MS. W (Gwentian Code, Cleop., A14): hýný ymhoelo, f. 60a; yd ymhoelir, f. 97a, etc. MS. S (Add. MS. 22,356): yd ymhoilyr, f. 64b; ymchoylyt, f. 16b; nyd ymh6elant h6ynteu, f. 16b; yd ymch6oylir, f. 48b. Add. MS. 19,709: ymchoelut, f. 226; ymhoelut, f. 10a.

In the Mabinogion and Triads, B. of Herg. (Oxford, 1887), I found at least 39 times ymchoelut, -yt (Infin.), -a6d, -es, etc.; once ymhoela6d, col. 672, and probably never ymchwelut.

In the parts edited by R. Williams from Llyfr. Gwyn Rhydderch: in Charlemagne's Voyage, only ymchoelut (p. 2, -es, 4), and ymhoelut, p. 17, occur. In Turpin's Chronicle, ymchwelut, -ws, -assant, etc., occur 17 times; ymchoelut, etc., 3 times (ymchoel, p. 87; ymchoylei, p. 46; -wys, p. 44); ymhoelut, 22 times (ymhoylut, p. 43; ymhoyl, p. 82, etc.).—In Bown o Hamtwn: ymchwelut, p. 127; ymchoelawd, p. 188; ymhoelut (ymhoyly, 2nd sing., p. 138; ymoel, p. 138), 30 times. In Purdan Padric: ymchwelut, etc., p. 192(2); amhaelaud, p. 196; ymhoelut, 10 times.-In Yst. Gulat Ieuan Vend.: ymhoelyt, p. 328, -u, 2nd sing., p. 328; ymhoel, pp. 328, 334 (Select. from Hgt. MSS., vol. ii).

Salesbury, N. T.: ymchwelyt, f. 3a; ymchoelyt, f. 1716; nad ymchoelent at, marg. y dychwelasant, f. 3b; ydd ymchoelodd, f. 26b, etc. In later MSS. ymchwelyd usually occurs, also in those which contain regularly doedyd.

§ 44. Like dywedyd (see § 42), ymchweilyd occurs very seldom. I found it in only two MSS., viz., in Cleop., B5 [3rd sing. ymchweil, f. 68b; Inf. dychwelut, f. 1536; ymchelut, f. 1386; dychelut adref, f. 152a; Imp. ymchwelet, f. 107a, etc.; but odyno ýdymchwoyl, f. 62a (like MS. S); and yd ymchweilent, f. 63a; a ymchweilir, ff. 65a, 67b; ýd ýmchweilant, f. 67a; a ymchweilassant, f. 90a; a ymchweylws, f. 100b; a hýnný oll a datymchweilies ynnev, f. 108b; a dychweiliassant, f. 141a; pan ymchweilit, f. 129a], and in the prose text Buchedd yr anrhydeddus Bedr, in Add. MS. 14,979 (16th cent.): ymchweylût a dwedûd, f. 158b; onid ymchweûlai, f. 162a; ymchwevlyd, f. 163a; ymchweylût, ib.; besides ymchwelodd, f. 1576; etc. In the later MS. doydud and dyweydyt occur (see § 42; as to û, cf. ib. onaddûnt, f. 155a; dûallt, gan eû brodûr, f. 159b; etc.

§ 45. It results from the above quoted examples that ymhoelyd is a Middle-Welsh form surviving in the South Western dialects, whilst doedyd occurs not earlier than about the 16th century, and is not seldom met with in MSS. which for other reasons should be attributed to the Powysian dialects. The phonetic processes which are apparently common to both must therefore be held to be different ones, and cannot be further discussed here, though the phonetics of the modern dialects, which are up to the present very incompletely known, might perhaps illustrate them. Dyweidyd and ymchweilyd probably contain the ei of the 3rd sing. dyweid, ymchweil, which was analogically transferred into them at the time when the 3rd sing, dywed, ymchwel, formed after the model of cymmer, began to replace dyweid, ymchweil. At that time the coexistence of dywed and dyweid caused dyweidaf, like dywedaf, etc., to be formed. The modern deidyd shows the group dw before vowels, simplified by the elimination of w; we may remember dôd from *dfod, besides dwad (dyfod). The 3rd sing. Pret. of dywedyd is dyfod, dywad, dywawd (Davies, Gramm.). On these forms see Rhys, Rev. Celt., vi, p. 17, who observes that dywad, still used in Gwynedd, is formed "with the modern preference for wa over wo". Like gwared-gwarawd-gwares, an s-form of dywed -dywawd existed also (dywes-), on which see above, § 32.

I. THE VERB SUBSTANTIVE.

§ 46. The forms of wyf, oeddwn, compounded with yd-, ytt-, and with the preceding verbal particles ydd-, yr- (Zeuss, • 2p. 551), show some phonetic and dialectal peculiarities not mentioned in Zeuss. Yd-, ytt- in some South-Welsh MSS. from the 15th cent. downwards, and in the later colloquial

language, generally appear as od-, ott-, sometimes wd-.1 Cf. from Brut Gr. ab Arthur, in MS. Cleop., B5 (the only MiddleWelsh MS. in which I found these forms): val yr ottoedwn, f. 60a; thra ottoydit, f. 73a; y gwyr a ottoed, f. 90a; ýr ottoedynt, f. 60b; a phan ottoýdynt, f. 10a; marw o nychdod ir odwyf, Robin Ddu, 1460, quoted Add. MS. 14,944, f. 134a. In the Gwentian part (Huet) of Sal. N.T., very seldom: yr odoedd, f. 386a. On forms occurring in the Gwentian Add, MS. 14,921 see my Beitr., § 41 (odwi, ody, ni dody, ny dodi, nydody, odyn, ny dodyn, odoedd, odys). In Marchawg Crwydrad (printed from a South-Welsh MS.): lle nad ody, lle na bu a lle na bydd byth, p. 2; er nad ody aur ac arian ond..., p. 148; y rhai a odynt yn eu harfer, p. 145; ody duw yn 'wyllysu, p. 144; nid ody yn gwneuthur, p. 143; yr arglwyddes Fenws yr honn ody yr holl garedig gariadon yn foddlon iddi, p. 141; petheu. ... drwy yr hwn y byd odys yn ei drallodi. L. Morris, Add. MS. 14,923, f. 1326: y in SouthWelsh pronounced as o in ôdym, North-Welsh ydym, ôdi, NorthWelsh ydyw.

In modern dialects:

Dimetian: a odi chi yn meddwl, Ser. Cymru, i, p. 429; a odi Mari.. ..., p. 231; a odi chi yn dost? p. 251, etc. Gwentian: Rhys, Lectures, 2p. 45, gives oti, from a part of Gwent, that in which the media are pronounced in such a manner as to be commonly written tenues, cf. jocal = diogel, in Neath, etc. Aberdare: odi (Y Gweithiwr). Llanelly, Cyf.yr Aelwyd: ydw, p. 68; ydy, ydach chi, p. 20; besides ody ?—answered by odyn, p. 68. Y Tyw. a'r G.: odi e'n ddrwg, ii, p. 66; oti chi, i, p. 154; otuchi, pp. 94, 117; neu otu chi, p. 117; otuch I, p. 94 (ydu chi, ib.); otus ddim . . . p. 96. Monmouthshire: Y Bedyddiwr: otw, otw, viii, p. 174; nag oti, p. 107; odi, p. 174; nag odyn, p. 174; etc.

Wd.: Seren Cymru: yr wdw i, iii, p. 264; wdw I, i, p. 232; Gwron Cymreig: rwdwi, wdw i, 20, 5, 52; ib. wes, wedd, rwedd, trwed for oes, etc. Carnarvonshire: mi rwdwi, Yr Aru., 9, 4, 57; yr wdw i, ib.; rwdwi yn deud, 11, 12, 56; ac yr wdw ina yn cael . . . etc. On similar phonetics in other words see my Beitr.,

1 yd before consonants occurs frequently in the old poems and in the so-called Gogynfeirdd. Cf. from Dr. Davies' MS. (Add. MS. 14,869), e.g., yd vernir, f. 55a, yd rotir, y gelwir, yd gedwir, yt (= ydd) ergryner f. 56a, mal yd glywir f. 70b (Cynddelw), etc.

§§ 40-48; cf., e.g., bychan, bachan, bochan, bwchan (Wogan, Vwghan, in the English part of Llyfr Achau, pp. 55, 56); to the forms of the Old Welsh Catguocaun, given l. c., § 40: Cydwgan, Kwdogan, Kodwgan, Kwdwgan, Kodogan; add. Kydogan, Ll. Achau, p. 22, Kadogan, p. 15, Kadwgan, p. 15; etc.

§ 47. Davies, Dict. and Gramm., p. 182, gives ydd- instead of yr as the Dimetian form. So also Pughe, in Coxe's Monmouthshire: Gwent. yddoedd = Venedot. yroedd; Dosp. Edeyrn, § 824: yd-, ydd- are "more generally" used in SouthWelsh, y-, yr- in North-Welsh. This dialectal difference may be seen in the two texts of Gruff. ab Cynan's Life used in Myv. Arch., vol. ii; cf. ac yna ydd oeddynt, 2p. 723 (SouthWelsh), yr- (North-Welsh); ydd oedd, 2p. 724: yr- (NorthWelsh); ac yny lle ydd annogasant wynteu, p. 724: yr annogasant hwytheu; also pp. 730, 731 : ydd erchis-yr archodd ; etc. But yr is not wanting in South-Welsh texts and modern dialects, at least as they are written in Seren Cymru, Y Gwron Cymreig, etc. These different forms are due to the generalisation of mutations incurred before certain consonants, like the different forms of the article in Welsh and Breton. Yd-, notwithstanding its frequent use in the form of od- in South-Welsh, is said by Davies, Gramm., and others to be more generally used in NorthWelsh. Cf. Davies, Gramm.: Pass. ydys; poet., Dimet., Powysian, ys; poet. also ydis. Hughes, 1822: SouthWelsh wyf rather than ydwyf. Rowlands, Gramm., p. 76: South-Welsh wyf, pl. ym, ych, ynt; North-Welsh ydwyf, etc. Y Geninen, iii, 19, Glamorganshire: shwd i chi sut yr ydych chwi.

=

§ 48. The problem of the origin of wyf and oeddwn in Welsh, ouf and off, oann and en, etc., in Breton and Cornish, has been very differently treated by Rhys, Lectures, 2p. 234, R. C., vi, p. 49, n. 1, and by Stokes, Kuhn's Zeitschr., xxviii, p. 101 seq. I am of the first opinion of the former, that these forms belong to the root es- ; ys Ir. is; 3rd pl. int, ynt *enti for

=

*s-enti, caused by *e- s- m . . . of the 1st pl. From the very early period when the vowel of the root, lost in the plural, was reintroduced in the 1st and 2nd pers. pl.-in which persons it was afterwards altered, as the Cornish and Breton forms show-the singular is supposed to have assumed thematic flexion: *esemi, *esesi. Perhaps the assumption of another analogical process, of the retransgression of the e of the plural into the singular, or of the *ē (from the augment e +e of es-) in the imperfect, would also meet the difficulties here existing : *esmi gave *em or *ym (hence -m in buum bûm, I was, etc.); *esi gave ei, *esti gave ys, which has been kept; *e-esmi (or *e-em) or *ēm, *ēsi or *ēi, *ēsti or *ēs gave *wym (altered later into wyf, like all other verbs), *wy (later wyt, with t of the pronoun), oes, which still exists. An argument in favour of the theory of thematic flexion is furnished by the 3rd sing. yd-y, yd-i, Corn. us-y, Bret. ed-y, if -y is from *eset, and yw, Corn. yu, eu, Bret. eu, is to be explained by the supposition. of an affixed pronominal element (Rhŷs). However, *ēset would give *wy; and nyw-: nwy-, y'w (do), etc. caution us against denying the possibility of *wy becoming yw under certain yet unknown conditions,-at a very remote period, it is true: cf. the Corn. and Bret. forms.-These forms are analogically transformed during their history in Welsh itself. J. D. Rhŷs, Gramm., 1595, gives the pl. ym, yn (like all later plurals, following the pron. ni), ych and ywch, ynt. Ywch, ydywch, yttywch are ascribed by Davies to the Dimetian dialect and the poets. Dosp. Ed., § 653, gives even 1st plur. Dimet. wyn, 2nd. ywch, ywch. The modern South-Welsh dialects, however, use ych. Ydem, ydech, are the common North-Welsh forms, identical in termination with the modern Pres. Sec. in -em, -ech, the only terminations of this tense used in these dialects. So probably when -em was generalised in the Pres. Sec. (see § 17) ydym followed this model, like wrthachi for wrthywch, etc. See Obs. on the Pron., § 40.