one of the names of Bala Lake, now called in Welsh ILyn Tegid, and formerly sometimes ILyn Aerfen.

While touching at p. 286 on Gwaen ILifon with its ILyn Pencraig as one of those claiming to be the ILyn ILïon of the Triads, it was hinted that Lion was but a thinner form of Lifon. Here one might mention perhaps another ILifon, for which, however, no case could be made. I allude to the name of the residence of the Wynns descended from Gilmin Troedđu, namely, Glyn ILifon, which means the river ILifon's Glen; but one could not feel surprised if the neighbouring Lyfni, draining the lakes of Nantfle, should prove to have once been also known as a ILifon, with the Nantlle waters conforming by being called ILyn ILifon. But however that may be, one may say as to the flood caused by the bursting of any such lake, that the notion of the universality of the catastrophe was probably contributed by the author of Triad iii. 13, from a non-Welsh source. He may have, however, not invented the vessel in which he places Dwyfan and Dwyfach: at all events, one version of the story of the Fan Fach represents the Lake Lady arriving in a boat. As to the writer of the other Triad, iii. 97, he says nothing about Dwyfan and his wife, but borrows Nefyd Naf Neifion's ship to save all that were to be saved; and here one may probably venture to identify Nefyd with Nemed1, genitive Nemid, a name borne in Irish legend by a rover who is represented as one of the early colonizers of Erin. As to the rest, the name Neifion by itself is used in Welsh for Neptune and the sea, as in the following couplet of D. ab Gwilym's poem lv :

Nofiad a wnaeth hen Neifion
O Droia fawr draw i Fón.

It is old Neptune that has swam
From great Troy afar to Mona.

'See my Hibbert Lectures, pp. 649-50.

In the same way Môr Neifion, 'Sea of Neifion,' seems to have signified the ocean, the high seas.

To return to the Triad about Dwyfan and Dwyfach, not only does it make them from being water divinities into a man and woman, but there is no certainty even that both were not feminine. In modern Welsh all rivers are treated as feminine, and even Dyfrdwyf has usually to submit, though the modern bard Tegid, analysing the word into Dwfr Dwyf, Water of the Divinity or Divine Water,' where dwfr, 'water,' could only be masculine, addressed ILyn Tegid thus, p. 78:

Drwyot, er dydiau'r Drywon,
Y rhwyf y Dyfrdwyf ei don.
This question, however, of the gender of river names,
or rather the sex which personification ascribed them,
is a most difficult one. If we glance at Ptolemy's
Geography written in the second century, we find in
his account of the British Isles that he names more than
fifty of our river mouths and estuaries, and that he
divides their names almost equally into masculine and
feminine. The modern Welsh usage has, it is seen,
departed far from this, but not so far the folklore: the
afanc is a male, and we have a figure of the same sex
appearing as the father of the lake maiden in the Fan
Fach story, and in that of ILyn Du'r Arđu ; the same, too,
was the sex of the chief dweller of ILyn Cwm ILwch;
the same remark is applicable also to the greatest
divinity of these islands—the greatest, at any rate, so far
as the scanty traces of his cult enable one to become
acquainted with him. As his name comes down into
legend it belongs here, as well as to the deities of
antiquity, just as much, in a sense, as the Dee. I refer
to Nudons or Nodons, the remains1 of whose sanctuary

Through thee, from the days of the Druids,
The Dwfr Dwyf impels his wave.

A full account of them will be found in a volume devoted to them, and entitled Roman Antiquities at Lydney Park, Gloucestershire, being a post

were many years ago brought to light on a pleasant hill in Lydney Park, on the western banks of the Severn. In the mosaic floor of the god's temple there is a coloured inscription showing the expense of that part of the work to have been defrayed by the contributions (ex stipibus) of the faithful, and that it was carried out by two men, of whom one appears to have been an officer in command of a naval force guarding the coasts of the Severn Sea. In the midst of the mosaic inscription is a round opening in the floor of nine inches in diameter and surrounded by a broad band of red enclosed in two of blue. This has given rise to various speculations, and among others that it was intended for libations. The mosaics and the lettering of the inscriptions seem to point to the third century as the time when the sanctuary of Nudons was built under Roman auspices, though the place was doubtless sacred to the god long before. In any case it fell in exactly with the policy of the more astute of Roman statesmen to encourage such a native cult as we find traces of in Lydney Park.

One of the inscriptions began with D. M. Nodonti, 'to the great god Nudons,' and a little bronze crescent intended for the diadem of the god or of one of his priests gives a representation of him as a crowned, beardless personage driving a chariot with four horses; and on either side of him is a naked figure supposed to represent the winds, and beyond them on each of the two sides is a triton with the fore feet of a horse. The god holds the reins in his left hand, and his right uplifted grasps what may be a sceptre or possibly a whip, while the whole equipment of the god recalls in humous work of the Rev. W. Hiley Bathurst, with Notes by C. W. King, London, 1879. See also an article entitled 'Das Heiligtum des Nodon,' by Dr. Hübner in the Jahrbücher des Vereins von Alterthumsfreunden im Rheinlande, lxvii. pp. 29-46, where several things in Mr. King's book are criticized.

some measure the Chariot of the Sun. Another piece of the bronze ornament shows another triton with an anchor in one of his hands, and opposite him a fisherman in the act of hooking a fine salmon. Other things, such as oars and shell trumpets, together with mosaic representations of marine animals in the floor of the temple, compel us to assimilate Nudons more closely with Neptune than any other god of classical mythology.

The name of the god, as given in the inscriptions, varies between Nudons and Nodens, the cases actually occurring being the dative Nodonti, Nodenti, and Nudente, and the genitive Nodentis, so I should regard ō or й as optional in the first syllable, and o as preferable, perhaps, to e in the second, for there is no room for reasonably doubting that we have here to do with the same name as Irish Nuadu, genitive Nuadat, conspicuous in the legendary history of Ireland. Now the Nuadu who naturally occurs to one first, was Nuadu Argetlám or Nuadu of the Silver Hand, from argat, 'silver, argentum,' and lám, 'hand.' Irish literature explains how he came to have a hand made of silver, and we can identify with him on Welsh ground a ILuđ ILawereint ; for put back as it were into earlier Brythonic, this would be Lūdo(ns) Lām'-argentios: that is to say, a reversal takes place in the order of the elements forming the epithet out of ereint (for older ergeint), 'silvern, argenteus,' and ttaw, for earlier lama, 'hand.' Then comes the alliterative instinct into play, forcing Nūdo(ns) Lāmargentio(s) to become Lūdo(ns) Lāmargentio(s), whence the later form, Luc Lawereint, derives regularly1. Thus we have in Welsh the name Lûd, fashioned into that form under the influence of the epithet, whereas elsewhere it is Nud, which occurs as a man's name in the pedi grees, while an intermediate form was probably Nūdos

1 See my Hibbert Lectures, pp. 122, 125.

or Nudo, of which a genitive NVDI occurs in a postRoman inscription found near Yarrow Kirk in Selkirkshire. It is worthy of note that the modification of Nudo into Ludo must have taken place comparatively early-not improbably while the language was still Goidelic-as we seem to have a survival of the name in that of Lydney itself.

It is very possible that we have Ludo, ILud, also in Porthlud, which Geoffrey of Monmouth gives, iii. 20, as the Welsh for Ludesgata or Ludgate, in London, which gate, according to him, was called after an ancient king of Britain named Lud. He seems to have been using an ancient tradition, and there would be nothing improbable in the conjecture that Geoffrey's Lud was our ILuc, and that the great water divinity of that name had another sanctuary on the hill by the Thames, somewhere near the present site of St. Paul's Cathedral, and occupying a post as it were prophetic of Britain's rule of the water-ways in later times.

Perhaps as one seems to find traces of Nudons from the estuary of the Thames to that of the Severn and thence to Ireland, one may conclude that the god was one of the divinities worshipped by the Goidels. With regard to the Brythonic Celts, there is nothing to suggest that he belonged also to them except in the sense of his having been probably adopted by them from the Goidels. It might be further suggested that the Goidels themselves had in the first instance adopted him from the pre-Celtic natives, but in that case a goddess would have been rather more probable 1. In fact in the case of the Severn we seem to have a trace of such a goddess in the Sabrina, Old Welsh Habren, now Hafren, so called after a princess whom Geoffrey, ii. 5, represents drowned in the river: she may have been

1 On this subject, see The Welsh People, especially pp. 54-61.