[11] As it stands now this would be unmutated Césel Gýfarch, ‘Cyfarch’s Nook,’ but there never was such a name. There was, however, Elgýfarch or Aelgýfarch and Rhygýfarch, and in such a combination as Césel Elgýfarch there would be every temptation to drop one unaccented el. [↑]
[12] Owing to some oversight he has ‘a clean or a dirty cow’ instead of cow-yard or cow-house, as I understand it. [↑]
[13] Cwta makes cota in the feminine in North Cardiganshire; the word is nevertheless only the English cutty borrowed. Du, ‘black,’ has corresponding to it in Irish, dubh. So the Welsh word seems to have passed through the stages dyv, dyw, before yw was contracted into û, which was formerly pronounced like French û, as proved by the grammar already mentioned (p. 22) of J. D. Rhys, published in London in 1592; see p. 33, to which my attention has been called by Prof. J. Morris Jones. In Old or pre-Norman Welsh m did duty for m and v, so one detects dyv as dim in a woman’s name Penardim, ‘she of the very black head’; there was also a Penarwen, ‘she of the very blonde head.’ The look of Penardim having baffled the redactor of the Branwen, he left the spelling unchanged: see the (Oxford) Mabinogion, p. 26. The same sort of change which produced du has produced cnu, ‘a fleece,’ as compared with cneifio, ‘to fleece’; ỻuarth, ‘a kitchen garden,’ as compared with its Irish equivalent lubhghort. Compare also Rhiwabon, locally pronounced Rhuabon, and Rhiwaỻon, occurring sometimes as Rhuaỻon. But the most notable rôle of this phonetic process is exemplified by the verbal nouns ending in u, such as caru, ‘to love,’ credu, ‘to believe,’ tyngu, ‘to swear,’ in which the u corresponds to an m termination in Old Irish, as in sechem, ‘to follow,’ cretem, ‘belief,’ sessam or sessom, ‘to stand.’ [↑]
[14] In medieval Welsh poetry this name was still a dissyllable; but now it is pronounced Ỻŷn, in conformity with the habit of the Gwyndodeg, which makes into porfŷđ what is written porfeyđ, ‘pastures,’ and pronounced porféiđ in North Cardiganshire. So in the Ỻeyn name Sarn Fyỻteyrn the second vocable represents Maelteyrn, in the Record of Carnarvon (p. 38) Mayltern̄: it is now sounded Myỻtyrn with the second y short and accented. Ỻeyn is a plural of the people (genitive Ỻaën in Porth Dinỻaën), used as a singular of their country, like Cymru = Cymry, and Prydyn. The singular is ỻain, ‘a spear,’ in the Book of Aneurin: see Skene, ii. 64, 88, 92. [↑]
[15] It is also called dolur byr, or the ‘short disease’; I believe I have been told that it is the disease known to ‘the vet.’ as anthrax. [↑]
[16] Here the writer seems to have been puzzled by the mh of Amheirchion, and to have argued back to a radical form Parch; but he was on the wrong tack—Amheirchion comes from Ap-Meirchion, where the p helped to make the m a surd, which, with the syllabic accent on the succeeding vowel, became fixed as mh, while the p disappeared by assimilation. We have, later on, a similar instance in Owen y Mhaxen for Owen Amhacsen = O. ap Macsen. Another instance will be found at the opening of the Mabinogi of Branwen, to wit, in the word prynhawngweith, ‘once on an afternoon,’ from prynhawn, ‘afternoon,’ for which our dictionaries substitute prydnawn, with the accent on the ultima, though D. ab Gwilym used pyrnhawn, as in poem xl. 30. But the ordinary pronunciation continues to be prynháwn or pyrnháwn, sometimes reduced in Gwyneđ to pnawn. Let me add an instance which has reached me since writing the above: In the Archæologia Cambrensis for 1899, pp. 325–6, we have the pedigree of the Ameridiths from the Visitation of Devonshire in 1620: in the course of it one finds that Iuan ap Merydeth has a son Thomas Amerideth, who, knowing probably no Welsh, took to writing his patronymic more nearly as it was pronounced. The line is brought down to Ames Amerideth, who was created baronet in 1639. Amerideth of course = Ap Meredyđ, and the present member of the family who writes to the Archæologia Cambrensis spells his patronymic more correctly, Ameridith; but if it had survived in Wales it might have been Amheredyđ. For an older instance than any of these see the Book of Taliessin, poem xlix (= Skene, ii. 204), where one reads of Beli Amhanogan, ‘B. ab Mynogan.’ [↑]
[17] This is pronounced Rhiwan, though probably made up of Rhiw-wen, for it is the tendency of the Gwyndodeg to convert e and ai of the unaccented ultima into a, and so with e in Glamorgan; see such instances as Cornwan and casag, p. 29 above. It is possibly a tendency inherited from Goidelic, as Irish is found to proceed in the same way. [↑]
[18] I may mention that some of the Francises of Anglesey are supposed to be descendants of Frazers, who changed their name on finding refuge in the island in the time of the troubles which brought there the ancestor of the Frazer who, from time to time, claims to be the rightful head of the Lovat family. [↑]
[19] According to old Welsh orthography this would be written Moudin, and in the book Welsh of the present day it would have to become Meuđin. Restored, however, to the level of Gallo-Roman names, it would be Mogodunum or Magodunum. The place is known as Casteỻ Moeđin, and includes within it the end of a hill about halfway between Ỻannarth and Lampeter. [↑]
[20] For other mentions of the colours of fairy dress see pp. 44, 139 above, where red prevails, and contrast the Lake Lady of Ỻyn Barfog clad in green, p. 145. [↑]