CHAPTER VI

THE FOLKLORE OF THE WELLS

... Iuvat integros accedere fontes.--Lucretius.

It is only recently [158] that I heard for the first time of Welsh instances of the habit of tying rags and bits of clothing to the branches of a tree growing near a holy well. Since then I have obtained several items of information in point: the first is a communication received in June, 1892, from Mr. J. H. Davies, of Lincoln College, Oxford--since then of Lincoln's Inn--relating to a Glamorganshire holy well, situated near the pathway leading from Coychurch to Bridgend. It is the custom there, he states, for people suffering from any malady to dip a rag in the water, and to bathe the affected part of the body, the rag being then placed on a tree close to the well. When Mr. Davies passed that way, some three years previously, there were, he adds, hundreds of such shreds on the tree, some of which distinctly presented the appearance of having been very recently placed there. The well is called Ffynnon Cae Moch, 'Swine-field Well,' which can hardly have been its old name; and a later communication from Mr. Davies summarizes a conversation which he had about the well, on December 16, 1892, with Mr. J. T. Howell, of Pencoed, near Bridgend. His notes run thus:--'Ffynnon Cae Moch, between Coychurch and Bridgend, is one mile from Coychurch, one and a quarter from Bridgend, near Tremains. It is within twelve or fifteen yards of the high-road, just where the pathway begins. People suffering from rheumatism go there. They bathe the part affected with water, and afterwards tie a piece of rag to the tree which overhangs the well. The rag is not put in the water at all, but is only put on the tree for luck. It is a stunted, but very old tree, and is simply covered with rags.' A little less than a year later, I had an opportunity of visiting this well in the company of Mr. Brynmor-Jones; and I find in my notes that it is not situated so near the road as Mr. Howell would seem to have stated to Mr. Davies. We found the well, which is a powerful spring, surrounded by a circular wall. It is overshadowed by a dying thorn tree, and a little further back stands another thorn which is not so decayed: it was on this latter thorn we found the rags. I took off a twig with two rags, while Mr. Brynmor-Jones counted over a dozen other rags on the tree; and we noticed that some of them had only recently been suspended there: among them were portions undoubtedly of a woman's clothing. At one of the hotels at Bridgend, I found an illiterate servant who was acquainted with the well, and I cross-examined him on the subject of it. He stated that a man with a wound, which he explained to mean a cut, would go and stand in the well within the wall, and there he would untie the rag that had been used to tie up the wound and would wash the wound with it: then he would tie up the wound with a fresh rag and hang the old one on the tree. The more respectable people whom I questioned talked more vaguely, and only of tying a rag to the tree, except one who mentioned a pin being thrown into the well or a rag being tied to the tree.

My next informant is Mr. D. J. Jones, a native of the Rhonda Valley, in the same county of Glamorgan. He was an undergraduate of Jesus College, Oxford, when I consulted him in 1892. His information was to the effect that he knows of three interesting wells in the county. The first is situated within two miles of his home, and is known as Ffynnon Pen Rhys, or the Well of Pen Rhys. The custom there is that the person who wishes his health to be benefited should wash in the water of the well, and throw a pin into it afterwards. He next mentions a well at Llancarvan, some five or six miles from Cowbridge, where the custom prevails of tying rags to the branches of a tree growing close at hand. Lastly, he calls my attention to a passage in Hanes Morganwg, 'The History of Glamorgan,' written by Mr. D. W. Jones, known in Welsh literature as Dafyd Morganwg. In that work, p. 29, the author speaks of Ffynnon Marcros, 'the Well of Marcros,' to the following effect:--'It is the custom for those who are healed in it to tie a shred of linen or cotton to the branches of a tree that stands close by; and there the shreds are, almost as numerous as the leaves.' Marcros is, I may say, near Nash Point, and looks on the map as if it were about eight miles distant from Bridgend. Let me here make it clear that so far we have had to do with four different wells [159], three of which are severally distinguished by the presence of a tree adorned with rags by those who seek health in those waters; but they are all three, as the reader will have doubtless noticed, in the same district, namely, the part of Glamorganshire near the main line of the Great Western Railway.

There is no reason, however, to think that the custom of tying rags to a well tree was peculiar to that part of the Principality. One day, in looking through some old notes of mine, I came across an entry bearing the date of August 7, 1887, when I was spending a few days with my friend, Chancellor Silvan Evans, at Llanwrin Rectory, near Machynlleth. Mrs. Evans was then alive and well, and took a keen interest in Welsh antiquities and folklore. Among other things, she related to me how she had, some twenty years before, visited a well in the parish of Llandrillo yn Rhos, namely Ffynnon Eilian, or Elian's Well, between Abergele and Llandudno, when her attention was directed to some bushes near the well, which had once been covered with bits of rags left by those who frequented the well. This was told Mrs. Evans by an old woman of seventy, who, on being questioned by Mrs. Evans concerning the history of the well, informed her that the rags used to be tied to the bushes by means of wool. She was explicit on the point, that wool had to be used for the purpose, and that even woollen yarn would not do: it had to be wool in its natural state. The old woman remembered this to have been the rule ever since she was a child. Mrs. Evans noticed corks, with pins stuck in them, floating in the well, and her informant remembered many more in years gone by; for Elian's Well was once in great repute as a ffynnon reibio, or a well to which people resorted for the kindly purpose of bewitching those whom they hated. I infer, however, from what Mrs. Evans was told of the rags, that Elian's Well was visited, not only by the malicious, but also by the sick and suffering. My note is not clear on the point whether there were any rags on the bushes by the well when Mrs. Evans visited the spot, or whether she was only told of them by the caretaker. Even in the latter case it seems evident that this habit of tying rags to trees or bushes near sacred wells has only ceased in that part of Denbighshire within this century. It is very possible that it continued in North Wales more recently than this instance would lead one to suppose; indeed, I should not be in the least surprised to learn that it is still practised in out of the way places in Gwyned, just as it is in Glamorgan: we want more information.

I cannot say for certain whether it was customary in any of the cases to which I have called attention to tie rags to the well tree as well as to throw pins or other small objects into the well; but I cannot help adhering to the view, that the distinction was probably an ancient one between two orders of things. In other words, I am inclined to believe that the rag was regarded as the vehicle of the disease of which the ailing visitor to the well wished to be rid, and that the bead, button, or coin deposited by him in the well, or in a receptacle near the well, formed alone the offering. In opposition to this view Mr. Gomme has expressed himself as follows in Folk-Lore, 1892, p. 89:--'There is some evidence against that, from the fact that in the case of some wells, especially in Scotland at one time, the whole garment was put down as an offering. Gradually these offerings of clothes became less and less till they came down to rags. Also in other parts, the geographical distribution of rag-offerings coincides with the existence of monoliths and dolmens.' As to the monoliths and dolmens, I am too little conversant with the facts to risk any opinion as to the value of the coincidence; but as to the suggestion that the rag originally meant the whole garment, that will suit my hypothesis admirably. In other words, the whole garment was, as I take it, the vehicle of the disease: the whole was accursed, and not merely a part. But Mr. Gomme had previously touched on the question in his presidential address (Folk-Lore for 1892, p. 13); and I must at once admit that he succeeded then in proving that a certain amount of confusion occurs between things which I should regard as belonging originally to distinct categories: witness the inimitable Irish instance which he quotes:--'To St. Columbkill--I offer up this button, a bit o' the waistband o' my own breeches, an' a taste o' my wife's petticoat, in remimbrance of us havin' made this holy station; an' may they rise up in glory to prove it for us in the last day.' Here not only the button is treated as an offering, but also the bits of clothing; but the confusion of ideas I should explain as being, at least in part, one of the natural results of substituting a portion of a garment for the entire garment; for thereby a button or a pin becomes a part of the dress, and capable of being interpreted in two senses. After all, however, the ordinary practices have not, as I look at them, resulted in effacing the distinction altogether: the rag is not left in the well; nor is the bead, button, or pin attached to a branch of the tree. So, in the main, it seemed to me easier to explain the facts, taken altogether, on the supposition that originally the rag was regarded as the vehicle of the disease, and the bead, button, or coin as the offering. My object in calling attention to this point was to have it discussed, and I am happy to say that I have not been disappointed; for, since my remarks were published [160], a paper entitled Pin-wells and Rag-bushes was read before the British Association by Mr. Hartland, in 1893, and published in Folk-Lore for the same year, pp. 451-70. In that paper the whole question is gone into with searching logic, and Mr. Hartland finds the required explanation in one of the dogmas of magic. For 'if an article of my clothing,' he says, 'in a witch's hands may cause me to suffer, the same article in contact with a beneficent power may relieve my pain, restore me to health, or promote my general prosperity. A pin that has pricked my wart ... has by its contact, by the wound it has inflicted, acquired a peculiar bond with the wart; the rag that has rubbed the wart has by that friction acquired a similar bond; so that whatever is done to the pin or the rag, whatever influences the pin or the rag may undergo, the same influences are by that very act brought to bear, upon the wart. If, instead of using a rag, or making a pilgrimage to a sacred well, I rub my warts with raw meat and then bury the meat, the wart will decay and disappear with the decay and dissolution of the meat.... In like manner my shirt or stocking, or a rag to represent it, placed upon a sacred bush, or thrust into a sacred well--my name written upon the walls of a temple--a stone or a pellet from my hand cast upon a sacred image or a sacred cairn--is thenceforth in continual contact with divinity; and the effluence of divinity, reaching and involving it, will reach and involve me.' Mr. Hartland concludes from a large number of instances, that as a rule 'where the pin or button is dropped into the well, the patient does not trouble about the rag, and vice versa.' This wider argument as to the effluence of the divinity of a particular spot of special holiness seems to me conclusive. It applies also, needless to say, to a large category of cases besides those in question between Mr. Gomme and the present writer.

So now I would revise my position thus:--I continue to regard the rag much as before, but treat the article thrown into the well as the more special means of establishing a beneficial relation with the well divinity: whether it could also be viewed as an offering would depend on the value attached to it. Some of the following notes may serve as illustrations, especially those relating to the wool and the pin:--Ffynnon Gwynwy, or the Well of Gwynwy, near Llangelynin, on the river Conwy, appears to be partly in point; for it formerly used to be well stocked with crooked pins, which nobody would touch lest he might get from them the warts supposed to attach to them, whence it would appear that a pin might be regarded as the vehicle of the disease. There was a well of some repute at Cae Garw, in the parish of Pistyll, near the foot of Carnguwch, in Lleyn, or West Carnarvonshire. The water possessed virtues to cure one of rheumatism and warts; but, in order to be rid of the latter, it was requisite to throw a pin into the well for each individual wart. For these two items of information, and several more to be mentioned presently, I have to thank Mr. John Jones, better known in Wales by his bardic name of Myrdin Fard, and as an enthusiastic collector of Welsh antiquities, whether in the form of manuscript or of unwritten folklore. On the second day of the year 1893 I paid him a visit at Chwilog, on the Carnarvon and Avon Wen Railway, and asked him many questions: these he not only answered with the utmost willingness, but he also showed me the unpublished materials which he had collected. I come next to a competition on the folklore of North Wales at the London Eistedfod in 1887, in which, as one of the adjudicators, I observed that several of the competitors mentioned the prevalent belief, that every well with healing properties must have its outlet towards the south (i'r dê). According to one of them, if you wished to get rid of warts, you should, on your way to the well, look for wool which the sheep had lost. When you had found enough wool you should prick each wart with a pin, and then rub the wart well with the wool. The next thing was to bend the pin and throw it into the well. Then you should place the wool on the first whitethorn you could find, and as the wind scattered the wool, the warts would disappear. There was a well of the kind, the writer went on to say, near his home; and he, with three or four other boys, went from school one day to the well to charm their warts away. For he had twenty-three on one of his hands; so that he always tried to hide it, as it was the belief that if one counted the warts they would double their number. He forgets what became of the other boys' warts, but his own disappeared soon afterwards; and his grandfather used to maintain that it was owing to the virtue of the well. Such were the words of this writer, whose name is unknown to me; but I guess him to have been a native of Carnarvonshire, or else of one of the neighbouring districts of Denbighshire or Merionethshire. To return to Myrdin Fard, he mentioned Ffynnon Cefn Lleithfan, or the Well of the Lleithfan Ridge, on the eastern slope of Mynyd y Rhiw, in the parish of Bryncroes, in the west of Lleyn. In the case of this well it is necessary, when going to it and coming from it, to be careful not to utter a word to anybody, or to turn to look back. What one has to do at the well is to bathe the warts with a rag or clout which has grease on it. When that is done, the clout with the grease has to be carefully concealed beneath the stone at the mouth of the well. This brings to my mind the fact that I noticed more than once, years ago, rags underneath stones in the water flowing from wells in Wales, and sometimes thrust into holes in the walls of wells, but I had no notion how they came there.

On the subject of pin-wells I had in 1893, from Mr. T. E. Morris, of Portmadoc, barrister-at-law, some account of Ffynnon Faglan, or Baglan's Well, in the parish of Llanfaglan, near Carnarvon. The well is situated in an open field to the right of the road leading towards the church, and close to it. The church and churchyard form an enclosure in the middle of the same field, and the former has in its wall the old stone reading FILI LOVERNII ANATEMORI. My friend derived information from Mrs. Roberts, of Cefn y Coed, near Carnarvon, as follows:--'The old people who would be likely to know anything about Ffynnon Faglan have all died. The two oldest inhabitants, who have always lived in this parish of Llanfaglan, remember the well being used for healing purposes. One told me his mother used to take him to it, when he was a child, for sore eyes, bathe them with the water, and then drop in a pin. The other man, when he was young, bathed in it for rheumatism; and until quite lately people used to fetch away the water for medicinal purposes. The latter, who lives near the well, at Tan y Graig, said that he remembered it being cleaned out about fifty years ago, when two basinfuls of pins were taken out, but no coin of any kind. The pins were all bent, and I conclude the intention was to exorcise the evil spirit supposed to afflict the person who dropped them in, or, as the Welsh say, dadwitsio. No doubt some ominous words were also used. The well is at present nearly dry, the field where it lies having been drained some years ago, and the water in consequence withdrawn from it. It was much used for the cure of warts. The wart was washed, then pricked with a pin, which, after being bent, was thrown into the well. There is a very large and well-known well of the kind at C'lynnog, Ffynnon Beuno, "St. Beuno's Well," which was considered to have miraculous healing powers; and even yet, I believe, some people have faith in it. Ffynnon Faglan is, in its construction, an imitation, on a smaller scale, of St. Beuno's Well at C'lynnog.'

In the cliffs at the west end of Lleyn is a wishing-well called Ffynnon Fair, or St. Mary's Well, to the left of the site of Eglwys Fair, and facing Ynys Enlli, or Bardsey. Here, to obtain your wish, you have to descend the steps to the well and walk up again to the top with your mouth full of the water; and then you have to go round the ruins of the church once or more times with the water still in your mouth. Viewing the position of the well from the sea, I should be disposed to think that the realization of one's wish at that price could not be regarded as altogether cheap. Myrdin Fard also told me that there used to be a well near Criccieth Church. It was known as Ffynnon y Saint, or the Saints' Well, and it was the custom to throw keys or pins into it on the morning of Easter Sunday, in order to propitiate St. Catherine, who was the patron of the well. I should be glad to know what this exactly meant.

Lastly, a few of the wells in that part of Gwyned may be grouped together and described as oracular. One of these, the big well in the parish of Llanbedrog in Lleyn, as I learn from Myrdin Fard, required the devotee to kneel by it and avow his faith in it. When this had been duly done, he might proceed in this wise: to ascertain, for instance, the name of the thief who had stolen from him, he had to throw a bit of bread into the well and name the person whom he suspected. At the name of the thief the bread would sink; so the inquirer went on naming all the persons he could think of until the bit of bread sank, when the thief was identified. How far is one to suppose that we have here traces of the influences of the water ordeal common in the Middle Ages? Another well of the same kind was Ffynnon Saethon, in Llanfihangel Bachellaeth parish, also in Lleyn. Here it was customary, as he had it in writing, for lovers to throw pins (pinnau) into the well; but these pins appear to have been the points of the blackthorn. At any rate, they cannot well have been of any kind of metal, as we are told that, if they sank in the water, one concluded that one's lover was not sincere in his or her love.

Next may be mentioned a well, bearing the remarkable name of Ffynnon Gwyned, or the Well of Gwyned, which is situated near Mynyd Mawr, in the parish of Abererch: it used to be consulted in the following manner:--When it was desired to discover whether an ailing person would recover, a garment of his would be thrown into the well, and according to the side on which it sank it was known whether he would live or die.

Ffynnon Gybi, or St. Cybi's Well, in the parish of Llangybi, was the scene of a somewhat similar practice; for there, girls who wished to know their lovers' intentions would spread their pocket-handkerchiefs on the water of the well, and, if the water pushed the handkerchiefs to the south--in Welsh i'r dê--they knew that everything was right--in Welsh o dê--and that their lovers were honest and honourable in their intentions; but, if the water shifted the handkerchiefs northwards, they concluded the contrary. A reference to this is made by a modern Welsh poet, as follows:--

Ambell dyn, gwaeldyn, a gyrch I bant gorís Moel Bentyrch, Mewn gobaith mai hen Gybi Glodfawr syd yn llwydaw'r lli.

Some folks, worthless [161] folks, visit A hollow below Moel Bentyrch, In hopes that ancient Kybi Of noble fame blesses the flood.

The spot is not far from where Myrdin Fard lives; and he mentioned, that adjoining the well is a building which was probably intended for the person in charge of the well: it has been tenanted within his memory. Not only for this but also for several of the foregoing items of information am I indebted to Myrdin; and now I come to Mrs. Williams-Ellis, of Glasfryn Uchaf, who tells me that one day not long ago, she met at Llangybi a native who had not visited the place since his boyhood: he had been away as an engineer in South Wales nearly all his life, but had returned to see an aged relative. So the reminiscences of the place filled his mind, and, among other things, he said that he remembered very well what concern there was one day in the village at a mischievous person having taken a very large eel out of the well. Many of the old people, he said, felt that much of the virtue of the well was probably taken away with the eel. To see it coiling about their limbs when they went into the water was a good sign: so he gave one to understand. As a sort of parallel I may mention that I have seen the fish living in Ffynnon Beris, not far from the parish church of Llanberis. It is jealously guarded by the inhabitants, and when it was once or twice taken out by a mischievous stranger he was forced to put it back again. However, I never could get the history of this sacred fish, but I found that it was regarded as very old [162]. I may add that it appears the well called Ffynnon Fair, 'Mary's Well,' at Llandwyn, in Anglesey, used formerly to have inhabiting it a sacred fish, whose movements indicated the fortunes of the love-sick men and maidens who visited there the shrine of St. Dwynwen [163]. Possibly inquiry would result in showing that such sacred fish have been far more common once in the Principality than they are now.

The next class of wells to claim our attention consists of what I may call fairy wells, of which few are mentioned in connexion with Wales; but the legends about them are of absorbing interest. One of them is in Myrdin Fard's neighbourhood, and I questioned him a good deal on the subject: it is called Ffynnon Grassi, or Grace's Well, and it occupies, according to him, a few square feet--he has measured it himself--of the south-east corner of the lake of Glasfryn Uchaf, in the parish of Llangybi. It appears that it was walled in, and that the stone forming its eastern side has several holes in it, which were intended to let water enter the well and not issue from it. It had a door or cover on its surface; and it was necessary to keep the door always shut, except when water was being drawn. Through somebody's negligence, however, it was once on a time left open: the consequence was that the water of the well flowed out and formed the Glasfryn Lake, which is so considerable as to be navigable for small boats. Grassi is supposed in the locality to have been the name of the owner of the well, or at any rate of a lady who had something to do with it. Grassi, or Grace, however, can only be a name which a modern version of the legend has introduced. It probably stands for an older name given to the person in charge of the well; to the one, in fact, who neglected to shut the door; but though the name must be comparatively modern, the story, as a whole, does not appear to be at all modern, but very decidedly the contrary.

So I wrote in 1893; but years after my conversation with Myrdin Fard, my attention was called to the fact that the Glasfryn family, of which the Rev. J. C. Williams-Ellis is the head, have in their coat of arms a mermaid, who is represented in the usual way, holding a comb in her right hand and a mirror in her left. I had from the first expected to find some kind of Undine or Liban story associated with the well and the lake, though I had abstained from trying the risky effects of leading questions; but when I heard of the heraldic mermaid I wrote to Mr. Williams-Ellis to ask whether he knew her history. His words, though not encouraging as regards the mermaid, soon convinced me that I had not been wholly wrong in supposing that more folklore attached to the well and lake than I had been able to discover. Since then Mrs. Williams-Ellis has taken the trouble of collecting on the spot all the items of tradition which she could find: she communicated them to me in the month of March, 1899, and the following is an abstract of them, preceded by a brief description of the ground:--

The well itself is at the foot of a very green field-bank at the head of the lake, but not on the same level with it, as the lake has had its waters lowered half a century or more ago by the outlet having been cut deeper. Adjoining the field containing the well is a larger field, which also slopes down to the lake and extends in another direction to the grounds belonging to the house. This larger field is called Cae'r Ladi, 'the Lady's Field,' and it is remarkable for having in its centre an ancient standing stone, which, as seen from the windows of the house, presents the appearance of a female figure hurrying along, with the wind slightly swelling out her veil and the skirt of her dress. Mr. Williams-Ellis remembers how when he was a boy the stone was partially white-washed, and how an old bonnet adorned the top of this would-be statue, and he thinks that an old shawl used to be thrown over the shoulders.

Now as to Grassi, she is mostly regarded as a ghostly person somehow connected with the lake and the house of Glasfryn. One story is to the effect, that on a certain evening she forgot to close the well, and that when the gushing waters had formed the lake, poor Grassi, overcome with remorse, wandered up and down the high ground of Cae'r Ladi, moaning and weeping. There, in fact, she is still at times to be heard lamenting her fate, especially at two o'clock in the early morning. Some people say that she is also to be seen about the lake, which is now the haunt of some half a dozen swans. But on the whole her visits appear to have been most frequent and troublesome at the house itself. Several persons still living are mentioned, who believe that they have seen her there, and two of them, Mrs. Jones of Talafon, and old Sydney Griffith of Tydyn Bach, agree in the main in their description of what they saw, namely, a tall lady with well marked features and large bright eyes: she was dressed in white silk and a white velvet bonnet. The woman, Sydney Griffith, thought that she had seen the lady walking several times about the house and in Cae'r Ladi. This comes, in both instances, from a young lady born and bred in the immediate neighbourhood, and studying now at the University College of North Wales; but Mrs. Williams-Ellis has had similar accounts from other sources, and she mentions tenants of Glasfryn who found it difficult to keep servants there, because they felt that the place was haunted. In fact one of the tenants himself felt so unsafe that he used to take his gun and his dog with him to his bedroom at night; not to mention that when the Williams-Ellises lived themselves, as they do still, in the house, their visitors have been known to declare that they heard the strange plaintive cry out of doors at two o'clock in the morning.

Traces also of a very different story are reported by Mrs. Williams-Ellis, to the effect that when the water broke forth to form the lake, the fairies seized Grassi and changed her into a swan, and that she continued in that form to live on the lake sixscore years, and that when at length she died, she loudly lamented her lot: that cry is still to be heard at night. This story is in process apparently of being rationalized; at any rate the young lady student, to whom I have referred, remembers perfectly that her grandfather used to explain to her and the other children at home that Grassi was changed into a swan as a punishment for haunting Glasfryn, but that nevertheless the old lady still visited the place, especially when there happened to be strangers in the house. At the end of September last Mrs. Rhys and I had the pleasure of spending a few days at Glasfryn, in the hope of hearing the plaintive wail, and of seeing the lady in white silk revisiting her familiar haunts. But alas! our sleep was never once disturbed, nor was our peace once troubled by suspicions of anything uncanny. This, however, is negative, and characterized by the usual weakness of all such evidence.

It is now time to turn to another order of facts: in the first place may be mentioned that the young lady student's grandmother used to call the well Ffynnon Grâs Siôn Gruffud, as she had always heard that Grâs was the daughter of a certain Siôn Gruffyd, 'John Griffith,' who lived near the well; and Mrs. Williams-Ellis finds that Grâs was buried, at a very advanced age, on December 14, 1743, at the parish church of Llangybi, where the register describes her as Grace Jones, alias Grace Jones Griffith. She had lived till the end at Glasfryn, but from documents in the possession of the Glasfryn family it is known that in 1728 Hugh Lloyd of Trallwyn purchased the house and estate of Glasfryn from a son of Grace's, named John ab Cadwaladr, and that Hugh Lloyd of Trallwyn's son, the Rev. William Lloyd, sold them to Archdeacon Ellis, from whom they have descended to the Rev. J. C. Williams-Ellis. In the light of these facts there is no reason to connect the old lady's name very closely with the well or the lake. She was once the dominant figure at Glasfryn, that is all; and when she died she was as usual supposed to haunt the house and its immediate surroundings; and if we might venture to suppose that Glasfryn was sold by her son against her will, though subject to conditions which enabled her to remain in possession of the place to the day of her death, we should have a further explanation, perhaps, of her supposed moaning and lamentation.

In the background, however, of the story, one detects the possibility of another female figure, for it may be that the standing stone in Cae'r Ladi represents a woman buried there centuries before Grace ruled at Glasfryn, and that traditions about the earlier lady have survived to be inextricably mixed with those concerning the later one. Lastly, those traditions may have also associated the subject of them with the well and the lake; but I wish to attach no importance to this conjecture, as we have in reserve a third figure of larger possibilities than either Grace or the stone woman. It needs no better introduction than Mrs. Williams-Ellis' own words: 'Our younger boys have a crew of three little Welsh boys who live near the lake, to join them in their boat sailing about the pool and in camping on the island, &c. They asked me once who Morgan was, whom the little boys were always saying they were to be careful against. An old man living at Tal Llyn, "Lake's End," a farm close by, says that as a boy he was always told that "naughty boys would be carried off by Morgan into the lake." Others tell me that Morgan is always held to be ready to take off troublesome children, and somehow Morgan is thought of as a bad one.' Now as Morgan carries children off into the pool, he would seem to issue from the pool, and to have his home in it. Further, he plays the same part as the fairies against whom a Snowdonian mother used to warn her children: they were on no account to wander away from the house when there was a mist, lest the fairies should carry them to their home beneath Llyn Dwythwch. In other words, Morgan may be said to act in the same way as the mermaid, who takes a sailor down to her submarine home; and it explains to my mind a discussion which I once heard of the name Morgan by a party of men and women making hay one fine summer's day in the neighbourhood of Ponterwyd, in North Cardiganshire. I was a child, but I remember vividly how they teased one of their number whose 'style' was Morgan. They hinted at dreadful things associated with the name; but it was all so vague that I could not gather that his great unknown namesake was a thief, a murderer, or any kind of ordinary criminal. The impression left on my mind was rather the notion of something weird, uncanny, or non-human; and the fact that the Welsh version of the Book of Common Prayer calls the Pelagians Morganiaid, 'Morgans,' does not offer an adequate explanation. But I now see clearly that it is to be sought in the indistinct echo of such folklore as that which makes Morgan a terror to children in the neighbourhood of the Glasfryn Lake.

The name, however, presents points of difficulty which require some notice: the Welsh translators of Article IX in the Prayer Book were probably wrong in making Pelagians into Morganiaid, as the Welsh for Pelagius seems to have been rather Morien [164], which in its oldest recorded form was Morgen, and meant sea-born, or offspring of the sea. In a still earlier form it must have been Morigenos, with a feminine Morigena, but when the endings came to be dropped both vocables would become Morgen, later Morien. I do not remember coming across a feminine Morgen in Welsh, but the presumption is that it did exist. For, among other things, I may mention that we have it in Irish as Muirgen, one of the names of the lake lady Liban, who, when the waters of the neglected well rushed forth to form Lough Neagh, lived beneath that lake until she desired to be changed into a salmon. The same conclusion may be drawn from the name Morgain or Morgan, given in the French romances to one or more water ladies; for those names are easiest to explain as the Brythonic Morgen borrowed from a Welsh or Breton source, unless one found it possible to trace it direct to the Goidels of Wales. No sooner, however, had the confusion taken place between Morgen and the name which is so common in Wales as exclusively a man's name, than the aquatic figure must also become male. That is why the Glasfryn Morgan is now a male, and not a female like the other characters whose rôle he plays. But while the name was in Welsh successively Morgen and Morien, the man's name was Morcant, Morgant, or Morgan [165], so that, phonologically speaking, no confusion could be regarded as possible between the two series. Here, therefore, one detects the influence, doubtless, of the French romances which spoke of a lake lady Morgain, Morgan, or Morgue. The character varied: Morgain le Fay was a designing and wicked person; but Morgan was also the name of a well disposed lady of the same fairy kind, who took Arthur away to be healed at her home in the Isle of Avallon. We seem to be on the track of the same confusing influence of the name, when it occurs in the story of Geraint and Enid; for there the chief physician of Arthur's court is called Morgan Tut or Morgant Tut, and the word tut has been shown by M. Loth to have meant the same sort of non-human being whom an eleventh-century Life of St. Maudez mentions as quidam dæmon quem Britones Tuthe appellant. Thus the name Morgan Tut is meant as the Welsh equivalent of the French Morgain le Fay or Morgan la Fée [166]; but so long as the compiler of the story of Geraint and Enid employed in his Welsh the form Morgan, he had practically no choice but to treat the person called Morgan as a man, whether that was or was not the sex in the original texts on which he was drawing. Of course he could have avoided the difficulty in case he was aware of it, if he had found some available formula in use like Mary-Morgant, said to be a common name for a fairy on the island of Ouessant, off the coast of Brittany.

Summarizing the foregoing notes, we seem to be right in drawing the following conclusions:--(1) The well was left in the charge of a woman who forgot to shut it, and when she saw the water bursting forth, she bewailed her negligence, as in the case of her counterpart in the legend of Cantre'r Gwaelod. (2) The original name of the Glasfryn 'Morgan' was Morgen, later Morien. (3) The person changed into a swan on the occasion of the Glasfryn well erupting was not Grassi, but most probably Morgen. And (4) the character was originally feminine, like that of the mermaid or the fairies, whose rôle the Glasfryn Morgan plays; and more especially may one compare the Irish Muirgen, the Morgen more usually called Líban. For it is to be noticed that when the neglected well burst forth she, Muirgen or Líban, was not drowned like the others involved in the calamity, but lived in her chamber at the bottom of the lake formed by the overflowing well, until she was changed into a salmon. In that form she lived on some three centuries, until in fact she was caught in the net of a fisherman, and obtained the boon of a Christian burial. However, the change into a swan is also known on Irish ground: take for instance the story of the Children of Lir, who were converted into swans by their stepmother, and lived in that form on Loch Dairbhreach, in Westmeath, for three hundred years, and twice as long on the open sea, until their destiny closed with the advent of St. Patrick and the first ringing of a Christian bell in Erin [167].

The next legend was kindly communicated to me by Mr. Wm. Davies already mentioned at p. 147 above: he found it in Cyfaill yr Aelwyd [168], "The Friend of the Hearth," where it is stated that it belonged to David Jones' Storehouse of Curiosities, a collection which does not seem to have ever assumed the form of a printed book. David Jones, of Trefriw, in the Conwy Valley, was a publisher and poet who wrote between 1750 and 1780. This is his story: 'In 1735 I had a conversation with a man concerning Tegid Lake. He had heard from old people that near the middle of it there was a well opposite Llangower, and the well was called Ffynnon Gywer, "Cower's Well," and at that time the town was round about the well. It was obligatory to place a lid on the well every night. (It seems that in those days somebody was aware that unless this was done it would prove the destruction of the town.) But one night it was forgotten, and by the morning, behold the town had subsided and the lake became three miles long and one mile wide. They say, moreover, that on clear days some people see the chimneys of the houses. It is since then that the town was built at the lower end of the lake. It is called Y Bala [169], and the man told me that he had talked with an old Bala man who had, when he was a youth, had two days' mowing of hay [170] between the road and the lake; but by this time the lake had spread over that land and the road also, which necessitated the purchase of land further away for the road; and some say that the town will yet sink as far as the place called Llanfor--others call it Llanfawd, "Drown-church," or Llanfawr, "Great-church," in Penllyn.... Further, when the weather is stormy water appears oozing through every floor within Bala, and at other times anybody can get water enough for the use of his house, provided he dig a little into the floor of it.'

In reference to the idea that the town is to sink, together with the neighbouring village of Llanfor, the writer quotes in a note the couplet known still to everybody in the neighbourhood as follows:--

Y Bala aeth, a'r Bala aiff, A Llanfor aiff yn Llyn.

Bala old the lake has had, and Bala new The lake will have, and Llanfor too.

This probably implies that old Bala is beneath the lake, and that the present Bala is to meet the like fate at some time to come. This kind of prophecy is not very uncommon: thus there has been one current as to the Montgomeryshire town of Pool, called, in Welsh, Trallwng or Trallwm, and in English, Welshpool, to distinguish it from the English town of Pool. As to Welshpool, a very deep water called Llyn Du, lying between the town and the Castell Coch or Powys Castle, and right in the domain of the castle, is suddenly to spread itself, and one fine market day to engulf the whole place [171]. Further, when I was a boy in North Cardiganshire, the following couplet was quite familiar to me, and supposed to have been one of Merlin's prophecies:--

Caer Fyrdin, cei oer fore; Daear a'th lwnc, dw'r i'th le.

Carmarthen, a cold morn awaits thee; Earth gapes, and water in thy place will be.

In regard to the earlier half of the line, concerning Bala gone, the story of Ffynnon Gywer might be said to explain it, but there is another which is later and far better known. It is of the same kind as the stories related in Welsh concerning Llynclys and Syfadon; but I reserve it with these and others of the same sort for chapter vii.

For the next legend belonging here I have to thank the Rev. J. Fisher, a native of the parish of Llandybïe, who, in spite of his name, is a genuine Welshman, and--what is more--a Welsh scholar. The following are his words:--'Llyn Llech Owen (the last word is locally sounded w-en, like oo-en in English, as is also the personal name Owen) is on Mynyd Mawr, in the ecclesiastical parish of Gors Lâs, and the civil parish of Llanarthney, Carmarthenshire. It is a small lake, forming the source of the Gwendraeth Fawr. I have heard the tradition about its origin told by several persons, and by all, until quite recently, pretty much in the same form. In 1884 I took it down from my grandfather, Rees Thomas (b. 1809, d. 1892), of Cil Coll Llandebïe--a very intelligent man, with a good fund of old-world Welsh lore--who had lived all his life in the neighbouring parishes of Llandeilo Fawr and Llandybïe.

'The following is the version of the story (translated) as I had it from him:--There was once a man of the name of Owen living on Mynyd Mawr, and he had a well, "ffynnon." Over this well he kept a large flag ("fflagen neu lech fawr": "fflagen" is the word in common use now in these parts for a large flat stone), which he was always careful to replace over its mouth after he had satisfied himself or his beast with water. It happened, however, that one day he went on horseback to the well to water his horse, and forgot to put the flag back in its place. He rode off leisurely in the direction of his home; but, after he had gone some distance, he casually looked back, and, to his great astonishment, he saw that the well had burst out and was overflowing the whole place. He suddenly bethought him that he should ride back and encompass the overflow of the water as fast as he could; and it was the horse's track in galloping round the water that put a stop to its further overflow. It is fully believed that, had he not galloped round the flood in the way he did, the well would have been sure to inundate the whole district and drown all. Hence the lake was called the Lake of Owen's Flag, "Llyn Llech Owen."

'I have always felt interested in this story, as it resembled that about the formation of Lough Neagh, &c.; and, happening to meet the Rev. D. Harwood Hughes, B.A., the vicar of Gors Lâs (St. Lleian's), last August (1892), I asked him to tell me the legend as he had heard it in his parish. He said that he had been told it, but in a form different from mine, where the "Owen" was said to have been Owen Glyndwr. This is the substance of the legend as he had heard it:--Owen Glyndwr, when once passing through these parts, arrived here of an evening. He came across a well, and, having watered his horse, placed a stone over it in order to find it again next morning. He then went to lodge for the night at Dyllgoed Farm, close by. In the morning, before proceeding on his journey, he took his horse to the well to give him water, but found to his surprise that the well had become a lake.'

Mr. Fisher goes on to mention the later history of the lake: how, some eighty years ago, its banks were the resort on Sunday afternoons of the young people of the neighbourhood, and how a Baptist preacher put an end to their amusements and various kinds of games by preaching at them. However, the lake-side appears to be still a favourite spot for picnics and Sunday-school gatherings. Mr. Fisher was quite right in appending to his own version that of his friend; but, from the point of view of folklore, I must confess that I can make nothing of the latter: it differs from the older one as much as chalk does from cheese. It would be naturally gratifying to the pride of local topography to be able to connect with the pool the name of Owen Glyndwr; but it is worthy of note that this highly respectable attempt to rationalize the legend wholly fails, as it does not explain why there is now a lake where there was once but a well. In other words, the euhemerized story is itself evidence corroborative of Mr. Fisher's older version, which is furthermore kept in countenance by Howells' account, p. 104, where we are told who the Owen in question was, namely, Owen Lawgoch, a personage dear, as we shall see later, to the Welsh legend of the district. He and his men had their abode in a cave on the northern side of Mynyd Mawr, and while there Owen used, we are informed, to water his steed at a fine spring covered with a large stone, which it required the strength of a giant to lift. But one day he forgot to replace it, and when he next sought the well he found the lake. He returned to his cave and told his men what had happened. Thereupon both he and they fell into a sleep, which is to last till it is broken by the sound of a trumpet and the clang of arms on Rhiw Goch: then they are to sally forth to conquer.

Now the story as told by Howells and Fisher provokes comparison, as the latter suggests, with the Irish legend of the formation of Lough Ree and of Lough Neagh in the story of the Death of Eochaid McMaireda [172]. In both of these legends also there is a horse, a kind of water-horse, who forms the well which eventually overflows and becomes Lough Ree, and so with the still larger body of water known as Lough Neagh. In the latter case the fairy well was placed in the charge of a woman; but she one day left the cover of the well open, and the catastrophe took place--the water issued forth and overflowed the country. One of Eochaid's daughters, named Líban, however, was not drowned, but only changed into a salmon as already mentioned at p. 376 above. In my Arthurian Legend, p. 361, I have attempted to show that the name Líban may have its Welsh equivalent in that of Llïon, occurring in the name of Llyn Llïon, or Llïon's Lake, the bursting of which is described in the latest series of Triads, iii. 13, 97, as causing a sort of deluge. I am not certain as to the nature of the relationship between those names, but it seems evident that the stories have a common substratum, though it is to be noticed that no well, fairy or otherwise, figures in the Llyn Llïon legend, which makes the presence of the monster called the afanc the cause of the waters bursting forth. So Hu the Mighty, with his team of famous oxen, is made to drag the afanc out of the lake.

There is, however, another Welsh legend concerning a great overflow in which a well does figure: I allude to that of Cantre'r Gwaelod, or the Bottom Hundred, a fine spacious country supposed to be submerged in Cardigan Bay. Modern euhemerism treats it as defended by embankments and sluices, which, we are told, were in the charge of the prince of the country, named Seithennin, who, being one day in his cups, forgot to shut the sluices, and thus brought about the inundation, which was the end of his fertile realm. This, however, is not the old legend: that speaks of a well, and lays the blame on a woman--a pretty sure sign of antiquity, as the reader may judge from other old stories which will readily occur to him. The Welsh legend to which I allude is embodied in a short poem in the Black Book of Carmarthen [173]: it consists of eight triplets, to which is added a triplet from the Englynion of the Graves. The following is the original with a tentative translation:--

Seithenhin sawde allan. ac edrychuirde varanres mor. maes guitnev rytoes.

Boed emendiceid y morvin aehellygaut guydi cvin. finaun wenestir [174] mor terruin.

Boed emendiceid y vachteith. ae . golligaut guydi gueith. finaun wenestir mor diffeith.

Diaspad mererid y ar vann caer. hid ar duu y dodir. gnaud guydi traha trangc hir.

Diaspad mererid . y ar van kaer hetiv. hid ar duu y dadoluch. gnaud guydi traha attreguch.

Diaspad mererid am gorchuit heno. ac nimhaut gorlluit. gnaud guydi traha tramguit.

Diaspad mererid y ar gwinev kadir kedaul duv ae gorev. gnaud guydi gormot eissev.

Diaspad mererid . am kymhell heno y urth uyistauell. gnaud guydi traha trangc pell.

Bet seithenhin synhuir vann rug kaer kenedir a glan. mor maurhidic a kinran.

Seithennin, stand thou forth And see the vanguard of the main: Gwydno's plain has it covered.

Accursed be the maiden Who let it loose after supping, Well cup-bearer of the mighty main.

Accursed be the damsel Who let it loose after battle, Well minister of the high sea.

Mererid's cry from a city's height, Even to God is it directed: After pride comes a long pause.

Mererid's cry from a city's height to-day, Even to God her expiation: After pride comes reflection.

Mererid's cry o'ercomes me to-night, Nor can I readily prosper: After pride comes a fall.

Mererid's cry over strong wines, Bounteous God has wrought it: After excess comes privation.

Mererid's cry drives me to-night From my chamber away: After insolence comes long death.

Weak-witted Seithennin's grave is it Between Kenedyr's Fort and the shore, With majestic Mor's and Kynran's.

The names in these lines present great difficulties: first comes that of Mererid, which is no other word than Margarita, 'a pearl,' borrowed; but what does it here mean? Margarita, besides meaning a pearl, was used in Welsh, e.g. under the form Marereda [175], as the proper name written in English Margaret. That is probably how it is to be taken here, namely, as the name given to the negligent guardian of the fairy well. It cannot very well be, however, the name belonging to the original form of the legend; and we have the somewhat parallel case of Ffynnon Grassi, or Grace's Well; but what old Celtic name that of Mererid has replaced in the story, I cannot say. In the next place, nobody has been able to identify Caer Kenedyr, and I have nothing to say as to Mor Maurhidic, except that a person of that name is mentioned in another of the Englynion of the Graves. It runs thus in the Black Book, fol. 33a:--

Bet mor maurhidic diessic unben. post kinhen kinteic. mab peredur penwetic.

The grave of Mor the Grand, ... prince, Pillar of the ... conflict, Son of Peredur of Penwedig.

The last name in the final triplet of the poem which I have attempted to translate is Kinran, which is otherwise unknown as a Welsh name; but I am inclined to identify it with that of one of the three who escaped the catastrophe in the Irish legend. The name there is Curnán, which was borne by the idiot of the family, who, like many later idiots, was at the same time a prophet. For he is represented as always prophesying that the waters were going to burst forth, and as advising his friends to prepare boats. So he may be set, after a fashion, over against our Seithenhin synhuir vann, 'S. of the feeble mind.' But one might perhaps ask why I do not point out an equivalent in Irish for the Welsh Seithennin, as his name is now pronounced. The fact is that no such equivalent occurs in the Irish story in question, nor exactly, so far as I know, in any other.

That is what I wrote when penning these notes; but it has occurred to me since then, that there is an Irish name, an important Irish name, which looks as if related to Seithenhin, and that is Setanta Beg, 'the little Setantian,' the first name of the Irish hero Cúchulainn. The nt, I may point out, makes one suspect that Setanta is a name of Brythonic origin in Irish; and I have been in the habit of associating it with that of the people of the Setantii [176], placed by Ptolemy on the coast of what is now Lancashire. Whether any legend has ever been current about a country submerged on the coast of Lancashire I cannot say, but the soundings would make such a legend quite comprehensible. I remember, however, reading somewhere as to the Plain of Muirthemhne, of which Cúchulainn, our Setanta Beg, had special charge, that it was so called because it had once been submarine and become since the converse, so to say, of Seithennin's country. The latter is beneath Cardigan Bay, while the other fringed the opposite side of the sea, consisting as it did of the level portion of County Louth. On the whole, I am not altogether indisposed to believe that we have here traces of an ancient legend of a wider scope than is represented by the Black Book triplets, which I have essayed to translate. I think that I am right in recognizing that legend in the Mabinogi of Branwen, daughter of Llyr. There we read that, when Brân and his men crossed from Wales to Ireland, the intervening sea consisted merely of two navigable rivers, called Lli and Archan. The story-teller adds words to the effect, that it is only since then the sea has multiplied its realms [177] between Ireland and Ynys y Kedyrn, or the Isle of the Keiri, a name which has already been discussed: see pp. 279-83.

These are not all the questions which such stories suggest; for Seithennin is represented in later Welsh literature as the son of one Seithyn, associated with Dyfed; and the name Seithyn leads off to the coast of Brittany. For I learn from a paper by the late M. le Men, in the Revue Archéologique for 1872 (xxiii. 52), that the Île de Sein is called in Breton Enez-Sun, in which Sun is a dialectic shortening of Sizun, which is also met with as Seidhun. That being so, one would seem to be right in regarding Sizun as nearly related to our Seithyn. That is not all--the tradition reminds one of the Welsh legend: M. le Men refers to the Vie du P. Maunoir by Boschet (Paris, 1697) p. 126, and adds that, in his own time, the road ending on the Pointe du Raz opposite the Île de Sein passed 'pour être l'ancien chemin qui conduisait à la ville d'Is (Kaer-a-Is, la ville de la partie basse).' It is my own experience, that nobody can go about much in Brittany without hearing over and over again about the submerged city of Is. There is no doubt that we have in these names distant echoes of an inundation story, once widely current in both Britains and perhaps also in Ireland. With regard to Wales we have an indication to that effect in the fact, that Gwydno, to whom the inundated region is treated as having belonged, is associated not only with Cardigan Bay, but also with the coast of North Wales, especially the part of it situated between Bangor and Llandudno [178]. Adjoining it is supposed to lie submerged a once fertile district called Tyno Helig, a legend about which will come under notice later. This brings the inundation story nearer to the coast where Ptolemy in the second century located the Harbour of the Setantii, about the mouth of the river Ribble, and in their name we seem to have some sort of a historical basis for that of the drunken Seithennin [179]. I cannot close these remarks better than by appending what Professor Boyd Dawkins has recently said with regard to the sea between Britain and Ireland:--

'It may be interesting to remark further that during the time of the Iberian dominion in Wales, the geography of the seaboard was different to what it is now. A forest, containing the remains of their domestic oxen that had run wild, and of the indigenous wild animals such as the bear and the red deer, united Anglesey with the mainland, and occupied the shallows of Cardigan Bay, known in legend as "the lost lands of Wales." It extended southwards from the present sea margin across the estuary of the Severn, to Somerset, Devon, and Cornwall. It passed northwards across the Irish Sea off the coast of Cheshire and Lancashire, and occupied Morecambe Bay with a dense growth of oak, Scotch fir, alder, birch, and hazel. It ranged seawards beyond the ten-fathom line, and is to be found on most shores beneath the sand-banks and mud-banks, as for example at Rhyl and Cardiff. In Cardigan Bay it excited the wonder of Giraldus de Barri [180].'

To return to fairy wells, I have to confess that I cannot decide what may be precisely the meaning of the notion of a well with a woman set carefully to see that the door or cover of the well is kept shut. It will occur, however, to everybody to compare the well which Undine wished to have kept shut, on account of its affording a ready access from her subterranean country to the residence of her refractory knight in his castle above ground. And in the case of the Glasfryn Lake, the walling and cover that were to keep the spring from overflowing were, according to the story, not water-tight, seeing that there were holes made in one of the stones. This suggests the idea that the cover was to prevent the passage of some such full-grown fairies as those with which legend seems to have once peopled all the pools and tarns of Wales. But, in the next place, is the maiden in charge of the well to be regarded as priestess of the well? The idea of a priesthood in connexion with wells in Wales is not wholly unknown.

I wish, however, before discussing these instances, to call attention to one or two Irish ones which point in another direction. Foremost may be mentioned the source of the river Boyne, which is now called Trinity Well, situated in the Barony of Carbury, in County Kildare. The following is the Rennes Dindsenchas concerning it, as translated by Dr. Stokes, in the Revue Celtique, xv. 315-6:--'Bóand, wife of Nechtán son of Labraid, went to the secret well which was in the green of Síd Nechtáin. Whoever went to it would not come from it without his two eyes bursting, unless it were Nechtán himself and his three cup-bearers, whose names were Flesc and Lám and Luam. Once upon a time Bóand went through pride to test the well's power, and declared that it had no secret force which could shatter her form, and thrice she walked withershins round the well. (Whereupon) three waves from the well break over her and deprive her of a thigh [? wounded her thigh] and one of her hands and one of her eyes. Then she, fleeing her shame, turns seaward, with the water behind her as far as Boyne-mouth, (where she was drowned).' This is to explain why the river is called Bóand, 'Boyne.' A version to the same effect in the Book of Leinster, fol. 191a, makes the general statement that no one who gazed right into the well could avoid the instant ruin of his two eyes or otherwise escape with impunity. A similar story is related to show how the Shannon, in Irish Sinann, Sinand, or Sinend, is called after a woman of that name. It occurs in the same Rennes manuscript, and the following is Stokes' translation in the Revue Celtique, xv. 457:--'Sinend, daughter of Lodan Lucharglan son of Ler out of Tir Tairngire (Land of Promise, Fairyland), went to Connla's Well, which is under sea, to behold it. That is a well at which are the hazels and inspirations (?) of wisdom, that is, the hazels of the science of poetry, and in the same hour their fruit and their blossom and their foliage break forth, and these fall on the well in the same shower, which raises on the water a royal surge of purple. Then the salmon chew the fruit, and the juice of the nuts is apparent on their purple bellies. And seven streams of wisdom spring forth and turn there again. Now Sinend went to seek the inspiration, for she wanted nothing save only wisdom. She went with the stream till she reached Linn Mna Feile, "the Pool of the Modest Woman," that is Bri Ele--and she went ahead on her journey; but the well left its place, and she followed it [181] to the banks of the river Tarr-cáin, "Fair-back." After this it overwhelmed her, so that her back (tarr) went upwards, and when she had come to the land on this side (of the Shannon) she tasted death. Whence Sinann and Linn Mna Feile and Tarr-cain.'

In these stories the reader will have noticed that the foremost punishment on any intruder who looked into the forbidden well was the instant ruin of his two eyes. One naturally asks why the eyes are made the special objects of the punishment, and I am inclined to think the meaning to have originally been that the well or spring was regarded as the eye of the divinity of the water. Should this prove well founded it looks natural that the eyes, which transgressed by gazing into the eye of the divinity, should be the first objects of that divinity's vengeance. This is suggested to me by the fact that the regular Welsh word for the source of a river is llygad, Old Welsh licat, 'eye,' as for instance in the case of Licat Amir mentioned by Nennius, § 73; of Llygad Llychwr, 'the source of the Loughor river' in the hills behind Carreg Cennen Castle; and of the weird lake in which the Rheidol [182] rises near the top of Plinlimmon: it is called Llyn Llygad y Rheidol, 'the Lake of the Rheidol's Eye.' By the way, the Rheidol is not wholly without its folklore, for I used to be told in my childhood, that she and the Wye and the Severn sallied forth simultaneously from Plinlimmon one fine morning to run a race to the sea. The result was, one was told, that the Rheidol won great honour by reaching the sea three weeks before her bigger sisters. Somebody has alluded to the legend in the following lines:--

Tair afon gynt a rifwyd Ar dwyfron Pumlumon lwyd, Hafren a Gwy'n hyfryd ei gwed, A'r Rheidol fawr ei hanrhyded.

Three rivers of yore were seen On grey Plinlimmon's breast, Severn, and Wye of pleasant mien, And Rheidol rich in great renown.

To return to the Irish legends, I may mention that Eugene O'Curry has a good deal to say of the mysterious nuts and 'the salmon of knowledge,' the partaking of which was synonymous with the acquisition of knowledge and wisdom: see his Manners and Customs of the Ancient Irish, ii. 142-4. He gives it as his opinion that Connla's Well was situated somewhere in Lower Ormond; but the locality of this Helicon, with the seven streams of wisdom circulating out of it and back again into it, is more intelligible when regarded as a matter of fairy geography. A portion of the note appended to the foregoing legend by Stokes is in point here: he traces the earliest mention of the nine hazels of wisdom, growing at the heads of the chief rivers of Ireland, to the Dialogue of the Two Sages in the Book of Leinster, fol. 186b, whence he cites the poet Néde mac Adnai saying whence he had come, as follows:--a caillib .i. a nói collaib na Segsa ... a caillib didiu assa mbenaiter clessa na súad tanacsa, 'from hazels, to wit, from the nine hazels of the Segais ... from hazels out of which are obtained the feats of the sages, I have come.' The relevancy of this passage will be seen when I add, that Segais was one of the names of the mound in which the Boyne rises; so it may be safely inferred that Bóand's transgression was of the same nature as that of Sinand, to wit, that of intruding on sacred ground in quest of wisdom and inspiration which was not permitted their sex: certain sources of knowledge, certain quellen, were reserved for men alone.

Before I have done with the Irish instances I must append one in the form it was told me in the summer of 1894: I was in Meath and went to see the remarkable chambered cairns on the hill known as Sliabh na Caillighe, 'the Hag's Mountain,' near Oldcastle and Lough Crew. I had as my guide a young shepherd whom I picked up on the way. He knew all about the hag after whom the hill was called except her name: she was, he said, a giantess, and so she brought there, in three apronfuls, the stones forming the three principal cairns. As to the cairn on the hill point known as Belrath, that is called the Chair Cairn from a big stone placed there by the hag to serve as her seat when she wished to have a quiet look on the country round. But usually she was to be seen riding on a wonderful pony she had: that creature was so nimble and strong that it used to take the hag at a leap from one hill-top to another. However, the end of it all was that the hag rode so hard that the pony fell down, and that both horse and rider were killed. The hag appears to have been Cailleach Bhéara, or Caillech Bérre, 'the Old Woman of Beare,' that is, Bearhaven, in County Cork [183]. Now the view from the Hag's Mountain is very extensive, and I asked the shepherd to point out some places in the distance. Among other things we could see Lough Ramor, which he called the Virginia Water, and more to the west he identified Lough Sheelin, about which he had the following legend to tell:--A long, long time ago there was no lake there, but only a well with a flagstone kept over it, and everybody would put the flag back after taking water out of the well. But one day a woman who fetched water from it forgot to replace the stone, and the water burst forth in pursuit of the luckless woman, who fled as hard as she could before the angry flood. She continued until she had run about seven miles--the estimated length of the lake at the present day. Now at this point a man, who was busily mowing hay in the field through which she was running, saw what was happening and mowed the woman down with his scythe, whereupon the water advanced no further. Such was the shepherd's yarn, which partly agrees with the Boyne and Shannon stories in that the woman was pursued by the water, which only stopped where she died. On the other hand, it resembles the Llyn Llech Owen legend and that of Lough Neagh in placing to the woman's charge only the neglect to cover the well. It looks as if we had in these stories a confusion of two different institutions, one being a well of wisdom which no woman durst visit without fatal vengeance overtaking her, and the other a fairy well which was attended to by a woman who was to keep it covered, and who may, perhaps, be regarded as priestess of the spring. If we try to interpret the Cantre'r Gwaelod story from these two points of view we have to note the following matters:--Though it is not said that the moruin, or damsel, had a lid or cover on the well, the word golligaut or helligaut, 'did let run,' implies some such an idea as that of a lid or door; for opening the sluices, in the sense of the later version, seems to me out of the question. In two of the Englynion she is cursed for the action implied, and if she was the well minister or well servant, as I take finaun wenestir to mean, we might perhaps regard her as the priestess of that spring. On the other hand, the prevailing note in the other Englynion is the traha, 'presumption, arrogance, insolence, pride,' which forms the burden of four out of five of them. This would seem to point to an attitude on the part of the damsel resembling that of Bóand or Sinand when prying into the secrets of wells which were tabu to them. The seventh Englyn alludes to wines, and its burden is gormod, 'too much, excess, extravagance,' whereby the poet seems to lend countenance to some such a later story as that of Seithennin's intemperance.

Lastly, the question of priest or priestess of a sacred well has been alluded to once or twice, and it may be perhaps illustrated on Welsh ground by the history of Ffynnon Eilian, or St. Elian's Well, which has been mentioned in another context, p. 357 above. Of that well we read as follows, s. v. Llandrillo, in the third edition of Lewis' Topographical Dictionary of Wales:--'Fynnon Elian, ... even in the present age, is frequently visited by the superstitious, for the purpose of invoking curses upon the heads of those who have grievously offended them, and also of supplicating prosperity to themselves; but the numbers are evidently decreasing. The ceremony is performed by the applicant standing upon a certain spot near the well, whilst the owner of it reads a few passages of the sacred Scriptures, and then, taking a small quantity of water, gives it to the former to drink, and throws the residue over his head, which is repeated three times, the party continuing to mutter imprecations in whatever terms his vengeance may dictate.' Rice Rees, in his Essay on the Welsh Saints (London, 1836), p. 267, speaks of St. Elian as follows: 'Miraculous cures were lately supposed to be performed at his shrine at Llanelian, Anglesey; and near to the church of Llanelian, Denbighshire, is a well called Ffynnon Elian, which is thought by the peasantry of the neighbourhood to be endued with miraculous powers even at present.'

Foulkes, s. v. Elian, in his Enwogion Cymru, published in Liverpool in 1870, expresses the opinion that the visits of the superstitious to the well had ceased for some time. The last person supposed to have had charge of the well was a certain John Evans, but some of the most amusing stories of the shrewdness of the caretaker refer to a woman who had charge of the well before Evans' time. A series of articles on Ffynnon Eilian appeared in 1861 in a Welsh periodical called Y Nofelyd, printed by Mr. Aubrey at Llanerch y Med, in Anglesey. The articles in question were afterwards published, I am told, as a shilling book, which I have not seen, and they dealt with the superstition, with the history of John Evans, and with his confessions and conversion. I have searched in vain for any account in Welsh of the ritual followed at the well. When Mrs. Silvan Evans visited the place, the person in charge of the well was a woman, and Peter Roberts, in his Cambrian Popular Antiquities, published in London in 1815, alludes to her or a predecessor of hers in the following terms, p. 246:--'Near the Well resided some worthless and infamous wretch, who officiated as priestess.' He furthermore gives one to understand that she kept a book in which she registered the name of each evil wisher for a trifling sum of money. When this had been done, a pin was dropped into the well in the name of the victim. This proceeding looks adequate from the magical point of view, though less complicated than the ritual indicated by Lewis. This latter writer calls the person who took charge of the well the owner; and I have always understood that, whether owner or not, he or she used to receive gifts, not only for placing in the well the names of men who were to be cursed, but also from those men for taking their names out again, so as to relieve them from the malediction. In fact, the trade in curses seems to have been a very thriving one: its influence was powerful and widespread.

Here there is, I think, very little doubt that the owner or guardian of the well was, so to say, the representative of an ancient priesthood of the well. That priesthood dated its origin probably many centuries before a Christian church was built near the well, and coming down to later times we have unfortunately no sufficient data to show how the right to such priesthood was acquired, whether by inheritance or otherwise; but we know that a woman might have charge of St. Elian's Well.

Let me cite another instance, which I unexpectedly discovered some years ago in the course of a ramble in quest of early inscriptions. Among other places which I visited was Llandeilo Llwydarth, near Maen Clochog, in the northern part of Pembrokeshire. This is one of the many churches bearing the name of St. Teilo in South Wales: the building is in ruins, but the churchyard is still used, and contains two of the most ancient post-Roman inscriptions in the Principality. If you ask now for 'Llandeilo' in this district, you will be understood to be inquiring after the farm house of that name, close to the old church; and I learnt from the landlady that her family had been there for many generations, though they have not very long been the proprietors of the land. She also told me of St. Teilo's Well, a little above the house: she added that it was considered to have the property of curing the whooping-cough. I asked if there was any rite or ceremony necessary to be performed in order to derive benefit from the water. Certainly, I was told: the water must be lifted out of the well and given to the patient to drink by some member of the family. To be more accurate, I ought to say that this must be done by somebody born in the house. Her eldest son, however, had told me previously, when I was busy with the inscriptions, that the water must be given to the patient by the heir, not by anybody else. Then came my question how the water was lifted, or out of what the patient had to drink, to which I was answered that it was out of the skull. 'What skull?' said I. 'St. Teilo's skull,' was the answer. 'Where do you get the saint's skull?' I asked. 'Here it is,' was the answer, and I was given it to handle and examine. I know next to nothing about skulls; but it struck me that it was the upper portion of a thick, strong skull, and it called to my mind the story of the three churches which contended for the saint's corpse. That story will be found in the Book of Llan Dâv, pp. 116-7, and according to it the contest became so keen that it had to be settled by prayer and fasting. So, in the morning, lo and behold! there were three corpses of St. Teilo--not simply one--and so like were they in features and stature that nobody could tell which were the corpses made to order and which the old one. I should have guessed that the skull which I saw belonged to the former description, as not having been much thinned by the owner's use of it; but this I am forbidden to do by the fact that, according to the legend, this particular Llandeilo was not one of the three contending churches which bore away in triumph a dead Teilo each. The reader, perhaps, would like to take another view, namely, that the story has been edited in such a way as to reduce a larger number of Teilos to three, in order to gratify the Welsh weakness for triads.

Since my visit to the neighbourhood I have been favoured with an account of the well as it is now current there. My informant is Mr. Benjamin Gibby of Llangolman Mill, who writes mentioning, among other things, that the people around call the well Ffynnon yr Ychen, or the Oxen's Well, and that the family owning and occupying the farm house of Llandeilo have been there for centuries. Their name, which is Melchior (pronounced Melshor), is by no means a common one in the Principality, so far as I know; but, whatever may be its history in Wales, the bearers of it are excellent Kymry. Mr. Gibby informs me that the current story solves the difficulty as to the saint's skull as follows:--The saint had a favourite maid servant from the Pembrokeshire Llandeilo: she was a beautiful woman, and had the privilege of attending on the saint when he was on his death-bed. As his end was approaching he gave his maid a strict and solemn command that in a year's time from the day of his burial at Llandeilo Fawr, in Carmarthenshire, she was to take his skull to the other Llandeilo, and to leave it there to be a blessing to coming generations of men, who, when ailing, would have their health restored by drinking water out of it. So the belief prevailed that to drink out of the skull some of the water of Teilo's Well ensured health, especially against the whooping-cough. The faith of some of those who used to visit the well was so great in its efficacy, that they were wont to leave it, he says, with their constitutions wonderfully improved; and he mentions a story related to him by an old neighbour, Stifyn Ifan, who has been dead for some years, to the effect that a carriage, drawn by four horses, came once, more than half a century ago, to Llandeilo. It was full of invalids coming from Pen Clawd, in Gower, Glamorganshire, to try the water of the well. They returned, however, no better than they came; for though they had drunk of the well, they had neglected to do so out of the skull. This was afterwards pointed out to them by somebody, and they resolved to make the long journey to the well again. This time they did the right thing, we are told, and departed in excellent health.

Such are the contents of Mr. Gibby's Welsh letter; and I would now only point out that we have here an instance of a well which was probably sacred before the time of St. Teilo: in fact, one would possibly be right in supposing that the sanctity of the well and its immediate surroundings was one of the causes why the site was chosen by a Christian missionary. But consider for a moment what has happened: the well paganism has annexed the saint, and established a belief ascribing to him the skull used in the well ritual. The landlady and her family, it is true, neither believe in the efficacy of the well, nor take gifts from those who visit the well; but they continue, out of kindness, as they put it, to hand the skull full of water to any one who perseveres in believing in it. In other words, the faith in the well continues in a measure intact, while the walls of the church have long fallen into utter decay. Such is the great persistence of some primitive beliefs; and in this particular instance we have a succession which seems to point unmistakably to an ancient priesthood of a sacred spring.

NOTES

[1] As to the spelling of Welsh names, it may be pointed out for the benefit of English readers that Welsh f has the sound of English v, while the sound of English f is written ff (and ph) in Welsh, and however strange it may seem to them that the written f should be sounded v, it is borrowed from an old English alphabet which did so likewise more or less systematically. Th in such English words as thin and breath is written th, but the soft sound as in this and breathe is usually printed in Welsh dd and written in modern Welsh manuscript sometimes like a small Greek delta: this will be found represented by d in the Welsh extracts edited by me in this volume.--J. R.

[2] 'Blaensawde, or the upper end of the river Sawde, is situate about three-quarters of a mile south-east from the village of Llandeusant. It gives its name to one of the hamlets of that parish. The Sawde has its source in Llyn y Fan Fach, which is nearly two miles distant from Blaensawde House.'

[3] The rendering might be more correctly given thus: 'O thou of the crimped bread, it is not easy to catch me.'--J. R.

[4] 'Mydfai parish was, in former times, celebrated for its fair maidens, but whether they were descendants of the Lady of the Lake or otherwise cannot be determined. An old pennill records the fact of their beauty thus:--

Mae eira gwyn Ar ben y bryn, A'r glasgoed yn y Ferdre, Mae bedw mân Ynghoed Cwm-brân, A merched glân yn Mydfe.

Which may be translated,

There is white snow On the mountain's brow, And greenwood at the Verdre, Young birch so good In Cwm-brân wood, And lovely girls in Mydfe.'

[5] Similarly this should be rendered: 'O thou of the moist bread, I will not have thee.'--J. R.

[6] In the best Demetian Welsh this word would be hwedel, and in the Gwentian of Glamorgan it is gwedel, mutated wedel, as may be heard in the neighbourhood of Bridgend.--J. R.

[7] This is not generally accepted, as some Welsh antiquarians find reasons to believe that Dafyd ap Gwilym was buried at Strata Florida.--J. R.

[8] This is not quite correct, as I believe that Dr. C. Rice Williams, who lives at Aberystwyth, is one of the Medygon. That means the year 1881, when this chapter was written, excepting the portions concerning which the reader is apprised of a later date.--J. R.

[9] Later it will be seen that the triban in the above form was meant for neither of the two lakes, though it would seem to have adapted itself to several. In the case of the Fan Fach Lake the town meant must have been Carmarthen, and the couplet probably ran thus:

Os na cha'i lonyd yn ym lle, Fi foda dre' Garfyrdin.

[10] Llwch is the Goidelic word loch borrowed, and Llyn Cwm y Llwch literally means the Lake of the Loch Dingle.

[11] I make no attempt to translate these lines, but I find that Mr. Llewellyn Williams has found a still more obscure version of them, as follows:--

Prw med, prw med, prw'r gwartheg i dre', Prw milfach a malfach, pedair llualfach, Llualfach ac Acli, pedair lafi, Lafi a chromwen, pedair nepwen, Nepwen drwynog, brech yn llyn a gwaun dodyn, Tair bryncethin, tair cyffredin, Tair caseg du, draw yn yr eithin; Dewch i gyd i lys y brenin.

[12] The Ty-fry is a house said to be some 200 years old, and situated about two miles from Rhonda Fechan: more exactly it is about one-fourth of a mile from the station of Ystrad Rhonda, and stands at the foot of Mynyd yr Eglwys on the Treorky side. It is now surrounded by the cottages of colliers, one of whom occupies it. For this information I have to thank Mr. Probert Evans.

[13] It is to be borne in mind that the sound of h is uncertain in Glamorgan pronunciation, whether the language used is Welsh or English. The pronunciation indicated, however, by Mr. Evans comes near enough to the authentic form written Elfarch.

[14] In the Snowdon district of Gwyned the call is drwi, drwi, drw-i bach, while in North Cardiganshire it is trwi, trwi, trw-e fach, also pronounced sometimes with a surd r, produced by making the breath cause both lips to vibrate--tR'wi, tR'wi, which can hardly be distinguished from pR'wi, pR'wi. For the more forcibly the lips are vibrated the more difficult it becomes to start by closing them to pronounce p: so the tendency with R' is to make the preceding consonant into some kind of a t.

[15] This is the Welsh form of the borrowed name Jane, and its pronunciation in North Cardiganshire is Siân, with si pronounced approximately like the ti of such French words as nation and the like; but of late years I find the si made into English sh under the influence, probably, to some extent of the English taught at school. This happens in North Wales, even in districts where there are still plenty of people who cannot approach the English words fish and shilling nearer than fiss and silling. Siôn and Siân represent an old importation of English John and Jane, but they are now considered old-fashioned and superseded by John and Jane, which I learned to pronounce Dsiòn and Dsiên, except that Siôn survives as a family name, written Shone, in the neighbourhood of Wrexham.

[16] This term dafad (or dafaden), 'a sheep,' also used for 'a wart,' and dafad (or dafaden) wyllt, literally 'a wild sheep,' for cancer or epithelioma, raises a question which I am quite unable to answer: why should a wart have been likened to a sheep?

[17] The name is probably a shortening of Cawellyn, and that perhaps of Cawell-lyn, 'Creel or Basket Lake.' Its old name is said to have been Llyn Tardenni.

[18] Tyn is a shortening of tydyn, which is not quite forgotten in the case of Tyn Gadlas or Tyn Siarlas (for Tydyn Siarlys), 'Charles' Tenement,' in the immediate neighbourhood. Similarly the Anglesey Farm of Tyn yr Onnen used at one time to be Tydyn yr Onnen in the books of Jesus College, Oxford, to which it belongs.

[19] That is the pronunciation which I have learnt at Llanberis, but there is another, which I have also heard, namely Derwenyd.

[20] Ystrad is the Welsh corresponding to Scotch strath, and it is nearly related to the English word strand. It means the flat land near a river.

[21] Betws (or Bettws) Garmon seems to mean Germanus's Bede-hus or House of Prayer, but Garmon can hardly have come down in Welsh from the time of the famous saint in the fifth century, as it would then have probably yielded Gerfon and not Garmon: it looks as if it had come through the Goidelic of this country.

[22] One of the rare merits of our Welsh bards is their habit of assuming permanent noms de plume, by means of which they prevent a number of excellent native names from falling into utter oblivion in the general chaos of Anglo-Hebrew ones, such as Jones, Davies, and Williams, which cover the Principality. Welsh place-names have similarly been threatened by Hebrew names of chapels, such as Bethesda, Rehoboth, and Jerusalem, but in this direction the Jewish mania has only here and there effected permanent mischief.

[23] The Brython was a valuable Welsh periodical published by Mr. Robert Isaac Jones, at Tremadoc, in the years 1858-1863, and edited by the Rev. Chancellor Silvan Evans, who was then the curate of Llangïan in Lleyn: in fact he was curate for fourteen years! His excellent work in editing the Brython earned for him his diocesan's displeasure, but it is easier to imagine than to describe how hard it was for him to resign the honorarium of £24 derived from the Brython when his stipend as a clergyman was only £92, at the same time that he had dependent on him a wife and six children. However much some people affect to laugh at the revival of the national spirit in Wales, we have, I think, got so far as to make it, for some time to come, impossible for a Welsh clergyman to be snubbed on account of his literary tastes or his delight in the archæology of his country.

[24] This parish is called after a saint named Tegái or Tygái, like Tyfaelog and Tysilio, and though the accent rests on the final syllable nothing could prevent the grammarian Huw Tegai and his friends from making it into Tégai in Huw's name.

[25] For can they now usually put Ann, and Mr. Hughes remembers hearing it so many years ago.

[26] I remember seeing a similar mound at Llanfyrnach, in Pembrokeshire; and the last use made of the hollow on the top of this also is supposed to have been for cock-fights.

[27] My attention has also been called to freit, frete, freet, fret, 'news, inquiry, augury,' corresponding to Anglo-Saxon freht, 'divination.' But the disparity of meaning seems to stand in the way of our ffrit being referred to this origin.

[28] The Oxford Mabinogion, p. 63; Guest, iii. 223.

[29] See the Itinerarium Kambriæ, i. 2 (pp. 33-5), and Celtic Britain, p. 64.

[30] As for example in the Archæologia Cambrensis for 1870, pp. 192-8; see also 1872, pp. 146-8.

[31] Howells has also an account of Llyn Savadhan, as he writes it: see his Cambrian Superstitions, pp. 100-2, where he quaintly says that the story of the wickedness of the ancient lord of Syfadon is assigned as the reason why 'the superstitious little river Lewenny will not mix its water with that of the lake.' Lewenny is a reckless improvement of Mapes' Leueni (printed Lenem); and Giraldus' Clamosum implies an old spelling Llefni, pronounced the same as the later spelling Llyfni, which is now made into Llynfi or Llynvi: the river so called flows through the lake and into the Wye at Glasbury. As to Safadan or Syfadon, it is probably of Goidelic origin, and to be identified with such an Irish name as the feminine Samthann: see Dec. 19 in the Martyrologies. To keep within our data, we are at liberty to suppose that this was the name of the wicked princess in the story, and that she was the ancestress of a clan once powerful on and around the lake, which lies within a Goidelic area indicated by its Ogam inscriptions.

[32] These were held, so far as I can gather from the descriptions usually given of them, exactly as I have seen a kermess or kirchmesse celebrated at Heidelberg, or rather the village over the Neckar opposite that town. It was in 1869, but I forget what saint it was with whose name the kermess was supposed to be connected: the chief features of it were dancing and beer drinking. It was by no means unusual for a Welsh Gwyl Fabsant to bring together to a rural neighbourhood far more people than could readily be accommodated; and in Carnarvonshire a hurriedly improvised bed is to this day called gwely g'l'absant, as it were 'a bed (for the time) of a saint's festival.' Rightly or wrongly the belief lingers that these merry gatherings were characterized by no little immorality, which made the better class of people set their faces against them.

[33] Since the editing of this volume was begun I have heard that it is intended to publish the Welsh collection which Mr. Jones has made: so I shall only give a translation of the Edward Llwyd version of the afanc story: see section v. of this chapter.

[34] This word is not in Welsh dictionaries, but it is Scotch and Manx Gaelic, and is possibly a remnant of the Goidelic once spoken in Gwyned.

[35] Our charlatans never leave off trying to make this into Tryfaen so as to extract maen, 'stone,' from it. They do not trouble themselves to find out whether it ever was Tryfaen or not: in fact they rather like altering everything as much as they can.

[36] Ystrádllyn, with the accent on the penult, is commonly pronounced Strállyn, and means 'the strand of the lake,' and the hollow is named after it Cwm Strállyn, and the lake in it Llyn Cwm Strállyn, which literally means 'the Lake of the Combe of the Strand of the Lake'--all seemingly for the luxury of forgetting the original name of the lake, which I have never been able to ascertain.

[37] So Mr. Jones puts it: I have never heard of any other part of the Principality where the children are usually baptized before they are eight days old.

[38] I cannot account for this spelling, but the ll in Bellis is English ll, not the Welsh ll, which represents a sound very different from that of l.

[39] Where not stated otherwise, as in this instance, the reader is to regard this chapter as written in the latter part of the year 1881.

[40] See Giraldus' Itinerarium Kambriæ, i. 8 (pp. 75-8); some discussion of the whole story will be found in chapter iii of this volume.

[41] Dr. Moore explains this to be cabbages and potatoes, pounded and mixed with butter or lard.

[42] It would be interesting to know what has become of this letter and others of Llwyd's once in the possession of the canon, for it is not to be supposed that the latter ever took the trouble to make an accurate copy of them any more than he did of any other MSS.

[43] There is also a Sarn yr Afanc, 'the Afanc's Stepping Stones,' on the Ogwen river in Nant Ffrancon: see Pennant's Tours in Wales, iii. 101.

[44] The oxen should accordingly have been called Ychain Pannog; but the explanation is not to be taken seriously. These oxen will come under the reader's notice again, to wit in chapter x.

[45] The lines are copied exactly as given at p. 189 (I. vi. 25-30) of The Poetical Works of Lewis Glyn Cothi, edited for the Cymmrodorion by Gwallter Mechain and Tegid, and printed at Oxford in the year 1837.

[46] This, I should say, must be a mistake, as it contradicts all the folklore which makes the rowan an object of dread to the fairies.

[47] See Choice Notes from 'Notes and Queries' (London, 1859), p. 147.

[48] It is more likely that it is a shortening of Llyn y Barfog, meaning the Lake of the Bearded One, Lacus Barbati as it were, the Bearded One being somebody like the hairy monster of another lake mentioned at p. 18 above, or him of the white beard pictured at p. 127.

[49] So far from afanc meaning a crocodile, an afanc is represented in the story of Peredur as a creature that would cast at every comer a poisoned spear from behind a pillar standing at the mouth of the cave inhabited by it; see the Oxford Mabinogion, p. 224. The corresponding Irish word is abhac, which according to O'Reilly means 'a dwarf, pigmy, manikin; a sprite.'

[50] I should not like to vouch for the accuracy of Mr. Pughe's rendering of this and the other Welsh names which he has introduced: that involves difficult questions.

[51] The writer meant the river known as Dyfi or Dovey; but he would seem to have had a water etymology on the brain.

[52] This involves the name of the river called Disynni, and Diswnwy embodies a popular etymology which is not worth discussing.

[53] It would, I think, be a little nearer the mark as follows:--

Come thou, Einion's Yellow One, Stray-horns, the Particoloured Lake Cow, And the Hornless Dodin: Arise, come home.

But one would like to know whether Dodin ought not rather to be written Dodyn, to rhyme with Llyn.

[54] Hywel's real name is William Davies, Tal y Bont, Cardiganshire. As adjudicator I became acquainted with several stories which Mr. Davies has since given me permission to use, and I have to thank him for clues to several others.

[55] Or Llech y Deri, as Mr. Williams tells me in a letter, where he adds that he does not know the place, but that he took it to be in the Hundred of Cemmes, in North-west Pembrokeshire. I take Llech y Derwyd to be fictitious; but I have not succeeded in finding any place called by the other name either.

[56] Perhaps the more usual thing is for the man returning from Faery to fall into dust on the spot: see later in this chapter the Curse of Pantannas, which ends with an instance in point, and compare Howells, pp. 142, 146.

[57] B. Davies, that is, Benjamin Davies, who gives this tale, was, as I learn from Gwynionyd, a native of Cenarth. He was a schoolmaster for about twelve years, and died in October, 1859, at Merthyr, near Carmarthen: he describes him as a good and intelligent man.

[58] This is ordinarily written Cenarth, the name of a parish on the Teifi, where the three counties of Cardigan, Pembroke, and Carmarthen meet.

[59] The name Llan Dydoch occurs in the Bruts, A.D. 987 and 1089, and is the one still in use in Welsh; but the English St. Dogmael's shows that it is derived from that of Dogfael's name when the mutation consonant f or v was still written m. In Welsh the name of the saint has been worn down to Dogwel, as in St. Dogwell's near Fishguard, and Llandogwel in Llanrhudlad parish in Anglesey: see Reece's Welsh Saints, p. 211. It points back to an early Brythonic form Doco-maglos, with doco of the same origin as Latin dux, ducis, 'a leader,' and maglo-s = Irish mal, 'a lord or prince.' Dogfael's name assumes in Llan Dydoch a Goidelic form, for Dog-fael would have to become in Irish Doch-mhal, which, cut down to Doch with the honorific prefix to, has yielded Ty-doch; but I am not clear why it is not Ty-doch. Another instance of a Goidelic form of a name having the local preference in Wales to this day offers itself in Cyfelach and Llan Gyfelach in Glamorganshire. The Welsh was formerly Cimeliauc (Reece, p. 274). Here may also be mentioned St. Cyngar, otherwise called Docwinnus (Reece, p. 183), but the name occurs in the Liber Landavensis in the genitive both as Docunn-i and Docguinni, the former of which seems easily explained as Goidelic for an early form of Cyngar, namely Cuno-caros, from which would be formed To-chun or Do-chun. This is what seems to underlie the Latin Docunnus, while Docguinni is possibly a Goidelic modification of the written Docunni, unless some such a name as Doco-vindo-s has been confounded with Docunnus. In one instance the Book of Llan Dâv has instead of Abbas Docunni or Docguinni, the shorter designation, Abbas Dochou (p. 145), which one must not unhesitatingly treat as Dochon, seeing that Dochou would be in later book Welsh Dochau, and in the dialect of the district Docha; and that this occurs in the name of the church of Llandough near Cardiff, and Llandough near Cowbridge. The connexion of a certain saint Dochdwy with these churches does not appear at all satisfactorily established, but more light is required to help one to understand these and similar church names.

[60] This name which may have come from Little England below Wales, was once not uncommon in South Cardiganshire, as Mr. Williams informs me, but it is now mostly changed as a surname into Davies and Jones! Compare the similar fortunes of the name Mason mentioned above, p. 68.

[61] I have not succeeded in discovering who the writer was, who used this name.

[62] This name as it is now written should mean 'the Gold's Foot,' but in the Demetian dialect aur is pronounced oer, and I learn from the rector, the Rev. Rhys Jones Lloyd, that the name has sometimes been written Tref Deyrn, which I regard as some etymologist's futile attempt to explain it. More importance is to be attached to the name on the communion cup, dating 1828, and reading, as Mr. Lloyd kindly informs me, Poculum Eclyseye de Tre-droyre. Beneath Droyre some personal name possibly lies concealed.

[63] Y Ferch o Gefn Ydfa ('The Maid of Cefn Ydfa'), by Isaac Craigfryn Hughes, published by Messrs. Daniel Owen, Howell & Co., Cardiff, 1881.

[64] In a letter dated February 9, 1899, he states, however, that as regards folklore the death of his father at the age of seventy-six, in the year 1889, had been a great loss to him; for he adds that he was perfectly familiar with the traditions of the neighbourhood and had associated with older men. Among the latter he had been used to talk with an old man whose father remembered Cromwell passing on his way to destroy the Iron Works of Pant y Gwaith, where the Cavaliers had had a cannon cast, which was afterwards used in the engagement at St. Fagan's.

[65] This term is sometimes represented as being Bendith eu Mamau, 'their Mother's Blessing,' as if each fairy were such a delightful offspring as to constitute himself or herself a blessing to his or her mother; but I have not found satisfactory evidence to the currency of Bendith eu Mamau, or, as it would be pronounced in Glamorgan, Béndith i Máma. On the whole, therefore, perhaps one may regard the name as pointing back to the Celtic goddesses known in Gaul in Roman times as the Mothers.

[66] On Pen Craig Daf Mr. Hughes gives the following note:--It was the residence of Dafyd Morgan or 'Counsellor Morgan,' who, he says, was executed on Kennington Common for taking the side of the Pretender. He had retreated to Pen y Graig, where his abode was, in order to conceal himself; but he was discovered and carried away at night. Here follows a verse from an old ballad about him:--

Dafyd Morgan ffel a ffol, Taffy Morgan, sly and daft, Fe aeth yn ol ei hyder: He did his bent go after: Fe neidod naid at rebel haid He leaped a leap to a rebel swarm, Pan drod o blaid Pretender. To arm for a Pretender.

[67] A tòn is any green field that is used for grazing and not meant to be mown, land which has, as it were, its skin of grassy turf unbroken for years by the plough.

[68] On this Mr. Hughes has a note to the effect that the whole of one milking used to be given in Glamorgan to workmen for assistance at the harvest or other work, and that it was not unfrequently enough for the making of two cheeses.

[69] Since this was first printed I have learnt from Mr. Hughes that the first cry issued from the Black Cauldron in the Taff (o'r Gerwyn Du ar Daf), which I take to be a pool in that river.

[70] The Fan is the highest mountain in the parish of Merthyr Tydfil, Mr. Hughes tells me: he adds that there was on its side once a chapel with a burial ground. Its history seems to be lost, but human bones have, as he states, been frequently found there.

[71] The above, I am sorry to say, is not the only instance of this nasty trick associating itself with Gwent, as will be seen from the story of Bwca'r Trwyn in chapter x.

[72] This chapter, except where a later date is suggested, may be regarded as written in the summer of 1883.

[73] Trefriw means the town of the slope or hillside, and stands for Tref y Riw, not tref y Rhiw, which would have yielded Treffriw, for there is a tendency in Gwyned to make the mutation after the definite article conform to the general rule, and to say y law, 'the hand,' and y raw, 'the spade,' instead of what would be in books y llaw and y rhaw from yr llaw and yr rhaw.

[74] Why the writer spells the name Criccieth in this way I cannot tell, except that he was more or less under the influence of the more intelligible spelling Crugcaith, as where Lewis Glyn Cothi. I. xxiv, sang

Rhys ab Sion â'r hysbys iaith, Gwr yw acw o Grugcaith.

This spelling postulates the interpretation Crug-Caith, earlier Crug y Ceith, 'the mound or barrow of the captives,' in reference to some forgotten interment; but when the accent receded to the first syllable the second was slurred almost out of recognition, so that Crug-ceith, or Cruc-ceith, became Crúceth, whence Crúcieth and Cricieth. The Bruts have Crugyeith the only time it occurs, and the Record of Carnarvon (several times) Krukyth.

[75] Out of excessive fondness for our Arthur English people translate this name into Arthur's Seat instead of Idris' Seat; but Idris was also somebody: he was a giant with a liking for the study of the stars. But let that be: I wish to say a word concerning his name: Idris may be explained as meaning 'War-champion,' or the like; and, phonologically speaking, it comes from Iud-rys, which was made successively into Id-rys, Idris. The syllable iud meant battle or fight, and it undergoes a variety of forms in Welsh names. Thus before n, r, l, and w, it becomes id, as in Idnerth, Idloes, and Idwal, while Iud-hael yields Ithel, whence Ab Ithel, anglicized Bethel. At the end, however, it is yd or ud, as in Gruffud or Gruffyd, from Old Welsh Grippiud, and Maredud or Meredyd for an older Marget-iud. By itself it is possibly the word which the poets write ud, and understand to mean lord; but if these forms are related, it must have originally meant rather a fighter, soldier, or champion.

[76] There is a special similarity between this and an Anglesey story given by Howells, p. 138: it consists in the sequence of seeing the fairies dance and finding money left by them. Why was the money left?

[77] It was so called by the poet D. ab Gwilym, cxcii. 12, when he sang:

I odi ac i luchio To bring snow and drifting flakes Odiar lechwed Moel Eilio. From off Moel Eilio's slope.

[78] This is commonly pronounced 'Y Gath Dorwen,' but the people of the neighbourhood wish to explain away a farm name which could, strangely enough, only mean 'the white-bellied cat'; but y Garth Dorwen, 'the white-bellied garth or hill,' is not a very likely name either.

[79] The hiring time in Wales is the beginning of winter and of summer; or, as one would say in Welsh, at the Calends of Winter and the Calends of May respectively. In North Cardiganshire the great hiring fair was held at the former date when I was a boy, and so, as I learn from my wife, it was in Carnarvonshire.

[80] In a Cornish story mentioned in Choice Notes, p. 77, we have, instead of ointment, simply soap. See also Mrs. Bray's Banks of the Tamar, pp. 174-7, where she alludes to H. Cornelius Agrippa's statement how such ointment used to be made--the reference must, I think, be to his book De Occulta Philosophia Libri III (Paris, 1567), i. 45 (pp. 81-2).

[81] See the Mabinogion, pp. 1-2; Evans' Facsimile of the Black Book of Carmarthen, fol. 49b-50a; Rhys' Arthurian Legend, pp. 155-8; Edmund Jones' Spirits in the County of Monmouth, pp. 39, 71, 82; and in this volume, pp. 143, 203, above. I may mention that the Cornish also have had their Cwn Annwn, though the name is a different one, to wit in the phrase, 'the Devil and his Dandy-dogs': see Choice Notes, pp. 78-80.

[82] 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.

[83] Owing to some oversight he has 'a clean or a dirty cow' instead of cow-yard or cow-house, as I understand it.

[84] 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'; lluarth, 'a kitchen garden,' as compared with its Irish equivalent lubhghort. Compare also Rhiwabon, locally pronounced Rhuabon, and Rhiwallon, occurring sometimes as Rhuallon. 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.'

[85] In medieval Welsh poetry this name was still a dissyllable; but now it is pronounced Llyn, in conformity with the habit of the Gwyndodeg, which makes into porfyd what is written porfeyd, 'pastures,' and pronounced porféid in North Cardiganshire. So in the Lleyn name Sarn Fyllteyrn the second vocable represents Maelteyrn, in the Record of Carnarvon (p. 38) Mayltern: it is now sounded Mylltyrn with the second y short and accented. Lleyn is a plural of the people (genitive Llaën in Porth Dinllaën), used as a singular of their country, like Cymru = Cymry, and Prydyn. The singular is llain, 'a spear,' in the Book of Aneurin: see Skene, ii. 64, 88, 92.

[86] 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.

[87] 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 Gwyned 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 Meredyd, 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 Amheredyd. 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.'

[88] 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.

[89] 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.

[90] According to old Welsh orthography this would be written Moudin, and in the book Welsh of the present day it would have to become Meudin. Restored, however, to the level of Gallo-Roman names, it would be Mogodunum or Magodunum. The place is known as Castell Moedin, and includes within it the end of a hill about halfway between Llannarth and Lampeter.

[91] For other mentions of the colours of fairy dress see pp. 44, 139 above, where red prevails, and contrast the Lake Lady of Llyn Barfog clad in green, p. 145.

[92] This name means the Bridge of the Blessed Ford, but how the ford came to be so called I know not. The word bendigaid, 'blessed,' comes from the Latin verb benedico, 'I bless,' and should, but for the objection to nd in book Welsh, be bendigaid, which, in fact, it is approximately in the northern part of the county, where it is colloquially sounded Pont Rhyd Fyndiged, Fydiged, or even Fdiged, also Pont Rhyd mdiged, which represents the result of the unmutated form Bdiged coming directly after the d of rhyd. Somewhat the same is the case with the name of the herb Dail y Fendigaid, literally 'the Leaves of the Blessed' (in the feminine singular without any further indication of the noun to be supplied). This name means, I find, 'hypericum androsæmum, tutsan,' and in North Cardiganshire we call it Dail y Fyndiged or Fdiged, but in Carnarvonshire the adjective is made to qualify dail, so that it sounds Dail Bydigad or Bdigad, 'Blessed Leaves.'

[93] I am far from certain what y nos, 'the night,' may mean in such names as this and Craig y Nos, 'the Rock of the Night' (p. 254 above), to which perhaps might be added such an instance as Blaen Nos, 'the Point of (the?) Night,' in the neighbourhood of Llandovery, in Carmarthenshire. Can the allusion be merely to thickly overshadowed spots where the darkness of night might be said to lurk in defiance of the light of day? I have never visited the places in point, and leading questions addressed to local authorities are too apt to elicit misleading answers: the poetic faculty is dangerously rampant in the Principality.

[94] Dâr is a Glamorgan pronunciation, metri gratiâ of what is written daear, 'earth': compare d'ar-fochyn in Glamorgan for a badger, literally 'an earth pig.' The dwarf's answer was probably in some sort of verse, with dâr and iâr to rhyme.

[95] Applied in Glamorgan to a child that looks poorly and does not grow.

[96] In Cardiganshire a conjurer is called dyn hysbys, where hysbys (or, in older orthography, hyspys) means 'informed': it is the man who is informed on matters which are dark to others; but the word is also used of facts--Y mae 'r peth yn hysbys, 'the thing is known or manifest.' The word is divisible into hy-spys, which would be in Irish, had it existed in the language, so-scese for an early su-squestia-s, the related Irish words being ad-chiu, 'I see,' pass. preterite ad-chess, 'was seen,' and the like, in which ci and ces have been equated by Zimmer with the Sanskrit verb caksh, 'to see,' from a root quas. The adjective cynnil applied to the dyn hyspys in Glamorgan means now, as a rule, 'economical' or 'thrifty,' but in this instance it would seem to have signified 'shrewd,' 'cunning,' or 'clever,' though it would probably come nearer the original meaning of the word to render it by 'smart,' for it is in Irish conduail, which is found applied to ingenious work, such as the ornamentation on the hilt of a sword. Another term for a wizard or conjurer is gwr cyfarwyd, with which the reader is already familiar. Here cyfarwyd forms a link with the kyvarwyd of the Mabinogion, where it usually means a professional man, especially one skilled in story and history; and what constituted his knowledge was called kyvarwydyt, which included, among other things, acquaintance with boundaries and pedigrees, but it meant most frequently perhaps story; see the (Oxford) Mabinogion, pp. 5, 61, 72, 93. All these terms should, strictly speaking, have gwr--gwr hyspys, gwr cynnil, and gwr cyfarwyd--but for the fact that modern Welsh tends to restrict gwr to signify 'a husband' or 'a married man,' while dyn, which only signifies a mortal, is made to mean man, and provided with a feminine dynes, 'woman,' unknown to good Welsh literature. Thus the spoken language is in this matter nearly on a level with English and French, which have quite lost the word for vir and anêr.

[97] Rhyd y Gloch means 'the Ford of the Bell,' in allusion, as the story goes, to a silver bell that used in former ages to be at Llanwonno Church. The people of Llanfabon took a liking to it, and one night a band of them stole it; but as they were carrying it across the Taff the moon happened to make her appearance suddenly, and they, in their fright, taking it to be sunrise, dropped the bell in the bed of the river, so that nothing has ever been heard of it since. But for ages afterwards, and even at the present day indeed, nothing could rouse the natives of Llanfabon to greater fury than to hear the moon spoken of as haul Llanfabon, 'the sun of Llanfabon.'

[98] It was peat fires that were usual in those days even in Glamorgan.

[99] See Hartland's Science of Fairy Tales, pp. 112-6.

[100] In no other version has Mr. Reynolds heard cwcwll wy iâr, but either plisgyn or cibyn wy iâr, to which I may add masgal from Mr. Craigfryn Hughes' versions. The word cwcwll usually means a cowl, but perhaps it is best here to treat cwcwll as a distinct word derived somehow from conchylium or the French coquille, 'a shell.'

[101] The whole passage will be found in the Itinerarium Kambriæ, i. 8 (pp. 75-8), and Giraldus fixes the story a little before his time somewhere in the district around Swansea and Neath. With this agrees closely enough the fact that a second David, Dafyd ab Geralld or David Fitzgerald, appears to have been consecrated Bishop of St. David's in 1147, and to have died in 1176.

[102] The words in the original are: Nec carne vescebantur, nec pisce; lacteis plerumque cibariis utentes, et in pultis modum quasi croco confectis.

[103] Perhaps it is this also that suggested the name Eliodorus, as it were Hêliodôros; for the original name was probably the medieval Welsh one of Elidyr = Irish Ailithir, ailither, 'a pilgrim': compare the Pembrokeshire name Pergrin and the like. It is curious that Elidyr did not occur to Glasynys and prevent him from substituting Elfod, which is quite another name, and more correctly written Elfod for the earlier El-fodw, found not only as Elbodu but also Elbodug-o, Elbodg, Elbot and Elfod: see p. 117 above.

[104] For one or two more instances from Wales see Howells, pp. 54-7. Brittany also is a great country for death portents: see A. Le Braz, Légende de la Mort en Basse-Bretagne (Paris, 1893), also Sébillot's Traditions et Superstitions de la Haute-Bretagne (Paris, 1882), i. pp. 270-1. For Scotland see The Ghost Lights of the West Highlands by Dr. R. C. Maclagan in Folk-Lore for 1897, pp. 203-256, and for the cognate subject of second sight see Dalyell's Darker Superstitions of Scotland, pp. 466-88.

[105] Another word for the toeli is given by Silvan Evans as used in certain parts of South Wales, namely, tolaeth or dolath, as to which he mentions the opinion that it is a corruption of tylwyth, a view corroborated by Howells using, p. 31, the plural tyloethod; but it could not be easily explained except as a corruption through the medium of English. Elias Owen, p. 303, uses the word in reference to the hammering and rapping noise attending the joinering of a phantom coffin for a man about to die, a sort of rehearsal well known throughout the Principality to every one who has ears spiritually tuned. Unfortunately I have not yet succeeded in locating the use of the word tolaeth, except that I have been assured by a Carmarthen man that it is current in Welsh there as toleth, and by a native of Pumsant that it is in use from Abergwili up to Llanbumsant.

[106] See, for instance, pp. 200, 221, 228.

[107] Mrs. Williams-Ellis of Glasfryn writes to me that the place is now called Bwlch Trwyn Swncwl, that it is a gap on the highest part of the road crossing from Llanaelhaearn to Pistyll, and that it is quite a little mountain pass between bleak heather-covered hillsides, in fact a very lonely spot in the outskirts of the Eifl, and with Carnguwch blocking the horizon in the direction of Cardigan Bay.

[108] For this I am indebted to Mr. Gwenogvryn Evans' Report on MSS. in the Welsh Language, i. 585 k. The words were written by Williams about the beginning of the seventeenth century, and his û does not mean w. He was, however, probably thinking of cawr, cewri, and such instances as tawaf, 'taceo,' and tau, 'tacet.' At all events there is no trace of u in the local pronunciation of the name Tre'r Ceiri. I have heard it also as Tre' Ceiri without the definite article; but had this been ancient one would expect it softened into Tre' Geiri.

[109] See the Oxford Mabinogion, pp. 110, 113, and 27-9, 36-41, 44, also 309, where a Triad explains that the outposts were Anglesey, Man, and Lundy. But the other Triads, i. 3 = iii. 67, make them Orkney, Man, and Wight, for which we have the older authority of Nennius. § 8. The designation Tair Ynys Brydain, 'The Three Isles of Prydain,' was known to the fourteenth-century poet, Iolo Goch: see his works edited by Ashton, p. 669.

[110] For Prydyn in the plural see Skene's Four Ancient Books of Wales, ii. 209, also 92, where Pryden is the form used. In modern Welsh the two senses of Cymry are distinguished in writing as Cymry and Cymru, but the difference is merely one of spelling and not very ancient.

[111] So Geoffrey (i. 12-15) brings his Trojans on their way to Britain into Aquitania, where they fight with the Pictavienses, whose king he calls Goffarius Pictus.

[112] Cadarn and cadr postulate respectively some such early forms as catrno-s and cadro-s, which according to analogy should become cadarn and cadr. Welsh, however, is not fond of dr; so here begins a bifurcation: (1) retaining the d unchanged cadro-s yields cadr, or (2) dr is made into dr, and other changes set in resulting in the ceir of ceiri, as in Welsh aneirif, 'numberless,' from eirif, 'number,' of the same origin as Irish áram from *ad-rim = *ad-rima, and Welsh eiliw, 'species, colour,' for ad-liw, in both of which i follows d combinations; but that is not essential, as shown by cader, cadair, for Old Welsh cateir, 'a chair,' from Latin cat[h]edra. The word that serves as our singular, namely cawr, is far harder to explain; but on the whole I am inclined to regard it as of a different origin, to wit, the Goidelic word caur, 'a giant or hero,' borrowed. The plural cewri or cawri is formed from the singular cawr, which means a giant, though, associated in the plural with ceiri, it has sometimes to follow suit with that vocable in connoting dress.

[113] The most important of these are the old Breton kazr, now kaer, 'beautiful or pretty,' and old Cornish caer of the same meaning; elsewhere we have, as in Greek, the Doric kekadmai and kekadmenos, to be found used in reference to excelling or distinguishing one's self; also kosmos, 'good order, ornament,' while in Sanskrit there is the theme çad, 'to excel or surpass.' The old meaning of 'beautiful,' 'decorated,' or 'loudly dressed,' is not yet lost in the case of ceiri.

[114] For the text see the Oxford Mabinogion, pp. 193-4, and for comparisons of the incident see Nutt's Holy Grail, p. 154 et seq.; and Rhys' Arthurian Legend, pp. 75-6. A more exact parallel, however, is to be mentioned in the next chapter.

[115] This chapter was written mostly in 1891.

[116] The spelling there used is phynnodderee, to the perversity of which Cregeen calls attention in his Dictionary. In any case the pronunciation is always approximately fun-ó-dur-i or fun-ód-ri, with the accent on the second syllable.

[117] I am inclined to think that the first part of the word fenodyree is not fynney, the Manx word for 'hair,' but the Scandinavian word which survives in the Swedish fjun, 'down.' Thus fjun-hosur (for the fjun-hosa suggested by analogy) would explain the word fenodyree, except its final ee, which is obscure. Compare also the magic breeks called finn-brækr, as to which see Vigfusson's Icelandic Dict. s. v. finnar.

[118] Cumming's Isle of Man (London, 1848), p. 30, where he refers his readers to Waldron's Description of the Isle of Man: see pp. 28, 105.

[119] See Windisch's Irische Grammatik, p. 120.

[120] The Manx word for the rowan tree, incorrectly called a mountain ash, is cuirn, which is in Mod. Irish caorthann, genitive caorthainn, Scotch Gaelic caorunn; but in Welsh books it is cerdin, singular cerdinen, and in the spoken language mostly cerdin, cerding, singular cerdinen, cerdingen. This variation seems to indicate that these words have possibly been borrowed by the Welsh from a Goidelic source; but the berry is known in Wales by the native name of criafol, from which the wood is frequently called, especially in North Wales, coed criafol, singular coeden griafol or pren criafol. The sacredness of the rowan is the key to the proper names Mac-Cáirthinn and Der-Cháirthinn, with which the student of Irish hagiology is familiar. They mean the Son and the Daughter of the Rowan respectively, and the former occurs as Maqui Cairatini on an Ogam inscribed stone recently discovered in Meath, not very far from the Boyne.

[121] I am sorry to say that it never occurred to me to ask whether the shooting was done with such modern things as guns. But Mr. Arthur Moore assures me that it is always understood to be bows and arrows, not guns.

[122] Edited by Oswald Cockayne for the Master of the Rolls (London, 1864-6): see more especially vol. ii. pp. 156-7, 290-1, 401; vol. iii. pp. 54-5.

[123] Mr. Moore is not familiar with this term, but I heard it at Surby, in the south; and I find buidseach and buidseachd given as Highland Gaelic words for a witch and witchcraft respectively.

[124] See Stokes' Goidelica, p. 151.

[125] This chapter was written in 1891, except the portions of it which refer to later dates indicated.

[126] See the Stokes-O'Donovan edition of Cormac (Calcutta, 1868), pp. 19, 23.

[127] Sir John Sinclair's Statistical Account of Scotland, xi. 620; Pennant's Tour in Scotland in 1769 (3rd edition, Warrington, 1774), i. 97, 186, 291; Thomas Stephens' Gododin, pp. 124-6; and Dr. Murray in the New English Dictionary, s. v. Beltane.

[128] In my Hibbert Lectures on Celtic Heathendom, pp. 517-21.

[129] As to the Thargelia and Delia, see Preller's Griechische Mythologie, i. 260-2, and A. Mommsen's Heortologie, pp. 414-25.

[130] See section H of the Report of the Liverpool Meeting of the British Association in 1896, pp. 626-56.

[131] It is my impression that it is crowned with a small tumulus, and that it forms the highest ground in Jurby, which was once an island by itself. The one between Ramsey and Bride is also probably the highest point of the range. But these are questions which I should like to see further examined, say by Mr. Arthur Moore or Mr. Kermode.

[132] Cronk yn Irree Laa, despite the gender, is the name as pronounced by all Manxmen who have not been misled by antiquarians. To convey the other meaning, referring to the day watch, the name would have to be Cronk ny Harrey Laa; in fact, a part of the Howe in the south of the island is called Cronk ny Harrey, 'the Hill of the Watch.' Mr. Moore tells me that the Jurby cronk was one of the eminences for 'Watch and Ward'; but he is now of opinion that the high mountain of Cronk yn Irree Laa in the south was not. As to the duty of the inhabitants to keep 'Watch and Ward' over the island, see the passage concerning it extracted from the Manx Statutes (vol. i. p. 65) by Mr. Moore in his Manx Surnames, pp. 183-3; also my preface to the same work, pp. v-viii.

[133] Quoted from Oliver's Monumenta de Insula Manniæ, vol. i. (Manx Society, vol. iv) p. 84: see also Cumming's Isle of Man, p. 258.

[134] See the New English Dictionary, s. v. 'Allhallows.'

[135] This comes near the pronunciation usual in Roxburghshire and the south of Scotland generally, which is, as Dr. Murray informs me, Hunganay without the m occurring in the other forms to be mentioned presently. But so far as I have been able to find, the Manx pronunciation is now Hob dy naa, which I have heard in the north, while Hob ju naa is the prevalent form in the south.

[136] See my Hibbert Lectures, pp. 514-5; and as to hiring fairs in Wales see pp. 210-2 above.

[137] See Robert Bell's Early Ballads (London, 1877), pp. 406-7, where the following is given as sung at Richmond in Yorkshire:--

To-night it is the New-Year's night, to-morrow is the day, And we are come for our right, and for our ray, As we used to do in old King Henry's day. Sing, fellows, sing, Hagman-heigh. If you go to the bacon-flick, cut me a good bit; Cut, cut and low, beware of your maw; Cut, cut and round, beware of your thumb, That me and my merry men may have some. Sing, fellows, sing, Hagman-heigh. If you go to the black-ark bring me X mark; Ten mark, ten pound, throw it down upon the ground, That me and my merry men may have some. Sing, fellows, sing, Hagman-heigh.

[138] The subject is worked out in Nicholson's Golspie, pp. 100-8, also in the New English Dictionary, where mention is made of a derivation involving calendæ, which reminds me of the Welsh call for a New-Year's Gift--Calennig! or C'lennig! in Arfon 'Y Ngh'lennig i! 'My Calends gift if you please!'

[139] On being asked, after reading this paper to the Folk-Lore Society, who was supposed to make the footmarks in the ashes, I had to confess that I had been careless enough never to have asked the question. I have referred it to Mr. Moore, who informs me that nobody, as I expected, will venture on any explanation by whom the footmarks are made.

[140] This seems to imply the application of the same adjective, some time or other, to clean water and a handsome man, just as we speak in North Cardiganshire of dwr glân, 'clean water,' and bachgen glân, 'a handsome boy.'

[141] In Phillips' Book of Common Prayer this is called Lá nolick y biggy, 'Little Nativity Day,' and Lá ghian blieny, 'The Day of the Year's End,' meaning, of course, the former end of the year, not the latter: see pp. 55, 62, 66.

[142] See my Hibbert Lectures, pp. 514-5, and the Brython, ii. 20, 120: an instance in point occurs in the next chapter.

[143] This has been touched upon in my Hibbert Lectures, p. 676; but to the reasons there briefly mentioned should be added a reference to the position allotted to intercalary months in the Norse calendar, namely, at the end of the summer half, that is, as I think, at the end of the ancient Norse year.

[144] My paper was read before the Folk-Lore Society in April or May, 1891, and Miss Peacock's notes appeared in the journal of the Society in the following December: see pp. 509-13.

[145] See Choice Notes, p. 76.

[146] See the third edition of Wm. Nicholson's Poetical Works (Castle-Douglas, 1878), pp. 78, 81.

[147] See p. 321 above and the references there given; also Howells' Cambrian Superstitions, p. 58.

[148] Pomponius Mela De Chorographia, edited by Parthey, iii, chap. 6 (p. 72); see also my Hibbert Lectures, pp. 195-6, where, however, the identification of the name Sena with that of Sein should be cancelled. Sein seems to be derived from the Breton Seidhun, otherwise modified into Sizun and Sun: see chap. vi below.

[149] See my Hibbert Lectures, pp. 195-7; also my Arthurian Legend, pp. 367-8, where a passage in point is cited at length from Plutarch De Defectu Oraculorum, xviii. (= the Didot edition of Plutarch's works, iii. 511); the substance of it will be found given likewise in chap. viii below.

[150] For an allusion to the traffic in winds in Wales see Howells, p. 86, where he speaks as follows:--'In Pembrokeshire there was a person commonly known as the cunning man of Pentregethen, who sold winds to the sailors, after the manner of the Lapland witches, and who was reverenced in the neighbourhood in which he dwelt, much more than the divines.'

[151] This may turn out to be all wrong; for I learn from the Rev. John Quine, vicar of Malew, in Man, that there is a farm called Balthane or Bolthane south of Ballasalla, and that in the computus (of 1540) of the Abbey Tenants it is called Biulthan. This last, if originally a man's name, would seem to point back to some such a compound as Beo-Ultán. In his Manx Names, p. 138, Mr. Moore suggests the possibility of explaining the name as bwoailtyn, 'folds or pens'; but the accentuation places that out of the question. See also the Lioar Manninagh, iii. 167, where Mr. C. Roeder, referring to the same computus passage, gives the name as Builthan in the boundary inter Cross Jvar Builthan. This would be read by Mr. Quine as inter Cross Ivar et Biulthan, 'between Cross-Ivar and Bolthane.' For the text of the boundary see Johnstone's edition of the Chronicon Manniæ (Copenhagen, 1786), p. 48, and Oliver's Monumenta de Insula Manniæ, vol. i. p. 207; see also Mr. Quine's paper on the Boundary of Abbey Lands in the Lioar Manninagh, iii. 422-3.

[152] I say 'approximately,' as, more strictly speaking, the ordinary pronunciation is Sndaen, almost as one syllable, and from this arises a variant, which is sometimes written Stondane, while the latest English development, regardless of the accentuation of the Anglo-Manx form, which is Santon, pronounced Sántn, makes the parish into a St. Ann's! For the evidence that it was the parish of a St. Sanctán see Moore's Names, p. 209.

[153] The Athenæum for April 1, 1893, p. 415. I may here remark that Mr. Borlase's note on do fhagaint is, it seems to me, unnecessary: let do fhagaint stand, and translate, not 'I leave' but 'to leave.' The letter should be consulted for curious matter concerning Croagh Patrick, its pagan stations, cup-markings, &c.

[154] Since this paper was read to the Folk-Lore Society a good deal of information of one kind or another has appeared in its journal concerning the first-foot: see more especially Folk-Lore for 1892, pp. 253-64, and for 1893, pp. 309-21.

[155] This was written at the beginning of the year 1892.

[156] With this compare what Mr. Gomme has to say of a New Year's Day custom observed in Lanarkshire: see p. 633 of the Ethnographic Report referred to at p. 103 above, and compare Henderson, p. 74.

[157] Old-fashioned grammarians and dictionary makers are always delighted to handle Mrs. Partington's broom: so Kelly thinks he has done a fine thing by printing guee, 'prayer,' and gwee, 'cursing.'

[158] 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.

[159] 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.

[160] In Folk-Lore for 1893, pp. 58-9.

[161] 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.

[162] 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, Llanberis. 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 tydyn, '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.

[163] See Ashton's Iolo Goch, p. 234, and Lewis' Top. Dict.

[164] See my Hibbert Lectures, p. 229, and the Iolo MSS., pp. 42-3, 420-1.

[165] 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 enw Morgant vchot y gelwir Morgannwc. Ereill a dyweit. Mae o en&wwelsh; 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.

[166] M. Loth's remarks in point will be found in the Revue Celtique,