sorcerers,’ in Joyce’s Keating’s History of Ireland, pp. 122–3. ↑

10 For all about the Children of Lir, and about Liban and Lough Neagh, see Joyce’s Old Celtic Romances, pp. 4–36, 97–105. ↑

11 On my appealing to Cadrawd, one of the later editors, he has found me the exact reference, to wit, volume ix of the Cyfaiỻ (published in 1889), p. 50; and he has since contributed a translation of the story to the columns of the South Wales Daily News for February 15, 1899, where he has also given an account of Crymlyn, which is to be mentioned later. ↑

12 Judging from the three best-known instances, y bala meant the outlet of a lake: I allude to this Bala at the outlet of Ỻyn Tegid; Pont y Bala, ‘the Bridge of the bala,’ across the water flowing from the Upper into the Lower Lake at Ỻanberis; and Bala Deulyn, ‘the bala of two lakes,’ at Nantỻe. Two places called Bryn y Bala are mentioned s. v. Bala in Morris’ Celtic Remains, one near Aberystwyth, at a spot which I have never seen, and the other near the lower end of the Lower Lake of Ỻanberis, as to which it has been suggested to me that it is an error for Bryn y Bela. It is needless to say that bala has nothing to do with the Anglo-Irish bally, of such names as Ballymurphy or Ballynahunt: this vocable is in English bailey, and in South Wales beili, ‘a farm yard or enclosure,’ all three probably from the late Latin balium or ballium, ‘locus palis munitus et circumseptus.’ Our etymologists never stop short with bally: they go as far as Balaklava and, probably, Ballarat, to claim cognates for our Bala. ↑

13 Cadrawd here gives the Welsh as ‘2 bladur … 2 đyđ o wair,’ and observes that the lacuna consists of an illegible word of three letters. If that word was either sef, ‘that is,’ or neu, ‘or,’ the sense would be as given above. In North Cardiganshire we speak of a day’s mowing as gwaith gwr, ‘a man’s work for a day,’ and sometimes of a gwaith gwr bach, ‘a man’s work for a short day.’ ↑

14 See By-Gones for May 24, 1899. The full name of Welshpool in Welsh is Traỻwng Ỻywelyn, so called after a Ỻywelyn descended from Cuneđa, and supposed to have established a religious house there; for there are other Traỻwngs, and at first sight it would seem as if Traỻwng had something to do with a lake or piece of water. But there is a Traỻwng, for instance, near Brecon, where there is no lake to give it the name; and my attention has been called to Thos. Richards’ Welsh-English Dictionary, where a traỻwng is said to be ‘such a soft place on the road (or elsewhere) as travellers may be apt to sink into, a dirty pool.’ So the word seems to be partly of the same derivation as go-ỻwng, ‘to let go, to give way.’ The form of the word in use now is Traỻwm, not Traỻwng or Traỻwn. ↑

15 See the Book of the Dun Cow, fo. 39^a–41^b and Joyce’s Old Celtic Romances, pp. 97–105; but the story may now be consulted in O’Grady’s Silva Gadelica, i. 233–7, translated in ii. 265–9. On turning over the leaves of this great collection of Irish lore, I chanced, i. 174, ii. 196, on an allusion to a well which, when uncovered, was about to drown the whole locality but for a miracle performed by St. Patrick to arrest the flow of its waters. A similar story of a well bursting and forming Lough Reagh, in County Galway, will be found told in verse in the Book of Leinster; fo. 202^b: see also fo. 170^a, and the editor’s notes, pp. 45, 53. ↑

16 See Evans’ autotype edition of the Black Book of Carmarthen, fos. 53^b, 54^a, also 32^a: the punctuation is that of the MS. In the seventh triplet kedaul is written k^eadaul, which seems to mean kadaul corrected into kedaul; but the a is not deleted, so other readings are possible. ↑

17 In the Iolo MSS., p. 89, finaun wenestir is made into Ffynon-Wenestr and said to be one of the ornamental epithets of the sea; but I am convinced that it should be rather treated as ffynnon fenestr with wenestir or fenestr mutated from menestr, which meant a servant, attendant, cup-bearer: for one or two instances see Pughe’s Dictionary. The word is probably, as suggested by M. Loth in his Mots Latins, p. 186. the old French menestre, ‘cup-bearer,’ borrowed. Compare the mention of Nechtán’s men having access to the secret well in Sid Nechtáin, p. 390 below, and note that they were his three menestres or cup-bearers. ↑

18 See the Cymmrodor, viii. 88 (No. xxix), where a Marereda is mentioned as a daughter of Madog son of Meredyđ brother to Rhys Gryg. ↑

19 There is another reading which would make them into Segantii, and render it irrelevant—to say the least of it—to mention them here. ↑

20 See the Mabinogion, p. 35: the passage has been mistranslated in Lady Charlotte Guest’s Mabinogion, iii. 117. ↑

21 See my Arthurian Legend, pp. 263–4. ↑

22 I do not profess to see my way through the difficulties which the probable etymological connexion between the names Setantii, Setanta, Seithyn, and Seithennin implies. But parts of the following string of guesses may be found to hold good:—Seithyn is probably more correct than Seithin, as it rhymes with cristin = Cristyn (in Cristynogaeth: see Silvan Evans’ Geiriadur, s. v., and Skene’s Four Ancient Books, ii. 210); and it might be assumed to be from the same stem as Seizun; but, supposing it to represent an earlier Seithynt, it would equate phonologically with Setanta, better Setinte, of which the genitive Setinti actually occurs, as a river name, in the Book of the Dun Cow, fo. 125^b: see my Hibbert Lectures, p. 455, and see also the Revue Celtique, xi. 457. It would mean some such an early form Setn̥ti̯o-s, and Seithenhin, another derivative from the same stem, Setn̥tīno-s. But the retention of n before t in Setinte proves it not to be unconnected with Seithyn, but borrowed from some Brythonic dialect when the latter was pronounced Seithn̥ti̯o-s. If this be anywhere nearly right one has to assume that the manuscripts of Ptolemy giving the genitive plural as Σεταντίων or Σεγαντίων should have read Σεκταντίων, unless one should rather conjecture Σεγταντίων with cht represented by gt as in Ogams in Pembrokeshire: witness Ogtene and Maqui Quegte. This conjecture as to the original reading would suggest that the name was derived from the seventh numeral sechtn̥, just as that of the Galloway people of the Novantæ seems to be from the ninth numeral. Ptolemy’s next entry to the Harbour of the Setantii is the estuary of the Belisama, supposed to be the Mersey; and next comes the estuary of the Σετεία or Σεγεία, supposed to be the Dee. Now the country of the Setantii, when they had a country, may have reached from their harbour near the mouth of the Ribble to the Seteia or the Dee without the name Seteia or Segeia having anything to do with their own, except that it may have influenced the latter in the manuscripts of Ptolemy’s text. Then we possibly have a representative of Seteia or Segeia in the Saidi or Seidi, sometimes appended to Seithyn’s name. In that case Seithyn Saidi, in the late Triad iii. 37, would mean Seithyn of Seteia, or the Dee. A Mab Saidi occurs in the Kulhwch story (Mabinogion, p. 106), also Cas, son of Saidi (ib. 110); and in Rhonabwy’s Dream Kadyrieith, son of Saidi (ib. 160); but the latter vocable is Seidi in Triad ii. 26 (ib. 303). It is to be borne in mind that Ptolemy does not represent the Setantii as a people in his time: he only mentions a harbour called after the Setantii. So it looks as if they then belonged to the past—that in fact they were, as I should put it, a Goidelic people who had been conquered and partly expelled by Brythonic tribes, to wit, by the Brigantes, and also by the Cornavii in case the Setantii had once extended southwards to the Dee. This naturally leads one to think that some of them escaped to places on the coast, such as Dyfed, and that some made for the opposite coast of Ireland, and that, by the time when the Cúchulainn stories came to be edited as we have them, the people in question were known to the redactors of those stories only by the Brythonic form of their name, which underlies that of Setanta Beg, or the Little Setantian. Those of them who found a home on the coast of Cardigan Bay may have brought with them a version of the inundation story with Seithennin, son of Seithyn, as the principal figure in it. So in due time he had to be attached to some royal family, and in the Iolo MSS., pp. 141–2, he is made to descend from a certain Plaws Hen, king of Dyfed, while the saints named as his descendants seem to have belonged chiefly to Gwyneđ and Powys. ↑

23 See the Professor’s Address on the Place of a University in the History of Wales, delivered at Bangor at the opening ceremony of the Session of 1899–1900 (Bangor, 1900), p. 6. The reference to Giraldus is to his Itin. Kambriæ, i. 13 (p. 100), and the Expugnatio Hibernica, i. 36 (p. 284). ↑

24 Instead of ‘she followed it’ one would have expected ‘it followed her’; but the style is very loose and rough. ↑

25 As a ‘Cardy’ I have here two grievances, one against my Northwalian fellow countrymen, that they insist on writing Rheidiol out of sheer weakness for the semivowel i̯; and the other against the compilers of school books on geography, who give the lake away to the Wye or the Severn. I am told that this does not matter, as our geographers are notoriously accurate about Natal and other distant lands; so I ought to rest satisfied. ↑

26 Professor Meyer has given a number of extracts concerning her in his notes to his edition of The Vision of Mac Conglinne (London, 1892), pp. 131–4, 208–10, and recently he has published The Song of the Old Woman of Beare in the Otia Merseiana (London, 1899), pp. 119–28, from the Trinity College codex, H. 3, 18, where we are told, among other things, that her name was Digdi, and that she belonged to Corcaguiny. The name Béara, or Bérre, would seem to suggest identification with that of Bera, daughter of Eibhear, king of Spain, and wife of Eoghan Taidhleach, in the late story of The Courtship of Moméra, edited by O’Curry in his Battle of Magh Leana (Dublin, 1855); but the other name Digdi would seem to stand in the way. However none of the literature in point has yet been discovered in any really old manuscript, and it may be that the place-name Berre, in Caillech Bérri, has usurped the place of the personal name Béra, whose antiquity in some such a form as Béra or Méra is proved by its honorific form Mo-mera: see O’Curry’s volume, p. 166, and his Introduction, p. xx. ↑