in its older form chwaer. Chwa-er from *sve(s)er- is a third possibility. Chwaer had its old *ves treated like gwaeanwyn, O.-W. guiannuin, containing *vesant-. Though guiannuin1 (in Merioneth gwinwyn, Rhys') and chwiorydd exist, we have gwaeanwyn and chwaer, containing "wa, "va, from *vo for *ve; but it is not clear for what reason the syllable *ves became either *vi(s) or *vo(s), wa(s), although it is true the next syllable contains different vowels in both cases; chwaer from *svo(s)er, *sve(s)er, chwior- from *svi(s)or-, *sve(s)or-. Chwiorydd as an isolated form was altered by analogy; cf. chwaeored, MS. Cl. B 5, f. 249a, col. 2 (Dares Phrygius); brodur a chwayorydd, Add. MS. 14,987, f. 35b (Araith y Trwstan). In The Red Dragon, ii, p. 420, chwaeriorydd is mentioned. This form is actually printed in Llyfr Achau (Her. Vis., ii, p. 12) as chwaerorydd. Pughe gives chweirydd, probably to be considered like meusydd, pl. of maes.

[8.] The ending -awr, later -or, is almost completely lost in later Welsh; it is frequent with certain nouns, as llafn, byddin, ysgwydd, gwaew, etc., in the older poems, those printed by Skene and in the Myv. Arch., but it hardly ever occurs in Middle W. prose texts (cf., however, yr ieuawr, Ll. Gw. Rh., p. 7). From O.-W. cf. poulloraur (*pugillār- + -ār-), M. Cap.3 In Breton this ending is still common as -er, -ier; cf. er méneïer in a Tréc. poem, Rev. de Bret., 1st ser., v, p. 408 (meneïo, 4, p. 170), etc., in W. mynyddoedd. The Welsh ending is mentioned in Dosp. Ed., § 489, omitted by Zeuss, but exemplified in E. Evans' Stud. in Cymr. Phil., § 12.4 On its Irish equivalents opinions have been advanced in Three Mid.-Ir. Hom., p. 135, and Togail Troi (LL.), Introd.

It is possible that or has survived in certain plurals for which a new singular was made by means of -yn or -en-in

1 Ovid Glosses, 40b; Stokes, Cambrica, p. 236.

2 Lect.2, p. 27.

3 Capella Glosses, No. 28 (Arch. Camb., 4th Ser., iv, p. 7). 4 lb., p. 141.

VOL. IX.

T

marwor, marwar, marfor, 'embers', for instance-but as a distinct plural termination it occurs in later Welsh only in gwaew, gwaewawr, whence gwaewor, gwaewar, like gwatwor, gwatwar, marwar, chwaer, etc.

Cf. gбayawaбr (sic), B. of Tal., xliv (Sk., ii, 199); gwaywar, Ll. Gw. Rh., pp. 59, 67, 69; g6aewar, B. of Herg., cols. 698 (R. B. Mab., p. 82), 1093, etc.

Even here it is replaced by -yr; cf. gostбng g6aewyr, B. of Herg., col. 1093; y gweiwyr, Wms., Hgt. MSS., ii, p. 304; gwewyr, Y S. Gr., p. 230 (gwaywyr et vulgo gwewyr, Davies, Dict.). This -yr can only be the ending -yr (-r) of brodyr, pl. of brawd, analogically transferred. Besides brodyr, broder less often occurs, rhyming with full syllables in -er (passives, etc.) in poets, as in L. Glyn Cothi, pp. 42-3 (see p. 269, infra), etc.

[9.] Tai and lloi are older S.-W. plurals, commonly used in S. W., but replaced by teiau, lloiau in N. W. Cf. Hughes, 1823, p. 29; Y Traeth., iii, p. 9; D. S. Evans, Llyth., s. vv.; etc.

Ll. Gw. Rh., pp. 94, 97, 110: Dagr-eu-oed; Add. MS. 19,709, f. 58b, etc.: blod-eu-oed, llys-eu-oed. Llysieuoedd, Cl. B 5, f. 1026. These forms should be compared with Breton ones like bot-o-ier, bot-o-io, Ernault, De l'urgence, etc., p. 14; the additional plural-ending seems to convey a more collective

sense.

Cf. also or dyбededigyon llysseue hynny, Add. MS. 14,912, f. 77a; o oll dy6ededigyon llyssy oedd hynny, f. 79 (y for eu, S.-W.); llysewun, llisiav, Add. MS. 15,049 (17th cent.), ff. 4a, 20a; llissewyn, B. of Herg., col. 436. In Neath, lly sewyn, pl. llysa (ib., blotyn, pl. blota).

On eu ew, cf. g6neuthur creu yr moch, B. of Herg., col. 754; y creu, col. 766 (R. B. Mab., p. 63, 1. 3; 78, 11. 8, 11, 16): crewyn; gieu giewyn; ceneu: cenawon, etc.

On the words meaning 'day' see Rhŷs, Hibb. Lect., pp. 116-118. Cf. die6-ed, Add. MSS. 19,709, f. 25b; 22,356, f. 12a; diewoed, Cl. B 5, f. 235b, col. 1 ; die-oed, f. 244b, col. 2; dieuoedd in Davies' Dict.; ew: eu, as in the words above given.

A plural ending is also often incorporated in singulatives formed from English loan-words; cf. sklait-s, sing. sklǝit-s-an ('slate'), tatws, sing. täsan, etc., Sweet, p. 437. On Breton analogies see Rev. Celt., vi, pp. 388-9.

It must not be assumed that the suffixes -en and -an are of the same phonetic value in all words which exhibit them in the present state of the language. In a few such words -aen is the older form of the suffix, as is confirmed by other Brythonic languages. So agalen, ogalen : agalayn, 'cos', MS. Vesp. E 11 of the Latin Laws (Owen, p. 851). Cf. O.Corn. ocoluin, gl. 'cos' (Stokes, Cambr., p. 241); llamhystaen (=llymysten); see Zeuss, G. C.2, p. 291. Cf. also croessaeniait, MS. L, p. 182 (= croesan, -iaid); maharaen, MS. A, p. 135, pl. meheryn, p. 160; maharen, MS. K, pl. meherein, MSS. C, D, B, E; in the Latin Laws: maharayn, Hgt. MS., p. 791; maharain, Harl. MS., p. 862; maharaen, MSS. Calig. A 13, f. 1836, Tit. D 2, f. 53b. Cf. O.-W. maharuin (bis), B. of St. Chad, pp. 18, 19.1 (Lib. Land., pp. 272-3, nos. [3]

and [4])

This reduction of the final unstressed syllable may be compared to that in gallel: gallael, cafel: cafael, gadel: gadael, etc., but the details of the process are not all clear. Is maharuin an error for *maharain, or perhaps (cf. ocoluin) the representative of a later *maharwyn, like mollwyn, -od (which might be an imitation of it), gwanwyn, etc.? In other words have we again -an and -ain? Cf. rhiain and rhian, adain and adan (also aden), celain and celan, all fem., and all forming plurals in -edd and -ydd. These are stems in - in which both the

=

1 Myharan, myheryn, 'wether, ram,' Sp., Dict.3 L. Morris, Add. MS. 14,923, f. 134a, says: S.-W. cig maharen N.-W. cig mollt (Bret. maout, Ir. molt). Richards, Dict.: "maharen, in N. W. and in some parts of S. W. as Glam., a ram; in Powis and in the greatest part of S. W., . . a weather"; Arch. Brit.: Dimet., Powis., 'vervex'; Venedot., 'aries', p. 285, col. 2, s. v. 'sheep'; Richards, Dict., s. v. yspawd, 'a shoulder': N. W. 'Spawd môllt. 'Spold gweddar, S. W. [Dimet.]. Palfais gweddar in Monm. Glam, and Brec. Y Gwyl., vi (1828), p. 207: S.-W. hwrdd = N.-W. myharan. Palfais, in Neath palfish, is the shoulder-blade. In Neath, mynharan is a ram, given also in the Cambr. Journ., iv, p. 208 (minharan = hwrdd, fem, dafad), from Monm, and Glam. Cf. Sanskr. mēsha; mynharan contains myn, ‘kid', introduced by popular etymology.

·

nominative affected by the termination *- and the other cases not affected by it have been generalised. Ir. anner (fem.) is represented in Welsh by anner, anneir, 'heifer'; 'bucula, junix', Davies, Dict.1 (anner, anneirod, Sp., Dict.3).

[10.] So far these notes on nominal declension; to which some details on dialectal differences in the gender of nouns, in nouns of relationship, etc., may be appended.

As masc. in N.-W. and fem. in S.-W. I find mentioned: troed, effaith, ysgrif, rhif, nifer, clust, sain, munyd, man, golwg, ystyr, gradd (Dosp. Ed., § 471; dwy droed, Williams Pant y Celyn, Y Traeth., 1870, etc.); hanes, ciniaw, gwniadur, cyflog, clôd, clorian, gàr, Rowland, Gramm., p. 39; N.-W. pellen y pen glin, S.-W. padell y benlîn, D. S. Evans, Llyth., s. v. 'penlîn'; alarch, Id., Welsh Dict.

Braich, fem., is, according to Rhys, Loanwords, s. v. ' brachium',2 a masc. in Salesbury's language and still so in Carnarvonshire as the spur of a mountain. The same author has recorded the older comm. gender of dyn and the fem. gender of haul (still preserved in some parts, e.g., about Ystradmeurig in Cardiganshire, Rev. Celt., vi, p. 40), and drawn certain conclusions from them (Hibb. Lect., pp. 92, n. 1, 572, n. 2).

Cwpan, pennill, pontbren, canwyllbren, canrif, clust are given by Rowland, l. c., as masc. in S.-W. and fem, in N.-W.

1 According to Y Geninen, iii, p. 19, Glamorgansh, y dreisiad (in Neath trishad)=N.-W. yr aner; also heffar ('heifer'), Yr Arw., Feb. 20, 1859. Anner is the Dimetian word, Cambr. Journ., iii, p. 252. Treisiad is also the S.-W. word for N.-W. bustach, 'steer' (Davies, Dict.). L. Morris, Add. MS. 14,944, f. 28a: llo-dyniewed-bustach―ych are the successive yearly names used in Anglesey. Enderic, Juv. Gl. (Stokes, Cambr., p. 206); W. Lleyn's Vocab.: enderig bustach. Dyniewed, dyniawed, pl. dyniewyd (dinewyt, Z.2, p. 282), 'steer, heifer', Sp., Dict.3 It is the old Cornish deneuoit, juvencus', Corn. Vocab., f. 9a. It seems to have been submitted to popular etymology, if the following notice, taken from Y Cylchgr. (Abertawy), 1853, p. 17, be genuine : "deunawiad", fal y geilw gwŷr Morganwg eidion blwydd a hanner oed (alluding to deunaw, i.e., 18 months old).

2 Arch. Camb., 4th Ser., iv, P. 269.

=

[11.] Nouns of relationship differ greatly in the dialects.

Cf. Hanes y Ffydd, 1677: N.-W. nain = S.-W. mam gu; do., Y Gwyl., vi (1828), p. 207.

R. Morris, Add. MS. 14,945, f. 247b: hendaid in Anglesey = tad cu in Powys; also gorhendaid and hên hendaid = hendad cu. L. Morris, Add. MS. 14,923, f. 1336: S.-W. mam gu a thad cu = N.-W. nain a thaid. Id., ibid., 15,025, f. 80b in Powys tad dâ, mam ddâ, 'grandfather, -mother'; tad cû, mam gû, 'great-grandfather'; etc.; nain, 'great-great-grandmother'; mam wen, 'stepmother'.

=

Y Traeth., iii, p. 12: Vened, taid, nain = Powys. tad da, mam dda S.-W. tad cu, mam gu; Vened, tad yn nghyfraith (mam, mab, merch yn ngh.) Powys. and S.-W. chwegrwn (chwegr, daw, gwaudd); Vened, tad yn nghyfraith (mam, etc., yn ngh.) also = Powys, tad gwyn (mam wen, etc.) = S.-W. llysdad, llysfam, etc.

I may quote from L. Glyn Cothi's Poems, p. 210: Tir yr hynaiv, trwy raniad | A rhan o dir yr hen dad; | Tai'r gorhendad, a'r tad da | Tai'r ewythyr vàl Troia. (He uses broder, besides brodyr; cf. Y tri broder lle gosoder | Yr aur doder ar wyr dedwydd, p. 42; broder-Rhosser, p. 43; Dau vroder, ryw amser, oedd | A wnaeth Ruvain a'i threvoedd, p. 433, etc.)

Daw, pl. dawon, Davies, Gramm., p. 40; and dofion.

Altrou, victricus', and altruan, 'noverca', Corn. Vocab., are W. alltraw, 'sponsor', and elltrewen, 'stepmother'; Davies, Dict., has elldrewyn, from W. Lleyn's Vocab.

Cefnder, etc., are pronounced at Neath: centar, pl. cenderwydd; cnithtar, pl. cnithterwydd; cyferddar (= cyfyrder); ib.: whār, whiorydd; brawd, brotyr; tad yn nghyfrith.

6

In Breton tad coz is 'grandfather', tad cuñ, as in Powys, ' greatgrandfather'; tad you, 'great-great-grandfather'. Tad caer, fatherin-law' and 'stepfather' (tadec in Vann.), is in imitation of the French name; the Welsh tad gwyn shows the same idea. It is lestad, as in S.-W., in the dialect of Léon. Deuñ, gouhez are W. daw, gwaudd. Tadek, gourtadik exist in Batz (Ernault, Batz Dialect, p. 34); cf. W. gorhendad.

A list of all possible degrees of relationship, with the Welsh names, will be found in the Cambrian Journal. It was sent to Wales by Lewis Morgan and filled up by Williams ab Ithel, Though Morgan's circular, showing the kind of researches he was making, is printed before the list, Ab Ithel gave him only the literary words, and omitted the dialectal forms, different as they are. Morgan's great work is well known, and so is the German book of