THE WELSH PEOPLE

THE

WELSH PEOPLE

CHAPTERS ON THEIR

ORIGIN, HISTORY, LAWS, LANGUAGE, LITERATURE AND CHARACTERISTICS

JOHN RHYS, M.A.,

PRINCIPAL OF JESUS COLLEGE, AND PROFESSOR OF CELTIC XN THE UNIVERSITY OF OXFORD,

DAVID BRYNMOR-JONES, LL.B.,

BENCHER OF THE HON. SOCIETY OF THE MIDDLE TEMPLE, KING'S COUNSEL, AND MEMBER OF PARLIAMENT.

(WITH TWO MAPS)

FOURTH EDITION.

LONDON

T. FISHER UNWIN I Adelphi Terracenbsp;igo6

First Edition, April, 1900.

Second Edition (Revised), August, 1900.

Third Edition (with further Revisions), April, 1902. Fourth Edition (Corrected), March, 1906.

IA U rights reset ved. |

WE DEDICATE THIS BOOK

TO

THE MEMORY OE

THE LATE

HENRY BARON ABERDARE

OF DUFFRYN

AND OF THE LATE

THOMAS EDWARD ELLIS

OF CYNLAS

IN RECOGNITION OF THE PUBLIC SERVICES RENDERED BY THEM TO THEIRnbsp;NATIVE LAND.

PREFACE TO THE FOURTH EDITION.

We have in this edition simply corrected some few errors that had escaped our notice.

JOHN RHYS,

D. BRYNMOR-JONES.

tZ'H February^ 1906.

PREFACE TO THE THIRD EDITION.

In issuing a third edition of this book, we desire to say that we have tried to profit by the criticisms passed uponnbsp;our work. We have not, however, deemed it expedientnbsp;at present to enlarge its scope by dealing with topics tonbsp;which it has been alleged, rightly or wrongly, we have notnbsp;given sufficient space or attention. Errors which havenbsp;been pointed out by reviewers or which we have ourselvesnbsp;discovered have been, we believe, duly corrected. We alsonbsp;wish to express our gratitude for the kindness with whichnbsp;these chapters have been received by all those who arenbsp;interested in the past and present condition of the Welshnbsp;People.

JOHN RHYS,

D. BRYNMOR-JONES.

X'jth April, igo2.

PREFACE.

The following chapters concerning the Welsh people consist partly of extracts from the Report of the Royalnbsp;Commission on Land in Wales and Monmouthshire, andnbsp;partly of matter which we have written since that bodynbsp;finished its work.

Chapters I., IV., X., XL, XII., and XIII. are based upon the Report, and were all (except some paragraphs innbsp;Chapter XII., for which the Commission was indebted tonbsp;Mr. Lleufer Thomas, one of the secretaries) originallynbsp;drafted by us, and were adopted with many changes by allnbsp;the Commissioners. As now published they have, however, been greatly added to and altered, and though, as theynbsp;appeared in the Report, they were signed by all ournbsp;colleagues, we cannot hold them responsible for them innbsp;their present form.

Chapters II., III., V., VI., VII., and VIII. are new.

The greater part of Chapter IX. was written by Mr. Frederic Seebohm, LL.D., one of the Commissioners, thenbsp;well-known author of �The English Village Community�nbsp;and other works. This part of the Report appears to us sonbsp;valuable a contribution to economic history that it ought tonbsp;be made more accessible to students than it is at present.nbsp;We accordingly republish it here with Mr. Seebohm�snbsp;kind consent.

For permission to reprint such paragraphs of the Report

PREFACE.

as we might deem necessary for this work, we are indebted to the Lords Commissioners of the Treasury, whose consentnbsp;was courteously signified through the Right Honourablenbsp;R. W. Hanbury, M.P.

In the Introduction which follows we give some necessary preliminary information, and explain briefly thenbsp;scope and character of the work.

We have to acknowledge our obligation for help of various kinds to Professor Morris Jones, M.A.; Mr. Henrynbsp;Owen, B.C.L.; Mr. Cadwaladr Davies; Mr. Edward Owen,nbsp;F.S.A.; Lieut-Col. Morgan, pf Brynbriattu, Swansea;nbsp;Principal Viriamu Jones, F.R.S. ; Chancellor Trevornbsp;Parkins ; Mr. Lleufer Thomas, M.A.; Mr. Cecil Owen,nbsp;M.A.; Mr. T. E. Morris, LL.M.; and many other friends.

JOHN RHYS, p. BRYNMOR-JONES.

Sf. David's Day, 1900.

NOTE TO THE READER.

In this book where we refer to �the Commission,� we mean, unless the context shows the contrary, the Royalnbsp;Commission on Land in Wales and Monmouthshire,nbsp;appointed on March 27, 1893. The Commissionersnbsp;were:�Earl Carrington, G.C.M.G. (chairman); Lordnbsp;Kenyon ; Sir John Talbot Dillvvyn Llewelyn, Bart., M.P. ;nbsp;Mr. Edwin Grove, chairman of the Monmouthshire Countynbsp;Council; Mr. John Morgan Griffiths, of Penally; Mr.nbsp;Richard Jones, of Pertheirin; Mr. Frederic Seebohm,nbsp;M.A., LL.D. ; and ourselves. The secretaries were Mr.nbsp;Lleufer Thomas and Mr. C. E. Owen.

The Minutes of the evidence taken by the Commission are contained in five Blue-books, the references to whichnbsp;are as follows:�Vol. I., (1894) C�7439 gt; Vol. II., (1894)nbsp;C-7439; Vol. III., (1895) C-7661; Vol. IV., (189s)nbsp;C�7757; Vol. V., (1896) C�8222. In these volumes thenbsp;questions put to the witnesses, with their answers, arenbsp;numbered consecutively from the beginning to the end.nbsp;So in referring to the evidence we do so by giving thenbsp;number of the question and answer in these Minutes.

When we have occasion to refer to the Report of this Commission (which was signed and delivered on August 26,nbsp;1896, and is Parly. Paper (1896) C�8221), we use thenbsp;word � Report � only; and the words � Appendix tonbsp;Report� mean the Appendix to that Report (Parly. Papernbsp;(1896) C�8242).

NOTE TO THE READER.

It has been found impossible to insure uniformity in the spelling of Welsh names in this work, as we have notnbsp;thought it advisable to depart altogether from the spellingsnbsp;occurring in the documents consulted. This may occasionnbsp;some inconvenience to the reader, but if he should happennbsp;to be a student of English history he will readily recognisenbsp;in it an inconvenience with which he has had to strugglenbsp;in his own field of study. As regards, however, the historynbsp;of Wales and the Welsh, the case is somewhat aggravatednbsp;by the fact that not only have we to deal with namesnbsp;belonging to widely different centuries, carrying with themnbsp;phonetic modifications, but that in not a few instancesnbsp;a name may have besides several Welsh spellings, severalnbsp;English ones too. Take, for example, that of Gruffu^ ornbsp;Gruffyd, of which the most usual English spelling is Griffith:nbsp;the index shows a still greater variety. It will simplifynbsp;matters for the English reader if he will bear in mind thenbsp;following points of Welsh orthography and phonology :�

(1) nbsp;nbsp;nbsp;C has always the sound of k, and formerly bothnbsp;c and k were used, though the present Welsh alphabetnbsp;does not recognise k.

(2) nbsp;nbsp;nbsp;G has never the English sound of j or dzh as innbsp;John or ]ames.

(3) nbsp;nbsp;nbsp;F is sounded v, and both letters were formerly used,nbsp;but V is not included in the modern alphabet.

(4) nbsp;nbsp;nbsp;Dd (printed D, d in this work) has the sound ofnbsp;th in the English words this and that, while th is confinednbsp;to the sound of the same digraph in thick and thin.

(5) nbsp;nbsp;nbsp;LI (here printed L, ft) represents the surd force ofnbsp;unilateral /, and its sound stands to that of / as thatnbsp;of th to d or of/// to V. It is a single and simple consonant,nbsp;though Englishmen sometimes seem to hear it as thl,nbsp;which has now and then been their way of representingnbsp;it, as, for instance, in �The Record of Carnarvon,� innbsp;Thlandrethloh for Eandritto.

NOTE TO THE READER.

(6) nbsp;nbsp;nbsp;R is trilled as in Italian, and in rA it is a surdnbsp;strengthened by the aspirate.

(7) nbsp;nbsp;nbsp;5 is never sounded s.

(8) nbsp;nbsp;nbsp;W may be either a vowel or a consonant, that isnbsp;English 00 (approximately) and w; a similar remark isnbsp;in part applicable to the sound of i

(9) nbsp;nbsp;nbsp;(7 is sounded nearly like a thickish English z, asnbsp;in the word bit, and so under certain circumstances is_y.nbsp;Thus Gruffyd or Gruffwd is sounded very nearly asnbsp;Indicated by the English spelling Griffith, provided thenbsp;final th have its soft sound. The inconsistency betweennbsp;GruffyA and MareduA was observed too late to be corrected : the preferable spelling is MaredyA or MeredyA�nbsp;the early forms were Gripp-iud and Marget-iud.

(10) nbsp;nbsp;nbsp;The syllabic accent is ordinarily on the penultimate,nbsp;except where the two last syllables of a word have beennbsp;run into one in the more modern stages of the Welshnbsp;language: then the word is of course a perispomenon.

J. R.

D. B.-J.

bi^,. �-^_gt;i, fot m bas fSmisil. rb �� hi-i-'hj � gt;�. jd)

,: nbsp;nbsp;nbsp;-.nbsp;nbsp;nbsp;nbsp;�nbsp;nbsp;nbsp;nbsp;17nbsp;nbsp;nbsp;nbsp;-vi-'i -'il i;:.-:-gt;;-a--rs�ir-

.� .' j, � nbsp;nbsp;nbsp;.nbsp;nbsp;nbsp;nbsp;, 7; -nbsp;nbsp;nbsp;nbsp;'i.-^ ,b.?lf.--nbsp;nbsp;nbsp;nbsp;. --vsrt-f ��'� lt;gt;);

iti jsib jmiK-mm a to ipvov s -r-i' � � -�' vi ;7! '74 (ci

70 7-. ': - O . gt;-� ; nbsp;nbsp;nbsp;JViJfj rti

;48'|.7^3^S*Af S,�t-|;=bh�Si; !cijf:Ui� t� (Q) '

;'niEi-)'J3 -ILhlH: ;.7 �f,.; .-�� b-;ow -��� igi

�s nbsp;nbsp;nbsp;�^bni;-

�gt;dj bobi'/inq (�At�f �'' o-o'lfK:-7 :� vofi^jrifeo'jfnt vb i' 7 �lijO :f't .)!nbsp;nbsp;nbsp;nbsp;773' � ..Xfj r'V / i77 .

a-f; l�i7

INTRODUCTION.

The Dominion or Principality of Wales may be described as a broad indented peninsula situated in the south-westernnbsp;part of Great Britain. Its greatest length from north tonbsp;south is about 135 miles, and its breadth from east to westnbsp;ranges from about 35 to 95 miles. On the north it isnbsp;bounded by the Irish Sea and the estuary of the Dee,nbsp;on the west by St. George�s Channel, on the south bynbsp;the Bristol Channel, and on the east by Cheshire, Shropshire, Herefordshire, and Monmouthshire. The easternnbsp;boundary was definitely fixed by the operation of thenbsp;St. 27 Henry VI11, c. 26,^ though some small variationsnbsp;have subsequently taken place.

The Principality is divided, as is the case with England, into counties, which, including Monmouthshire, are thirteennbsp;in number. The names of these counties with their Welshnbsp;equivalents are :��

Anglesey Carnarvonshirenbsp;Denbighshirenbsp;Flintshire .

Merionethshire Montgomeryshirenbsp;Brecknockshirenbsp;Cardiganshire

' See below, pp, 368-74.

INTRODUCTION.

Carmarthenshire nbsp;nbsp;nbsp;.nbsp;nbsp;nbsp;nbsp;.nbsp;nbsp;nbsp;nbsp;Sir Gaerfyrdin.

Glamorganshire nbsp;nbsp;nbsp;.nbsp;nbsp;nbsp;nbsp;. Sir Forgannwg.

Pembrokeshire nbsp;nbsp;nbsp;.nbsp;nbsp;nbsp;nbsp;.nbsp;nbsp;nbsp;nbsp;Sir Benfro.

Radnorshire nbsp;nbsp;nbsp;.nbsp;nbsp;nbsp;nbsp;. Sir Faesyfed.

Monmouthshire nbsp;nbsp;nbsp;.nbsp;nbsp;nbsp;nbsp;.nbsp;nbsp;nbsp;nbsp;Sir Fynwy.

The first six form what is generally known as North Wales, and the remainder South Wales. Their boundariesnbsp;preserve, to some extent, the ancient divisions of thenbsp;Principality. There are also two large county boroughs�nbsp;Cardiff and Swansea.

Monmouthshire is technically an English county, but it is often for administrative purposes, and sometimes bynbsp;legislation, treated as part of Wales, or grouped with somenbsp;of the Welsh counties.^ This has been due partly to the

' Monmouthshire was constituted into a county or shire by 27 Henry VIII. c. 26, out of territory that was expressly stated to be part of the Dominionnbsp;of Wales, but it was, however, by the same statute made subject to thenbsp;courts at Westminster, while the rest of Wales was granted separatenbsp;jurisdiction, which lasted until its abolition in 1830, See infra, p. 373.

The position of the county at present is in many respects anomalous. It forms part of the Oxford Circuit, and, therefore, is included in England innbsp;most matters concerned with the administration of law, but in the divisionnbsp;of the country into county court circuits, the whole of the county, alongnbsp;with Cardiff and Crickhowell (in Wales) and Ross (in Herefordshire), isnbsp;grouped into what may be regarded as a Welsh circuit (No. 24). Its inclusion in Wales for executive purposes has been the general, though notnbsp;universal, rule. It is so recognised by the Registrar-General for statisticalnbsp;purposes, by the Local Government Board for poor-law purposes, and bynbsp;the Home Office for the purposes of the Mines Regulation Acts, the Factorynbsp;and Workshop Acts, and the Quarries Act. In all matters educational,nbsp;Wales and Monmouthshire have been treated as a unit distinct fromnbsp;England, e.g., by the appointment of Commissioners to inquire into thenbsp;stated Education in Wales in 1846, and of the Departmental Committee tonbsp;inquire into Intermediate and Higher Education in Wales in 1880-81, as wellnbsp;as by the subsequent passing of the Intermediate Education Act of 1889. Itnbsp;was similarly dealt with by the Schools Inquiry Commission of 1864, whichnbsp;issued a separate report (vol. xx.) on the schools of � Monmouthshire andnbsp;Wales ; by the Court of Chancery in Schemes for the re-organization ofnbsp;Charitable Trusts; by the Oxford University Commissioners in Statutesnbsp;made undrar the Universities of Oxford and Cambridge Act, 1877; and bynbsp;the Privy Council in the granting of charters to the University College of

INTRODUCTION.

similarity of its development to that of the adjoining county of Glamorgan, and partly to the fact that a largenbsp;proportion of its inhabitants are Welsh as to their origin,nbsp;language, and habits. In any historical inquiry as to thenbsp;Cymric people it must be looked on as Welsh, and so innbsp;this work we generally use the term Wales as includingnbsp;Monmouthshire.

These thirteen counties are divided into hundreds, poor-law unions, highway districts, sanitary districts, andnbsp;parishes for the purposes of local government; into pettynbsp;sessional divisions, county court districts, and circuits fornbsp;the administration of justice; and into borough and countynbsp;constituencies for the appointment of parliamentary representatives. There are also ecclesiastical divisions similar tonbsp;those that exist in England.^ The most ancient politicalnbsp;division of Wales about which we have any sure knowledgenbsp;is that into cantrefs and cymwds. This must not be

South Wales and Monmouthshire, and to the University of Wales. Both the Education Department and the Charity Commissioners, in their super-intendence and inspection of elementary and intermediate schools respectively, also include Monmouthshire in Wales, and reports specially dealingnbsp;with Wales are reprinted, in separate book-form, from the annual generalnbsp;reports made by these two Departments. Monmouthshire was, however,nbsp;treated as a part of England (and not of Wales) in the Welsh Sundaynbsp;Closing Act of i88i, as it has also been in nearly all matters agricultural,nbsp;e g., by the Royal Commission on Agriculture (1880); by the Royal Commission on Labour (in its inquiry as to the agricultural labourer in 1892-93),nbsp;and also, we believe, by the Royal Commission on Agricultural Depressionnbsp;(1893-6), but not by the Commission on the Employment of Children,nbsp;Young Persons, and Women in Agriculture (1867). The Board of Agriculture also treats Monmouthshire as a part of England in the preparationnbsp;of their agricultural returns, as well as in all other matters.

* A list of the hundreds in each county, with the names of the parishes in each hundred, will be found in the Appendix to the Report, pp. 361nbsp;et seq. The student should compare this list with that of the Welshnbsp;parishes, according to cymwds printed in the quot; Myvyrian Archaiology,quot;nbsp;ii. 613-628. For a list of the poor-law unions, see the Appendixnbsp;cited above, pp. 378-403, and of the highway districts, see the samenbsp;Appendix, pp. 404-408. Much information as to Wales and its placenbsp;names is given in Carlisle's quot; Topography of Wales,� Lond. 1811.

W.P. nbsp;nbsp;nbsp;I

INTRODUCTION.

confounded with the distribution of the Cymric land among the regal or princely families which resulted in smallnbsp;kingdoms or principalities. It was regarded as old in thenbsp;time of Howel Da (907-950), and the boundaries of thenbsp;cantrefs and cymwds were then well enough ascertained fornbsp;practical purposes. We think it useful to reproduce thenbsp;map of Wales according to cantrefs and cymwds made bynbsp;William Owen (known later as Dr. Owen Pughe) aboutnbsp;the end of the eighteenth century. We ought to saynbsp;that this map cannot be taken as representing boundariesnbsp;with absolute accuracy, but it may be found of some servicenbsp;as showing the geographical relations of the various areas.

There is yet another division of Welsh land distinct in origin, and based on different conceptions from that intonbsp;cantrefs and cymwds, which must be mentioned, though itnbsp;has now little or no practical political significance�thatnbsp;into seigniories, lordships-marchers, lordships, manors, andnbsp;fees. These feudal divisions were the result of the Normannbsp;Conquest of South Wales and the Marches, and of thenbsp;final conquest of the Principality of Edward I. They arenbsp;often conterminous with other areas.

We think it well to add a few observations as to the population of Wales at different periods, for withoutnbsp;bearing in mind the number of persons concerned it isnbsp;not possible to appreciate historical events correctly, or tonbsp;see the past conditions of things in true perspective.

Beginning with the known and proceeding to the conjectural statistics, we find that according to the census returns of 1891 the population of England and Wales wasnbsp;29,002,525, and that of Wales (including Monmouthshire)nbsp;was 1,776,405 ; so that the population of the latter area

1 As to the lordships, etc., see the quot; Memorandum on the Lordships and Manors of Wales and Monmouthshirequot; in the Appendix to the Report,nbsp;pp. 437 et seq. In the Principality, in the limited sense�the land acquirednbsp;by Edward I.�the cymwds were treated as in effect lordships that passed bynbsp;conquest into the hands of the Crown. See below, p. 281, n. i, 305, 256, 403.

INTRODUCTION.

was about one-sixteenth of that of the former.^ Coming to the first complete and systematic enumeration ofnbsp;England and Wales�the census of i8oi�we find that thenbsp;figures for England and Wales are 8,892,536, and fornbsp;Wales and Monmouthshire 586,634. From them it appearsnbsp;that the population of Wales and Monmouthshire was anbsp;little over one-fifteenth of that of the whole of Englandnbsp;and Wales.^

Before 1801 no direct and trustworthy enumeration of the inhabitants of this country had taken place, butnbsp;numerous estimates had been made from time to time.nbsp;Three such estimates of the population towards the closenbsp;of the seventeenth century have been made at differentnbsp;times by three statisticians, acting without concert, and thenbsp;results differ but slightly. Perhaps the best known ofnbsp;these computations is that made by Gregory King. Thenbsp;basis of his calculation is the number of houses returned innbsp;1690 by the officers who made the last collection of thenbsp;hearth money, and the conclusion at which he arrived wasnbsp;that the population of England was nearly 5,500,000. Wenbsp;find that the number of houses in Wales (excludingnbsp;Monmouthshire) according to the Hearth Book of 1690nbsp;was 77,921. Assuming that on a general average everynbsp;house so returned contained five persons, the population ofnbsp;Wales amounted to 389,605. If we add this number tonbsp;the population of England, we get a total of 5,889,605, ornbsp;nearly six millions. So the population of Wales (excludingnbsp;Monmouthshire) was then a little more than one-fifteenthnbsp;of that of England and Wales, and if Monmouthshire benbsp;added on this basis to Wales, the population of Walesnbsp;plus Monmouthshire would come out at nearly one-fourteenth of that of England and Wales. ^

' App. to Report, p. 272. nbsp;nbsp;nbsp;^ Ibid.

* See King�s �Natural and Political Observationsquot; (1696). A second calculation was made on the basis of returns made to William III. as to the

INTRODUCTION.

For periods earlier than the end of the seventeenth century, estimates of the population of this country arenbsp;still more speculative. McCulloch puts the population ofnbsp;England and Wales at the time of the Domesday surveynbsp;at 2,150,000.^ Thorold Rogers came to the conclusionnbsp;that from the fourteenth to nearly the end of the sixteenth century, it could not have exceeded two to two-and-a-half millions. He bases this opinion partly on thenbsp;postulate that the number of persons in a country, chieflynbsp;subsisting on one kind of grain, will be almost exactlynbsp;equal to the number of quarters of wheat which is annuallynbsp;produced, and on the estimate that during that time thenbsp;maximum produce of wheat in any one year could notnbsp;have been more than the higher figure just given ; andnbsp;partly on the direct evidence of taxing rolls, and especiallynbsp;records of poll taxes.^ We have seen that the ratio ofnbsp;the population of Wales and Monmouthshire to that ofnbsp;the whole of England and Wales, on Gregory King�snbsp;estimate, was in 1690, 1:14. If we assume that rationbsp;to have been constant (except for slight and temporarynbsp;variations) from the beginning of the fourteenth centurynbsp;to the time of Elizabeth, and adopt the higher limit givennbsp;number of the adherents of the different religious denominations, and it putsnbsp;the population of England and Wales at 5,200,000. (Macaulay, quot; Hist, ofnbsp;England,quot; c. 3.) The third important estimate is that of Finlaison, who,nbsp;after subjecting the ancient parochial registers to modern tests, calculatesnbsp;that towards the close of the seventeenth century the number was a littlenbsp;under 5,200,000. (See Population Returns, 1831; Macaulay, ubi supra;nbsp;Lecky's quot;Hist, of England in the Eighteenth Century,� i., p, 197;nbsp;Macpherson�s ��Annals of Commerce,� ii. 68, 634, 674, and iii. 134).

gt; McCulloch�s �Statistical Account of the British Empirequot; (1847), vol. i. p. 396. Mr. York Powell estimates the population in the area covered bynbsp;the survey at 2,000,000; quot; Social England,� vol. i. p. 240. Mr. A. L. Smithnbsp;puts it at the same figure : Ibid., p. 357.

2 �Dictionary of English History� (Lond. 1885), art. Population. See too, Rogers' quot; History of Agriculture and Prices in England,� vol. i. pp. 50nbsp;et seq.; and quot; England�s Industrial and Commercial Supremacy,� pp. 44-64.

INTRODUCTION.

by Thorold Rogers�2,500,000, the population of Wales and Monmouthshire could not during that period havenbsp;been greater than about 178,000.

We are, however, not justified in assuming that the ratio 1:14 truly represents the proportion of the inhabitants ofnbsp;Wales and Monmouthshire to that of England and Walesnbsp;at the time of the Edwardian conquest in 1282, or even innbsp;the sixteenth century. We could do so safely only if wenbsp;found that the economic development of the two areas hadnbsp;proceeded at equal rates. The facts, however, indicatenbsp;that this was not the case before the end of the great war.nbsp;For instance, the total acreage of Wales and Monmouthshire is 5,121,013;^ in 1795 the waste area was 1,696,827nbsp;acres^�that is, more than one-third of the whole was wildnbsp;and uninclosed. The mountainous character and thenbsp;climatic conditions place the country at a disadvantagenbsp;in regard to the production of cereals as compared withnbsp;the greater part of England, and the progress of agriculturenbsp;was very slow until the beginning of this century. Thesenbsp;and other considerations� lead us to believe that thenbsp;population of Wales and Monmouthshire in comparisonnbsp;with that of England and Wales in 1282 was proportionallynbsp;smaller than it is to-day, and that the rate of its increasenbsp;has been slower than that of England and Wales as anbsp;whole. We think, therefore, that the population of Walesnbsp;and Monmouthshire at the end of the thirteenth centurynbsp;was not greater than 150,000,^ and if McCulloch�s estimatenbsp;as to the number at the time of the Domesday survey isnbsp;fairly accurate, it was still less in the eleventh century.

____________________ _____j ___o- ________ Cardiff (which had in

1891 a population of 128,915) had only 1,870 inhabitants according to the census of 1801.

Thorold Rogers placesit aslow as 131,040. quot; England�s Ind. and Comm Supremacy,� p. 48.

INTRODUCTION.

No attempt to estimate the population of Wales or of the western part of the island at times earlier than thenbsp;Norman Conquest can be made with success, for we havenbsp;no materials at all on which to base any exact calculation.nbsp;Giraldus, writing at the end of the twelfth century, says :�nbsp;�For as the mountains of Eryri could supply pasturage fornbsp;all the herds of cattle in Wales if collected together, sonbsp;could the Isle of M�n provide a requisite quantity of cornnbsp;for all the inhabitants ; on which account there is an oldnbsp;British proverb, � M�n mam Gymru,� that is, � M�n thenbsp;mother of Wales.' If the saying was then really old andnbsp;this explanation is correct, we may infer that in the far distantnbsp;days of Cuneda and Cadwaladr the population was verynbsp;scanty. In any case the notion which was formerly widelynbsp;entertained, and still lingers in some quarters, that Walesnbsp;was in early times very populous is quite unfounded.

It is, however, not only in regard to their numbers in the past that the view popularly taken of their own history bynbsp;the Welsh is erroneous. It is not easy to state briefly, andnbsp;at the same time quite accurately, the current theory ; butnbsp;we think it may be fairly expressed thus: That they arenbsp;the descendants of a great homogeneous nation callednbsp;Cymry or Britons (now referred to as Ancient Britons);nbsp;that in distant centuries they formed a mighty state ornbsp;empire, the dominions of which comprised not only Britainnbsp;but larger territories on the Continent as well; that theynbsp;were ruled over by a line of illustrious kings which stretchednbsp;up to Brutus, son of Aineas, and from him to Noah, whonbsp;ordered the world anew after the Deluge; but that they,nbsp;owing to unsuccessful wars, bad government, and all sortsnbsp;of mischances, lost not only their continental possessions,nbsp;but also the Crown of Britain ; and at last became confinednbsp;in what is now Cymru, and reduced into subjection throughnbsp;the � fraud and rapacity � of the Saxon.

* Giraldus Cambrensis, � Descriptio,quot; i. c. 6.

INTRODUCTION.

Now all this is not mere nonsense, and as to every proposition that goes to make up this bundle of historicalnbsp;ideas there is some sound basis of fact. It is true that thenbsp;determining element in the composition of the Cymry ofnbsp;what is now Wales was Brythonic; that the Brythonsnbsp;belonged to a Celtic race which, before Caesar�s time, hadnbsp;spread over more than half this island and a considerablenbsp;part of the Continent; that there were many British kingsnbsp;who were very important in their day, and who had longnbsp;and well-known pedigrees; and that a confederation ofnbsp;Celtic tribes�Brythonic and Goidelic�did most stubbornlynbsp;resist the Teutonic tribes which invaded the island andnbsp;settled in it after the departure of the Roman legions, andnbsp;that for a long time it maintained its domination in thenbsp;western half of Britain. Yet the representation of thenbsp;early history of the Welsh given by the theory we havenbsp;summarised makes a picture in which things and personsnbsp;are exaggerated and distorted, and by adopting it the chainnbsp;of events is thrown out of gear�that is, if our conclusionsnbsp;are correct. For we feel bound to repeat, in regard to thenbsp;Brythons and the Cymry, what a famous Greek historiannbsp;said about the Hellenes, that judging from the evidencenbsp;which we are able to trust, �after most careful inquiry,�nbsp;we should imagine that � past ages were not great eithernbsp;in wars or anything else.�

So far as we can make out, the beginning of the history of the Cymry, considered as a separate and independentnbsp;nation, must be associated with the migration into what isnbsp;now North Wales of a Brythonic tribe, whose chief wasnbsp;Cuneda Wledig, and which came from the North. Thisnbsp;invasion took place not long after the time when the Romannbsp;occupation ceased. Before this, however, there were bothnbsp;Goidels and Brythons in Wales. A glance at the mapnbsp;of Roman Britain below will show the relative positionsnbsp;of these two Celtic races. The former were settled over

INTRODUCTION.

a part of North Wales and nearly the whole of the South, while the latter had spread over the central area to thenbsp;coast; but we must add that there were in the same partsnbsp;besides these Celts certain non-Aryan elements, whichnbsp;must be looked on as aboriginal, though more or lessnbsp;completely assimilated by the first Celtic conquerors.nbsp;This was the position when Cuneda, who seems to havenbsp;assumed the authority of the Roman military officernbsp;called the Dux Britannice, established his rule over Wales,nbsp;and united the Celtic tribes of the west of Britain intonbsp;a kind of confederation under his leadership, which wasnbsp;soon forced to defend the land against divers streams ofnbsp;Teutonic settlers, and which under his successors for anbsp;long time struggled to retain its supremacy.

It was during this contest that the term Cymro (which means compatriot), became a national name coveringnbsp;the members of all the Celtic tribes and kindreds whonbsp;acknowledged the over-lordship of the line of Cuneda.nbsp;Such, if the matter is looked at, not through the mists ofnbsp;Neo-Druidism or the bright yet delusive atmosphere ofnbsp;mediaeval romance, but in the clearer light of the evidencenbsp;afforded by inscriptions, language, laws, and reasonablynbsp;trustworthy chroniclers, seems to us the true conclusionnbsp;as to the origin of the Cymry. If we are right, the Welshnbsp;people of to-day have the satisfaction of knowing that theynbsp;are not the decayed and disconsolate remnant of a oncenbsp;great nation, but that in the main they are the descendantsnbsp;of Celtic races which though absorbed into the Englishnbsp;polity, after a prolonged struggle for independence, havenbsp;steadily progressed by the side of their conquerors innbsp;regard to all that goes to make up civilisation, and bynbsp;combining an obstinate vitality with a certain happy powernbsp;of adapting themselves to new circumstances, have succeeded in retaining their language and some of the bestnbsp;characteristics of their ancestors.

INTRODUCTION.

We are well aware that there are grave imperfections in our treatment of many of the problems we discuss.nbsp;We have, however, tried to give them true solutions, evernbsp;keeping before our minds the motto of the new-bornnbsp;University of Wales�Goreu awen gwirioned (Optima musanbsp;Veritas). That we have, as to every point, succeeded wenbsp;do not of course assert; and indeed we advance our conclusions on controverted questions in no dogmatic spirit,nbsp;but in tentative fashion, though we cannot always be sayingnbsp;so in the text. Nor do we present this collection ofnbsp;chapters as a history of the Welsh people, but rather asnbsp;a contribution to such a work, which may be useful tonbsp;students at our national colleges and to others who arenbsp;seriously interested in things Welsh. Yet there is a connecting thread of purpose running through the book, asnbsp;will be seen by a brief description of the subject-matternbsp;of each chapter.

In the first, second, and third chapters we deal with the ethnology and origin of the Cymry, and in order tonbsp;justify and explain our views, discuss minutely some ofnbsp;the questions connected with the so-called � Piets � andnbsp;the distribution of tribes in this island during the Romannbsp;occupation. Having shown that the Cymry emerge as anbsp;separate nation under the rule of Cuneda and his descendants when that occupation ceased, we pass on to statenbsp;very briefly in the fourth chapter their history down to thenbsp;death of Cadwaladr, when their kingdom in its morenbsp;extensive sense came to an end. In the next (the fifth)nbsp;chapter we treat of the history of Wales from that timenbsp;to the Norman Conquest of England. Then we stop tonbsp;describe the legal organisation and social condition of thenbsp;Cymry in the tenth and the immediately succeedingnbsp;centuries. In the seventh chapter we describe the way innbsp;which the greater part of Wales was gradually conquerednbsp;by the Normans, and sketch the history of the last and

W.P.

INTRODUCTION.

greatest Cymric principality to its transference by conquest to Edward I. From this event the history of the Welsh innbsp;regard to wars, foreign policy, and general affairs becomesnbsp;so merged into that of Great Britain that it is hardlynbsp;susceptible of separate treatment in a continuous narrativenbsp;form. They have, however, a particular history as to manynbsp;of the institutions, conditions, and activities, that createnbsp;or maintain the life of a nation. It is with some of thesenbsp;things that our eighth, ninth, tenth, eleventh, and twelfthnbsp;chapters are concerned. quot; �^ere we try to show how, fornbsp;nearly all purposes of government, Wales has becomenbsp;organised in the same way as England ; how the oldnbsp;Cymric tribal notions of land-holding and administrationnbsp;(which became by natural and easy stages very like tonbsp;those of the feudal system) gradually disappeared undernbsp;the influence of Norman-English officials, and by degreesnbsp;developed into the land tenure of to-day ; how a religiousnbsp;movement commencing in the sixteenth century culminated in a great revival in the eighteenth, and broughtnbsp;about the predominance of Nonconformity in the Welshnbsp;counties, the preservation and growth of Cymraeg, andnbsp;an intellectual renaissance; and how this movement in itsnbsp;turn created a demand for schools and colleges, which hasnbsp;resulted in the formation of a system of Welsh publicnbsp;education as perfect as any to be found in the Unitednbsp;Kingdom. After that we pass on to give some informationnbsp;as to the language and literature of the Welsh ; and finallynbsp;in the thirteenth chapter we attempt to exhibit the characteristic features of the most typical classes of the populationnbsp;of the Principality in our own day.

THE WELSH PEOPLE.

CHAPTER I.

THE ETHNOLOGY OF ANCIENT WALES.

In this chapter we propose to state briefly what comparative philology and ethnology have to say as to the races of ancient Britain, especially those to be found innbsp;what is now Wales, and what admixture has taken placenbsp;there in later times.

Archaeologists who have studied the contents of the aneient barrows or burial mounds of this country find thatnbsp;the human remains which they detect in them belong tonbsp;more than one race. The barrows of the earliest date arenbsp;long, and yield long skulls, while the round barrows, whichnbsp;are later, show the remains of a short-skulled people ; butnbsp;the round barrows sometimes contain long skulls as wellnbsp;as short ones, a fact which suggests that the conquerorsnbsp;began early to intermarry with the conquered population.nbsp;Looking at the same order of questions from the point ofnbsp;view of language, one may say, that the first race as to whosenbsp;presence in this country in ancient times there can be nonbsp;manner of doubt is the Celtic ; and it is also a matternbsp;admitting of no philological doubt, that the Celts of thenbsp;British Isles were Aryans, speaking related languagesnbsp;which fall into two groups, the Goidelic and the Brythonic.nbsp;The Goidelic group embraces at the present day the Gaelic

W.P. nbsp;nbsp;nbsp;B

THE WELSH PEOPLE, (chap, l)

of Ireland, of the Isle of Man, and of Scotland. The Brythonic group, on the other hand, is now representednbsp;by Welsh, and the Armoric dialects of Brittany ornbsp;ILydaw. To this group also belonged old Cornish, whichnbsp;has been extinct as a spoken language for somewhat overnbsp;a century.

These two groups of Goidelic and Brythonic languages may be regarded as variations of two ancient tongues,nbsp;Goidelic and Brythonic respectively, which differed fromnbsp;one another somewhat in the same way as Latin andnbsp;the Umbro-Samnite dialects of ancient Italy. Thus thenbsp;Goidelic for who is in Manx Gaelic quei, qiwi, but in Welshnbsp;pivy: compare Latin qul, quis, Oscan pis, accusative pirn.nbsp;Similarly the Goidelic for five is in Manx queig, but innbsp;WLish. ptimp (and pimfi) ; compare Latin qidnque, whencenbsp;the derivatives quintus, � fifth,� Quintus (for quinctus,nbsp;Quinctus), and Quinctius, Quintius, which we have as annbsp;Oscan name in no/Ajmcs, Latinised Pomptius, and Pontius,nbsp;as in the well-known name of the Samnite Pontius Pilate.nbsp;This distinction of qu and p is, it is needless to say, onlynbsp;one of the differences which must have existed betweennbsp;early Goidelic and early Brythonic ; but it has the advantage of forming a conspicuous and decisive mark wherevernbsp;it happens to occur. Nobody, however, supposes that qunbsp;and p are equally original here : qu is the older, and wherenbsp;its equivalent occurs as/gt;, this last is to be regarded as anbsp;simplification of the qu, but the simplification appears tonbsp;date very early. Thus a Roman inscription^ at Hexhamnbsp;was set up in honour of a god called therein Apollininbsp;Mapono, where Mapmo may be regarded as cognate with

' Pimp, � fise,� a.ni pimpket, �fifth,� occur among the old Welsh glosses in the Bodleian manuscript, Auct. F. 4�32 (� Gram. Celtica,� p. 1060, andnbsp;�Transactions of the Philological Society,� 1860-1, p. 237), and the regularnbsp;spelling in modern Welsh would be py/np and pymed, which represent thenbsp;actual pronunciation in the spoken language of most of South Wales.

� See the Berlin �Corpus Inscriptiomun I.atinaruin,� vol. vii., No. 1345.

ETHNOLOGY OF ANCIENT WALES. 3

a Brythonic or Gaulish nominative, Mapono-s} a word which in old Welsh would have been mapon, now mabon,nbsp;� a boy or youth,� derived from the simple form map, nownbsp;mab, � a boy or son.� This is in Irish macc or mac, for annbsp;early Goidelic maqua-s, genitive maqui, which occursnbsp;frequently in the Ogam inscriptions of Britain and Ireland.^nbsp;All this, however, only takes us back at most to the Romannbsp;occupation, but, as a matter of fact, we have no data fromnbsp;a time when Brythons and Gauls had not already madenbsp;qu into p; and the question arises which of the two groups

* In Gaulish the nominatives corresponding to those in us and um (like dominus, regnum in Latin) ended in os and on respectively, as in Greek, andnbsp;where the declension is fairly certain we shall write them so ; but Latin authorities using us in two declensions leave us sometimes unable to decide betweennbsp;the o and u steins. In that case we shall follow the Latin in using us.

^ Speaking more precisely, Ogam inscriptions occur (i) m Wales, mostly South Wales, but North Wales has two, one in Denbighshire and one atnbsp;least in Carnarvonshire ; (2) in Devon and Cornwall, and one remarkablenbsp;instance occurs at Silchester, in Hampshire; (3) in Ireland, mostly in the southnbsp;of that country ; (4) in the Isle of Man; and (5) in Scotland, including Orkneynbsp;and Shetland, but these are mostly late in comparison with the bulk of thenbsp;others. The older Ogam characters consist of scores or notches on thenbsp;edge of the stones used, and the following is the alphabet of the most ancientnbsp;monuments, with the continuous line representing the edge of the stonenbsp;inscribed :�

B, I,

/d, d, t.

The classification of the vowels into broad and slender suggests that the inventor was a grammarian ; and the group which stands second in the usualnbsp;arrangement was probably the first to be fixed, as it is found that'.4, d, t, c, qu,nbsp;represent the initials of the Goidelic words for I, 2, 3, 4, 5, in the oldest formsnbsp;which can be inferred for them, thus : �a hbina-, a duou, a ttn-, a ccetuor,nbsp;a qq^nqtie. See Rhys�s � Outlines of the Phonology of Manx Gaelic � (in vol.nbsp;xxxiii. of the Publications of the Manx Society), p. 73, also pp. 41, 58, 59, 60nbsp;88. 102, 178.

B 2

THE WELSH PEOPLE, (chap, i.)

of Celts, Brythons or Goidels, came here first, and also whence they came. The answer to this latter questionnbsp;must be, that in all probability they first came from thenbsp;nearest part of the Continent, from the land where the bulknbsp;of the Celts dwelt in the time of Caesar, namely, Gaul,nbsp;comprising ancient France, Belgium, Switzerland, andnbsp;North Italy, also parts of Spain. To the question of thenbsp;order of their coming the answer is sufficiently indicatednbsp;by the relative positions of the peoples speaking Goidelicnbsp;and Brythonic respectively at the present day. For it maynbsp;be regarded as fairly certain that those who are foundnbsp;driven furthest to the west were the earlier comers, namelynbsp;the Goidels.

Goidelic was a phase of the language of the Celtae in Caesar�s restricted use of that word, and we may, in thisnbsp;context, call the language Goidelo-Celtic orCeltican butnbsp;we must suppose its place, at any rate as a dominantnbsp;speech on the Continent, to have been taken by Gaulishnbsp;some time anterior to Caesar�s Gallic wars. Gaulishnbsp;belonged to the same group as Brythonic ; or, to be morenbsp;exact, Brythonic may be treated as the Gaulish spoken innbsp;Britain, as we shall see presently. Both may be, however,nbsp;included under the term Galatic or Galato-Celtic.

The ancient distinction of speech between the Celts implies a corresponding difference of race and institutions,nbsp;a difference existing indeed long before Celts of anynbsp;description are mentioned in connection with these islands.nbsp;Perhaps few matters of prehistoric archaeology have recently been more discussed than the distinction betweennbsp;Celticans or Celts of the older stock and the Galaticnbsp;warriors by whom they were encroached upon. Thenbsp;two peoples are found to have differed in their mannernbsp;of disposing of their dead, and each had weapons

* See the �Transactions of the (London) Philological Society � for 1891-3, pp. 104, 105.

ETHNOLOGY OF ANCIENT WALES.

characteristic of its own civilisation. The interments with the most important remains of the older stock arenbsp;found mostly in the neighbourhood of the Alps, including the upper portions of the basin of the Danubenbsp;and the plains of North ItalyThis older Celtic worldnbsp;began, about the sixth century B.C., to be invaded bynbsp;the Galatic Celts, whose home may be inferred to havenbsp;consisted of Central and Northern Germany and of Belgium;nbsp;and the remains of these Galatic Celts are to be studied innbsp;the great burial-places between the Seine, the Marne, andnbsp;the Rhine�in the country, in short, from which theynbsp;invaded Britain. It has been surmised that this movementnbsp;was begun by the Brythons between the time of Pytheas,nbsp;in the fourth century B.C., and the visits of Julius Caesar.nbsp;The latter mentions, ii. 4, a certain Diviciacos, king ofnbsp;the Suessiones, a Belgic people which has left its name tonbsp;Soissons, as the most powerful prince in Gaul and as rulingnbsp;also over Britain. This was, moreover, late enough to benbsp;within the memory of men living in Caesar�s time. Wenbsp;have also Caesar�s general statement as to the settlers fromnbsp;Gaul, that they belonged to the Belgm or the tribesnbsp;inhabiting Belgium, by which he meant, roughly speaking,nbsp;the tract between the Seine and the Rhine and betweennbsp;the sea and the tributary waters of the Marne and thenbsp;Moselle; also that their settlements here were known by thenbsp;names which they already had on the Continent. This isnbsp;borne out by those names themselves where they happennbsp;to be recorded. Thus the Atrebates, whose chief townnbsp;was Calleva, supposed to be the site of Silchester, in Hampshire, probably came from the Atrebates of the Continent,nbsp;where that name of theirs has been worn down to Arras,nbsp;in the Pas de Calais. Then, as to the tribe called Belgm,

* For the most recent and comprehensive account of this chapter in Celtic archteology, see Bertrand and Reinach�s volumes � Nos Origines,� especiallynbsp;vol. ii., entitled �Les Celtesdans les Vall�es du P6 et du Danubequot; (Paris, 1894),nbsp;pp. ii., 42, 180, i8l. To this work, with its numerous illustrations, we arenbsp;much indebted.

6 nbsp;nbsp;nbsp;THE WELSH PEOPLE, (chap, i.)

who appear to have once been important enough to impose their name on a whole group of peoples on the Continent,nbsp;they seem to have come over here in a body, leaving to thenbsp;Germans their old home in a country now indicated bynbsp;such a place-name as that of Billig, in the governmentnbsp;of Cologne, and that of WdiSSamp;r-Billig, near where thenbsp;Sauer joins the waters of the Moselle. In Britain thenbsp;people called Belgae took possession of a tract whichnbsp;included the sites of the towns now called Winchester andnbsp;Bath. Similarly the Paris! of the south-east of what isnbsp;now Yorkshire occupied a country which is remarkablenbsp;for its interments, containing as they do remains of thenbsp;iron war-chariots used by the conquerors. They werenbsp;probably a division of the Parish, who have left their namenbsp;to the city of Paris, on the Seine. The first of the Belgicnbsp;peoples to cross over to this country were probably thosenbsp;who dwelt nearest to it, namely, on the other side of thenbsp;Straits of Dover. These appear to have been callednbsp;Brittani or Brittones, whom Pliny ^ seems to have foundnbsp;so called in the valley of the Somme : the name of a townnbsp;of theirs is duly perpetuated by that of the village ofnbsp;Bretagne, near the mouth of that river. In our countrynbsp;the history of their name was probably the following :nbsp;from being exclusively that of the first settlers it came tonbsp;be extended to the successive hordes, so that at the lastnbsp;it actually denoted all the settlers here of Belgic descent;nbsp;and we have probably to look for the Brittani proper undernbsp;the geographical appellation of Cantii, or people of Cantionnbsp;�� Kent.� In its form of Brittones it yields in Welsh thenbsp;name Brytlw^i, �a Briton or Welshman,� and in its form ofnbsp;Brittani it yields the Irish plural Bretain (genitive Bretan),nbsp;� Brythons or Britons, also their country.� The prevalent

^ See Detlefsen�s Pliny, � Historia Naturalis,*�iv. I06; A Scaldi incolunt extera Texuandri pluribus nominibus, dein Menapi^ Morini ora Marsacisnbsp;jxincti pago qui Chersiacus vocatur, Britannia Ambiani^ Bellovaci^ Bassi,

ETHNOLOGY OF ANCIENT WALES.

spelling in Latin appears to have been Britanni; but the related forms, and among them the French Bretagne, shownbsp;that it ought to have been written with tt.

Without enumerating the Belgic tribes which constituted the Brythonic group of Celts in this country, suffice it tonbsp;say that the whole of the coast on the east and south wasnbsp;probably occupied by them from the Isle of Wight to thenbsp;Firth of Forth. But how far inland they had overrun thenbsp;country by the time of the Roman occupation, it is impossible to say with any precision. We have, however, innbsp;the Itinerary of Antoninus, evidence of their presence asnbsp;far west as Staffordshire, to wit, in the name of thenbsp;station of Pennocriicion, which is now represented letternbsp;for letter by Penkridge. The form in the Itinerary is thenbsp;dative-ablative Pennocmcio, derived from the stem penno,nbsp;represented in Welsh by pen, �end, extremity, head,� earlynbsp;Goidelic quenna, whence Irish ceann, � end, head,� and criicio,nbsp;Welsh crug, � a heap or tumulus.� So the whole compound,nbsp;had it been Goidelic, would have been Qiiennacrucio innbsp;the Itinerary. But as that is not the case, we know thatnbsp;we have here to do with Brythons, not with Goidels, Thenbsp;spot was comprised probably in the territory of the Cornovii,nbsp;who may accordingly be supposed to have been Brythons.nbsp;Behind them towards the west were the Ordovices, whonbsp;were also probably Brythons, though we have no exactlynbsp;similar evidence to prove it; but we can account best fornbsp;the facts which we have at our disposal, if we supposenbsp;the Ordovices to have divided what is now Wales into twonbsp;parts, as it were with a wedge, and to have reached thenbsp;shores of Cardigan Bay. Beyond the Ordovices, towardsnbsp;the south, the country was occupied by the Silures and thenbsp;Demetm, neither of whom can be supposed to have beennbsp;Brythons, for their territory supplies inscriptions in Goidelicnbsp;dating some time after the departure of the Romans.nbsp;Similarly beyond the Ordovices in a northern and north-

THE WELSH PEOPLE, (chap, i.)

western direction in the country called Venedot, later Gwyned, there were likewise peoples who cannot have beennbsp;Brythons, as is proved more or less explicitly by inscriptionsnbsp;found there showing traces of Goidelic.

We have some aid to the understanding of the early history of the tract of country which we may loosely callnbsp;Mid-Wales in the dialects of Welsh actually spoken there.nbsp;Welsh dialects are commonly treated as three: (i) the Gwyn-dodeg, or Venedotian Welsh ; (2) the Powyseg, or Welsh ofnbsp;Powys ; and (3) the Southwalian, consisting of the closelynbsp;connected Gwenliwyseg (Gwentian or Silurian Welsh) andnbsp;the Demetian or Welsh of Dyfed. But this is of little usenbsp;as regards Mid-Wales, for the facts are briefly as follows :�nbsp;The Demetian reaches northwards as far as the stream ofnbsp;the Wyrai at ILanrhystud, a little south of Aberystwyth,nbsp;in Cardiganshire. North of the Wyrai and of Strata Floridanbsp;the dialect is very distinct from the Demetian, and stillnbsp;more so from the contiguous dialect of the counties ofnbsp;Merioneth and Montgomery; but, on the other hand, itnbsp;resembles that of Penftyn as represented by Bala and othernbsp;places in the neighbourhood of Bala Lake. We may addnbsp;that the same dialect embraces, without any sudden variation, the Welsh of the Dee valley and some of the northernnbsp;portion of Montgomeryshire. The inference to be drawnnbsp;is probably this: at one time a uniform dialect prevailednbsp;in the region of Mid-Wales comprising North Cardiganshire, Radnorshire, Montgomeryshire, and the south-east ofnbsp;Denbighshire. Its northern boundary was probably thenbsp;river Mawdach, the northern watershed of the Dee, andnbsp;eventually the Clwyd and the Dee Estuary. The areanbsp;indicated may be said to represent the conquests andnbsp;settlements of the Ordovices. Later, however�not longnbsp;probably after the departure of the Romans from Britain�nbsp;another people came on the scene and established its ownnbsp;dialect in a great part of the Ordovic territory, but so as to

ETHNOLOGY OF ANCIENT WALES.

leave the marginal districts of North Cardiganshire and Penityn in the continued use of their old Ordovic speech.nbsp;The intruding dialect, which severed them, is now no othernbsp;than the Welsh of Powys, which prevails in Montgomeryshire and on the coast of Merionethshire from the rivernbsp;Dovey to Dolgeliey and Harlech. It was introducednbsp;probably by Cuneda Wledig and his Sons, that is to say,nbsp;by the people whose princes are collectively known tonbsp;Welsh history and hagiology by that designation.

Now Cuneda was the ancestor to whom the kings of Gwyned traced their origin, and most of the best-knownnbsp;saints of Wales are represented as his descendants. Thenbsp;legend ^ of Cuneda represents him and his sons coming tonbsp;Wales from the North, where he appears to have been atnbsp;the head of a force of cavalry defending the Roman wall,nbsp;and to have filled the post of the military leader, who,nbsp;during the Roman occupation, used to be known as thenbsp;Dux Britannim or Dux Britan7iiarum. The legend furthernbsp;represents Cuneda and his Sons as engaged in Wales innbsp;the expulsion of the Goidels; so we may suppose that thenbsp;Ordovices had been hard pressed by the Goidels on bothnbsp;sides of them, and that they had appealed for aid tonbsp;Cuneda, who accordingly sent a force under the commandnbsp;of his sons to combat the Goidels. However that maynbsp;be, Cuneda�s men must have permanently settled in thenbsp;country, and so did his sons, as we are told, except thenbsp;eldest, who is said to have died some time before in Manaunbsp;Guotodin, a district near the Firth of Forth. The share ofnbsp;the eldest son was given to his son Meirion, from whomnbsp;it was called Cantref Meirion, � the Hundred of Meirion,�nbsp;which in its turn has given its name to Meiriony� and thenbsp;county of Merioneth. This Hundred of Meirion, as the

* See San Marte�s � Nennius,� � 62, and the Harleian MS. 3859, published by Mr. Egerton Phillimore in ��Y Cymmrodor,� ix. 182, also Rhys�s � Celticnbsp;Britain,� pp. 118-21.

THE WELSH PEOPLE, (chap, i.)

dominion of the eldest branch of the Cuneda family, is probably what was sometimes called simply Y Cantref}nbsp;quot;the Hundred.� It was also that which an old tract onnbsp;boundaries calls Cantref Oriwyf, which possibly meantnbsp;the Hundred of the Ordovices, extending from the Doveynbsp;to the Mawdach, the southern boundary of Ardudwy, whichnbsp;belonged to Gwyned.^ Now Meirion, grandson of Cuneda,nbsp;as chief of the family, assigns to his uncles, the sons ofnbsp;Cuneda, their respective territories, and the list of themnbsp;purports to represent the Brythonic conquests made by thenbsp;family in various directions, including the carving out ofnbsp;Keredig�s kingdom of Keredigion, which is, roughly speaking, the present county of Cardigan. Other members ofnbsp;the family have different parts of Gwyned, including Monnbsp;or Anglesey, assigned to them in this legend. One of thenbsp;branches of the Cuneda family established in Gwynednbsp;became in the sixth century so powerful, to wit in thenbsp;person of Maglocunos or Maelgwn Gwyned, that not onlynbsp;Wales, but also Cumbria, was, to a certain extent, forced tonbsp;own his sway; and it was perhaps in his time that thenbsp;Hundred of Meirion first became a part of Gwyned.

We have suggested that the Brythons came over to Britain between the time of Pytheas and that of Julius Csesar.nbsp;Their invasions probably spread over a considerable periodnbsp;of time, but we should perhaps not be far wrong in ascribingnbsp;the mass of them to the second century before our era.�nbsp;When, it may be asked, did the other Celts, the Goidels,

� See the �lolo MSS.� (Handovery, 1848), pp. 85, S6, translated at pp. 475-7, but with serious errors; also pp. 120, 519 ; and Rhys�s �Celticnbsp;Britain,� pp. 308, 309.

* See the Mabinogi of Math, in which Math has the disposal of Ardudwy, probably as king of Gwyned, Oxford Mabinogion, p. 73.

� See M. d�Arbois de Jubainville�s � Premiers Habitants de 1�Europe,� vol. ii., p. 295, where he expresses himself to the same effect, also in thenbsp;� Revue Celtique,� vol. xiii., p. 402 ; see also Holder�s � Alt-Celtischer Sprach-schatz,� S.V., Brittani. Both, however, make the Brythons conquer this islandnbsp;from the Piets, not from Goidels.

ETHNOLOGY OF ANCIENT WALES. ii

whom they found here, arrive in this country ? It is impossible to give any precise answer to such a question, but it may be supposed that the Goidels came over not later thannbsp;the great movements which took place in the Celtic worldnbsp;of the Continent in the sixth and fifth centuries before ournbsp;erad We mean the movements which resulted in the Celtsnbsp;reaching the Mediterranean and penetrating into Spain,nbsp;while others of the same family began to press towards thenbsp;east of Europe, whence some of them eventually crossed tonbsp;Asia Minor and made themselves a home in the countrynbsp;called after them Galatia. On the whole, we dare notnbsp;suppose the Goidels to have come to Britain much laternbsp;than the sixth century B.C.; rather should we say that theynbsp;probably began to arrive in this country very much earlier.nbsp;Before the Brythons came the Goidels had presumablynbsp;occupied most of the island south of the firths of the Clydenbsp;and Forth. So when the Brythons arrived and began tonbsp;press the Goidels in the west some of the latter may havenbsp;crossed to Ireland : possibly they had begun still earliernbsp;to settle there. The portion of Ireland which they firstnbsp;occupied was probably the tract known as the kingdomnbsp;of Meath, approximately represented now by the diocesenbsp;of that name ; but settlements may have also been madenbsp;by them at other points on the coast.

We have next to consider the question whether the first Celtic comers, the Goidels, were also the first inhabitantsnbsp;of this country. This may be briefly answered to thenbsp;effect that there seems to be no reason to think so, or evennbsp;to suppose that it may not have been uninterruptedlynbsp;inhabited from a time before it ceased to form a continuousnbsp;portion of the continent of Europe. By what race is anbsp;much harder question. Indeed, there is a previous question

* See the � Premiers Habitants de 1�Europe,� vol. i., p. 262, and Zimmer�s �Mutterrecht der Pikten,quot; in the �Zeitschrift fiir Rechtsgeschichte,quot; vol. xv.nbsp;(Rom. Abth.), pp. 233, 234. For later views see Read�s British Museumnbsp;Bronze Age Guide.

THE WELSH PEOPLE, (chap, i.)

which might reasonably be asked, namely, Was it a single race or several ? This cannot be answered, but it wouldnbsp;clearly be a waste of conjecture to suppose the pre-Goidelicnbsp;inhabitants to have belonged to more than one race, untilnbsp;at any rate evidence is found to compel us to that conclusion. So we rest satisfied for the present to treat them asnbsp;if belonging to a single race ; and we proceed to consider,nbsp;very briefly, the nature of the relation in which that race isnbsp;likely to have found itself placed as regards the new-comers.nbsp;It is but natural to suppose that the Goidels, when theynbsp;arrived, subjugated the natives, and made slaves of themnbsp;and drudges. From the first the fusion of the two racesnbsp;may have begun to take place, but to what extent it proceeded it would be impossible to say. It is, however,nbsp;fairly certain that the process of fusion between the Goidelicnbsp;and native elements must have been quickened by thenbsp;advent of a third and hostile element, the Brythonic. Fornbsp;it must have been to the advantage of the Goidels to havenbsp;induced the natives to make common cause with themnbsp;against the intruders; and under the pressure exerted bynbsp;the Brythons the fusion of the two other nations may havenbsp;been so complete as to produce a new people of mixednbsp;Goidelic and native origin. To be more correct, perhaps,nbsp;we ought to have restricted the term Goidelic to that mixednbsp;nationality, and applied some other designation, such asnbsp;Celtican, to the early Celtic invaders of the island beforenbsp;they mixed with the Aborigines of these islands. Then asnbsp;to the Brythons, coming later as they did, they had thenbsp;Goidels between them and the Aborigines, and they were notnbsp;likely to come in contact on any large scale with the latternbsp;before they had been to a considerable extent Celticised.^

1 Except perhaps in the North, where, for example, the Picto-Brythons of Fortrenn, with their headquarters eventually at Forteviot, on the banks of thenbsp;Earn, seem to have spoken a kind of Brythonic. But that dialect is unknownnbsp;with the exception of a few words like Peanfahd, given by Bede (� Hist.nbsp;Eccles.� i. 12) as the vernacular for Latinpmnae valli.�

ETHNOLOGY OF ANCIENT WALES. 13

Accordingly, supposing the Aborigines not to have been Aryans, one might expect the language of the resultantnbsp;Goidelic people to show more non-Aryan traits than thenbsp;language of the Brythons : as a matter of fact, this provesnbsp;to be the case.

The Goidels have already been represented as a mixed race, and when later this mixed Goidelic population becamenbsp;one people with the Brythons, the result was still morenbsp;composite ; and one may say that the Welsh people of thenbsp;present day is made up of all three elements: the Aboriginal,nbsp;the Goidelic, and the Brythonic. And it would be unsafenbsp;to assume that the later elements predominate; for thenbsp;Celtic invaders, both Goidels and Brythons, may havenbsp;come in comparatively small numbers, not to mention thenbsp;fact that the Aboriginal race, having been here possiblynbsp;thousands of years before the first Aryan arrived, may havenbsp;had such an advantage in the matter of acclimatisation,nbsp;that it alone survives in force. This is now supposed to benbsp;the case with France, whose people, taken in the bulk, arenbsp;neither Frankish nor Celtic so much as the representativesnbsp;of the non-Aryan populations which the first Aryans foundnbsp;there. It thus becomes a matter of interest for us to knownbsp;all we can about the earliest inhabitants of this country.nbsp;Now the question of the origin of that race is, according tonbsp;one view taken of it, inseparably connected with the Pictishnbsp;question ; and the most tenable hypothesis may be said tonbsp;be, that the Piets were non-Aryans, whom the first Celticnbsp;migrations found already settled here. The Piets appearnbsp;to have retained their language and institutions latest onnbsp;the east coast of Scotland in portions of the region betweennbsp;Clackmannan and Banff. But Irish literature alludes tonbsp;Piets here and there in Ireland, and that in such a way asnbsp;to favour the belief that they were survivals of a race holdingnbsp;^possession at one time of the whole country. If the Piets werenbsp;not Aryans, we could hardly suppose them to have been

THE WELSH PEOPLE, (chap, i.)

able to acquire possession of extensive tracts of these islands after the arrival of such a powerful and warlike race as thenbsp;early Aryans. The natural conclusion is, that the Pietsnbsp;were here before the Aryans came, that they were, in fact,nbsp;the Aborigines.

Now something is known of the manners and customs of the ancient Piets ; for one of them at least was so remarkable as to attract the attention of the ancient authors whonbsp;mention the peoples of this country. It was the absencenbsp;among them of the institution of marriage as known tonbsp;men of Aryan race. This is illustrated by the history ofnbsp;the Piets in later times, especially in the case of their kings,nbsp;for it is well known that a Pictish king could not benbsp;succeeded by a son of his own, but usually by a sister�snbsp;son. The succession was through the mother, and it pointsnbsp;back to a state of society which, previous to the conversion of the Piets to Christianity, was probably based onnbsp;matriarchy as distinguished from marriage and maritalnbsp;authority. Accordingly the Greeks and Romans who havenbsp;touched on the manners and customs of the Piets shownbsp;clearly that they could not understand the relations of thenbsp;sexes among peoples of that race, except as mere licencenbsp;and wanton promiscuity. Among others may be mentionednbsp;Dion Cassius, who in writing (Ixxvi. i6) about the wars ofnbsp;the Emperor Severus introduces, for the evident benefit ofnbsp;Roman women, a Pictish lady, who replies to the stricturesnbsp;of Julia, the Emperor�s wife, on Pictish morality, to thenbsp;effect that she thought the Pictish custom the better, since,nbsp;as she said, Pictish ladies openly consorted with the bestnbsp;warriors of the race, while Roman matrons privily committed adultery with the vilest of men. Further thenbsp;Pictish succession cannot have always been confined to thenbsp;Pictland of the North,^ for the ancient literature of Ireland

^ Since writing the above we have come across a passage showing that the same kind of succession once prevailed at Tara : see the Place-name Story of

ETHNOLOGY OF ANCIENT WALES.

abounds in allusions to heroes who are usually described with the aid of the mother�s name. Take the case ofnbsp;Conchobar mac Nessa, � Conor son of Nessa � (his mother),nbsp;Diarmait Ua Duibne, �Dermot descendant of Dubinn� (hisnbsp;ancestress), not to mention the gods called Tuatha D�nbsp;Datiann., �the tribes of the goddess Danu or Donu,� whichnbsp;Donu appears also in the old literature of Wales, to witnbsp;as Don in the names of such personages as � Gwydionnbsp;son of Don,� � Arianrhod daughter of Don,� and others ofnbsp;the same family, all placed by the Mabinogion on thenbsp;northern coast of Gwyned. This kind of nomenclaturenbsp;implies the Pictish succession as its origin, and probablynbsp;all that such origin implied. If the Aryans ever hadnbsp;this kind of custom, a view which is not universallynbsp;accepted, it was probably so very far back that wenbsp;could not with any confidence invoke it to explainnbsp;these designations and others like them ; so we arenbsp;inclined to regard them as having originated in non-Aryan surroundings.

The same conclusion as to the probable non-Aryan origin of the Piets is warranted by facts of another order,nbsp;namely, those of speech; but the Pictish question isnbsp;rendered philologically difficult by the scantiness of thenbsp;remains of the Pictish language. It would seem to havenbsp;been rapidly becoming overloaded with loan-words fromnbsp;Goidelic and Brythonic when we first hear anything aboutnbsp;it. So, failing to recognise this borrowing of words bynbsp;the Piets, some have been led to regard Pictish as a kind of

Druim Criaich, edited with a translation by Stokes in the �Revue Celtique,� xvi. 148-50. The storyteller undertakes to explain the peculiarity of thenbsp;succession: he first relates how the three sons of the Irish king, Eochaidnbsp;Feidlech, rebelled against their father, and how they fell in the conflict. Henbsp;then adds words to the following effect;��Then before nightfall their threenbsp;heads came to Druim Criaich, and there Eochaid uttered the word, that fromnbsp;that time forward no son should overtake the lordship of Tara after his fathernbsp;unless some one came between them.�

THE WELSH PEOPLE, (chap, i.)

Gaelic, and some as a dialect akin to Welsh. The point to have been decided, however, was not whether Gaelic ornbsp;Welsh explains certain words said to have been in usenbsp;among the Piets, but whether there does not remain anbsp;residue to which neither Gaelic nor Welsh, nor, indeed,nbsp;any Aryan tongue whatsoever, can supply any sort of key.nbsp;This has of late begun to be perceived,^ and all the morenbsp;clearly now that the ancient inscriptions found in thenbsp;Pictland of the North have been more carefully studied.*�nbsp;The whole group of inscriptions, however, is a very smallnbsp;one, and it shows the manifold influence of Gaelic andnbsp;Norse, especially in Shetland, for Pictish cannot havenbsp;become extinct for some time after the earlier visitsnbsp;of the Norsemen to our coasts. Among those inscriptionsnbsp;and fragments of inscriptions, there are two or three whichnbsp;may be said to be fairly legible; and one of them isnbsp;punctuated word by word. Nevertheless the adherentsnbsp;to the view that Pictish is Celtic and Aryan have innbsp;vain been challenged to produce a convincing translation.nbsp;Neither Gaelic nor Welsh seems to be of any materialnbsp;avail in the effort, and one may confidently surmise thatnbsp;any other Aryan language will be found of still less use, ifnbsp;possible. This being so, it is not too much to say that thenbsp;theory of the non-Aryan origin of the Pictish languagenbsp;holds the field at present.

Precarious as the Pictish inscriptions must be admitted to be, they have supplied the key to the interpretation ofnbsp;certain other inscriptions, to wit, in Wales, Cornwall, andnbsp;Ireland. Thus a short one in minuscules at St. Vigeans,nbsp;near Arbroath, reads, Drosteri ipe uoret ett forms, which

1 nbsp;nbsp;nbsp;See the �Revue Celtique,� vi. 398, 399, and Zimmer�s article in thenbsp;�Zeitschrift f�r Rechtsgeschichte,� p. 217.

2 nbsp;nbsp;nbsp;Lists of them will be found in Lord Southesk�s �Ogams at Brodie,nbsp;Aquhollie, Golspie, and Newton� (Edinburgh, 1886), p. 38, and in thenbsp;� Proceedings of the Society of Antiquaries of Scotland,� vol. xxvi.,nbsp;pp. 267-304, to which may now be added vol. xxxii., pp, 324-98.

ETHNOLOGY OF ANCIENT WALES.

maybe for the present rendered �Drost�s offspring,^ Uoret, for Fergus.� Similarly a well - known monument atnbsp;Newton, in Aberdeenshire, bears in Ogam what may benbsp;provisionally read Vorrenn ipuai Osir, which may, in thenbsp;same way, be interpreted as � Vor�s offspring Osir,� thatnbsp;is to say, Osir son of Vor or Vaur, for the name is readnbsp;Vaur in the script on the same stone. From these onenbsp;learns how to construe the following, found at ILanfaglannbsp;near Carnarvon �

FILI LOVERNII ANATEMORI.

That meant, no doubt, � (the monument) of Lovernias�s son, Anatemoras,� that is, of Anatemoras son of Lovernias.nbsp;This inscription was written in Latin, and Anateniori wasnbsp;the Celtic genitive of a proper name which in Brythonicnbsp;would have been Anatj(omaros, Welsh eneid-fawr, � great-souled,� fj.ijaX�xjnjxp'i. From this cannot be severed thenbsp;following, which is to be seen at Helston, in Cornwall;nbsp;Cnegimii fili Genaius? Though meant to be Latin, it hasnbsp;to be construed according to the grammar of a differentnbsp;idiom, for fili is here treated as the crude stem of thenbsp;word, so that fiUGeiiai-tis is to be regarded as doing dutynbsp;for filius and Genahcs in apposition, and with only one casenbsp;termination. The whole means Genaius filius Cnegumi,nbsp;but the syntax is not that of an Aryan language. It isnbsp;familiar, however, in agglutinative languages like Basque,nbsp;and it occurs in our inscriptions too frequently to benbsp;regarded as a slip. Thus we have it both in Latin and innbsp;Goidelic on a tombstone found at Clydai, in Pembroke-

* It will be seen later (p. 50) that a more probable rendering than �offspring� would be �kin� or �nephew,�,and so probably with the vocable poi of certain Irish Ogams.

^ See H�bner�s � Inscriptiones Britanni:e Christiance,� No. 147.

^ Hiibner, No. 5.

W.P. nbsp;nbsp;nbsp;C

THE WELSH PEOPLE, (chap, i.)

shire: the Latin reads ETTERNI FILI VICTORS wherefiliVictor is treated as one noun in the nominative,nbsp;and enough of the Ogam remains to show that the construction in the Goidelic version was identical. The wholenbsp;meant �^Eternus�s son Victor,� that is, Victor son ofnbsp;.Eternus. Another instance of the same agglutinativenbsp;syntax occurs on a stone at Eglwys Cymun, in the southwest of Carmarthenshire: it reads in Latin,¹�

AVITORIA FILIA CVNIGNI,

and in Goidelic Inigena Cunigni Avittoriges, � (the monument) of Avitoria, daughter of Cunignas.� Here the nouns in apposition are inigena, quot;daughter,� and Avit-toriga ; so the genitive ending es is applied to the latternbsp;alone, while the genitive of the father�s name is insertednbsp;between the two feminines. From Ireland may be mentioned an inscription at Dunloe, near Killarney, whichnbsp;runs thus : Maqui Ttal maqui Vorgos maqui mucoi Toicac,nbsp;� (the monument) of MacTail, son of Fergus, son � amp;c. Innbsp;correct Goidelic this should have been Maqui Ttali maquinbsp;Vorgossos^ maqui, amp;c.; but Ttal and maqui, Vorgos andnbsp;maqui, being respectively in apposition, have only onenbsp;mark of the genitive each. The same construction isnbsp;shown in one of the northern Ogam inscriptions of Ireland :nbsp;it stands at a spot some twelve miles north-east ofnbsp;Omagh, in Tyrone, and reads: Dotoatt maqui Nan . . .^nbsp;� (the monument) of Dotoatt, son of Nainnidh.� Herenbsp;Dotoatt-maqui must be construed as an agglutination with

1

nbsp;nbsp;nbsp;Hiibner, No. no.

- See the � Archseologia Cambrensis � for 1893, p. 285.

It is not necessarily Vorgossos, as Vorgos may have represented a genitive which in a later form occurs as Forgo or Forco.

* nbsp;nbsp;nbsp;See the �Journal of the Royal Society of Antiquaries of Ireland� for 1895,nbsp;p. 104.

ETHNOLOGY OF ANCIENT WALES. ig

only one genitive termination, to wit, the t at the end of maqui, �son�s, filii'.'

The distribution of inscriptions with this non-Aryan syntax suggests that the British Isles were once inhabitednbsp;by a people speaking a non-Aryan language, and that,nbsp;while that people learned the vocabulary of an Aryannbsp;language, it continued the syntax of its previous speech.nbsp;This was so decidedly the case, that we trace it not onlynbsp;in the Goidelic which that people definitely adopted, butnbsp;also in the Latin which its learned men now and thennbsp;wrote. There is nothing incredible in this, as habits ofnbsp;pronunciation and the syntax peculiar to a language arenbsp;most persistent and difficult to eradicate, even when carefulnbsp;teaching is directed to that end, as anybody will admitnbsp;who knows anything of the difficulties of teaching Welshnbsp;boys idiomatic English. One sees accordingly how thenbsp;Goidelic of the west of Britain may have been profoundlynbsp;modified by the pronunciation and syntax of the non-Aryan language of the Aborigines; but to what extentnbsp;the Brythonic conquerors of Mid-Wales may have clearednbsp;the latter area of its ancient inhabitants, whether mostlynbsp;Goidelic or native, it is impossible to say. Beyond thosenbsp;conquests, however, the old inhabitants of the Venedotiannbsp;north on the one hand, and those of the Siluro-Demetiannbsp;south on the other, are not likely to have been displacednbsp;on any considerable scale. We are accordingly at libertynbsp;to regard the Ordovic territory of Mid-Wales as the mostnbsp;thoroughly Brythonic. It might appear at first sight anbsp;remarkable corroboration of this view, that Mid-Walesnbsp;shows no inscriptions in Ogam at all or any inscriptionsnbsp;whatsoever in which Goidelic can be traced. But as tonbsp;the absence of Goidelic traces in Mid-Wales, it must be atnbsp;once explained, that this region has very few inscriptionsnbsp;at all to show corresponding to the post-Roman ones to benbsp;found more to the north and more to the south. That

c 2

THE WELSH PEOPLE, (chap. I.)

very fact, however, is not without a significance of its own, for it seems to show that the burial customs of the ancientnbsp;Brythons of Mid-Wales differed from those of the othernbsp;peoples; add to this that Mid -Wales has few or nonbsp;cromlechs to show. Possibly the inhabitants buried theirnbsp;dead in barrows, where there was less inducement tonbsp;indulge in writing than in the case of peoples who putnbsp;up stones as memorials of their departed.

It may be objected that these arguments as to Mid-Wales are mostly negative, but there is at least one argument of a more positive kind�an argument basednbsp;on the actual pronunciation of Powysian Welsh. In thatnbsp;dialect the vowel a, whether short or long, has a narrownbsp;pronunciation resembling that of the narrow a in thenbsp;standard English pronunciation of such words as man andnbsp;bad. This vowel may be indicated as a, and we have it innbsp;such words as cam, �crooked,� mam, � mother,� pan, �when,quot;nbsp;and car, �a cousin.� Further, cam and car tend to become,nbsp;and have, in fact, extensively become, kiam and kiar, combinations from which the analogy of other languages wouldnbsp;lead us to expect eventually some such forms as tskqm andnbsp;tshqr, or even sham and shtir, that is, provided this dialectnbsp;of Welsh continued long enough a spoken language notnbsp;too much restrained by the yoke of the standard spelling.nbsp;The change here indicated is just what has happened innbsp;France: the Gauls appear to have pronounced their anbsp;narrow, and when they adopted Latin they could probablynbsp;not help continuing their old pronunciation with its q.nbsp;The result has been that French has made Latin wordsnbsp;like nassus, �nose,� and pratum, �a meadow,� into nez andnbsp;pr�, and caput, � head,� into chef, pronounced with tsh as innbsp;its English form of chief, and later, with ch � English shnbsp;as in standard French at the present day. The same wasnbsp;the case with such a word as Latin castra, � a camp,� whichnbsp;some of the Gauls and the Brythons of this country seem

ETHNOLOGY OF ANCIENT WALES. 21

to have pronounced kiastra, which the ancestors of the English borrowed and made into ceaster, whence thenbsp;modern Chester, pronounced Tshester. The initiative in allnbsp;these changes is the pronunciation of a as q, which is anbsp;sound with a tendency to become e; and the most naturalnbsp;explanation of the fact that q and iq occur in Mid-Walesnbsp;Welsh is, that some of the Brythons, like the other Celts ofnbsp;Northern Gaul, had q as the sound of a in their languagenbsp;when they penetrated to the west of this island. To thenbsp;extent here indicated by the pronunciation, the Welsh ofnbsp;Powys is more like that of ancient Gaulish, that is to say,nbsp;it is more purely Brythonic, than any of the other dialects.nbsp;The reason for this is, probably, that the language of thenbsp;Ordovices had been modified by the gradual absorption ofnbsp;other nationalities into that tribe, as it extended its conquests towards the west, while the Sons of Cuneda camenbsp;from a district called Manau in the land of the Guotodin,nbsp;Ptolemy�s Otadini, or better Votadini. It was in or nearnbsp;their territory that the Roman wall reached the North Sea,nbsp;and to them also belonged the coast northwards to the Forth,nbsp;including what is now known as the Lothians. It isnbsp;possible that the Votadini were not the first Brythons tonbsp;occupy that strip of country; and it is probable that thenbsp;previous inhabitants were driven inland, either by them ornbsp;by previous invaders of kindred nationality. In either casenbsp;the Gaulish of the Votadini may have had the chance ofnbsp;remaining less influenced by the idioms of the country,nbsp;than can have been the case with tribes whose assimilationnbsp;of the peoples conquered by them in war proceeded on anbsp;larger scale and for a longer period of time.

South of the domain of Powysian Welsh we have Brecknockshire, giving a broad pronunciation to the vowel a, whether long or short. But in the Welsh of Monmouthshire and Glamorganshire as far as the Neath Valley and thenbsp;western limit, approximately, of the Custom of Glamorgan

THE WELSH PEOPLE, (chap, i.)

(p. 30), the short a alone is broad, long a being uniformly made narrow. This seems to indicate that at one time thatnbsp;part of the southern border of the Principality came undernbsp;the influence of some Brythonic people pressing westwards,nbsp;such, for anything known to the contrary, as the Dobunninbsp;near the mouth of the Severn may have been.

It is, however, not to be supposed that Brythonic was formed in Wales : we are compelled by the close similaritynbsp;of Welsh with Old Cornish and Breton to suppose thatnbsp;the language in all its essential features was formed beforenbsp;the Ordovices reached the shores of Cardigan Bay; andnbsp;we have the means of gauging to some extent how far itnbsp;has deviated from Gaulish. A certain number of sentencesnbsp;are extant in Gaulish, and they invariably observe thenbsp;ordinary usage of Aryan syntax in not placing the verbnbsp;before its subject: take for example the following�

Sfyo/rapos OviXXoveo^ roourtovs Na/nautraTis eiwpov (Toaiv vefirjTOv

That is, �Segomaros son of Villonos, magistrate of Nimes, made for (the goddess) Belesama this temple.� Or this,nbsp;Ratin brivatioin Frontii Tarbeisonios ieuru ; that is, � Pro-pugnaculum pontilium Fronto, Tarbeisoni filius, fecit.� ^ Itnbsp;is unfortunate that not one of the Gaulish sentences extantnbsp;happens to come from the ancient Belgium; but therenbsp;is no reason to suppose that the Gaulish of Belgic Gaulnbsp;differed in its syntax from the Gaulish of other parts ofnbsp;the country. On the other hand, the normal syntax ofnbsp;the Neo-celtic languages requires the verb to precede itsnbsp;subject, and the question arises how this important difference began. It might be suggested as an explanation,nbsp;that the earlier Celts mixed with a non-Aryan race, whosenbsp;language had this syntactic peculiarity of Neo-celtic asnbsp;regards the position of the verb, and that they thus evolved

* According to Stokes�s �Celtic Declension� (Gottingen, 1886), pp. 60, 67.

ETHNOLOGY OF ANCIENT WALES. 23

the Goidelic language. The next stage might similarly be supposed to be a mixing of the Brythons with the Goidelsnbsp;of the description just suggested, when it became thenbsp;turn of the latter to be conquered, the result being thatnbsp;Brythonic emerged, having indirectly acquired some ofnbsp;the linguistic peculiarities of the Aboriginal inhabitants ofnbsp;Gaul, of Britain, or of both. Whatever the real explanation may prove to be, it will, in all probability, have tonbsp;take for granted a racial amalgamation on a considerablenbsp;scale. But the linguistic conditions seem to us, it is needless to say, to postulate a pre-Celtic race whose languagenbsp;was characterized by the chief peculiarities distinguishingnbsp;Neo-cel tic from Gaulish.^

The foregoing remarks amount briefly to this: the Goidelic and the Aboriginal elements should be expectednbsp;in their greatest strength in the south and in the north ofnbsp;Wales, while Mid-Wales is marked out by the Gaulishnbsp;affinities of the Powys dialect, and by the absence ofnbsp;monuments betraying any traces of Goidelic influence, asnbsp;the home of the Brythonic element in the west of thenbsp;island. We are, however, unable to detect in the habitsnbsp;or physical characteristics of the people at the presentnbsp;day any salient features corresponding to these vanishingnbsp;landmarks. Thus the Aboriginal non-Aryan ideas as tonbsp;marriage might, conceivably, have survived long in thenbsp;modified form of a tendency to take somewhat too lenientnbsp;a view of immorality, but the statistics of illegitimacy innbsp;Wales do not represent its geographical distribution to benbsp;such as clearly to suggest any such permanence of influence.nbsp;On the other hand, men of purely Aryan descent arenbsp;supposed to have been, like the ancient Gauls and thenbsp;ancient Germans, inclined to be of light complexion andnbsp;tall stature, which would, perhaps, imply the requirement

gt; We are glad to be able to refer our readers to an elaborate treatment of this question in Appendix B.

THE WELSH PEOPLE, (chap, i.)

of more food than in the case of men of smaller build. We have, however, no statistics to show whether the people ofnbsp;Mid-Wales are on the whole taller or blonder than othernbsp;Welshmen, but the Land Commission heard some evidencenbsp;to the effect that the former fare better in the matter ofnbsp;food, especially in the county of Montgomery.

When and how the comparative homogeneity of the Welsh people was produced, it is impossible to say withnbsp;any approach to precision. First of all, however, thenbsp;distinctive customs of the Aborigines must have givennbsp;way gradually to those of their Goidelic masters, thoughnbsp;hardly without affecting the latter themselves; and thenbsp;native language yielded, no doubt, comparatively early tonbsp;Goidelic. As to Goidelic itself, its turn to go came in duenbsp;time. We have no evidence that it was spoken in anynbsp;part of Wales in the eighth or ninth century ; but it wasnbsp;probably not dead till well into the seventh. Besidesnbsp;having a language of their own, however, the Goidels mustnbsp;have had also their own laws ; and these, it would seem,nbsp;proved in some respects more tenacious than their language.nbsp;Welsh literature speaks of one great and conspicuous legislator of this island in early times. He is called Dyfnwalnbsp;Moel-Mut, but the name Dyfnwal, answering as it doesnbsp;exactly to the Irish Domhnall, Anglicised Donald, teachesnbsp;us nothing precise as to his race. On the other hand, thenbsp;epithet Moel-nncd, in its oldest form Moel Mut, cannot benbsp;other than Goidelic, and its historical form in Irish is Moelnbsp;Muaid, which we have in Ua Maol-muaidh, in Englishnbsp;spelling 0�Molloy. It should mean � the tonsured man ornbsp;slave of Muad,� in the same way that Mael-Patraic ornbsp;Midpatric was made into Calvus Patricii} In the vocable

* See Rhys�s �Goidelic Words in Brythonic� in the �Archseologia Cambrensis� for 1895, pp. 299, 300; Nigra�s �Reliquie Celtiche� (Turin,nbsp;1870), p. 19; Stokes�s �Goidelica� (London, 1872), pp. 86, gi ; and Rhys�snbsp;�Celtic Britain,� pp. 73-5.

ETHNOLOGY OF ANCIENT WALES.

Muaid, for an earlier Moti, we have the genitive of the name, possibly of some forgotten divinity, and the formula is notnbsp;Christian, but merely retained in use in Christian times.nbsp;This personage, whose full name proves him to have been nonbsp;thorough Brython, is the accredited legislator of the Cymry ;nbsp;and, according to one of the manuscripts of the Venedotiannbsp;version of the Laws of Howel, the former�s laws continuednbsp;in force to the time of Howel. But even Howel, descendednbsp;as he was from Cuneda, and representing the Brythonicnbsp;element as he did, thought it inexpedient to undo thenbsp;whole of the work of the Goidel. He left undisturbed hisnbsp;reckoning of measurement from the barley-corn up to thenbsp;acre and the mile, as he did also his divisions of the countrynbsp;into cantrevs and their subdivisions. The words are tonbsp;the following effect : ^ � And he [Dyfnwal] was a man ofnbsp;authority and wisdom ; and he (first) made good laws innbsp;this country, which laws continued in force till the time ofnbsp;Howel the Good. Afterwards Howel enacted new lawsnbsp;and annulled those of Dyfnwal; and (yet) Howel did notnbsp;disturb the measurements of lands in this island, but [letnbsp;them continue] as Dyfnwal left them ; for the latter wasnbsp;the best man at measuring.� Without entering into thenbsp;question how far the laws of the Goidels differed fromnbsp;those of the Brythons, and to what extent Howel reallynbsp;modified^ the laws obtaining till then in any part of hisnbsp;kingdom, we cannot help suggesting that the statement wenbsp;have quoted represents one of the last stages in the amalgamation of the two Celtic races in the west of the island.

* Compare Owen�s �Ancient Laws and Institutes of Wales,� i. 184, 185, where the different manuscripts have been diligently wrought into a patch-worknbsp;most difficult to unravel. For this, however, the editor was probably not sonbsp;much to blame as the perverse policy obtaining at the Public Record Office innbsp;his time.

- Possibly we have an instance in point in the law, said to have once obtained in Britain, that any animal transgressing should be forfeited to thenbsp;person injured. See the article � Mug-Eime � in Cormac�s Irish Glossarynbsp;(Dublin, 1862, and [translated by Stokes at] Calcutta, 1868).

THE WELSH PEOPLE, (chap, i.)

Something may be learned on the head of race amalgamation from the probable history of the national name of the Welsh, to wit, Cymro, �Welshman,� plural, Cymry.nbsp;This word Cymro stands for an earlier Cumbrox or Com-brox, parallel to the Gaulish Allobrox, plural Allobroges, anbsp;name applied by Gauls to another people whose countrynbsp;they conquered ; and just as Allo-brox meant an alien ornbsp;foreigner, Welsh aftfro, � foreigner,� so Coni-brox mustnbsp;have meant � one belonging to one�s own country, anbsp;compatriot.� The choice of this term as the nationalnbsp;name suggests that it was applied to men who did notnbsp;all belong to one and the same race, or speak the samenbsp;language. As the word is to be traced in Cumbra-land,nbsp;Cumber-land, its use must have extended to the Brythonsnbsp;of Strathclyde, which renders it probable that it hadnbsp;acquired some popularity before the end of the strugglenbsp;between the kings of Gwyned and the Anglian princes ofnbsp;Northumbria in the earlier half of the seventh century. Onnbsp;the other hand, as the name seems to have been unknownnbsp;not only in Brittany but also in Cornwall, it may benbsp;conjectured that it cannot have acquired anything likenbsp;national significance for any length of time before thenbsp;battle of Deorham in the year 577, when the West Saxonsnbsp;permanently severed the Celts west of the Severn fromnbsp;their kinsmen in the country consisting now of the countiesnbsp;of Gloucester, Somerset, Devon, and Cornwall. Thus it isnbsp;probable that the national significance of the term Cymronbsp;may date from the sixth century, and that it is to benbsp;regarded as the exponent of the amalgamation of thenbsp;Goidelic and Brythonic populations under the highnbsp;pressure of attacks from without by the Saxons andnbsp;the Angles.

Thus far of the races which may be said to have constituted the Welsh people ; but some mention may now be made of others that have entered into the composition later.

ETHNOLOGY OF ANCIENT WALES.

Among the earliest may be supposed a certain admixture introduced by the legions of ancient Rome, chiefly at suchnbsp;places as IscaSilurum or Carleon on the Usk,and Segontium,nbsp;the remains of which are partially visible at the Carnarvonnbsp;of the present day. After the departure of the Romansnbsp;there was probably nothing of any importance in the matternbsp;of foreign blood introduced till the visits of the Scandinaviannbsp;rovers from the eighth to the twelfth century. They maynbsp;have left small settlements here and there on the coast, as,nbsp;for instance, at Angle, in Pembrokeshire, and in the neighbourhood of the Point of Ayre and other places in Flintshire.nbsp;Their presence also at Fishguard and Solva (in Welshnbsp;Solfach) in Pembrokeshire, seems to be proved by thosenbsp;names, and perhaps the same remark might be made as tonbsp;Harlech, in Merionethshire. But the Scandinavians mustnbsp;have lost their idioms and distinctiveness in the languagenbsp;and nationality of their Celtic neighbours. The next accession of foreign elements came in the course of the Normannbsp;conquests ; but it is not easy to say to what extent the conquerors contributed in flesh and blood to Welsh nationality,nbsp;or even to ascertain to what extent they were Normans,nbsp;and not Bretons similar in race to the Welsh among whomnbsp;they arrived. But the descendants, whether of Normansnbsp;proper or of Bretons, became eventually absorbed in thenbsp;body of the Welsh people and adopted the Welsh language,nbsp;even in the Vale of Glamorgan, where the conquest by thenbsp;Normans was probably the most systematic and thoroughnbsp;in Wales.

Not quite so, however, with another race which the Normans are supposed to have established in the west:nbsp;we allude to the Flemings, who deserve in this context tonbsp;be mentioned at somewhat greater length. Even beforenbsp;the Norman conquest of England, Flemings seem to havenbsp;been brought to this country, as, for instance, by Tostig innbsp;his contest with his brother Harold for the crown. William

THE WELSH PEOPLE, (chap, i.)

the Conqueror married Matilda, daughter of the Count of Flanders, and he appears to have had Flemings in hisnbsp;employ in England. His son, William Rufus, had Flemishnbsp;mercenaries in his army in Normandy when he attackednbsp;his brother, and Stephen employed them in large numbersnbsp;in England. In fact, there is reason to believe that in thenbsp;time of the early Norman kings Flemings settled in considerable numbers in this country. They appear to havenbsp;been unpopular with both Normans and Saxons, and itnbsp;occurred to Henry I. to make use of them, first, as a checknbsp;on the Scotch, and afterwards on the Welsh. He settlednbsp;them first in waste lands on the Tweed, but later he is saidnbsp;to have transported them bag and baggage to the Hundrednbsp;of Roose in Pembrokeshire.^ It is observed that Roose isnbsp;remarkable for its comparative absence of Welsh place-names, and it may be concluded that the Flemings clearednbsp;it of what Welsh inhabitants there may have been there.nbsp;The settlers made themselves masters of the rest of Southnbsp;Pembrokeshire, but as more Welsh names survive there, itnbsp;is not probable that the new-comers made a clean sweepnbsp;of the previous inhabitants. The question how far thisnbsp;Flemish settlement was really Flemish and not English isnbsp;one of considerable difficulty. In case it was purely ornbsp;mainly Flemish, one is tempted to ask, why the languagenbsp;of the district is now a dialect of English any more thannbsp;that of Flanders, where Flemish shows no innate tendencynbsp;to become English. To this it has been replied, that thenbsp;Fleming of Pembrokeshire now speaks English for thenbsp;same general reason that the Dane of Lincolnshire speaksnbsp;English ; and it may be readily admitted that the influence

' The principal contemporary authorities for this are Florence of Worcester, Orderic, Alfred of Beverley, William of Malmesbury, and Brompton ; the wordsnbsp;of those authors and of others in point will be found brought together in anbsp;valuable paper contributed by Dr. Henry Owen, of Poyston, to the � Archseo-logia Cambrensis� for 1895 ; see, more particularly, pp. 98�100.

ETHNOLOGY OF ANCIENT WALES. 29

of the Church and of the castles^ in the district, combined with an inveterate hatred of the neighbouring Welsh, mustnbsp;have amply made up for the isolation from the body of thenbsp;English world. On the other hand, one^ of the greatestnbsp;authorities on English dialects has examined the linguisticnbsp;evidence and declared that it breaks down. At most, henbsp;thought, there could only have been a subordinate Flemishnbsp;element, which soon lost all traces of its original and butnbsp;slightly different dialect, while the principal element mustnbsp;have been Saxon, as in Gower and in the Irish baronies ofnbsp;Bargy and Forth, forming the south-east corner of Ireland.

Settlements of a still more obscure history were made here and there on the rest of the coast from St. Govan�snbsp;Head to the mouth of the Severn, but far the most important must have been the group which made most of thenbsp;peninsula of Gwyr or Gower into a non-Welsh district,nbsp;now known as English Gower, and in Welsh as Browyr,nbsp;that is, Bro- Wyr � the march or country of Gower.�* Gowernbsp;and South Pembrokeshire, which are mutually visible and

' See Mr. Ivor James�s �Welsh in the l6th and 17th Centuries� (Cardiff, 1887), P- 31-

^ We allude to the late Mr. Alexander J. Ellis, in a paper to which we shall have occasion to refer again. It is � On the Delimitation of the Englishnbsp;and Welsh Languages,� and published in the �Cymmrodor� for 1882, seep. 178.nbsp;See also Mr. Edward Laws�s evidence received at Pembroke by the Welshnbsp;Land Commission, Questions 28,994-29,032. Mr. Laws is practically of thenbsp;same opinion as Mr. Ellis; and Professor Rhys has recently submitted thenbsp;linguistic evidence adduced in Mr. Henry Owen�s paper, p. 106, to thenbsp;greatest living authority on the history of English sounds, namely, Dr. Henrynbsp;Sweet, and he finds no reason to qualify Mr. Ellis�s account of the matter.nbsp;On the other hand, Professor Joseph Wright, in the course of his editingnbsp;his English Dialect Dictionary, considers that he has come across words whichnbsp;unmistakably point to the Belgic mother country of the Flemings : he instancesnbsp;the interjection ackaji, and blease, �a blister.�

* For a summary of the evidence as to the Flemish origin, see the Rev. J. D. Davies�s � West Gower,� especially part i., chapter iv., entitled � Thenbsp;Colonization by the Flemings,� pp. 95�li5- See also an important articlenbsp;entitled �Anglia Transwalliana � in the � Saturday Review� for the 20th ofnbsp;May, 1876 ; it is, unless we are greatly mistaken, from the pen of the latenbsp;Professor Freeman.

THE WELSH PEOPLE, (chap, i.)

enjoy the same dialect of English, may be supposed to have been at one time in close communication with onenbsp;another by sea. The establishment of Flemings andnbsp;Englishmen in Gower and the geographical position ofnbsp;their country would naturally suggest a distinct lordship,nbsp;which we have as the Seigniory of Gower: it has beennbsp;referred to more than once in the evidence taken by thenbsp;Welsh Land Commission in Glamorganshire.^ A greatnbsp;part of the south of Glamorgan is called in Welsh Bronbsp;Morgannwg, �the march, margin, or country of Glamorgan,�nbsp;a term incorrectly rendered into English as � the Vale ofnbsp;Glamorgan.� Here the district of Llantwit Major has beennbsp;thought to show traces of Flemish settlements; but thenbsp;Vale is remarkable chiefly for being, as already suggested,nbsp;one of the earliest Norman conquests in Wales. This factnbsp;is rendered conspicuous at the present day by two verynbsp;different features of the country, the Norman architecturenbsp;of its churches and the possession by its farmers of thenbsp;tenant-right known as the Custom of Glamorgan, whichnbsp;excels any other customary tenure in Wales.

Lastly, Wales, situated as it is between England and Ireland, has always received additions to its populationnbsp;from both countries. As to England, the number ofnbsp;Englishmen settling in Wales has perhaps at no time beennbsp;equal to the number of Welshmen migrating to the largenbsp;towns of England. Irishmen have probably at all timesnbsp;been coming over to Wales, especially to the nearestnbsp;corner, namely, Pembrokeshire. Thus the Irish story ofnbsp;the D�isi tells us how some of those people left the partnbsp;of Ireland represented by the Baronies of the Decies in thenbsp;county of Waterford, and gave to Dyfed, a line of kingsnbsp;represented in the time of Gildas by Vortiporius, from whomnbsp;Elen, wife of Howel the Good in the loth century, was

� Qii. 5,095, 6,189-6,404, 6,418-6,422, 23,359, 27.975gt; 27,980, 28,014, 28,029, 28,037.

ETHNOLOGY OF ANCIENT WALES. 31

descended. To come down to a later time, we read in the history of Pembrokeshire by George Owen, who lived innbsp;the reign of Elizabeth, that the Anglo-Flemish portion ofnbsp;his native county was so overrun by Irishmen, that innbsp;some parishes the clergyman was found to be the onlynbsp;inhabitant who was not Irish.^ This, it is true, was annbsp;exceptional time, as it was at the end of the war knownnbsp;as Tyrone�s Rebellion, but many of the exiles must havenbsp;settled in Pembrokeshire. In fact, Mr. Henry Owen, thenbsp;learned editor of George Owen�s work, remarks� that thenbsp;descendants of those Irishmen can still be traced.

Reverting for a moment to the chief races constituting the Welsh people, the Celtic or Aryan consisting of Goidelsnbsp;and Brythons, and the non-Aryan consisting of the Aboriginal population, we may say that their relative proportionsnbsp;to one another may be treated as little disturbed bynbsp;immigrants from Ireland or even from England; for thenbsp;average Englishman is at most not much more Aryan thannbsp;the average Welshman. But the Scandinavian settlements,nbsp;so far as they went, must have gone to strengthen the Aryannbsp;element, and in a qualified sense the same may be said ofnbsp;the Norman conquests in Wales. Then as to the Anglo-Pdemish districts, the settler cannot be regarded as havingnbsp;to any large extent helped to modify the composition ofnbsp;the Welsh people, as he has partly resisted the temptationnbsp;to merge his national individuality in the amalgam of racesnbsp;around him. As it is, he is there conveniently situatednbsp;for the purposes of comparison, and in this connection henbsp;may be roughly described, at any rate so far as Pembrokeshire is concerned, as somewhat better fed than his Welshnbsp;neighbour, more plump and well-conditioned in point ofnbsp;personal appearance, more happy and contented with his

� See George Owen�s � Pembrokeshire,� p. 40.

* In the volume, to which reference has already been made, of the � Archseologia Cambrensis,� p. 103.

THE WELSH PEOPLE, (chap, i.)

lot generally, and less troubled with social or political ideas as to the future.

Should it then be asked what the Welsh of the present day are, Aryan or not Aryan, the answer must be, we think,nbsp;that, on the whole, they are not Aryan ; that, in fact, thenbsp;Aryan element forms, as it were, a mere sprinkling amongnbsp;them. This is by no means surprising, as will be seen onnbsp;comparing the case of France, to which we have alreadynbsp;alluded. For the French of the present day, with thenbsp;exception of the Teutonic element in the north-east ofnbsp;France, are, in the main, neither Gauls nor Aryans of anynbsp;description so much as the lineal representatives of thenbsp;inhabitants whom the Aryans found there. In fact, thenbsp;Gauls were not very numerous, even when they ruled thenbsp;whole country. It has been estimated, on the basis of thenbsp;particulars given by Caesar as to the numbers of the cavalrynbsp;which the different Gaulish tribes were able to place in thenbsp;field to meet the Roman legions, that the Gaulish aristocracy formed a surprisingly small proportion of a population whose numbers ranged somewhere between three andnbsp;six millions.! There seems to be no reason to suppose thatnbsp;the dominant Celts in this country were relatively morenbsp;numerous than in Gaul. They formed a ruling class, andnbsp;led their dependents in war, which was their business abovenbsp;all other things.

Coming down to later times, we may say that their descendants retained the position of privilege and leadingnbsp;in the contests with the Normans, who either measured

'See Roget de Belloguet�s � Ethnogenie Gauloise,� ii. 308-314, and Bertrand and Reinach�s � Celtes dans les Vall�es du P� et du Danube,�nbsp;ii. 41. See also M. d�Arbois de Jubainville�s � Premiers Habitants denbsp;l�Europe,� vol. ii., pp. 7-9, where the learned Frenchman estimated thenbsp;aggregate of Gauls, inclusive of women and children, at 60,000, a figurenbsp;which has ahvays struck us as somehow too low. It has since been boldlynbsp;challenged by Mr. W. H. (Bullock) Hall in his �Romans on the Riviera�nbsp;(London, 1898), pp. 3-5.

ETHNOLOGY OF ANCIENT WALES. 33

swords with them on the field of battle, or entered into family alliances with them, as best suited their purposesnbsp;for the time being. In any case, the Normans do notnbsp;appear to have thought it beneath them to intermarry withnbsp;the nobility of Wales, and that to an extent not to be,nbsp;in the case of England, inferred from the history of theirnbsp;treatment of Saxon or Anglian families. As late as thenbsp;Tudor period the able-bodied men of the Welsh familiesnbsp;took part in raids on the Marches and in the interminablenbsp;feuds which raged among them at home, in the coursenbsp;of which they waylaid one another or burnt each other�snbsp;residences about their owners� ears. Witness, for instance,nbsp;the state of Eifionyd, as represented by Sir John Wynne,nbsp;in his History of the Gwydir Family. But from thenbsp;moment that Wales was subjected to English law theynbsp;began to find their occupation gone, and probably tonbsp;dwindle in importance and power ; but it remained fornbsp;the Civil War which broke out under Charles I. to complete their ruin, since they ranged themselves nearly allnbsp;on the side of the King. Neither folly nor misfortune,nbsp;however, could loosen the attachment felt for them bynbsp;their dependents, an attachment which a perusal of thenbsp;evidence taken by the Welsh Land Commission wouldnbsp;show to be still strong among the tenants on the largernbsp;estates in Wales. There remained a difference of education, a difference of class, to mark off the squire and hisnbsp;family from the people on his land, but no consciousnbsp;distinction of race.

Nevertheless, if a competent ethnologist were to be sent round Wales to identify the individual men and womennbsp;who seemed to him to approach what he should considernbsp;the Aryan type, his report would probably go to shownbsp;that he found comparatively few such people, and thatnbsp;those few belonged chiefly to the old families of the landowning class : the vast majority he could only label as

W.p. nbsp;nbsp;nbsp;p

THE WELSH PEOPLE, (chap, i.)

probably not Celtic, not Aryan. To pronounce, from the point of view of race and history, on the social, political, ornbsp;religious proclivities of that majority is rendered difficult bynbsp;the fact that it is not easy to estimate correctly the influencenbsp;on Wales of movements originated or developed in England.nbsp;There is, however, one instructive instance : the Welshnbsp;people have largely deserted the Established Church, andnbsp;they have done so in favour of more democratic forms ofnbsp;religion, just as their kinsmen have done in Cornwall andnbsp;the Highlands of Scotland. Herein the Welsh can hardlynbsp;be said to have been merely following the example ofnbsp;England, as England cannot be considered to have givennbsp;a decisive lead in the matter. Religion, however, is notnbsp;the only domain in which the tendency of the Welsh isnbsp;democratic: it holds good of their attitude, on the whole,nbsp;as regards social and political questions. And this cannotnbsp;fail to be rendered more and more conspicuous by allnbsp;movements calculated to weaken the attachment of thenbsp;many for the class which supplied them with leaders innbsp;the past.

It may perhaps be convenient if we summarise here the views set forth in this chapter, somewhat as follows:�

The study of the skulls and other remains found in early interments in this country proves that it was inhabitednbsp;by more than one race at the time when the Romans camenbsp;here to conquer.

The study of language and institutions suggests the view, that the earliest inhabitants were of a non-Aryan race,nbsp;namely, that represented probably by the Piets of history.

In the fifth or the sixth century before our era, or perhaps earlier, the first Celtic settlers came and overrannbsp;most of the southern half of Britain. They were thenbsp;Aryan ancestors of the Goidels, whose language is nownbsp;represented by the Gaelic dialects of Ireland, Man, andnbsp;Scotland.

ETHNOLOGY OF ANCIENT WALES. 35

In the second or the third century B.C. there arrived invaders belonging to the other branch of the Celticnbsp;family, namely, the Brythons, and they conquered fromnbsp;the Goidels most of the country which the latter hadnbsp;conquered previously from the Aborigines.

In what is now Wales the Brythonic conquests were represented by the territory of the Ordovices, coveringnbsp;the whole of Mid-Wales as far as Cardigan Bay.

The Goidels to the north and south of the Ordovices were never systematically displaced, and their Goidelicnbsp;may have continued a living tongue down into the seventhnbsp;century.

Soon after the Romans left Britain the Ordovices received an accession of Brythonic blood in the troopsnbsp;led by Cuneda and his Sons, to whom may be traced thenbsp;political framework of Wales under the aspect which itnbsp;presents to the historian of the Norman Conquest.

Conquests there must have been, but the study of the languages in point goes to prove more, namely, intermixture : the Brythons mixed with the Goidels, who werenbsp;themselves an amalgam of the first Celtic settlers withnbsp;the Aborigines ; but all conscious distinction of race hadnbsp;probably been obliterated before the eleventh century.

The admixture of other blood, Scandinavian, Norman, Flemish, and English, has not greatly modified the race,nbsp;the predominant element in which has probably alwaysnbsp;been the substratum contributed by the earliest lords ofnbsp;the soil of these islands.^

' Since this chapter was written we have had occasion to read a remarkable book by the late Rev. W. D. Babington on �Fallacies of Race Theories asnbsp;applied to National Characteristics� (London: Longmans, Green amp; Co., 1895).nbsp;Among other things we may say that it confirms our view as to the mixture ofnbsp;races constituting each of the nations in the United Kingdom, and it disposesnbsp;of the stock generalisations framed to flatter the German at the expense ofnbsp;the Celt.

CHAPTER II.

THE PICTISH QUESTION.

The foregoing outlines will serve to suggest a picture of the ethnology of ancient Wales, and we have endeavourednbsp;not to crowd it with details. Some of these we now proceed to supply by elaborating a few of the points on whichnbsp;we have touched in passing. We begin by reverting tonbsp;the Pictish succession and metronymic designations likenbsp;that of Gwydion son of Dbn; and when one comes tonbsp;consider what the Pictish succession must have originallynbsp;meant, one cannot overlook Caesar�s statement, that somenbsp;of the inhabitants of Britain had their wives in common.nbsp;Professor Zimmer, on analysing Caesar�s chapter in point,nbsp;comes to the conclusion that the words were meant tonbsp;apply to non-Aryan inhabitants in the interior�those, innbsp;fact, whom Caesar represents as regarding themselvesnbsp;descended from the Aboriginal islanders in contrast to thenbsp;later comers, who, according to the same authority, did notnbsp;materially differ in their customs from the Gauls. Caesar�snbsp;words (v. 14) are to the following effect:�Uxores habentnbsp;deni duodenique inter se communes, et maxime fratres cumnbsp;fratribus parentesque cum liberis ; sed qui sunt ex Us natt,nbsp;eorum habentur liberi, qtw primum virgo qvceque deducta est.nbsp;The first sentence makes a clean sweep of the institution ofnbsp;marriage, and leaves no room for the idea of incest; ^ but the

� As to the origin of that idea see M. S. Reinach�s article � � La prohibition de 1�inceste et ses origines � in � L�Anthropologie, � vol. x., pp. 59-70.

THE PICTISH QUESTION.

second sentence seems to us to have been dictated by the Roman�s inability to realise a state of society exclusivelynbsp;based on birth. The idea of assigning the children each tonbsp;its own father, if not entirely due to the working of Caesar�snbsp;own mind, reads in this context like an advance towardsnbsp;Aryan habits. At any rate, we shall as we proceed findnbsp;traces of a stage of society betraying no perceptiblenbsp;tendency in that direction.^

The kind of social arrangement here in question suggests several curious points for consideration, and foremost amongnbsp;them this : who would be reckoned a man�s nearest of kinnbsp;Clearly one�s own brothers and sisters by the same mother ;nbsp;and looking backwards one�s nearest relatives would be hisnbsp;mother and his mother�s brothers and sisters similarly, whilenbsp;looking forwards it would be one�s sisters� children. So onenbsp;would naturally look for one�s heir and successor in one�snbsp;brother, and after him in a son of one�s sister. This isnbsp;the key to a good deal that is otherwise unintelligiblenbsp;in Celtic literature : let us take for instance the Mabinoginbsp;of Math, which has already been mentioned. There thenbsp;leading family ruling over Gwyned consists of the followingnbsp;persons :�

Math the king, who is called son of Mathonwy, about whom nothing is known.

D�n, Math�s sister, about whom equally little is known, except that she had the following children :

Gwydion, Gofannon, (Amaethon), Gilfaethzvy, and Efeyd, all called sons of D�n ; and one daughter called

Aranrot, or Arianrhod, daughter of D�n. Arianrhod had two sons, Dylan and Lew Lawgyffes.

Next to the king himself, Gwydion plays the most important role in Math�s realm, and the king teaches him the

' Such a stage of society, together with well-known stories about virgin mothers, points back to a savage state in which the male element had nevernbsp;been supposed necessary to conception.

THE WELSH PEOPLE, (chap, n.)

magic of which he was master : in fact, everything points to Gwydion as Math�s successor, though that is not statednbsp;in the story. In due time ILew ILawgyffes is representednbsp;succeeding to the kingdom of Gwyned. In other words.nbsp;Math is succeeded by his sister�s son, Gwydion, andnbsp;Gwydion is succeeded by his sister�s son, Lew. It isnbsp;tacitly assumed that Gwydion was the father of JLew ; butnbsp;the relationship between Gwydion and Arianrhod is nevernbsp;discussed, and the silence maintained on that point onlynbsp;becomes intelligible in the light of the social arrangementnbsp;here supposed.

Similarly in the case of the Mabinogi of Branwen: there we have Bendigeitvran, or Br�,n the Blessed, as king ofnbsp;Britain, and he has a brother, Manawydan, and a sister,nbsp;Branwen: they are called sons of ILyr and daughter ofnbsp;ILyr respectively, while their mother is named Penardim,^nbsp;daughter of Beli, son of Mynogan.

Now Branwen is given to wife to Matholwch, who reigns in Ireland, and there she has a son by him called Gwernnbsp;son of Matholwch ; but after some years have passed Brinnbsp;hears of his sister being harshly treated, and he makes annbsp;expedition to Ireland. He leaves behind him Cradawc,nbsp;or Caradog, his son, to take charge of this country, thenbsp;kingship of which is, however, seized in the meantime bynbsp;Caswaitawn, or Caswallon, son of Beli. For this Caswatton,

* She had two other sons, Nissien and Efnissien, whose father is called Euroswytt in the Mabinogi: he is said in one of the Triads (i. 50 = ii. 49)nbsp;to have, some time or other, taken ILyr prisoner. The form Penardim innbsp;the Mabinogi was an archaism ; and our narrator, had he understood it, wouldnbsp;have put it into his own spelling as Penardu, which would be in modernnbsp;Welsh Penardu or pen-arAu, meaning � Her of the Black Head.� Compare thenbsp;variant Dyf-lyn for Du-lyn �Black pool,� and Welsh u in verbal nouns likenbsp;credu � act of believing � and gohebu � act of corresponding,� as compared withnbsp;Old Irish cretem �belief,� sechem �act of following,� and sessom �standing.�nbsp;The subject is too large to dispose of here in passing, but the reader shouldnbsp;consult the learned articles of M. Ernault on quot;Les Formes de 1�Infinitifnbsp;Breton,� in Meyer and Stern�s ��Zeit. fur Celt. Philologie,� vol. ij.

THE PlCnSH QUESTION.

we are told, donned a coat of magic mail and slew Cara-dog�s men without disclosing who it was that did it; but he did not slay Caradog, as he was his relative. He is callednbsp;his nephew, son of his cousin; and we learn from annbsp;ancient triad cited in the story that he died of grief andnbsp;vexation at the slaughter of his men. In Ireland thenbsp;coming of Bran and his host created a great commotion,nbsp;but, thanks to the intercession of Branwen, the two kingsnbsp;Bran and Matholwch came to terms, and the concessionnbsp;made by the Irish was to give Matholwch�s kingdom tonbsp;Branwen�s son Gwern. The concession consisted in thenbsp;fact implied that Gwern could not, according to the usagenbsp;of Matholwch�s people, be Matholwch�s successor, as henbsp;would, according to the birth succession, be no recognisednbsp;relation of Matholwch�s at all, whereas, according to thenbsp;same rule, he would be Bran�s nearest of kin and his rightfulnbsp;successor, as son of his sister. The editor or narrator ofnbsp;the story as we have it does not show that he understoodnbsp;this, and it is he probably that is to be held responsiblenbsp;for an inconsistency which occurs in it. More than oncenbsp;he makes Caswafton son of Beli cousin to Bran andnbsp;Manawydan, though he treats them at the outset as sons ofnbsp;Penardim, and her as daughter of Beli.^ It all comes right,nbsp;however, if we treat Penardim, not as daughter of Beli, butnbsp;as his sister on the mother�s side: then Bran�s right tonbsp;succeed Beli, who is fabled to have been king of Britain,nbsp;becomes clear�he is the son of Beli�s sister. But an editor

^ We take it that the latest editor is responsible for this, but that he found Caswafton made son of Beli in the version which he was using of the Mabinogi,nbsp;and that he forgot or hesitated to alter the relationship as indicated at pp. 41nbsp;and 44. The passage where Penardim is made daughter of Beli is the openingnbsp;of the Mabinogi of Branwen (p. 26), and the second line of it is remarkable fornbsp;the words arderchatc 0 goron lundein, which have been translated from somenbsp;such a phrase as insignitiis diademate^ common enough, for example, innbsp;Geoffrey�s Latinity, as in Lib. ij. I, 20, iv. ii, vi. 4. In fact W'e suspect thatnbsp;the Mabinogion had not assumed the form in which we have them till Geoffrey�snbsp;time.

THE WELSH PEOPLE, (chap, ii.)

not familiar with this kind of succession would naturally think that he improved Bran�s position by making hisnbsp;mother daughter of Beli�that is to say, by making Brannbsp;a descendant of Beli.

The story originally implied the Pictish succession, and it is worthy of note that of the two men whom it representsnbsp;succeeding their respective fathers contrary to it, the one,nbsp;Caradog, dies of vexation,^ while the other, Caswahon, isnbsp;only enabled to secure by violence and magic a positionnbsp;which the tenor of the Mabinogi assumes to have rightfullynbsp;belonged to Manawydan after the death of his brothernbsp;Brin. In fact, the introduction of Caradog and Caswa�onnbsp;betrays the falsifying hand of a historian, for Cradawc,nbsp;as we have assumed, is merely a form of Caradawc, thenbsp;representative of Caratdcos, in Latin Caratacus�sometimesnbsp;distorted still into Caractacus�the name of the famousnbsp;but unsuccessful leader of the Silures and Ordovices againstnbsp;the Romans. As a matter of fact, he was no son of Brin,nbsp;nor was he of his Goidelic race, as he was a Brython. Yet thisnbsp;fiction has been widely accepted in Modern Welsh literature,nbsp;according to which Caradog and Brin his father, togethernbsp;with their families, were taken captives to Rome, wherenbsp;Brin and others of his family were converted to Christianity,nbsp;and on their return brought the Gospel to Britain. Thennbsp;as to Caswatfon, by him we are doubtless to understand

' The exact meaning of the word used is merely inferred, as it occurs only in this triad. The part relating to Caradog reads in the Red Book (Oxfordnbsp;Mabinogion, p. 4^)1 ^ h6nn6 uu y trydyd dyn a toms y gallon 0 niuyget,nbsp;where the two last words should probably be 0 anniuyget. It seems to meannbsp;�And that was one of the three who broke their hearts of vexation or grief.�nbsp;Another of the three making up the original triad was Ffaraon, who is thusnbsp;mentioned in � Hudd and ILevelys, �nbsp;nbsp;nbsp;nbsp;, p. 98: Trydyd crynweissat uu h6nn6

a torres y gallon [0] anniuiged, � That was the third chief guardian who broke his heart of grief.� It is there said that Dinas Emreis in Snowdon had beennbsp;previously known from Ffaraon as Dinas Ffaraon Dande, and the namenbsp;carries us back to an old world of legend now submerged. The third limbnbsp;of the triad, we are Sony to say, has never been discovered.

THE PICTISH QUESTION.

Cassivellaunos, the leader of the Brythonic tribes who had opposed Julius Caesar in the south-east of Britain when thatnbsp;general paid his second visit to our shores: Geoffrey, atnbsp;the end of his third book, chap, xx., introduces him asnbsp;Cassibellaunus, son of Hely [read Bely]. The way innbsp;which Caradog�s history is referred to an ancient triad,nbsp;suggests that we have in the allusion to his death a touchnbsp;of genuine tradition, based remotely on the real historynbsp;of Caratacos, and preserved in the west of the island.

Lastly, as to Beli son of Mynogan,^ his identity with Bellinus son of Minocannus in Nennius�s Historia Brit-tonum, where he is made the native leader against Juliusnbsp;Cassar, has been known for some time. And Professornbsp;Zimmer** has traced the Nennian Bellinus, filius Mino-canni, back through Orosius�s gibberish Minocynobel-linum Britannorum regis filium to Suetonius�s Adminio,nbsp;Cynobellini Brittannorum regis filio; and from the latternbsp;historian we learn that Adminios was a fugitive from Britain,nbsp;who gave himself up to the mad emperor Caligula. Sonbsp;much for the designation of Beli, or Beli Mawr, son ofnbsp;Mynogan; but we cannot follow Professor Zimmer innbsp;thinking that his unravelling of this tangle of errors disposes of Beli. For we conjecture that the words translated � Son of Mynogan � were not to be found in thenbsp;original of the Mabinogi, but that they were introducednbsp;by an editor who was acquainted with the Historia Brit-tonum of Nennius. What stood in the story previouslynbsp;was rather Beli Maur map Aun, An, or Anau, �Belinbsp;the Great, son of A.,� which occurs as Beli Ma6r ni. Anna,nbsp;� Beli the Great, son of Anna,� in one of the pedigrees in

* See Skene�s �Four Anc. Books of Wales,� ii. 204, 420 ; and San Marte�s � Nennius und Gildas,� pp. 40, 4I) � �9-

quot; See Zimmer�s �Nennius Vindicatus,� pp. 271-3 : his references are to Suetonius�s Caligula, cap. 44 et seq., and Orosius�s � Histor. advers. Paganos,quot;nbsp;Sgt; 5 gt; see also Evans�s �Coins of the Ancient Britons,� pp. 208, 284-348 ;nbsp;and Rhys�s �Celtic Britain � (2nd edition), p. 278.

THE WELSH PEOPLE, (chap, ii.)

Jesus College Manuscript 20, supposed to be of the thirteenth century. Further one reads two sentences respectingnbsp;.A.nna, as follows: yr anna honn oed verch y anthera�dyrnbsp;rufein. yr anna honno a dywedei wyr yr eifft y bot ynnbsp;gyfynnithder� y veir vor�yn} which seem to represent twonbsp;glosses from two different sources, as will be seen from thenbsp;rendering of them, � This Anna was daughter to thenbsp;emperor of Rome. That Anna used to be said by the mennbsp;of Egypt to be cousin to the Virgin Mary.� The latternbsp;statement is also made in the pedigree of Owen son ofnbsp;Howel the Good, who is traced back to Aballac son ofnbsp;Amalech quifuit Beli magni filius., et Anna mater ejus, quamnbsp;dicunt esse consobrinam Marice Virgim's, Matris Domininbsp;nostri Ihesu Christi?

The treatment of the ILyr pedigree in the matter of Penardim prepares us to understand the treatment of Annanbsp;in the pedigrees in question. The editor probably foundnbsp;Anna represented as Bell�s wife or as his mother, but notnbsp;feeling bound to say anything about her, he simply addednbsp;to the name of Beli words meaning �Son of Mynogan,� afternbsp;the example of Nennius. With regard to the Christiannbsp;Anna, the introduction of her name is due probably to annbsp;early confusion of it with that of Ana or Anu, genitivenbsp;Anann, who figures in Irish mythology as mater deorumnbsp;hibernensium? This name would be treated in Old Welshnbsp;as Ann or An (possibly Anau), according to the quantity

' See �Y Cymmrodor,� vUj. 84; also p. 85.

^ For the abbreviations used see Phillimore�s edition of MS. A. of the � Annales Cambrise,� in �Y Cymmrodor, � ix. 170: the genealogies seem tonbsp;have been compiled in the tenth century. Whether the scribe here meant onenbsp;to regard Anna as the mother of Beli or of Amalech is not clear ; but a littlenbsp;later (p. 174) he undoubtedly takes the latter view�the wrong view, in fact ;nbsp;for he there has � Amalech son of Beii et Anna.'�

3 See Stokes�s edition of O�Donovan�s translation of � Cormac�s Glossary � (Calcutta, 1868), s.v. Ana, p. 4; also p. 17, where an article is devoted tonbsp;another female figure, Buanann, mother of Irish heroes, just as Ana wasnbsp;mother of Irish gods.

THE PICTISH QUESTION.

of the initial vowel in the Goidelic Ana or Anu, which is not certain.

To return to Beli, we read in the Red Book story called Maxen�s Dream,^ that he was in possession of Britain untilnbsp;Maxen and his legions came and drove him and his sonsnbsp;on sea; and so closely does Beli appear associated with thenbsp;sea that an ancient verse calls the brine of the ocean Beli�snbsp;liquor.� This we cannot help regarding as a popular touchnbsp;not to be explained by any amount of learned bungling onnbsp;the part of Orosius or Nennius. We are led back to anbsp;legend in the west of Britain, which represented it enjoyingnbsp;a sort of a golden age which was only brought to an end bynbsp;the advent of the Romans. We learn that the king�s namenbsp;was Beli, and we infer that he was a Goidel,who had shipsnbsp;on the Irish Sea. We know from the Chronicles that thenbsp;name which was Beli in Welsh (borne by one of the kingsnbsp;of Gwyned and by others in historical times) was in Irishnbsp;Bile; further, there was an ancient Bile with whom wenbsp;should identify our Beli the Great, and Irish legendnbsp;represents him as the father of Mil, the leader of the lastnbsp;legendary conquest of Ireland and ancestor of all those ofnbsp;the Irish who called themselves Milesians after his name.nbsp;The story as we have it makes Bile king of Spain, and bynbsp;giving his son the name of Mil, genitive Miled, it brings usnbsp;to the Latin miles, genitive militis, � a soldier �: this seemsnbsp;to have been a synonym or translation of another name,nbsp;Galam or Golam, by which Mil was known and described innbsp;Irish as a man of bravery and valour. The identity of thenbsp;names Bile and Beli is, however, not all: the parallel isnbsp;closer than it looks at first sight. Mil, son and successor ofnbsp;Bile, conquers Ireland, which is divided between his two

^ See the Oxford Mab., p. 88.

* See the Book of Taliessin in Skene�s � Four Ancient Books of Wales,� �. 150, where one reads Gna6t r�yfyn heli Beli wira�i^ � � Familiar is the sight ofnbsp;oars in the brine of Beli�s liquor.� For a mistranslation of it see i. 300.

THE WELSH PEOPLE, (chap, ii.)

sons Eber and Airem, while the Mabinogi makes Beli�s heir and successor, Bran, obtain the practical disposal ofnbsp;Ireland. The narratives otherwise differ, owing chiefly tonbsp;the Welsh one having gone off into a story which may benbsp;regarded as forming a counterpart of that of the Nibelungennbsp;Slaughter in the literature of Teutonic lands.

The tradition about Beli must be regarded as belonging to the Goidels of Britain, and it was only by a tour de forcenbsp;that Caswatton, the leader of Brythons�that is to say, ofnbsp;the hereditary foes of the Goidel�could be made son ofnbsp;Beli; but it was natural enough that an editor of thenbsp;Mabinogion should wish to graft the later history of hisnbsp;country on the legendary glories of the past, and itnbsp;was a step in that direction to place Caswatton amongnbsp;the sons of Beli. Whether he set himself to do this, ornbsp;merely followed the example set by a previous writer,^nbsp;his readers, accepting the words which he has used,nbsp;could not help saying in effect: Yes, it was by craft andnbsp;violence that Caswatton secured supreme power; but henbsp;was after all son of the rightful king of Britain in hernbsp;golden age�that is, of Beli the Great. The same editor isnbsp;possibly also to be held responsible for the order of thenbsp;events, which is probably unhistorical, as we should rathernbsp;regard the aggressiveness of Caswalton�s race as one at leastnbsp;of the reasons for Bran�s going to Ireland. But the datingnbsp;of Caswatton�s conquests in Wales after Bran�s departurenbsp;for Ireland is to be explained by the confounding of thenbsp;naming of Cadwailon Lawhir with Caswatton�s ; for Welshnbsp;tradition insists on that Cadwatton, who was grandson ofnbsp;Cuneda and father of Maelgwn, as the final vanquisher of the

1 Such as Geoffrey or the writer of the pedigree already mentioned (p. 42;, in which we have the descent of Owen son of Howel traced back not only tonbsp;Maelgwn and Cuneda, but to Beli. The former portion seems to reach backnbsp;to Tacit only : then comes a very Pictish looking portion beginning with Cein,nbsp;son of Guorcein, son of Doli, son of Guordoli, and terminating with Beli andnbsp;Anna. See also the �Proceedings of the Antiq. of Scotland,� xxxij. 342.

THE PICTISH QUESTION.

Goidels in North Wales4 We may therefore still suppose that the Brythonic tribe of the Ordovices pushing on tonbsp;the shores of Cardigan Bay may have been the cause ofnbsp;an emigration of Goidels to the nearest coast of Ireland.nbsp;In fact, we have possibly a trace of this in the namenbsp;Eblanii, which Ptolemy gives to the inhabitants of thenbsp;coast north and south of the mouth of the Liffey, asnbsp;we seem to have closely related names in that of thenbsp;river Elan (a tributary of the Wye), and that of thenbsp;mountain region of Elenid, in which Giraldus� placesnbsp;the sources of the Severn and the Wye, of the Towy, thenbsp;Teifi, and the Ystwyth. On the Irish side it is significantnbsp;that the story of Mil�s two sons gives to Eber, the eldernbsp;brother and eponymous hero of the Iverni, the southern halfnbsp;of Ireland, and to the younger brother, Airem (genitivenbsp;Aireman or Eremon'), the northern half; and that it furthernbsp;represents Airem slaying Eber and taking the whole ofnbsp;Ireland to himself. The name Airem means ploughman,nbsp;and possibly conveys a reference to the triumphs of thenbsp;Aryan farmer over the ruder native. But even disregarding all such connotation of the name, we still have the factnbsp;that it has gathered round it legends reminding one of thenbsp;story of Arthur; and that the name Airem was borne by onenbsp;of the early kings of Tara, in Meath and the land of thenbsp;ancient Eblanii, the centre of Milesian rule over Ireland.

Let us now see in what way the custom of reckoning

* See Triad i. 49 = ii. 40 = iii. 27 and the �lolo MSS.,� pp. 78, 468. Compare Nennius, � 62, and note the confusion of names in the � loIo MSS.��nbsp;See also the Oxford Bruts, where Kat6aiia6n E.a6ir at p. 200, stands possiblynbsp;for the same man as Kas�alla�n at p. 232: in Geoffrey�s Latin they arenbsp;respectively Caduallo ix. 12, and Cassibellanus xi. 2.

^ See his �Descriptie Kambrim� (Rolls Office edition), pp. 119, 138, *70-3, 175, where the spelling is Elennyth and Elennith ; also the Oxfordnbsp;Mabinogion, p. 62, where it is written Elenit: Lewis Glyn Cothi wrotenbsp;Elenid, III. iv. 43, 4 (p. 184). As to the phonology of the equation suggestednbsp;in the text, we have the similar reduction of sbl into el in the Welsh adverbnbsp;^Icni (this year) from some form of the Welsh word blyned, � year.�

THE WELSH PEOPLE, (chap, ii.)

descent by birth alone has left its impress on the language and monuments of those among whom it prevailed. Ournbsp;attention is challenged in the first instance by inscriptionsnbsp;which suggest no father�s name; and the earliest of thatnbsp;class is probably the bronze tablet found not very long agonbsp;at Colchester and read as follows �

DEO . MARTI . MEDOCIO . CAMP ES[tr]IVM . ET VICTORIE ALEXANnbsp;DRI . PII FELICIS AVGVSTI . NOS[tr]Inbsp;DONVM . LOSSIO . VEDA . DE . SVOnbsp;POSVIT . NEPOS . VEPOGENI . CALEDO.

�To the god Mars Medocius of the Cainpes\tr\es and to the victory of our Alexander Pius Felix Augustus (this) giftnbsp;has been dedicated at his own expense by Lossio Veda,nbsp;Vepogenos�s nephew, a Caledonian.�

The god Medocius who is here equated with the Roman Mars is otherwise unknown, as is also the precise meaning in this instance of the Latin Campestres,nbsp;which usually has reference to the open field, and innbsp;particular to the Campus Martius in Rome: this has callednbsp;forth the suggestion that perhaps Lossio Veda was anbsp;gladiator. However that may be, he has taken care tonbsp;tell us that he was a Caledonian, which is for our purposenbsp;much the same as if he had called himself a Piet. We havenbsp;indirect evidence to the same effect in the vocables Vedanbsp;and Vepogeni, for both may be said to occur in the list ofnbsp;the Pictish kings. The former has there been read Uecla,nbsp;a spelling due doubtless to the difficulty of distinguishing

* See the � Proceedings � of the Society of Antiquaries, 2nd S. xiv. io8, 183 ; also � The Archasologia, � liv. 37. By reading Campesium one seemed to arrivenbsp;at a native name Campeses, recalling Campsie in Stirlingshire, and the Linn ofnbsp;Campsie on the Tay. As we have, however, to treat NOSI as nostri, we seemnbsp;to be bound to insert tr in CAMPESIVM likewise. For some further accountnbsp;of the bronze, together with a photograph of it, see the � � Pro. of the Antiq. ofnbsp;Scotland,� xxxij. 325-30.

THE PICTISH QUESTION.

in some kinds of handwriting between d and cl; and the latter has in the same document yielded a nominativenbsp;Vipoig. The two entries occur also significantly near onenbsp;another, as follows : ^ �

Vipoig namet xxx. ann. regnauit

Canuhilachama iiii. ann. regnauit.

Wradech uecla ii. ann. regnauit.

Lossio Veda, though showing no inclination to be over brief in describing himself, suggests no father�s name; and this isnbsp;the case with certain other inscriptions, such as the onenbsp;found on Winsford Hill, in Somerset, which reads^ merely�

CARATACI

NEPVS.

But what did nepos (or nepus) mean ? For the Romans the word is known to have meant a grandson, a descendant,nbsp;also a nephew, whether son of one�s brother or of one�snbsp;sister; but in a society with birth alone considered, onlynbsp;one of those meanings is admissible�namely, a sister�s son.nbsp;Thus CarataciNepuswovXA mean �Nephew ( = sister�s son)nbsp;of Caratacas.� Where the language used is Goidelic, thenbsp;place of nepos is supplied by avias, genitive avi, reduced innbsp;Modern Irish to ua or �, genitive ui, as for instance in thenbsp;following inscription from the Barony of Bere, co. Cork :nbsp;Maqui Decceddas avi Toranias � (The Stone) of Mac-Dechetnbsp;� Torna.� Or take the following, found at Dunbell, co.nbsp;Kilkenny: Navvallo avvi Genittad{c{\f � (The Stone) ofNuall

THE WELSH PEOPLE, (chap, ii.)

� GentichT In one remarkable instance the word used is the etymological counterpart of the Latin nepos, genitivenbsp;nepotis, in Irish Ogam niotta\s\, later nioth, niath with anbsp;nominative nie or nia, Welsh nei, now nai, �a nephew.�^nbsp;The stone was found near Gortatlea, between Killarneynbsp;and Tralee, and it reads thus, in two lines �

Dumeli maqui Glasiconas

Niotta CobranoEJ\

� (The Stone) of Dumel, son of Glasiuc, nephew of (the) Distributor.� We take Cohranor... to stand for com-rannori,nbsp;genitive of comrannorias, and to mean one who shares ornbsp;divides, probably in the sense of carving and dividing meatnbsp;at feasts and banquets. Among the Irish this was a positionnbsp;of distinction, claimed by the warrior who had performednbsp;most feats of valour. There is a well-known Irish talenbsp;entitled the Story of Mac Ddth�s Pig, which turns on anbsp;contest for the carving of that portentous beast by thenbsp;braves of Ulster and Connaught. Mac Dath� was king ofnbsp;the Leinstermen, but afraid of both Ulster and Connaught,nbsp;on account of a remarkable hound of his which they coveted.nbsp;P'earing trouble, he took his wife�s advice and cunninglynbsp;invited both the men of Connaught and the men ofnbsp;Ulster for the same day : then they would, he said, getnbsp;the hound.

' The p which appears in the Latin nepos disappears, according to rule, in the Celtic equivalent; hence Irish nie, genitive nioth. The Welsh setting outnbsp;from ne(p)ot-s made it into ne-o or ne-io, whence nei and nai: compare Heidrnbsp;� � thief,� from latrio, for the Latin latro. Other instances will be found in thenbsp;� Proceedings of the Society of Antiquaries of Scotland,� xxvi., 309. Bretonnbsp;has ni and ntz �nephew,� and nizez niece,�where the s of niz probablynbsp;stands for the earlier i = A of the nominative ne(p)oi-s. Compare Bretonnbsp;noz = Welsh nos �night,� from not-s = noct-s, reduced in Latin to nox.

2 See the �Journal of the Royal Society of Antiquaries of Ireland � for 1895, pp. 1-4 ; but our reading is checked by a rubbing and a sketch kindlynbsp;supplied by the Rev. P. Sweeny, A.M., Ballinacourty Rectory, Annascaul,nbsp;and by a recent examination of the stone by Professor Rhys.

THE PICTISH QUESTION.

They came, and were filled with surprise at meeting one another so unexpectedly. Presently they prepared to sitnbsp;down to feast on the host�s great swine, and then arose thenbsp;question who was to carve : it was agreed to give it to thenbsp;bravest. So each warrior who had confidence in his recordnbsp;declared what that was, whereupon rose another and putnbsp;him down by enumerating greater feats of his own. Thisnbsp;went on for some time, when at length it looked as thoughnbsp;the honour of carving would fall for certain to Cet macnbsp;Matach, a Connaught hero, and he had taken up the knifenbsp;to begin the carving when a belated Ultonian, Conallnbsp;Cernach, hurries into the room and asks, Cia rannas d�ib ?nbsp;� Who is carving for you) � It was replied that Cetnbsp;was going to do it, and a contest of words takes placenbsp;between Cet and Conall, with the result that Cet reluctantly yields, with the remark that Conall would notnbsp;carve had Cet�s brother Anluan been present. �But henbsp;is present,� said Conall, who, after feeling in his girdle,nbsp;brought forth the bleeding head of Anluan and hurlednbsp;it in his brother�s face. Such was Conall�s excuse fornbsp;arriving late, and the passage is one of the most graphicnbsp;and savage in the whole range of old Irish literature.

We have taken Niotta Cobranari to mean Nepotis PartistcB, as describing either Dumeli or Glasiconas, but itnbsp;IS possible that it should rather be taken as an independentnbsp;proper name: at any rate, such names occur. Take, fornbsp;example, such a later instance as Nioth-Fruich}Niath-Froich,nbsp;Nat\}i\-Fraich, or Nad-Frdich, in which the first elementnbsp;owing to its proclitic position has suffered curtailment. It isnbsp;to be noticed that Fraech, genitive Fraich, was a separatenbsp;personal name of unknown signification, and that Niath-Fraich must have meant � Nephew of Fraech,� or, morenbsp;precisely speaking, � Son of Fraech�s Sister.� And this isnbsp;not mere inference, for we have the positive statement of

See Stokes�s �Patrick,� p. 331 ; also pp. 76, 194, 196, 214, 250, 468.

W.P.

THE WELSH PEOPLE, (chap, ii.)

Cormac�s Glossary that the meaning of the Mmrd nie, genitive niath, was mac sethar, � sister�s son,�^ and, as farnbsp;as we know, that was its only meaning. In harmony withnbsp;the foregoing interpretations, the St. Vigeans Stone (p. 17,nbsp;above) should be rendered � Drost�s nephew Voret fornbsp;Fergus,� rather than �Drost�s kin,Voret and Fergus.� Similarly in the case of the corresponding vocable� on thenbsp;Newton Stone (p. 17, above), and of the poi of certainnbsp;Ogam inscriptions in the south of Ireland.

The earliest existing manuscript of Adamnan�s Life of St. Columba dates from the beginning of the eighth century,nbsp;and it is found that he distinguishes as a rule between twonbsp;kinds of clan designations, (i) He mostly uses wherenbsp;the native usage of later times recognises �a or d, with thenbsp;plural nepotes rendered by ui or hid (Anglicised hy and o�),nbsp;as for instance in Nepos Lethani, called in Irish UaLiathain,nbsp;Anglicised Olethan. So with the plural, as in Nepotesnbsp;Nellis= Ua N�ill, � the Hy-Neill or O�Neills�; but some of

* See Stokes�s edition of O�Donoyun�s � Translation of Cormac�s Glossary,� p. I2I ; also Stokes�s paper on the � Bodleian Fragment of Cormac�s Glossary �nbsp;(read before the Royal Irish Academy, November 30th, 1871), p. 8.

^ Our last reading of it is ipuai, but we should treat ipe as a spelling of ipai, and equate it with ipuai; that is, unless it should prove more correct to regardnbsp;ipuai as a spelling of ipue. In either case we should treat both as accented onnbsp;the final syllable (like mucoi), and equate them with poi as a foreshortening ofnbsp;some such a vocable as ip�i or dp�i. In all instances poi appears affixednbsp;(contrary to the Celtic habit of prefixing) to the genitive of a personal name, asnbsp;in Broinienaspoi (Poi of Broiniu), Corbipoi (Poi of Corb), and lacinipoi (Poi ofnbsp;lacin). These come respectively from the counties of Cork, Kilkenny, andnbsp;Wicklow ; but the same formula must have been in use in the south-west ofnbsp;Britain. At any rate, it is thence we have to suppose it transported tonbsp;Brittany, where we have it in the well-known name of the king Erispoe in thenbsp;ninth century. It is made up of poi affixed to the genitive of a man�s namenbsp;which occurs now and then in the pedigrees in the Book of Leinster as Aires,nbsp;genitive Airiss as on fol. 326d, 353d, 356a, 3^3'^' See vol. xxxij. of thenbsp;� Proceedings of the Antiq. of Scotland� for 1897-8, where, in a paper entitled anbsp;� Revised Account of the Inscriptions of the Northern Piets,� Professor Rhysnbsp;has dealt with several of the questions touched upon in this chapter : for thisnbsp;tendering see more particularly pp, 347, 370.

THE PICTISH QUESTION.

the race are found styled Nieth-N�ill} where nieth is the etymological counterpart of nepotes. So we learn not onlynbsp;the equivalence of meaning of nieth and nepotes, but of bothnbsp;practically with treated in Modern Irish as meaning grandchildren or descendants. (2) Adamnan leaves untranslatednbsp;and undeclined in his Latin a certain word mocu, as innbsp;Mocu-Sogin, Mocu-Dalon, and Mocu-Alti. In Irish Ogamnbsp;inscriptions this is a very important word, and its mostnbsp;usual forms in them are moco, genitive mucoi; but in Irishnbsp;literature it appears as maccu, genitive maccui, which begannbsp;comparatively early to be regarded as made up of macc-ui andnbsp;meaning filius nepotis. As a matter of fact, however, it isnbsp;nothing of the kind, but a distinct word meaning race ornbsp;kin in the concrete. Thus a family or tribe called Mocunbsp;Runtir by Adamnan is called Ddl Runtir in the Tripartitenbsp;Life of Patrick,^ and members of it are said in the Book ofnbsp;Armagh to be de genere Runtir. As to this last, it is to benbsp;noticed that the little word de, though necessary in thenbsp;Latin, is not in the original Mocu Runtir, which literallynbsp;rendered would be genus Runtir, as if each individual of thenbsp;group personified the whole. The real explanation is thatnbsp;the Piets had not learnt to speak of any race apart fromnbsp;some individual member of it: to them � the kin of A. B.�nbsp;was � kin A. B.� Mocu here followed by a genitive isnbsp;Goidelic, while the Pictish inscription at Aboyne on thenbsp;Deeside has the words in apposition.*

In order to get over the difficulty Adamnan sometimes interposes the word gente, as in the instance Trenanum,nbsp;gente Mocuruntir, � Trenan, Mocu-Runtir by race ornbsp;family.� At other times he lets mocu do duty alone, asnbsp;in his mention de Erco fure Mocudruidi qui in Coloso

* See Skene in his �Chronicles of the Piets and Scots,quot; p. 352, where he copies the Annals of Ulster for A.D. 692.

See Stokes�s edition, p. 226; and as to the meaning of mocu, see Rhys�s Lectures on Welsh Philology,� pp. 408, 409.

See the �Proceedings of the Antiquaries of Scotland �� 1898-9, pp. 351-3,

E 2

THE WELSH PEOPLE, (chap, ii.)

insula comnianebat. The island was one of those now called Colonsay, and the clan to which Ere belonged tooknbsp;its name from a druid, somewhat like the Mactaggartsnbsp;and MaePhersons of later times, so named after ancestorsnbsp;in holy orders. When the great person in the past oi anbsp;family was a man and not a woman, the word offspring isnbsp;inadmissible in the rendering, and the nearest approach tonbsp;the original may be made perhaps by using � kin �: thusnbsp;Miliuc mocu Buain, the name of the king of Daln-Araidenbsp;(between Loch Neagh and Belfast Loch) who boughtnbsp;Patrick as a slave, might be rendered � Miliuc kin of Buan�;nbsp;and so with the following inscriptions from the county ofnbsp;Waterford, Catabor^ moco Viricorb\i\, � Cathbar kin of Fer-Corb,� and Gosoctas mucoi Macorbi, � the Monument ofnbsp;Guasacht kin of Macorb.� A wider choice of words isnbsp;permissible in a case like ad Insolas Maccuchor in thenbsp;Book of Armagh, as we might render it � to the Islandsnbsp;of the family or tribe of Cor��they are the isles at thenbsp;Skerries, off the north-west corner of Antrim ; and when anbsp;woman is the chief ancestral figure we are at liberty to usenbsp;a word meaning progeny and lineal descendants. But whatnbsp;is one to make of the double genitive maqui mucoi, whichnbsp;frequently occurs in ancient Ogam inscriptions, and mustnbsp;mean filii generis or filii gentis ? Take, for example, thenbsp;following from Corkaguiny in Kerry: Maqqui Ercciasnbsp;maqqui mucoi Dovinias? that is to say � (The Monument) ofnbsp;Mac Erce son of the kin of Dubinn,� where the ancestressnbsp;Dubinn (genitive Duibne)ffa.s givenher name to Corco Duibne,

� We are not certain whether we should read Catabor or Catabar ; but compare Ptolemy�s name�Vellabori�of a tribe in the south-west of Ireland, Velvor filia Broho of a somewhat late Cardiganshire inscription, and Falbhar,nbsp;a champion�s name mentioned in 0�Curry�s �Manners and Customs of thenbsp;Ancient Irish, � iii. 158.

2 From a rubbing supplied by the Rev. Edmond Barry, and recently verified by Professor Rhys. The stone is at Lord Ventry�s residence in the neighbourhood of Dingle.

THE PICTISH QUESTION.

Anglicised Corkaguiny, now the name of a barony in the west of Kerry. Or to come back to Wales, take the followingnbsp;Ogam at Brideit in the north of Pembrokeshire : Nettasagrunbsp;maqui mucoi Bred, which may be rendered {Monumentuni)nbsp;Nettasagrus filii generis Bred. These and the like inscriptions take us back without doubt to the words of Caesarnbsp;already cited : in fact they lead us back a little further, tonbsp;wit, to a stage antecedent to the consideration of the paternitynbsp;suggested by him. In Irish literature this state of societynbsp;is found surviving in a form which looks like polyandry, asnbsp;in the case of a king of Tara, supposed to have reignednbsp;about the beginning of the Christian era. He was knownnbsp;as Lugaid of the Red Stripes, and said to be the son ofnbsp;three brothers, the sons of Eochaid Feidlech.^ Or take thenbsp;Mac Lir family of Irish legend : one of its leading figuresnbsp;was the famous Manannan Mac Lir, and this is how he isnbsp;introduced in the opening verses of a poem in the well-known story of Bricriu�s Feast, in the Book of the Dunnbsp;Cow, fo. 50a:

F�gaid mac Iccchraidi Lir, do maigib �ogain Inbir !

Behold the son of the heroes of Lcr,

From the plains of Eogan of Inver !

It is, perhaps, relevant also to mention here that a state of things in which the children were the children of thenbsp;family, so to say, and owned no fathers in particular,nbsp;rendered necessary some arrangement of the nature ofnbsp;fosterage, an institution known to have been of vast importance among the ancient Goidels, including among themnbsp;the family of Pwytt, king of Dyfed, as mentioned in thenbsp;Mabinogi already cited.

A man who styles himself Nepos Vepogeni, or Son of Vepogen�s Sister, without naming her, leaves us no evidence

' See the � Book of Leinster,� fol. 124b, 151a ; the �Revue Celtique,� xvi. 148-50; 0�Mahony�s Keating, pp. 287-8, and the footnote on page 37 above.

THE WELSH PEOPLE, (chap, ii.)

that the community to which he belonged made much of its women. That community appears to have recognised nonbsp;paternity, but to have reckoned descent by birth alone ; itnbsp;is possible, however, that at a previous stage in its historynbsp;the family was constituted on strictly matriarchal lines.nbsp;At all events other cases occur, which seem favourable to thenbsp;belief in the former existence of matriarchy. Certain Ogamnbsp;inscriptions, for instance, have been found in the neighbourhood of Dingle, in Kerry, ending with the ancestress�snbsp;name, nominative Dovim[s], genitive Hovmia[s], reduced innbsp;Mediaeval Irish to Duhinn ornbsp;nbsp;nbsp;nbsp;genitivenbsp;nbsp;nbsp;nbsp;respec

tively, as already mentioned. In the next place, certain well-known characters in Irish literature are distinguished by the mother�s name, such as Conchobar son of Nessa, andnbsp;Fergus son of Roig, with which should be compared thenbsp;Welsh Gofannon son of D�n. Lastly, the legends of heroicnbsp;Erin picture the ladies sitting with their husbands at theirnbsp;banquets, and treated by them as their equals ; and sometimes courtship is represented in Irish story as initiated bynbsp;the woman, not to mention the doings of such personagesnbsp;as Queen Macha or Queen Maive. Supposing that proofnbsp;were to be found that Irish society began with matriarchy,nbsp;several things in Irish literature could be pointed out asnbsp;admitting of easy explanation as survivals; but we darenbsp;not reverse the argument and say that they admit of nonbsp;other explanation, and that we must therefore postulatenbsp;matriarchy.

To say the least of it, however, there is nothing to suggest that individual women might not enjoy greatnbsp;consideration among the early Goidels: there is muchnbsp;to the contrary, and in this connection a question offersnbsp;itself as to the nature of the theology evolved by a peoplenbsp;of the kind. Clearly, if they reckoned descent by birthnbsp;alone, and provided they were given to ancestor worship,nbsp;they must have had female divinities. Unfortunately it

THE PICTISH QUESTION.

happens that the whole range of Irish literature supplies extremely few references in express terms to divinities ofnbsp;any kind, and the few to be found are of the most meagrenbsp;and precarious description. In other words, the Irish pantheon forms but a very dim background to Irish history ;nbsp;but in that vanishing picture it is very remarkable that thenbsp;goddesses loom larger than the gods. Thus we have alreadynbsp;referred to Anu, said by Cormac to have been considerednbsp;the mother of the gods, and we pointed out traces of hernbsp;in Welsh pedigrees derived probably from Goidelic sourcesnbsp;in Britain. Cormac mentions also an analogous figurenbsp;whom we may call Buanu, genitive Buanann and to thisnbsp;latter he gives the position of mother or nurse of Irishnbsp;heroes, and of teacher who taught them feats of arms. Wenbsp;next come to the story of the Second Battle of Moytura,�nbsp;which mentions a people who invaded Ireland at differentnbsp;points, and bore the name of Fir Dornnann, or the Men ofnbsp;Domnu. They came from the west coast of Britain, wherenbsp;we shall presently find them to have borne the namenbsp;Dumnonii or Dumnonians. But the interest of their namenbsp;consists in the fact that the Irish form. Fir Domnann, is asnbsp;it were Viri Dumnonis, taken from that of a goddess Domnunbsp;(genitive Domnann). She was presumably considered to

* The name is given as Buanann, making probably a genitive Buanainne, but this is a comparatively late declension, superseding the older Buanu,nbsp;genitive Buanann. Cormac also gives Anu a. genitive, Anainne, s.v. dna,nbsp;and Danann is sometimes made into Danainne, while Danann or Donannnbsp;occasionally functions as nominative : see the � Book of Leinster,� fol. iia.

^ The story, which will be found published, with a translation by Stokes, in the �Revue Celtique,�xii. 52�130, treats the Fir Domnann as belonging to thenbsp;Fomori, a fabulous race of elves or demons whose name has been supposednbsp;by Stokes (pp. 128, 130) to be derived in part from the same source as thenbsp;latter syllable of the English word nightmare, to which we may add that innbsp;Scotch Gaelic stories the singular occurs as fomhair, meaning a giant. On thenbsp;other hand, popular etymology has associated the Fomori with the sea, muir,nbsp;as if the meaning had been that of a people who were fo muir, � up and downnbsp;the sea, all over the sea hence the term tended to mean invaders who camenbsp;over the sea, and sea rovers or pirates generally.

56 THE WELSH PEOPLE, (chap, ii.)

be their ancestress, and their leader is styled Indech mac D� Domnann, � Indech son of the goddess Domnu,� wherenbsp;mac, � son,� has probably to stand for a distant descendant.nbsp;At any rate, Domnu does not figure as intervening in thenbsp;story, and we may presume that many generations hadnbsp;passed away between her and Indech ; not to mention thatnbsp;Fir Domnann is probably to be equated with Dumnonii,nbsp;one of the most widely-spread designations of the Goidelsnbsp;of Britain.

Far more common, but just as little explicit, are the references to the goddess Danu, after whom were callednbsp;the Tuatha or Tuath D� Danann, �the goddess Danu�snbsp;Tribes or Tribe�; also Fir Dea^'We Goddess�s Men.�nbsp;No one of the leading figures in the many allusions to thenbsp;Tuatha D� Danann is styled Son or Daughter of Danu,nbsp;and as the people called after her are usually spoken of innbsp;the plural as tuatha, � tribes,� she was probably regarded asnbsp;belonging to a distant past. Here we have the advantagenbsp;of a Welsh identification : Danu is the D�n of the Mabinoginbsp;of Math son of Mathonwy; but at the stage in which D�nnbsp;is there found she is no goddess: she is briefly referred tonbsp;as sister to the king and mother of his successor, Gwydionnbsp;son of D�n, and of his brothers and sister, as already statednbsp;(p. 37). All this would have to be tumbled upside downnbsp;by those who seem to think that the Mabinogion have beennbsp;imported into Wales as Irish stories from Ireland.^ The

* Such appears to be the view taken by Professor Kuno Meyer in an article on Gael and Brython in the � Transactions of the Cymmrodorion Society,� 1895-6,nbsp;especially where he speaks (pp. 71-3) to � the deposits of Irish legendary lore �nbsp;which he finds, for example, in the Mabinogion. He instances the Irish storynbsp;of Mesce Ulad, in which a party of Ultonians are induced to be entertained innbsp;an iron house, the iron of which is concealed by the timber covering it bothnbsp;inside and out. When they are found to have drunk freely, their attendantsnbsp;leave them one by one, and the door is shut. Then fuel is piled up round thenbsp;iron house and set fire to. The story relates how the inmates at lengthnbsp;realised their position, and how some of them forced their way out. Now annbsp;iron house story is referred to in the Mabinogi of Branwen ; but, so far as the

THE FieriSH QUESTION.

view which recommends itself to us is that they are stories which were current among the Goidels of old in Britain,nbsp;and, in such instances as that here indicated, they representnbsp;a far earlier state of things than can be said of any Irishnbsp;story extant about the Tuatha D� Danann.

In one case we see, perhaps, a little more closely the deification in process: this takes us back for a moment to the barony of Corkaguiny and the name of the ancestressnbsp;Dubinn. A story,i which, as we have it, was committed tonbsp;writing in surroundings where Aryan ideas had begun tonbsp;prevail, makes Dubinn sister to Cairbre Muse, who appearsnbsp;to have been king in the west of Munster in the third century.nbsp;They had a son called Core Duibne, and the story relatesnbsp;how Cairbre�s realm was visited with bad seasons in consequence of the incest, and how Core Duibne had to be takennbsp;outside his father�s realm by the Druid who undertook thenbsp;boy�s education. Now several of the Ogam inscriptions ofnbsp;Corkaguiny, which may be said to belong to the fifth ornbsp;the sixth century, end with the name of the ancestress. Thusnbsp;one at Ballintaggart, near Dingle, reads: Maqqui laripinbsp;^naqqui Mucoi Dovvinias, � (The Stone) of larip son of thenbsp;Kin of Dubinn.� Another, preserved at Burnham House,nbsp;Lord Ventry�s residence in the same neighbourhood, hasnbsp;been read thus (p. 52) : Maqqui Erccias maqqui Mucoinbsp;Dovinias, �(The Stone) of Mac Erce, son of the Kin ofnbsp;Dubirm.� But the most remarkable one stands on a headland beyond Dunmore Head and looks out on the Atlanticnbsp;Ocean as if prophetically appealing to the Gaels beyond ;

brevity of the Welsh allows us to judge, it cannot have been the Mesce Ulati which the narrator had his story from : most likely the iron house had figurednbsp;in more than one tale. Further, as the iron house incident is there avowedlynbsp;Irish, one can hardly regard it as a very instinctive sample of �the depositsnbsp;of Irish legendary lore � in the Mabinogion: it is desirable to have morenbsp;instances, and of a less self-confessed description.

* See the Book of the Dun Cow, fob 54a ; and Rhys�s � Celtic Heathendom,� PP- 30S, 9-

THE WELSH PEOPLE, (chap, ii.)

it may have been the monument of a son of the chief commemorated on the last-mentioned stone, as it reads on the one edge, Ere ntaqqui Maqqui Ercias, � The Stone of Ere,nbsp;son of Mac Erce,� and on the other. Mu Dovinia, �of Mynbsp;Dubinn.� Here one would have expected a longer legendnbsp;making � Ere son of Mac Erce, son of the Kin of Mynbsp;Dubinn,� as that is probably how we are to construe. Butnbsp;what is most remarkable, if our reading should prove correct,^nbsp;is the use of the prefix mu or mo, which is familiar tonbsp;every student of Irish hagiology as a mark of respect andnbsp;affection prefixed to the names of certain saints. Thus innbsp;Corkaguiny we have first simply Dovinias, � Dubinn�s�; thennbsp;Mu Dovinia[s], � My Dubinn�s,� with the reverential prefix ;nbsp;and had not the deification been arrested by the advancenbsp;of Christian ideas we should have probably had the namenbsp;in a third stage: that is, Dubinn�s son would have beennbsp;known in Irish literature not as Core Duibne, but as Corenbsp;D� Duibne, or the progeny^ of the goddess Dubinn.

The folklore of Ireland from Meath to Beare Haven and Corkaguiny abounds with allusions to an old woman ofnbsp;fabulous age called Bera, B�ara, or B�irre, and she is probably to be identified with the Beara whom certain storiesnbsp;make the daughter of a king of Spain, and wife of Eogannbsp;M�r or Mog Nuadat, who, with Conn the Hundred-fighter,nbsp;is fabled to have divided Erin into a northern and anbsp;southern half between them in the second century. Butnbsp;in those stories Bera�s name is mostly given with the prefix

* It is only right to warn the reader that the reading of the Ogams on the same edge as Dovinia is contested by the Rev. Edmond Barry and bynbsp;Mr. Macalister. Professor Rhys, having become aware that the former readnbsp;it differently, took an opportunity of re-examining the stone in 1891, andnbsp;the result only confirmed him in his former opinion. Mr. Macalister�s remarksnbsp;in point will be found in his � Studies in Irish Epigraphy,� p. 56.

^ It is not known precisely what the word Core meant; nor is it evident that Corco in Corco-Duibne (which also occurs as Corea Duibne) is the plural ofnbsp;Core, as if core, corco, meant child, children, or the like, respectively. Seenbsp;� Pro. Soc. Antiq. Scotland,� xxxij. 355-7.

THE PICTISH QUESTION.

mo (or mu), as in the case of Mu-Dovinia, and then it is found written Mom�ra. The stories^ have not been foundnbsp;so far as we know in any very ancient manuscript; butnbsp;there appears to be no reason to suppose them to have begunnbsp;late. They would seem, however, to have been developednbsp;relatively so late that Bera has only succeeded in attainingnbsp;to the status of a witch or wise woman, of a nun or hag,nbsp;of a revered person and a giantess, not quite to that of anbsp;goddess, unless it be in Argyll, where she rules the storm.

Here also attention may be relevantly directed to the great place which women occupy in the legendary accountnbsp;of the early colonisations of Ireland. Take, for example,nbsp;Scota treated as chief ancestress of the Milesian Irish, andnbsp;as giving her name to all the Scots. She is not, at any ratenbsp;in this context, to be disposed of as a mere myth; for anbsp;cognate eponym of the other sex might have served equallynbsp;well for mythic purposes. The most remarkable instance,nbsp;perhaps, is the case of Cessair, said to have taken possessionnbsp;of Ireland before the Flood: her wanderings are made tonbsp;begin with Noah refusing her and hers room in the arknbsp;which he was building. She is represented landing at Dunnbsp;na m-Barc, � the Fortress of the Barks,� somewhere betweennbsp;Bantry and Tralee. The Irish historian Keating apologisesnbsp;for mentioning Cessair, and suggests it as his reason fornbsp;doing so, that he found her story in old books, such probablynbsp;as the Book of Leinster, folios 4, S, and those used by Dualdnbsp;mac Firbis, a well-known Irish antiquary of the earlier partnbsp;of the seventeenth century, who compiled from old manuscripts his annals known as the Chronicum Scotorum. Hisnbsp;first entry is under Anno Mundi iS99, and it runs thus: �In

* See O�Curry�s volume containing �The Battle of Magh Leana � and �The Courtship of Mom�ra,� pp. xx. 39, 166, and 31�, and comparenbsp;0�Flaherty�s �Ogygia,� p. 274, where he has Bera filia Ocha principisnbsp;Britonum Mannice, whatever that may have exactly meant; also Professornbsp;Kuno Meyer�s �Vision of Mac Conglinne,� pp. 131-4, 208-10, and Professornbsp;Uhys�s � Celtic Folklore,� p. 393.

THE WELSH PEOPLE, (chap, ii.)

this year the daughter of one of the Greeks came to Hibernia, whose name was Heriu, or Berba, or Cesar, andnbsp;fifty maidens, and three men with her. Ladhra was theirnbsp;conductor, who was the first that was buried in Hibernia.nbsp;This the antiquaries of the Scots do not relate.�^ This itnbsp;will be seen equates or co-ordinates Cessair with Eriu,nbsp;the eponym of �riu, genitive Erenn, � Ireland,� and withnbsp;Berha, which there is no sufficient reason for altering intonbsp;the better-known name Banba of another eponym of thenbsp;island. Berba is elsewhere only known as the name ofnbsp;the river Barrow, in which we seem accordingly to havenbsp;another ancestral name. The most reasonable view tonbsp;take of the legend of Cessair is, that it was a localnbsp;tradition of the Aborigines of the south-west of Ireland,nbsp;who by making Cessair the first coloniser asserted theirnbsp;own priority of possession to all other peoples in thenbsp;country. The synchronisers, not knowing what to make ofnbsp;this, accepted the alleged priority, and placed the wholenbsp;story before the Flood. Thereby they rid themselves ofnbsp;difficulties from two possible sources, to wit, the context ofnbsp;the story with the other events occupying their attention,nbsp;and the later fortunes of Cessair�s descendants. The legendnbsp;associates Cessair and her companions with various localitiesnbsp;in the south and west of Ireland, together with others lyingnbsp;so far north as Slieve Beagh in Fermanagh; not to mentionnbsp;that Ireland is occasionally found designated Cessair�snbsp;Island.** We gather, therefore, that Cessair may have beennbsp;the eponymous heroine of a race occupying the whole ofnbsp;the southern half of Ireland and more, together very possiblynbsp;with the nearest portions of the west and south-west of

' See the opening of the Four Masters� Annals of the Kingdom of Ireland, and the editor O�Donovan�s notes on the place-names involved; Joyce�snbsp;edition of Keating�s � History of Ireland,� part i., pp. 52-5 ; and Hennessy�snbsp;(Rolls edition of the) � Chronicum Scotorum,� pp. xxv.-xxxij. 2, 3.

^ See Tochmarc Monira [read Momera] in O�Curry�s � Battle of Magh Leana,� page 154. Compare � Rhys�s � Studies in Early Irish History,� p. 23.

THE PICTISH QUESTION.

Britain as its earlier home.^ However that may be, it proves to our satisfaction that to show a predilection fornbsp;ancestresses over ancestors was Ivernian : that it was alsonbsp;Aryan we are inclined to doubt, but that, where it has beennbsp;found among the peoples of these islands, it is to be tracednbsp;rather to the Aboriginal element in a mixed populationnbsp;promiscuously termed Celtic.

The influence of Christianity must have by degrees put an end to the social system to which we have been referring,nbsp;and this raises questions of great difficulty as to dates andnbsp;localities, on which we cannot enter. So we return to ournbsp;view, that if the reckoning of descent by birth alone wasnbsp;not Aryan, it must have been accepted by the Goidelicnbsp;Celts from the Aborigines, which would go far to prove thenbsp;numerical importance of the latter. It is known to havenbsp;been Pictish, but was it also Celtic and Aryan We are disposed to think that it was not, though we readily admit thatnbsp;the negative cannot be proved. Moreover it is right to saynbsp;that the following passage in chapter xxi. of the Germania *nbsp;of Tacitus is redolent of the same ancient menage : Sororumnbsp;filiis idem apud avunculum qui ad patrem honor. Quidamnbsp;sanctiorem artioremque hunc nexum sanguinis arbitranturnbsp;et in accipiendis obsidibus magis exigunt, tamquam etiam

' The name Cessair, genitive Cesra, admits of being regarded as derived from a stem, cestari, and should M. Salomon Reinach�s conjecture provenbsp;correct, that the Cassiterides originally meant the British Isles, and thatnbsp;tairo-lTepo�, � � tin, � was, like several other Greek names of metals, called simplynbsp;after the country or the people of the country in which it was found, ournbsp;Cessair would be found to supply a necessary link in the reasoning. Fornbsp;M. Reinach�s view see � L�Anthropologie � for 1892, pp. 275-81 ; also Rhys�snbsp;letter on Cassiterides m ihe �Academy,� October 5, 1895, pp. 272-3, and seenbsp;further pp. 298, 342, 366, 390, 414, 438. 524, 547-

Compare chapter viii., which treats female hostages as more efficacious in the case of the Germans�ut efficacius obllgentur animi civitatum^ quibusnbsp;inter obsides puella quoque nohiles imperantur ; and also a passage in Suetonius�snbsp;Augustus, 21, to the following effect : A quibusdam vero [the last peoplenbsp;mentioned seems to have been Germans] novufn genus ohsidunt^ feminastnbsp;^xtgere tentaverit^ quod negligere marium pignoraseniiebat.

THE WELSH PEOPLE, (chap, ii.)

animum firmius et domum latius teneant. �There is the same regard shown for the sons of the sisters by the unclenbsp;as by their father. Some think this tie of blood morenbsp;binding and closer, and insist on it more when they receivenbsp;hostages, on the theory that it restrains their impulses morenbsp;powerfully, and has a wider control over the family.�nbsp;The words are unfortunately so indefinite that we have nonbsp;clue to the identity of the tribes the historian had in view;nbsp;so it is impossible to say whether they were likely to havenbsp;been mixed with any Aboriginal race practising the samenbsp;customs and enjoying the same institutions as the Aboriginesnbsp;of the British Isles.^

We admit that the foregoing argument is not quite decisive, and we now turn to others of a more purelynbsp;linguistic nature, and leading to a more decided conclusion. So we revert for a moment to the dedicator of thenbsp;bronze tablet found at Colchester : we saw that he describesnbsp;himself as Lossio Veda, Nepos Vepogeni, Caledo, and thatnbsp;Veda appears as part of the name of one of the kings in

' Say somewhere between the mouth of the Rhine and that of the Elbe, where there was an amber coast, and where Tacitus would seem to have heardnbsp;of a people partly Celtic and partly Teutonic, whom he has mixed up with thenbsp;iEstii of the amber coast of the Baltic. Witness the following passage in thenbsp;Germania, 45 :�Ergo jam dextro Suebici maris litorc ALstiorum gtntesnbsp;adluuntur, quibus ritus habitusque Sueborum, lingua Britannica propior.nbsp;matrem deum venerantur. insigne superstitionis formas aprorum gestant: idnbsp;fro armis hominumque tuiela securum dece cultorem etiam inter hostis prcestat;nbsp;rants jerri, frequens fustium usus, frumenta ceterosque fructus patientius quamnbsp;pro solita Germanorum inertia laborant, sed et mare scrutantur, ac solinbsp;omnium sucimim, quod ipsi glesum vocant, inter vada atque in ipso litore legunt.nbsp;How well the allusion to the goddess would fit Goidelic surroundings neednbsp;not be dwelt upon ; and, as to the language, glesum is as easily explained bynbsp;means of Celtic as of Teutonic. Witness the Irish.glain, gloin, �glass ornbsp;crystal,� Welsh glcin, glain {gemma, tessera), for an older gles-inu-s ; whilenbsp;a language said to come nearer the Britannica would exactly describe thenbsp;position of Goidelic as compared with Brythonic. The sort of people whichnbsp;the Germania suggests might be Aborigines who had first become Goidels innbsp;speech and later Teutons, while retaining habits and customs which theynbsp;practised before they acquired any Aryan language at all.

THE PICTISH QUESTION.

'^he Pictish list, where also Vepogeni is to be found, curtailed, it is true, to Vepog, and written Vipoig. But while h is probable that Veda is not Celtic, it is certain thatnbsp;EepogenT is, and we compare it with Gaulish names likenbsp;^epus and Vepo-talos, of unknown meaning, and Matu-genos,

� well-born,� or Camido-genos, �offspring of Camulos.� Thus we seem to have in Vepogeni an early instance of thenbsp;Pictish habit of borrowing names and other words fromnbsp;the Celts. The interest of the present instance centresnbsp;in the way in which the name Vepogenos was treated. Thisnbsp;would be the Brythonic and Gaulish form, while in Goidelicnbsp;It would have been approximately Vequagenas. Now thenbsp;study of the laws of mutation of initial consonants innbsp;the Neo-cel tic languages goes to show that the ending ofnbsp;the nominative must have been dropped early, so that thenbsp;foregoing forms would be shortened to Vepogen and Vequa-gen. Another process of curtailment would be to drop thenbsp;thematic vowel of the first element in the compound, bringing the result approximately to Veh-gen and Fech-gen.nbsp;But the reduction of Vepogen to Vepog, which is whatnbsp;underlies Vipoig, is impossible on Celtic ground, whethernbsp;Brythonic or Goidelic, while Pictish offers a simple andnbsp;natural explanation. In that language it can be shownnbsp;that enn or en was a common ending of the genitive case,nbsp;so that Vepogen must in the long run have sounded to thenbsp;Piets as a genitive, whence was readily inferred a simplernbsp;form, Vepog, which we should call nominative in the casenbsp;of Aryan speech.

This leads us to consider the Pictish genitive somewhat further, and to mention another instance in Drosten, whichnbsp;luay be regarded as the genitive of the Pictish name Drost

� Vepogetios was, perhaps, the name represented by the abbreviation VEP the native coins of the Brythons north of the Humber, reading VEP COR F,nbsp;let us say, VEPOGENVS COROTICI FILIVS ; see Rhys�s � Celticnbsp;I'ritain, � p. 41 ; and p. xv., coin 5-

THE WELSH PEOPLE, (chap, ii.)

on a stone already mentioned as being at St. Vigeans, near Arbroath. Take also the well-known Newton Stone in Aberdeenshire, whichhas in Ogam the genitive, Vorrenn} of anamenbsp;which occurs as Vaur in the other lettering on the same stone.nbsp;This genitive occasionally appears in old Irish inscriptions,nbsp;such as one on the island of Valencia which reads, LogirPnbsp;maqui Erpenn, �The Monument of Lugar son of Erp.�nbsp;Had Erpenn been Goidelic, it should, in order to be onnbsp;a level with Logiri and maqui, have been Erpennas, or atnbsp;least Erpenna ; but the presence of the consonant p is verynbsp;fair evidence that the name is other than Goidelic. Thenbsp;Piets appear to have had genitives also in ann, on\n\ andnbsp;Instances of the first-mentioned occur in the mixednbsp;inscriptions found in the Shetlands, such as Meqqddrroann,nbsp;which might be rendered probably Filii Druidis; andnbsp;dattrrann on the same stone seems to be the Pictishnbsp;genitive of the Norse word for daughter? As to onn, thenbsp;Book of Deer mentions a grant of land to the Church ofnbsp;Aberdour, in which among other names of men occur anbsp;nominative Culii, and a genitive Cul�on : they are probablynbsp;cases of one and the same name.^ We have it now and

' See the � Proceedings of the Society of Antiquaries of Scotland,� vol. xxxij. pp. 360-4.

* This is a reading recently made out by Professor Rhys ; but it is not quite certain whether Logiri is to be equated with the Lugiroi Macm Lugir, or withnbsp;Loegairi, the genitive of Loegaire mac N�ill, the name of the king of Irelandnbsp;in whose time Dubthach Maccu Lugir was chief poet (Stokes�s � Goidelica,�nbsp;pp. 86, 126). Reeves in his Adamnan�s Vita Columba (p. 350) cites, afternbsp;Ussher,a passage the writer of which thought that macu, niocu, or mucoihaAnbsp;something to do with muc, �a swine,� so that in his hands the chief poet,nbsp;D. Maccu Lugir, becomes sululcus regis Loigeri plii Nil; but comparenbsp;Stokes�s �Patrick,� pp. 122, 324.

� � Pro. Soc. Antiq. Scotland,� xxvi. 297-300, where the dd of Meqq-ddrroann should probably be pronounced d, and NahhtvvddaBBs taken to mean NahhtvuddaeBs, as a more likely antecedent of the curtailed form Natdads ornbsp;Natdod's, which became historical as the name of the man who discoverednbsp;Iceland.

lt; From the same MS. one might quote Abber-deon (Aberdeen) but for the uncertainty that the Dee is the river implied and not the Don. The entries

THE PICTISH QUESTION.

then in the old inscriptions of Ireland, as, for instance, in an Ogam in the Kilkenny Museum, reading, in two linesnbsp;Mucoi Atr. . . � (The Monument) of the kin of A.,nbsp;Bivadon, \ namely, Bivad.

Perhaps, however, the order intended was the reverse, �The Monument of Bivad, kin of A.� Lastly, a remarkable instance of the genitive in inn occurs in the namenbsp;of the district called the Mearns, approximately thenbsp;county of Kincardine. Mearns is derived from a nativenbsp;name, Mag Gerginn, or Gergind^ which is also found asnbsp;Mag Cirgin, the Plain of Gerg, Greg, Giric or Ciric ; for thenbsp;name appears to have had several forms, between which itnbsp;is not easy to decide, not to mention that it has been confounded with that of St Ciricus. We have an unexpectednbsp;instance of this genitive in an Ogam inscription from thenbsp;townland of Ballinvoher in Corkaguiny, county Kerry. Itnbsp;reads, Coimagni maqui Vitalin, � The monument of Coemin

in Gaelic, including the names here in question, will be found printed and translated in Stokes�s � Goidelica,� pp. 106-IH ; and h propos of Vepogennbsp;and abbtr may be mentioned the old Welsh Morgant (xaoA. Welsh Morgan),nbsp;which appears borrowed as Morcunt, Morcunn, Morgainn; all three occurnbsp;in the genitive, and Morgunn in the nominative. Query, whether such Pictishnbsp;names as 7'alargan, Talorcen, Talorc, and kindred forms are not all adaptationsnbsp;of a Erythonic Talargent or Jalargant, � Silver-forehead�?

� � Book of Leinster,quot; fob 319c; Skene�s �Chronicles of the Piets and Scots,� p. 319 ; also Skene�s � Celtic Scotland,� i. 295, where he purports tonbsp;give Terra Circin from the Irish annalist Tigernach. For Gerg see O�Cuvry,nbsp;i. ccclxxv. ; in. 168, 307. The name Gerg occurs frequently in � The Tragicnbsp;tory of the Courtship of Gerg�s daughter Ferb � in the book of Leinster,nbsp;fob 253a�259b, where it is mostly nominative Gerg. genitive Ceirg (also Gerg).nbsp;His house was in Glen-Gerg in Ulster, but there was another Glenn-Gerg innbsp;Carlow : see the � Four Masters,� A.D. 1015. We have the genitive also possiblynbsp;m the patronymic of Munremur mac Gerrcind, an Ulster champion introducednbsp;to checkmate the Connaught magician Curoi mac Dairi, in the Tain B� Cualnge,nbsp;m the �Book of the Dun Cow,quot; fob 71b. Various forms of this namenbsp;occurred in Scotland, as will be seen under Grig in the index to Skene'snbsp;Chronicles of the Piets and Scots ; and some of them were stereotyped in thenbsp;hairie of the Mearns church, Eccles-greig or Eglis-girg, now called St. Cyrus,nbsp;edicated to St Ciricus : see Skene�s �Celtic Scotland,� i. 333, 4.

W.P. nbsp;nbsp;nbsp;*.

THE WELSH PEOPLE, (chap, ii.)

son of Fidlin,� where correct Goidelic would seem to require Coimagni maqui Vitalini. The scores are all certain, andnbsp;the third genitive never had a final i on the stone.^

All this has to do chiefly with the inflection of the names to which we have referred; and though it supplies convincingnbsp;evidence to the presence of some element other than Celtic,nbsp;whether Goidelic or Brythonic, Irish nomenclature providesnbsp;us with a still more sweeping argument to the same effect.nbsp;We allude to an important group of Irish names formednbsp;much in the same way as Hebrew names are representednbsp;chosen in the Old Testament. We begin with an instancenbsp;from a comparatively late manuscript, namely, that whichnbsp;contains the story of the Battle of Magh Leamhna :�Anbsp;certain Munster prince called Mogh N�id, � Slave of Ned,�nbsp;had a son called Eoghan M�r, or � Big Eoghan,� who wasnbsp;fostered by one of his nobles called Nuadha. One daynbsp;when Eoghan, in the company of the Druid, whose namenbsp;was Deargdamhsa, were watching the building of a rathnbsp;for Mogh N�id, the workmen came upon a stone which theynbsp;were unable to lift to its place, but the boy Eoghan M�rnbsp;went and lifted it at once, to everybody�s astonishment;nbsp;whereupon the workmen exclaimed, � This is a noble slavenbsp;that Nuadha has.� The Druid then said, � That name shallnbsp;be upon him for ever,� that is, Mogh Nuadhad., � the Slavenbsp;of Nuadha.�� It is right to say that N�d and Nuadha werenbsp;names of gods, the former a god of war of the ancientnbsp;Goidels, and the latter a god the remains of one of whosenbsp;temples have been found at Lydney, near the Severn,

' The stone is in the Irish National Museum in Dublin, where it has been examined repeatedly by Prof. Rhys, A paper on it by the Lord Bishop ofnbsp;Limerick will be found in the third series of the �Proceedings of the Royalnbsp;Irish Academy,� 1893, pp. 374-9, where he would identify Fidlin and Welshnbsp;Gwythelin with VUalin.

2 See Curry's � Battle of Magh Leana � (Dublin, 1855), pp. 1-3 : we have given the names in the late spelling in which Curry left them ; the older formsnbsp;would be Mug Net or Neit and Mug Nuadat,

THE PICTISH QUESTION.

together with a representation of his person careering in a chariot over the waves of the sea; and from inscriptions innbsp;honour of him we know that his name had in Roman timesnbsp;the form Nodens orNudens, genitive Nodentis, Nudentis,nbsp;or Nodontis, the termination of which is doubtless Latin dnbsp;in Welsh the name has become Nult and also Rutt. It isnbsp;needless to point out how such names as Mogh N�id andnbsp;Mogh Nuadhad resemble such instances as Abdiel, � Servant of El,� Abdastartus, � Servant of Astarte,� and othersnbsp;from Semitic lands.

To go back to an older manuscript, namely, the � Book of the Dun Cow,� which was written as we have it before thenbsp;end of the year i io6, we have there one of the oldest andnbsp;most weird of fairy tales, which relates how a fairy damselnbsp;came to entice Condla the Red, one of the two sons ofnbsp;Conn the Hundred-fighter, to go away with her to thenbsp;Land of the Living, and how Conn sent for Cordn, hisnbsp;Druid, to counteract her wiles. Cordn tried to do so, andnbsp;failed ; the youthful Condla leaps into the fairy�s glass boatnbsp;and away they sail till they are lost to the sight of Connnbsp;and his astonished friends. Before Conn had stirred fromnbsp;the spot his other son, called Art, came to them, when hisnbsp;father exclaimed, � Art is now solitary {oenfer, oenur), fornbsp;he has no brother.� � That is the word,� said the Druid ;nbsp;�that will be his name for ever, Oenfer!' ^ Here againnbsp;the Druid does his part, though in this instance he seemsnbsp;only to add an epithet, but Ar^, though common enoughnbsp;as an Irish name, probably meant as an appellative �a

� See tlie Berlin Corpus, vol. vii. Nos. 137-141 ; also the numerous plates with which are illustrated a posthumous work on the �Roman Antiquitiesnbsp;at Lydney Park,� by W. H. Batliurst, edited by C. W. King (London, 1879);nbsp;and a paper by Hubner on the Sanctuary of Nodens in the �JahrbUcher desnbsp;Vereins von Alterthumsfreunden im Rheinlande,� vol. Ixvii. pp. 29-46.

See the �Book of the Dun Cow,� fol. 120 : the story will also be found printed in Windisch�s � Kurzgefasste irische Grammatik mit Lesest�cken�nbsp;(Leipsic, 1879}, pp. 118-120.

F 2

THE WELSH PEOPLE, (chap, ii.)

bear,� in Welsh arth; and its original use in this story may have been that of a nickname which was allowed to standnbsp;as part of the new name.

The next two instances come also from the � Book of the Dun Cow, � but they belong to the Ultonian cycle of stories.nbsp;The first to be mentioned relates to Derdriu, the heroine ofnbsp;the story of the Exile of the Sons of Usnech. Even beforenbsp;she was born there were forebodings of the troubles beforenbsp;her, and before her country on her account. For whilenbsp;yet unborn she screamed in a way that alarmed Kingnbsp;Conchobar mac Nessa and his nobles as they sat banqueting in the house of her father and mother; andnbsp;Conchobar�s Druid, whose name was Cathbad, exclaimed onnbsp;hearing the scream which the child had given (derdrestar)nbsp;and said, � Verily it is a girl, and let Derdriu be her name.�nbsp;Here again it is the Druid of the party that fixes the namenbsp;of the child, and the relation between the deponent verbnbsp;derdrestar �nbsp;nbsp;nbsp;nbsp;and Derdriu, purports to explain how

he came to call her Derdriu. The rest of the story need not be reproduced here.^ The other story is one related ofnbsp;Cuchulainn�s infancy.** Conchobar and his nobles werenbsp;gone one evening to feast at the house of Culann the Smith,nbsp;who was a great man among them. Culann, when theynbsp;were all supposed to have arrived, had his gates closed, andnbsp;a famous war-dog of his was let loose, as was his wont, tonbsp;guard his possessions. The boy Cuchulainn, however, hadnbsp;been forgotten, and when he arrived he was attacked bynbsp;Culann�s watch-hound ; but, to everybody�s astonishment,nbsp;the boy killed the formidable beast. Culann afterwardsnbsp;complained loudly of his loss, whereupon Cuchulainn saidnbsp;that until the smith had another hound of the same breednbsp;reared to guard his possessions he would guard them : henbsp;would be himself Culann�s war-dog. Thereupon Cathbad

' For the text see Windisch�s � Irische Texte � (Leipsic, 1880), pp. 67-69.

* See the � Book of the Dun Cow,� fol. 6oa-6la.

THE PICTISH QUESTION.

the Druid said, � Let Cii-Chulainn (Culann�s Hound) be thy name �; for till then he had been known as Setanta Bee,nbsp;or the Little Setantian, in reference probably to his race;nbsp;for we know that he cannot have been altogether ofnbsp;Ultonian descent, as he and his father are represented asnbsp;never liable to the cess minden or couvade sickness ^ of thenbsp;Ultonians, while on the other hand the opposite coast ofnbsp;Britain had, according to Ptolemy, a Port of the Setantiinbsp;somewhere near the mouth of the Ribble. Before leavingnbsp;the incident which fixed Cuchulainn�s name, let us observenbsp;that the Druid Cathbad who gave it him was also the schoolmaster or tutor of the young nobles of the Ultonians : henbsp;had, we are told,* no less than one hundred at the same timenbsp;\^z.xvi\ngdruidecht or magic from him, andCuchulainn is foundnbsp;afterwards boasting that, in consequence of the teaching ofnbsp;Cathbad, he was an adept in � the arts of the god of magic,�nbsp;or whatever the term druidecht may have precisely meant.

We now turn to the Mabinogion, which represent, though doubtless not very closely, the stories of the Goidels ofnbsp;ancient Wales ; and there we have at least two instancesnbsp;in point. One of them relates how Lew Lawgyifes got thatnbsp;name given him by Gwydion from an exclamation madenbsp;by Arianrhod, the boy�s mother, when she saw him makingnbsp;a skilful hit at a wren. Thereby she unwittingly undid anbsp;destiny which she had put on her boy, that he should nevernbsp;have a name, whereby she had intended to protect hernbsp;own reputation as a maiden. This is from the Mabinogi ofnbsp;Math,� but the other occurs in that of Pwylt, Prince of Dyfed.nbsp;This latter Mabinogi relates how Rhiannon, his queen, had

� For the Irish account of this cess or suffering, see the �Berichte der k. Sachs. Gesellschaft der Wissenschaften, philologisch-historische Classe,� for 1884

(PP- 336-47), where Windisch discusses the question and gives some texts relating to it.

^ See the � Book of the Dun Cow,� fol. 6ia and fol. 124b (printed in Windisch�s Irische Texte, p. 325).

� Seethe Oxford Mabinogion, pp. 69-71; and Guest�s Mabinogion, iii. 233-6.

THE WELSH PEOPLE, (chap, ii.)

her firstborn kidnapped by witches immediately after his birth, and how the child, a wondrously fine boy, was foundnbsp;by Teyrnon, a chieftain of Gwent, and brought up by hisnbsp;wife as her own child. When Teyrnon, however, who hadnbsp;been one of Pwyft�s men, heard of Rhiannon�s trouble, itnbsp;struck him that the boy was unmistakably like Pwy�; sonbsp;he and his wife resolved to restore the boy to Pwylt andnbsp;Rhiannon at their court in Dyfed. When Teyrnon andnbsp;the boy were being entertained at table by Pwyli andnbsp;Rhiannon, with the nobles of their court, Teyrnon relatednbsp;the wonderful story of his finding the boy, and appealed tonbsp;all present to say whether they did not agree with him thatnbsp;the boy was Pwyli�s son. They responded with one accordnbsp;in the affirmative, whereupon Rhiannon exclaimed, � I callnbsp;heaven to witness that this, if true, would deliver me ofnbsp;my anxiety {pryderi)'.' Pendaran of Dyfed at once said,nbsp;� Well hast thou named thy son Pryderi, and the namenbsp;Pryderi son of Pwylt Penn Annwn befits him best.� Henbsp;had been called by Teyrnon and his wife �Gwri of thenbsp;Golden Hair,� but Pwylt insisted on his being now callednbsp;Pryderi according to his mother�s word when she got joyfulnbsp;news of him, and on his being fostered by Pendaran ofnbsp;Dyfed.^ Pendaran is not called a Druid�nobody is callednbsp;a Druid in the Mabinogion�but both he and Gwydionnbsp;remind one of the Druids of Irish tales ; and, furthermore,nbsp;the latter portion of the name of Penn-Daran, � Chiefnbsp;Daran,� is probably to be identified with that of the Irishnbsp;Druid, Dalan, in the story called the � Wooing of �tdin,�nbsp;where also the name Etain of the heroine happens to equatenbsp;letter for letter with the latter part of the name Rhiannon,nbsp;on the hypothesis of this last representing an earliernbsp;Rig-Antonf meaning � King�s Anton � or� Royal Anton.�

' See the Oxford Mabinogion, pp. 18-24; and Guest�s Mabinogion, hi. 60-70.nbsp;nbsp;nbsp;nbsp;,

- Tire text of the � Wooing of Etain � has been published in Windisch�s

THE PICTISH QUESTION.

These stories transport us into an atmosphere more like Semitic than Aryan, and we notice�(i) first, that the permanent name is drawn from some incident in one�s history,nbsp;and that as a rule it supersedes any name or nickname ofnbsp;one�s infancy; (2) secondly, that the name is fixed by thenbsp;Druid or by the foster-father and tutor ; (3) lastly, that innbsp;the Mabinogion the part played by the mother is regarded asnbsp;essential; and it suggests that the giving of the name wasnbsp;originally her exclusive right, while the man who took it upnbsp;only gave the transaction a certain stamp of ceremony andnbsp;publicity. Now this plan of naming men and women couldnbsp;not help resulting in names differing widely from thosenbsp;given under the Aryan system of nomenclature; but beforenbsp;proceeding to consider the latter, let us for a moment takenbsp;stock of the names we have been discussing. Their prevailing nature may be gathered from the following enumeration of the most common of them :�(i) First may benbsp;mentioned those formed with mug or mogk, � slave,� such asnbsp;Mogk Nuadhad, � the Slave of Nuadha.� (2) Those formednbsp;with mail, � cropped, tonsured,� are analogous, such as Mdil-Patraic, � the tonsured man (the slave) of Patrick,� callednbsp;in Latin Calvus Patricii. Names of this kind had anbsp;great vogue among Irish Christians, but the formula wasnbsp;doubtless pagan, and some of the names recorded appear tonbsp;be so, such as Mail-genn, the name of the third centurynbsp;Druid who is related to have caused demons to kill kingnbsp;Cormac mac Airt, because he had become a Christian.^nbsp;(3) The same view may be taken of names with gtlle^ suchnbsp;as Gilla-Muire^ � the gillie or servant of (the Virgin) Mary,�nbsp;Anglicised Gilmore. Names of this class also became verynbsp;common among Goidelic Christians, but we appear to havenbsp;a pagan instance in Gilvaethwy son of D�n, in the Mabinogi

Irische Texte,� pp. 117-130, and a letter by Prof. Rhys on the equations ere suggested will be found in the � Academy � for August 15th, 1896, p. iiS-

' See the Four Masters, A.n. 266.

THE WELSH PEOPLE, (chap, ii.)

of Mith. (4) Another vocable used for names of this kind was nia, genitive niath, � a champion,� as in Nia Segamain^nbsp;which appears in the genitive in Ogams as Neta Segamonas}nbsp;�the Champion of Segomo� Segomo being in Gaulishnbsp;theology a divinity equated with the Latin Mars. (5) Sonbsp;with fer, � a man, vir� as in Fer Tlachtga^ � the Man ofnbsp;Tlachtga.� (6) The same idea approximately was doubtlessnbsp;expressed hy cu, �hound,� the meaning intended being thatnbsp;of a watch-hound, champion, and protector, as in Cii-chulainn,nbsp;�Culann�s Hound,�nbsp;nbsp;nbsp;nbsp;� Corb�s Hound,� Cu-Chocriche,

� the Hound of the Frontier, the Watch-dog of the Boundary.� (7) Mac, � boy,� was used in much the same way, as for instance in Mac Naue, � Boy of the Boat or Ship,�nbsp;rendered by Adamnan Filius Navis,^ and Mac Tdil,^ supposed to mean � Boy of the Hatchet� or �Son of the Adze.�nbsp;(8) Feminine names with dcr are analogous, such as Her-Lugdach and Der-Fraich, meaning probably the maid ornbsp;woman of Lugaid and Fraech respectively. (9) Names withnbsp;niott, such as Niott- Vrecc, and Ni�th-Fruich or Nad-Fraich,nbsp;have already been sufficiently discussed.

The foregoing will suffice to show how this kind ot personal name forms a very striking feature of Goidelicnbsp;nomenclature, although the majority of Goidelic names arenbsp;of another description, remaining, as they do, true to thenbsp;Aryan system. On the other hand, it is not to be supposednbsp;that the above syntactic names are in any way the outcomenbsp;of a disintegration of Aryan names, or that as a group theynbsp;date later here than the Aryan ones. The reverse would be

' See Rhys�s � Celtic Heathendom,�p. 33; and Stokes�s �Celtic Declension,� p. 87. It is to be noticed, that, though the later forms of the word fornbsp;� champion,� nia, niath, become confused with the words for �nephew,� nia,nbsp;nioth (p. 48 above), the Ogmic spellings were respectively netta or neta, andnbsp;niotta.

2 See the � Book of Leinster,� fol. 326*^.

^ Reeves's Adamnan�s �Life of St. Columba,� preface, p. 9.

�* Stokes�s � Martyrology of Gorman,� June Ilth and October 9th.

THE PICTISH QUESTION.

nearer the truth. Take, for instance, those names which consist of a noun followed by a genitive, such as Mug N�itnbsp;or Mug Nuadat. How little we know about N�t or Nuadanbsp;it is needless to suggest; and as for Tlachtga in Per Tlachtganbsp;We know nothing, except that Irish story makes Tlachtganbsp;the daughter of a famous Druid called Mog Ruith, whom itnbsp;brings in contact with the Simon Magus of Christian legend;nbsp;und that it associates Tlachtga�s name with an ancient rdthnbsp;on the Hill of Ward (in Meath), where a fire used to benbsp;kindled at Allhallows, and distributed to the country roundnbsp;about.^ Then who knows anything about Corb ? And stillnbsp;we have not only Cu Chord, �Corb�s Hound,� and Per Corb,

� Corb�s Man,� but also Nia Corb, � Corb�s Champion,� Mac Corb, � Corb�s Boy or Son,� Mug Corb, � Corb�s Slave,� Artnbsp;Corb, � Corb�s Bear.� So we should probably not be farnbsp;wrong in supposing that Corb was a divinity, fetish, totem,nbsp;or ancestor of the Aborigines. In any case Corb must benbsp;regarded as antecedent to such personal names as Cu Chord,nbsp;Per Corb, and the like. In other terms, these names arenbsp;conglomerates involving elements derived from an ancientnbsp;system, and the obscurity that surrounds them precludes,nbsp;in most cases, one�s regarding the class of personal namesnbsp;in which they are present as late.

The Aryan system of personal names differed from the foregoing very strikingly. They may be classed undernbsp;two heads: first come the full names, consisting not ofnbsp;words in syntactic relation to one another, but of twonbsp;elements forming real compounds, such as, Sanskrit Candra-^aja, from candra, � shining, moon,� and rbja, � king,� Greeknbsp;d^ioyiv-q^, � descendant of Zeus,� Gaulish nej/vo-ovwSo?, meaningnbsp;� white-headed,� from penno-s, � head,� and vindos, � white,�nbsp;in Welsh Pen-wyn, and in Irish Cenn-fhinn, of the samenbsp;meaning. The number of words employed for the purposes of this composition does not appear to have been atnbsp;' For references see Rhys�s �Celtic Heathendom,� p. 515, note 2.

THE WELSH PEOPLE, (chap, ii.)

any time very great, but the best use was made of them when the Greeks formed not only�Iinrapxos but also�'Apxiir7ros;nbsp;and similarly in Old High German, Hariberht (Englishnbsp;Herbert) and Berhthari, Servian Milodrag^x\.�.Dragoniil,d.n�.nbsp;so in some of our inscriptions, such as the one at Laugh-arne in Carmarthenshire reading BARRIVENDI FILIVSnbsp;VENDVBARI4 The two names occur in later Irish asnbsp;Barrfhinn and Finnbharr respectively, in Welsh Berwynnbsp;and Gwynfar: both seem to have meant white-toppednbsp;or white-headed. These are instances of the Aryan fullnbsp;or compound name, but there was another class consistingnbsp;of the full name reduced to one of the elements in thenbsp;compound and supplied sometimes with hypocoristic ornbsp;endearing terminations. Thus in Greek,� for instance, wenbsp;find besides the compounds NiK�/xaxos, NiK�orpaTos, and thenbsp;like, shorter forms such as Nuceas, NikSs, N�ccuv, Nikevs, NI/ws,nbsp;NiVuXXos, and a good many more. Similarly, besides suchnbsp;names as Cadwallaon, Cadfael, Cadfan in Welsh, we havenbsp;the shorter ones suggested by them, Cadog and Catwg.nbsp;Many more might be added from all the Celtic languages,nbsp;but the foregoing will serve to show v^hat the Aryan systemnbsp;of names was, and how it would have taxed the ingenuitynbsp;of the cleverest Druid to select many incident names whichnbsp;at the same time should sound Aryan of the approvednbsp;type. Not only were the two systems different, they mustnbsp;have been incompatible, mutually destructive; and it isnbsp;needless to say that the habit of giving children incidentnbsp;names cannot have been developed in Aryan surroundings.nbsp;It is the less artificial of the two, and belongs to a rudernbsp;race ; and no evidence could well be more conclusive as tonbsp;the former presence in these Islands of a population ofnbsp;natives of non-Aryan origin.

� See Rhys�s � Lectures on Welsh Philology,� pp. 279, 388.

� For a discussion of the whole subject of Greek proper names, see Kick's �� Griechische Personennamen � (Gottingen, 1874),

CHAPTER III.

ROMAN BRITAIN.

The following chapter is intended mainly to do two things : to elucidate the map of Britain in the first centurynbsp;of our era which faces this page, and to lead up to thenbsp;history of Wales proper. The facts, where they are notnbsp;new or submitted to a fresh examination, are taken fromnbsp;Rhys�s � Celtic Britain,�' checked by Mr. F. Haverfield�snbsp;map of Roman Britain,� and the succinct account of thenbsp;Province with which his map is accompanied.

From what has already been said it will be seen that Pytheas, when he visited this country in the 4th centurynbsp;before our era, is not likely to have found any Brythonsnbsp;here: the inhabitants of the south of the Island consistednbsp;then of the Aborigines, with Goidels as the race rulingnbsp;over some or all of them. It is unfortunate that Pytheas�snbsp;account of his visit is not extant; abstracts, however, fromnbsp;his diary have come down through such channels as thenbsp;Works of Diodorus, Strabo, and Pliny. But the evidencenbsp;which principally concerns us is concentrated in a fewnbsp;proper names, such as Albion, Belertum, Brttanm, Cantium,nbsp;Ictis, Morimarnsani, and Pretanic Islands. Of thesenbsp;Cantium and Pretanic must be regarded as Brythonic,

' Published by the Society for Promoting Christian Knowledge, 3rd ed.,

2 Q X

of til nbsp;nbsp;nbsp;** Historical Atlas of Modern Europe from the Decline

� Eoman Empire� (Oxford, Clarendon Press, 1896), Plate XV.

THE WELSH PEOPLE, (chap, hi.)

Belgic, or Gaulish, and not Goidelic. The earliest existing work which alludes to the south-eastern portion of Britainnbsp;as Cantium is Csesar�s account of the Gallic War, and thennbsp;follows Diodorus with his KaVnov.i The name has thenbsp;same form in the pages of Strabo, who uses it in passagesnbsp;devoted to a criticism of the statements of Pytheas in such anbsp;way that it has been supposed that Pytheas had used it himself On this, however, one can only surmise that Pytheasnbsp;either employed a slightly different form of the word ornbsp;else that it reached him from a Brythonic or Belgic sourcenbsp;on the Continent. The same kind of remark would applynbsp;to Pretanic, if it could be shown that any such a term wasnbsp;known to Pytheas, for the Celtican or Goidelic form mustnbsp;be supposed to have been Qi^rtanic. The form, for example,nbsp;in which it occurs in Ptolemy�s Geography is ILptTaviKr) N^o-osnbsp;or IXperaviKal N^o-ai, � Pretanic Island or Islands,� the latternbsp;of which meant the British Isles generally, the largest ofnbsp;them being called Albion, �Albion, Britain,� and the nextnbsp;in size and importance 'lovepvCa, or Ivernia, � Ireland.� Thenbsp;collective name has its living cognates in the old Goidelicnbsp;Cruithni, �Piets,� Cruithnech,'W\dai)xquot; in Old Welsh Priten,nbsp;later Pryden, Prydyn and Pry dein, now Prydain? � Scotland,nbsp;Alba, or the Pictland of the North,� and Ynys Prydain,nbsp;� Great Britain,� literally � Prydain�s or Piets� Island.� Thusnbsp;the name of the Aborigines implied by these vocables wouldnbsp;have been in Greek orthography IIpcTavoi, with whichnbsp;eventually another and an unconnected name was confounded, namely that of the Brittani, and the confusion isnbsp;to be detected in the tt of nperraviK^, YlperTaviKai, and in thenbsp;e of Bperravot. The name of the Brittani was, as alreadynbsp;suggested, more usually and less correctly made in Latin

' See Cjesar�s Gallic War, v. 13, 14, 22; Diodorus, v. 21, 3; and Meineke�s Strabo, i. 4, 3 (C. 63), iv. 3, 3 (C. 193).

� �01 Priten see �Y Cymmrodor,� ix. 179: the other forms occur in the plural, meaning Piets, in the Books of Aneurin and Taliessin: see Skene, ii.nbsp;92, 209.

ROMAN BRITAIN.

orthography into Britanni, until at last it was ousted by the name as pronounced by those people themselves,nbsp;namely, Br�ttSnes. This last has regularly yielded thenbsp;Welsh Brython and French Bretons, while French Bretagnenbsp;similarly represents Brittania, not Britania or Britannia;nbsp;and it cannot be regarded as an accident that the Latinnbsp;Brittani corresponds exactly to the Mediaeval Irish pluralnbsp;Bretain, genitive Bretan. In other words the form Brittaninbsp;must have reached the Romans from the non-Brythonicnbsp;Celts of these Islands or of the Continent.

Let us now take the other names, (i) beginning with Pliny�s Albion, which is treated in Greek as quot;Kk^iov,nbsp;�AX/StW or �AXoinW, genitive �AX/Stoiros or �AXouiWos.^ Thisnbsp;name is unknown to the Brythonic dialects, except thatnbsp;Modern Welsh literature sometimes borrows Alban for Scotland; but it survives in the Goidelic dialects, namely, as Alba,nbsp;AIpa, and Elpa, genitive Alban (also Alba^). Traces of itsnbsp;application to the whole of Great Britain� before it camenbsp;to be confined to the northern portion of it occur in Irishnbsp;literature; and the fact of the Island being called Insulanbsp;Albionum in the Ora Maritima of Avienus^ makes it probable that the name is very ancient. Albion is supposed,nbsp;and perhaps rightly, to mean the White Country, in referencenbsp;to the appearance of the cliffs of the southern coast, and atnbsp;first it was applied presumably only to the south. There isnbsp;no evidence that the Brythons or Belgic Gauls used the word,nbsp;but rather that they translated it into their own tongue asnbsp;Caution; for some believe that also to have meant the

* nbsp;nbsp;nbsp;Pliny�s � Historia Naturalis,� iv. 30, i.

* nbsp;nbsp;nbsp;Rhys�s � Manx Phonology,� p. 85.

� See � Cormac�s Glossary,� s.v. �Mug-�ime,� and the instance in the Duan Albanach, quoted at the close of this chapter, p. 115 ; see also Stokes�s � Urkel-tischer Sprachschatz,� p. 21, where he explains the name to mean Whitenbsp;Land.

* nbsp;nbsp;nbsp;See the lines in question quoted and explained in Mullenhoff�s �Deutschenbsp;Altertumskunde,� i. 91.

THE WELSH PEOPLE, (chap. hi.

White Country.! Unlike the other name, however, Cantion remained confined to the south-east of the Island, so that itnbsp;has yielded the Welsh Caint and the English Kent.

(2) nbsp;nbsp;nbsp;Belerium is given twice by Diodorus as BcAcpiov,nbsp;which Ptolemy�s manuscripts give as BoX�piov, and it isnbsp;supposed to have meant some portion of the south-westernnbsp;peninsula, including probably the Land�s End. In Oldnbsp;Irish we have a word which comes very near it, namelynbsp;the neuter noun ��lre, (Modern Irish b�urld), which meantnbsp;a language, and unqualified, perhaps, an alien or foreignnbsp;tongue. This would explain how in Modern Irish it meansnbsp;�English�; and the inference suggested by the occurrencenbsp;of the name Belerion is, that it became current among thenbsp;Goidels at a time when the language of the Aborigines wasnbsp;still dominant over a certain area of the south-west of thenbsp;Island. Perhaps, however, it was merely meant to describenbsp;the extreme south-west as a tongue of land.

(3) nbsp;nbsp;nbsp;Ictis is the name recorded by Diodorus as given tonbsp;one of the islands at high tide to which the inhabitants ofnbsp;the south-western peninsula of Britain brought their tin fornbsp;sale to the merchants who traded with them. The samenbsp;name was known to Pliny, for his Insulam Mictim isnbsp;doubtless to be corrected into Insulam Ictim. But hisnbsp;account of Ictis differs from that of Diodorus, althoughnbsp;both are supposed to have drawn their information fromnbsp;Timaeus, a historian who was contemporary with Pytheas.nbsp;There seems to be no sufficient reason for identifying Ictisnbsp;with Veciis, � the Isle of Wight,� or to sever it from thenbsp;Irish name of the English Channel, namely, Muir n-Icht;-

' See Stokes�s � Urk. Sprachschatz,� p. 90, and Holder�s � Alt-celt. Sprach-schatz,� s.v. Cantion, canto : compare the Welsh word can, � white.� A certain school of English historians pretend that Cantium is in WelshCaint, �thenbsp;Kent,� and that it meant �the open country.� This interpretation comesnbsp;from Dr. W. Owen Pughe, but where the definite article has been foundnbsp;prefixed to this proper name we have not yet discovered. Both �the Caint�nbsp;and � the Gwent� figure among the curiosities of Guest�s � Origines Celticae �

* �Cormac�s Glossary,� j.z�. Mug-�ime.

ROMAN BRITAIN.

� the Sea of Icht, or Ictian Sea.� IcBs and Icht represent possibly a Celtic pronunciation of the same Aboriginalnbsp;word which the Romans made into Pictus} plural Picti;nbsp;for if the Celts learned the word sufficiently early they wouldnbsp;naturally treat it like any other word with the consonantp,nbsp;that is to say, they would get rid of that consonant as innbsp;their own words. It is probably a mistake to suppose thatnbsp;in this name we have the Latin pictus, �painted,� any morenbsp;than in the name of the Pictones in Gaul. Had it beennbsp;Latin it could hardly have been regarded as other than anbsp;kind of nickname, and no one would have expected thenbsp;Aborigines of Caithness and Sutherland to give the Norsemen who first reached their shores a Latin nickname asnbsp;their national designation. Rather must we suppose it annbsp;early name, which the Aborigines adopted, while the Celtsnbsp;sooner or later applied another name, Qurtani, Pretani,nbsp;Cruithni and Prydyn^ to them in Goidelic and Brythonicnbsp;respectively.�

� This may, perhaps, be regarded as confirmed by Ictium, given by Holder as an old name of a place now called LTsle-Jourdain in the D�p. of the Vienne,nbsp;covered by the eastern portion of the old province of Poitou. For Poitounbsp;represents an older Pictavi, another form of the name of the Pictones, and bothnbsp;claim close kinship with that of the Piets of this country. For the latter werenbsp;not only called Picti, but also Pictones (see Stokes�s �Annals ofTigernach�nbsp;in the �Revue Celtique,� xvii. 251, 253); and probably Pictores, which, undernbsp;the influence of the genitive plural Pictorum, is not uncommon, is everywherenbsp;to be corrected into Pictones. The Paris document, published by Skene at thenbsp;head of his collection of the Chronicles of the Piets and of the Scots, has, innbsp;that compilation, Pictavia and Pictaviam seven times. Skene�s v is meant tonbsp;represent the u of the manuscript; but on scrutinising the original (Latin, 4126)nbsp;in the Biblioth�que Nationale, we find only one instance which looks likenbsp;Pictauia. The others we should read Pictania, Pictaniam, with ni formednbsp;like m. The MS. appears to be a fourteenth century copy of an original of thenbsp;tenth. Add to this that the Life of St. Cadroe calls the Aborigines of Irelandnbsp;gentem Pictaneorum: see Skene�s � Piets and Scots,� p. 108; also p. 137, wherenbsp;the unusual form Piciinia is given.

* The words Cruithni and Prydyn have been regarded as derived, though the nasal has not been exactly accounted for, from the Irish and Welsh words fornbsp;�form or shape,� namely cruth andpryd respectively, and a reference in themnbsp;has been assumed to the forms or outlines of the beasts which the Piets are

THE WELSH PEOPLE, (chap, hi.)

(4) Morimarusam is said by Pliny, after Philemon, to have been the name of the northern ocean from the Cimbri�snbsp;country to a certain Cape Rubeas, and to have meant Deadnbsp;Sea. The passage, somewhat carelessly given by Pliny,nbsp;is repeated in a less ambiguous form by the later authornbsp;Solinus, xix. 2 :�Philemon a Cimbris ad promunturiumnbsp;Rubeas Morimarusam dicit vocari., hoc est Mortuum Mare:nbsp;ultra Rubeas quicquid est Cronimn nominal} Scholars arenbsp;not inclined to regard Morimarusam as a specimen of thenbsp;language of the Cimbri, whom they regard as a Teutonicnbsp;people, while on the other hand it admits of being explainednbsp;exactly as Celtic, Mori Marusam, which would make innbsp;Modern Irish Muir Marbh, Welsh Mor Marw, �Deadnbsp;Sea.� In the Latin of both Pliny and Solinus it looks likenbsp;an accusative feminine, but as the word mori, Irish muir,nbsp;was neuter like the Latin mare, it is probably to be treatednbsp;as accusative neuter; and the fact of Marusam endingnbsp;in am shows that we have here to do with Goidelic, asnbsp;Brythonic and Gaulish would have had on or om^ andnbsp;the Latin would have been given accordingly as Mori-Marusum. Pliny�s authority was a certain Philemon whonbsp;believed to have had tattooed 011 their persons. Should this prove tenable,nbsp;one could scarcely avoid treating Cruiihni and Prydyn as translations intonbsp;Goidelic and Brythonic of the word Piet regarded as the Latin pictus, � � painted. ��nbsp;It is needless to say that this would not help us to the meaning of Put as anbsp;word of the Pictish language to which it possibly belonged ; but the supposition here suggested as to Pretanic being merely a sort of translation of thenbsp;Latin pictus, would compel us to regard the first use of Pretanic as dating nonbsp;earlier than Caesar�s time and the spread of Latin in Northern Gaul. Thisnbsp;would simplify the question if the chronology should make it possible, whichnbsp;looks hardly probable.

t Pliny�s version runs thus, iv. 95 : Morimarusam [euni) a Cimbris vocari, hoc est, Mortuum Ma-e, inde usque ad promunturium Rubeas, ultra deindenbsp;Cronium. This we copy from Mtillenhoff�s �Deutsche Altertumskunde,��nbsp;i. 413, where the passage is discussed.

* The predilection of Goidelic for a instead of 0 as the thematic vowel is borne out by the most ancient Ogams of Britain and Ireland : thus thenbsp;genitive ending corresponding to Greek os (Latin is) is always as or a.

ROMAN BRITAIN.

appears to have lived in the last century before our Era and Philemon is supposed to have been using informationnbsp;obtained by Pytheas when he visited Britain.

With regard to these names Albion, Belerium, Ictis, and Mori-Marusam, it is probable that they were learnt in thisnbsp;country by Pytheas or some of the travellers who came herenbsp;after his time. In other words, we may treat them for whatnbsp;they are worth as evidence of the occupation of the southernnbsp;portions of Britain by a Celtican or Goidelic people at anbsp;time before the Brythons had obtained a footing on itsnbsp;shores. We have dwelt on these names at this point asnbsp;another view is sometimes put forward, that everythingnbsp;Goidelic in Britain is to be traced to invasions from Ireland,nbsp;and to a time subsequent to the second century of our era,nbsp;especially the later years of the Roman occupation andnbsp;those following the withdrawal of the Roman legions fromnbsp;the Island.^ That men from Ireland invaded Britain atnbsp;various points and at various times, and, further, that somenbsp;of them settled here, is not to be disputed. Take, fornbsp;instance, the case of the Dalriad Scots, who crossed fromnbsp;Ireland to Argyle in the fifth century, or that of the D�isinbsp;in the south-west of Wales at a still earlier date. This,nbsp;however, proves in no wise that there was not previouslynbsp;a Goidelic population in the west of the Island; it rathernbsp;favours the contrary supposition, for a native Goidelicnbsp;population might well be credited with having appealednbsp;to men of their own race and language in Ireland for aidnbsp;in their struggles with Brythonic tribes, and the responsenbsp;to such an appeal may have served as the beginning of anbsp;series of descents on the coasts of Wales and of the southnbsp;of England. To such invasions we maj^ possibly have to

' This view has been recently advocated by Professor Meyer in the � Transactions of the Hon. Society of Cymmrodorion � (64 Chancery Lane, London, 1897), 1895-6, pp. 55-86, where a number of facts illustrating the earlynbsp;intercourse between Wales and Ireland have been brought together in a verynbsp;interesting fashion,

W.P. nbsp;nbsp;nbsp;G

THE WELSH PEOPLE, (chap, hi.)

ascribe the destruction of such towns as Isca Silurum or Caerleon and Venta Silurum or Caer Went in Monmouthshire, and perhaps of Calleva or Silchester in Hampshire,nbsp;where an Ogam inscription^ testifying to the presence of anbsp;Goidel was discovered a short time ago. Nay, it is conceivable that Vortigern, whose name outside the Hengistnbsp;story is found to have been more at home in Ireland andnbsp;Brittany than in Wales, represented such an invasion, withnbsp;its influence reaching as far as Kent.

The advocates of the view to which we have referred appear to be at one with us as to the existence of a considerable Goidelic population in the west of Britain fromnbsp;the second century onwards, and also as to the influence ofnbsp;that Goidelic element on the subsequent history of thisnbsp;Island, especially that portion of it which constitutes thenbsp;Principality of Wales. The difference of view attaches tonbsp;the previous question, whence came the Goidelic elementnbsp;admitted to have been present here? Our hypothesis regardsnbsp;it as for the most part resident and as partly drawn fromnbsp;Ireland, while the other derives it wholly from Ireland.nbsp;The difficulty which we feel in estimating the respectivenbsp;merits of these hypotheses is enhanced by our lack of datanbsp;to enable us to judge of the attitude of the advocates ofnbsp;the hypothesis of the exclusive Irish origin with regard tonbsp;the question of the Aboriginal population. Nor can wenbsp;hope to understand their position till they indicate hownbsp;they suppose the Goidels of Ireland to have reached thatnbsp;country, also where and when they approximately thinknbsp;Goidelic nationality and Goidelic speech to have assumednbsp;their individuality. For our own part, we have alreadynbsp;sufficiently sketched our conjectures as to the Aboriginalnbsp;population ; and we have also indicated our conviction that

* See in the Arch^ologia, vol. liv., a paper by Mr. G. E. Fox and Mr. W. H. St. John Hope, entitled � Excavations on the Site of the Roman City atnbsp;Silchester, Hants,�� in 1893, pp. 35-9.

ROMAN BRITAIN.

Goidels and Brythons differed in speech before they left the Continent. We might probably add religion; for we understand Caesar (vi. 13) to represent Druidism as being on thenbsp;wane in Gaul, and as having originated in Britain, whithernbsp;those who wished to study it thoroughly had to resort. Butnbsp;as there is no convincing evidence to identify it with anynbsp;Brythonic tribe in this country, while there is evidence ofnbsp;its prevalence among the Goidels of M�n in the time ofnbsp;Agricola, and of its surviving in Ireland in that of Patrick,nbsp;and in the Pictland of the north in that of Columba, wenbsp;infer that it was a system evolved by the Continentalnbsp;Goidels, or rather accepted by them from the Aborigines.nbsp;When, however, the Goidels of Gaul were conquered by thenbsp;Galatic Celts, including the Belgic peoples, Druidism maynbsp;well have found it impossible to hold its own for any greatnbsp;length of time, though it may have continued to flourish innbsp;remote corners of Britain, which we take to be the realnbsp;meaning of the supposition that Britain was its nativenbsp;country.^

We now come to the question how the Goidels reached Ireland�that is to say, was it direct from the Continent ornbsp;across Britain In answer to this, we should say that thenbsp;first Celts to land in Ireland embarked probably on thenbsp;western shores of Britain ; in other words, they belonged tonbsp;a race which had conquered southern Britain from sea tonbsp;sea. In early ages the voyage from the nearest ports ofnbsp;the Continent to Ireland must have been a formidablenbsp;undertaking ; but by the time, let us say, of Caesar, it wasnbsp;probably well within the capacity of the mariners of thenbsp;Veneti and of the other tribes belonging to the Armoricnbsp;League. That in one instance at least this did take place

* Since this was written a most suggestive volume of �Nos Origines� has been published : it is the work of the veteran archaeologist, M. Bertrand, andnbsp;bears the titleof �La Religion des Gaulois, les Druides et le Druidisme �(Parir,nbsp;Leroux, 1897), pp. ix. 436, and numerous illustrations.

G 2

THE WELSH PEOPLE, (chap, hi.)

the following neglected indication is worthy of note : in the extreme north-west of Gaul�that is to say, on thenbsp;westernmost peninsula of Brittany�there was in Pytheas�snbsp;time a people called by him �Oo-Ttaiot, � Ostisei.� Anothernbsp;early form of their name appears to have been �flo-TtWcs ;nbsp;but later they were best known as �Oo-io-jaioi, the Osisminbsp;whom Catsar mentions among the allies of the Veneti.nbsp;Now a name which has all the appearance of being closelynbsp;connected with that of the Ostiaei or Ostiones is givennbsp;by Ptolemy to a people in the south of Ireland�namely, thenbsp;OvaBiaO � Usdias,� whose name in its turn is probably tonbsp;be identified with that of Ossory. In Irish this latter isnbsp;written Osraighe, and, roughly speaking, it means the countynbsp;of Kilkenny ; but the Ossorians formerly claimed againstnbsp;Munster the whole of the country from the Suir to thenbsp;Barrow, and from the mountains called Slieve Bloom innbsp;Queen�s County to the Meeting of the Three Waters nearnbsp;Waterford.^ Moreover, as Ptolemy represents the Usdiae asnbsp;reaching the coast, we should probably add to their territorynbsp;the greater part of the County Waterford, on which thenbsp;D�isi seized in the third century, together with the westernnbsp;portion of the Ossorians� countr}', north of the Suir, ofnbsp;which they got possession later.� The similarity of thenbsp;names Usdite and Ostiaei naturally leads one to supposenbsp;that some of the Ostiaei or Osismi sailed from Brittanynbsp;past the Land�s End to the coast between Youghai andnbsp;Waterford Harbour, and then gradually pushed inland,

' Ouo-Si'oi is considered the best reading, and it has been adopted in the text of Ptolemy, by C. Midler, in the Firmin Didot edition of 1883 ; and as to thenbsp;name niTTloioi see Midlenhofl�s � Deutsche Altertumskunde,� i. 373-5. If thenbsp;view suggested above should prove correct, one may propose that instead ofnbsp;correcting the ots rifiiovs of the MSS. of Strabo into 06s �Clffrt/xiovs, as Miillen-hoff does, it should rather be into ovs Ovo-ti/iIovs.

* See O�Donovan�s �Book of Rights,�� pp. I7gt; 18.

3 Ibid. pp. 49, 50; but for the whole story of the D�isi see the � Book of the Dun Cow,� fo. 53, 54, and 0�Curry�s �Manners and Customs of the Ancientnbsp;Irish,� ii. 205-8.

ROMAN BRITAIN.

taking possession of some of the best land in Ireland. Possibly this took place as late as Csesar�s Gallic War, andnbsp;one�s thoughts are naturally directed to the time when thenbsp;Osismi found their powerful allies, the Veneti, beingnbsp;crushed by him on land and sea. The archaeological discoveries of the future may perhaps supply evidence wherenbsp;we have at present only conjecture.

The south-east of Ireland seems also to have been occupied, at least in part, by settlers coming direct from the Continent. For next to the Usdiae come, according tonbsp;Ptolemy, a certain tribe of Brigantes, occupying the coast asnbsp;far as Carnsore Point, and above them he places a peoplenbsp;whom he calls the Coriondi or Coriondae, who probablynbsp;occupied a district to which the waters of Wexford Harbour and the River Slaney gave ready access. Then comenbsp;the Manapii, with their town called Manapia, and situatednbsp;somewhere near the mouth of a river called Modonnus, whichnbsp;may probably have been the Avonmore, at whose mouthnbsp;stands the town of Arklow, called in Med. Irish Inbernbsp;M�r, � the great River-mouth.� Beyond the Manapii comenbsp;the Cauci, occupying probably the north of the presentnbsp;County Wicklow, and extending, perhaps, towards thenbsp;mouth of the Liffey. Of these four tribes, the Manapiinbsp;point to the Menapii on the Lower Rhine as their mothernbsp;state, and as to the Cauci their name reminds one of that ofnbsp;the Teutonic people of the Chauchi, Chauci, or Cauchi, butnbsp;our Cauci are more likely to have been Celts�possibly Celtsnbsp;who had been under the Teutonic rule of the Chauci.' The

' In this connection it is worth while mentioning that there seems to have been a very ancient trade in amber between Britain and the coast, with its islands,nbsp;between the mouth of the Rhine and that of the Elbe. Some of the ambernbsp;found in the ancient burials south of the Thames seems to have been of thatnbsp;origin ; so one infers that there was active navigation between the country ofnbsp;the Chauci and the British Isles. On this amber question see Elton�s � Originsnbsp;of English History,� pp. 65-6. Holder cites Kiepert as believing the Caucinbsp;connected with the Lower Rhine.

THE WELSH PEOPLE, (chap, hi.)

Brigantes of Ireland were very probably of the same origin as the Brigantes of Britain; but we have no evidence tonbsp;help us to fix on their home on the Continent. We should,nbsp;perhaps, not be greatly mistaken in looking for it not farnbsp;from the territory of the Menapii; for instance, in the countrynbsp;at the mouth of the Rhine known as Insula Batavorum.nbsp;The Batavi and Caninefates whom we find in possession ofnbsp;it in the first century are believed to have been Teutonicnbsp;peoples ; but that the country had been formerly inhabitednbsp;by Celts is proved by Lugudunum {Batavorum), now Leyden,nbsp;which is as Celtic a name as that of the other Lugudunums^nbsp;in different parts of Gaul. Another Batavian town wasnbsp;known as Batavodurmn, which, in point of name, was atnbsp;least in part Celtic, and situated in the neighbourhood ofnbsp;Nymeguen, whilst a little higher up the Rhine seems tonbsp;have been a place called Burginatium, which looks like anbsp;sort of Teutonic translation of some such a Celtic namenbsp;as Brigantio, borne by a place once perhaps inhabited bynbsp;the Brigantes of whom we are in quest.

Now some, possibly, of the four tribes of this group were Brythonic or Belgic rather than Goidelic ; but we havenbsp;no means of tracing the influence of their language on thenbsp;Goidelic which became the language of Ireland. They arenbsp;also remarkable for the small scale of their territory: itnbsp;cannot have extended much beyond the limits of thenbsp;present counties of Wexford and Wicklow, and its area wasnbsp;probably far from covering them, as it presumably consisted of settlements surrounded by the Aboriginal population�that is, in case that part of the country had notnbsp;already been Goidelicised. In any case the territory ofnbsp;these Leinster tribes and that of the Usdise had been carvednbsp;probably in the first instance out of that of the Iverni,nbsp;who may accordingly be represented as an Aboriginal

* For a list of these see Holder�s � Alt-celtischer Sprachsdiatz,� s.v. ; see also his Batavoduron and Burginatio.

ROMAN BRITAIN.

population previously extending from the south-west of the Island to the mouth of the Liffey or the Boyne. Thisnbsp;would help to explain why Munster used formerly to claimnbsp;Ossory, which we have supposed to represent the Usdiae,nbsp;and why the Iverni gave their name to the whole Island�nbsp;that is to say, Ivernia, Latinised Hibernia; and similarlynbsp;in the case of the �lovepviKog ��Kcav�s, which still retains itsnbsp;name of the � Irish Sea.� When these names came intonbsp;vogue the Iverni must have been predominant in Ireland.nbsp;It is probably from the northern half of the Island thatnbsp;we have the name Scoiti, under which the first westernnbsp;invaders appear in the history of Roman Britain in thenbsp;fourth century. Before leaving Ireland we wish to mention two or three^ of the names identified in Ptolemy�snbsp;Geography. Foremost comes that of the people whom henbsp;calls OvcrXovvTioi or OvoXovvTioi,^ for the manuscripts differ.nbsp;His figures admit of our locating them around Armagh,nbsp;near which is the remarkable pre-historic fortress of Emainnbsp;Macha, now known as the Navan Fort, and we detect theirnbsp;name in the Irish Ulaid, Ultuf �the Ultonians, Ulster,� which

gt; We take them from an excellent paper on Ptolemy�s Map of Ireland, by Mr. G. H. Orpen, in the �Journal of the Royal Society of Antiquaries ofnbsp;Ireland,� 1894, pp. 115-128, with map.

quot; The editor of the Firmin Didot edition gives the preference to this reading, for the incredible reason, that the so-called Richard of Cirencester has Voluntiinbsp;or Volantii in Britain. To make OhoKoivTm fit as well as Ohlt;r\oiivTioi wenbsp;should have perhaps to spell it Ohxivrioi; but OhaKovvrm and the variantnbsp;Ov(rx6yTioi, which also occurs, would do ; however, the declension of the wordnbsp;in Irish suggests -toi rather than -riot.

� This was the accusative plural, and the dative was UUaib, contracted, no doubt, from Ulatu and Ulalaib respectively�no singular occurs : so thenbsp;genitive should have been Ulat and the nominative Ulait, but from Ultu andnbsp;Ultaib were, by false analogy probably, inferred Ulad and Ulaid. Comparenbsp;such words as ingnath or ingnad, � wonderful, a wonder,� pi. inganta, c�sad,nbsp;�a suffering or passion,� ac. plural cestu, and molad, �praise,� pi. moita andnbsp;moltha. The Irish Ulaid occurs in Welsh as Wleth: see the Book of Taliessin,nbsp;poem xiv. (Skene, i. 276, ii. 154), where Penren Wleth seems to mean somenbsp;headland called after the Ultonians or their country. The reduction of nt,nbsp;nc to tt (r), cc (r) is universal in Goidelic, and no certain instance from anbsp;previous stage has yet been discovered. Brigantes, for example, w.is Biythonic,

THE WELSH PEOPLE, (chap, hi.)

fits so well that there can be no serious doubt as to their identity. The name of Ptolemy�s �EpStvot survives in thatnbsp;of Lough Erne and of the ancient tribe of the Ernainbsp;associated by Irish legend with that water. His BoDoutVSanbsp;was undoubtedly the Boyne, and his Bipyos the Barrow, notnbsp;to mention his AaySpuva, which is probably to be correctednbsp;into ^afipmva, and to be identified with Sabrann} an old namenbsp;of the river Lee, on the banks of which, somewhere aboutnbsp;the position of Cork, stood Ivernis, which may be supposednbsp;to have been the capital of the Iverni in Ptolemy�s time.

We now at length come back to Britain and the disposition of its populations under the Romans. Here we have to deal first with the question, whom the Romans had tonbsp;contend with when they invaded the island. Caesar�snbsp;passage already mentioned (p. 5) as to the powerful Gaulnbsp;Diviciacos who ruled over Britain supplies a clue to thenbsp;answer. The statement of the Remi to Caesar referred tonbsp;a time which men then still alive remembered, and, sincenbsp;no hint as to a revolution is vouched, the probability isnbsp;that the empire of Diviciacos in this country subsistednbsp;and Uoluntii comes from a Pictish origin doubtless rather than a Celtic one.nbsp;Compare Pictish Uoret, -uorrann and -uorrn (Pro. of Antiq. Scot, xxxij. 347,nbsp;349, 372). On the other hand, not only Brythonic but also Pictish shows nonbsp;aversion to nt: take, for instance, the names of the thirty Pictish kingsnbsp;mentioned in the Pictish Chronicle as ruling over Erin and Britainnbsp;(Skene�s �Piets and Scots,� pp. 5-8); the list opens with Brute Pani,nbsp;followed by Brute Urpant, and contains others called Brude Gan^ and Brudenbsp;Urgant, Brude Cint and Brude Urcint, also such later names as Entifidichnbsp;and Custantin ( = Constantinus). Similarly Ptolemy places a tribe callednbsp;DecantcB in the northern reach of the Pictish country, while we have thenbsp;Decanti of the Arx Decantorum of the Annales Cambria, later Deganhwy ornbsp;Deganwy, near Llandudno, in North Wales : the Goidelic equivalent is foundnbsp;in the personal name which occurs in the genitive variously as DE CCE Tl,nbsp;DECHETI, and in Ogam Decceddas, Decedda, amp;c.

' This name is probably non-Celtic, and evidently identical with that of the Severn, in Welsh Hafren, from an earlier Sabrina, which our classical scholars,nbsp;so particular as to vowel quantity, are pleased to make into Sabrina. Comparenbsp;Irish salann, Welsh halen, �salt,� and Irish crann, Welshnbsp;nbsp;nbsp;nbsp;�a tree,

timber.�

ROMAN BRITAIN.

under his successors in Csesar�s time. But Diviciacos�s people were the Suessiones and the Remi; so we shouldnbsp;expect to find both of them represented in Britain, thoughnbsp;their names have not been detected. Now we know fromnbsp;a couple of ancient inscriptions^ that a favourite god of thenbsp;Remi was Camulos, whose name is the etymological equivalent of the German word himmel^ �heaven,� and may benbsp;regarded as a synonym or translation doubtless of thenbsp;god�s common Aryan name, which is represented by Zevsnbsp;in Greek, /ovis in Latin, and Dyaus in Sanskrit. This wasnbsp;the supreme god of the ancient Aryans, and the Celtsnbsp;made him their god of war; so some of them when theynbsp;settled in Britain called one of their fortresses Camulodunon,nbsp;� the town of Camulos.� This was near Colchester, in thenbsp;country of the Trinovantes, in whom we are accordinglynbsp;prepared to find the Remi we are seeking. The nextnbsp;neighbours of the Trinovantes were the Catuvellauni, innbsp;whom we probably have our insular Suessiones. At anynbsp;rate, the name of the Catuvellauni was also that which,nbsp;shortened into Catelauni or Catalauni, eventually becamenbsp;Chaalons and Chalons, the name of a well-known townnbsp;on the Marne, in a district usually assigned to the Remi,nbsp;But the fact is that the Remi and Suessiones formed a sortnbsp;of twin state, the boundary between whose lands we havenbsp;no data to enable us to draw. According to Caesar�snbsp;information, the Remi and the Suessiones regarded onenbsp;another as kinsmen: they lived under the same laws andnbsp;obeyed the same magistrates. But the Remi cultivatednbsp;the friendship of Caesar, while the Suessiones took part innbsp;various efforts made by the Gauls to throw off the Romannbsp;yoke; and when those efforts failed, the Remi came forwardnbsp;to intercede for the Suessiones and save them from ruin.-

' See the Beilin � Corpus Inscrip. Latinorum,� vi. No. 46, and HUbner�s �Exempla. Script. Epigraphicse, No. 198.�

? Cxsar, ii. 3, 12.

THE WELSH PEOPLE, (chap, hi.)

A somewhat similarly close relationship appears to have existed between their representatives in Britain�that is, ifnbsp;we are right in supposing these to have been the Trino-vantes and the Catuvellauni respectively. Thus thenbsp;Catuvellauni, under the lead of their chief Cassivellaunos,nbsp;strenuously opposed Caesar�s second invasion, while thenbsp;Trinovantes hastened to seek his protection, complainingnbsp;that Cassivellaunos had slain their king, whose son fled tonbsp;Caesar on the Continent. The feud between these kindrednbsp;peoples is perhaps to be detected also in the case of anbsp;certain prince named Dubnovellaunos, who in vain soughtnbsp;the aid of Augustus; at any rate, some of his coins seemnbsp;to identify him with the country of the Trinovantes.^

In any case the Catuvellauni and the Trinovantes between them may be regarded as the upholders of thenbsp;empire of Diviciacos, and for aught we know Cassivellaunosnbsp;may have been lineally descended from Diviciacos. Thenbsp;power and influence of the Catuvellauni is sufficientlynbsp;proved by the fact of their chief Cassivellaunos beingnbsp;entrusted with the conduct of the war against Caesar; andnbsp;Dion Cassius, speaking of the campaign of Aulus Plautius innbsp;43, represents the Catuvellauni as ruling over a part or thenbsp;whole of the people of the Dobunni on the Severn innbsp;Gloucestershire. On the other hand, the coins of Cuno-belinos their king, who died before that year, show himnbsp;occupying Camulodunon as his headquarters in the lifetimenbsp;of his father Tasciovant,*� who resided at Verlamion, now

' See Rhys�s �Celtic Britain,� pp. 27, 294.

� This name has a variety of forms on the coins ; the nominative occurs as Tasciovanus, Tasciovanius, and Tasciovans, and the genitive as Tasciovani,nbsp;Tasciiovanii, Tasciovantis, besides such abbreviations as Tasciov, Tasciav,nbsp;Tascio, Tascia, Taxcia, and Taxci; but there must have also been some such anbsp;form as Tacsivant-, for we find it represented in Welsh pedigrees by a form whichnbsp;must have been Techuant, written Tehvant, which is the explanation of the formnbsp;Teuhant in the Nennian genealogies: see �Y Cymmrodor,� ix. :74, 176,nbsp;also Tecmant for Techmant, with m for v (p. 174); it is written Tecwant innbsp;Jesus College M.S. ao. Ibid. viij. 84, and it is from Tehvant, by the easy

ROMAN BRITAIN.

Old Verulam, near St. Albans. Thus at the time when the Claudian invasion took place the people of the Catuvellauninbsp;held sway from the North Sea to the mouth of the Severn.nbsp;But to return to Tasciovant, the number of his coins seemsnbsp;to indicate that he had a long reign, terminating only innbsp;the earlier years of the Christian Era. Some of his coinsnbsp;suggest that his rule extended to Calleva of the Atrebates�nbsp;that is to say, Silchester, in the north of Hampshire. Henbsp;had two sons also whose coins are extant, namely, Cunobe-linos already mentioned, and Epaticcos, whose coins inducenbsp;one to believe that he held sway south of the Thames, innbsp;what is now SurreyBut the tribes south of the Thamesnbsp;appear to have for some time retained a certain independence under three princes named Tincommios, Epillos, andnbsp;Verica, each of whom styles himself on his coins �Sonnbsp;of Commios.� Their father may have been Commios thenbsp;Atrebat, who attempted to act as Caesar�s emissary innbsp;Britain, and who afterwards played the part of mediatornbsp;when Cassivellaunos sued for peace. The subsequent careernbsp;of this Commios in Gaul was a very chequered one. Henbsp;joined with his countrymen in various attempts to freenbsp;themselves from the yoke of Rome, and after narrowlynbsp;escaping Roman treachery he withdrew to Britain. Butnbsp;whether Tincommios and his brothers were the sons of thisnbsp;Commios or not, there is evidence that Tincommios, whonbsp;possibly ruled over the Regni, pursued a Romanising policy ;

misreading of A into �, that Geoffrey�s Tenuantius (iii. 20) arose. It is remarkable that we have in the pedigree of Rhun, son of Nwython (Cym-mrodor, ix. 176) the correct succession map. Caratauc. map. Cinbelin.nbsp;map. Teuhant, for filii Carataci filii Cunobelini Jilii Taxivanti; but thisnbsp;information which one obtains partly from the coins, above all the name ofnbsp;Tasciovant, is not to be got from any known author, whether Roman ornbsp;Greek. So we have here probably traces of a Welsh pedigree representing anbsp;genuine tradition reaching back beyond the beginning of the Christian era.

1 A coin with the letters CARA or CARAT is probably to be referred to Caratacos, the more famous brother: see Evans�s � Coins of the Ancient Britons, �nbsp;Supplement (London, 1890), p. SS3t and Plate XX. 8.

THE WELSH PEOPLE, (chap, hi.)

for Augustus having become emperor and styled himself on a coin Augustus Divi Filius, it was not long before Tincommios had coins of his own inscribed innbsp;Augustus�s Latin formula: Tin�ggt;niniius'\ Com\fni\ Fliliusl.

A passage in Tacitus�s Agricola, c. 14, is here in point. He says that certain cities were given to a certain kingnbsp;Cogidumnos, according to the received policy of Romenbsp;when she wanted tools for the enslaving of other nations,nbsp;and he adds that Cogidumnos continued faithful to thenbsp;Romans even within the historian�s own time. An inscription, dating from the time of Claudius or Vespasian andnbsp;found at Chichester, helps to localise Cogidumnos there.^nbsp;Thus we seem to have glimpses of a Romanising policynbsp;pursued among the kings of the Regni from the time ofnbsp;Tincommios to that of Tacitus. It originated probablynbsp;in fear and jealousy of the power of the Catuvellauni, andnbsp;in any case that is the key to the history of the Romannbsp;conquest in 43, for the legions seem to have met with nonbsp;serious resistance till they neared the Thames. The resistance then offered was organised by the Catuvellauni, whilenbsp;the Eceni in their rear, that is to say, in the district betweennbsp;the Trinovantes and the Wash, do not seem to have foughtnbsp;at all, for Tacitus* represents them as having entered intonbsp;alliance with Rome of their own free will. This makes itnbsp;appear all the more probable that we have the Eceni innbsp;the Cenimagni of a previous age, who head Caesar�s shortnbsp;list of tribes suing for peace. The names of the othersnbsp;were Segontiaci, Ancalites, Bibroci and Cassi, who were allnbsp;probably inhabitants of the southern side of the Thames,nbsp;and may have been forced or frightened into submission; butnbsp;it is hard to believe this of a warlike people like the Eceni,nbsp;located as they were beyond the Trinovantes and the

' See the Berlin � Corpus Inscr. Lat. � vii. No. 11 ; also Holder, s.v, � Cogidubnus.�nbsp;s See the Annals, xiv. 31,

ROMAN BRITAIN.

Catuvellauni. One is left to conclude that they, in case we are to identify them with the Cenimagni, only followednbsp;the example of the Trinovantes out of fear or jealousy ofnbsp;the aggressive policy of the Catuvellauni. How far northwards the power of the Catuvellauni extended it isnbsp;impossible to say, but it may have reached Uriconion, ornbsp;Wroxeter, on the Severn, near Shrewsbury, and probablynbsp;it took in the Coritavi or Coritani, whose country laynbsp;between the Trent and the North Sea, and contained thenbsp;towns of Lindon, now Lincoln, and Ratce, now Leicester.nbsp;It is significant as to the power of the Catuvellauni thatnbsp;Suetonius (Caligula, 44) calls their king Cunobelinos, whonbsp;died before the Claudian conquest, Britannorum Rex,nbsp;� King of the Brythons.quot;

Some further light is to be obtained on the disposition of these and neighbouring tribes from the history of thenbsp;conquest of the south of Britain. According to Dion Cassiusnbsp;a prince named Bericos, having been driven into exile bynbsp;troubles at home, sought for help from the Emperornbsp;Claudius, v'ho then made up his mind to conquer Britain.nbsp;Aulus Plautius was appointed leader of the expedition, andnbsp;the war was prosecuted with vigour for the first ten years ;nbsp;and Dion Cassius tells us that Aulus Plautius conquered twonbsp;of Cunobelinos�s sons, first Caratacos, and then his brothernbsp;Togodumnos, who had probably succeeded his father asnbsp;king. In the course of the war the Roman general�snbsp;lieutenant, Vespasian, afterwards emperor, appears to havenbsp;greatly distinguished himself, having, as it is said bynbsp;Suetonius (Vespasian, 4), reduced two of the most powerfulnbsp;tribes in the island, together w'ith more than twentytowns andnbsp;the Isle of Wight. We can only surmise that his operationsnbsp;were conducted chiefly against the Belgae and Dumnonii,nbsp;the latter of whom were probably Goidels. The Romans,nbsp;at any rate, appear to have lost little time in making theirnbsp;way to the Mendip Hills, where they had lead mines

THE WELSH PEOPLE, (chap, hi.)

worked. In the year 50 the command was taken by Ostorius Scapula, and he proceeded to deal with the nativesnbsp;with a strong hand. He took measures to hold in check allnbsp;the tribes living on his side of the Severn and the Trentnbsp;and possibly it was then the Roman town of Uriconium camenbsp;into existence, marking the limit reached by the provincenbsp;as well, perhaps, as that of the influence of the Catuvellauni.nbsp;But the determined attitude of Ostorius would seem tonbsp;have moved various widely severed tribes to take a sort ofnbsp;concerted action against him. The first to rebel werenbsp;the Eceni; and after humbling the Eceni he marched tonbsp;the coast of the other sea, namely, to the country of thenbsp;Deceangli, whose name survives in that of the deanery ofnbsp;Tegeingl, in Flintshire. The attraction of that district wasnbsp;its mineral wealth, and it is the inscription on some pigs ofnbsp;lead from there, now in the Grosvenor Museum at Chester,nbsp;that enables us approximately to fix the locality.^ From thenbsp;Deceangli, the general turned back to quell troubles causednbsp;by the Brigantes, who formed the dominant people in whatnbsp;is now the North of England. We next read of himnbsp;establishing a strong colony of veterans at Camulodunonnbsp;and undertaking to subdue the Silures, who occupied thenbsp;eastern half of the region between Cardigan Bay and thenbsp;Severn. The Silures were a warlike race, and they were atnbsp;this time led by Caratacos, one of the sons of Cunobelinos.nbsp;Caratacos had resisted the Romans from the beginning,nbsp;which meant some nine years of experience in war againstnbsp;them. When his own people, the Catuvellauni, succumbed,nbsp;he appears to have sought refuge among the Silures, to

1 See Tacitus�s Annals, xii. 31, where the passage in point has been emendated as follows by Dr. Henry Bradley ; Cundosque cis Trisantonam etnbsp;Sabrinam fluvios cohibere parat. See the �Academy,� April 28, 1883, p. 296.nbsp;Trisantona, which was probably the early form of the name of the Trent,nbsp;analyses itself into Tris-antona, and recalls Rhi-annon (for Rig-anton-), onnbsp;which see the � Academy,� Aug. 15, 1896, p. I15.

- See the � Academy,� Oct. 31, 1891, p. 390 ; see also pp. 412, 437.

ROMAN BRITAIN.

whom he must have been well known for his opposition to the Romans, not to mention that the portion of his father�snbsp;dominion entrusted to him may have been the western onenbsp;bordering on the country of the Silures. However, henbsp;shifted the war into the country of the Ordovices, who werenbsp;probably of the same Brythonic race as his own people,nbsp;while the Silures are more likely to have been Goidels.nbsp;The final battle was fought, as we gather, in the neighbourhood of the Breidin^ Hills, between Welshpool and thenbsp;English border. The legions prevailed, and Caratacos flednbsp;to the Brigantes, whose Queen Cartismandua gave him upnbsp;to the Romans to adorn a triumph in the streets of Rome ;nbsp;but the Silures continued unsubdued.

In the year 58, Nero sent here Suetonius Paulinus, who led his troops in 62 to Anglesey, and Tacitus�s account ofnbsp;the battle which they fought there is remarkable for thenbsp;Druids that figure in it. While Suetonius was reducing Monanbsp;and cutting down her sacred groves, Boudicca, queen of thenbsp;Eceni, widow of the king whose name is given as Prasutagus,nbsp;headed a determined revolt, which resulted in a terriblenbsp;slaughter of the Roman colony at Camulodunon. In time,nbsp;Suetonius supervened, and was able to avenge the Romannbsp;losses by inflicting others on a still larger scale. Nothingnbsp;worthy of note seems to have occurred till Vespasiannbsp;became emperor in 69 ; one of his generals effected thenbsp;reduction of the Brigantes in the years 69 and 70. Thenbsp;Silures were also at last conquered, and Julius Agricola,nbsp;who was sent here in 78, quickly crushed the Ordovicesnbsp;and led his troops as far as Anglesey. His subsequentnbsp;achievements in war took place mostly in the north ofnbsp;the island, where he is supposed to have drawn a line

' This word put back into its early form would probably make Bragidunon or Bragodunon, meaning possibly the � Hill-fort.� Dygen Freidin was the namenbsp;of a stream in the same neighbourhood : see Skene�s � Four Ancient Books ofnbsp;Wales,� ii. 277, and the � Myvyrian Arch.,� i. 193.

THE WELSH PEOPLE, (chap, hi.)

of fortifications from the Clyde to the Forth. One of his great victories was won at a place called Monsnbsp;Granpius or Graupius, over tribes who fought from warnbsp;chariots; and on his way back he took hostages from anbsp;people called by the otherwise unknown name of Boresti,nbsp;located somewhere between the Tay and the Forth. Thenbsp;fleet also co-operated with the land forces, and after passingnbsp;a winter at a certain Portus Truccuknsis or Trutulensis, itnbsp;performed the circumnavigation of Britain. Agricola wasnbsp;recalled in 85 or 86, but had he been allowed to go on henbsp;would probably have conquered Ireland : he had calculatednbsp;the cost, and he had an exiled chief ready to lead thenbsp;way. He was not only a great general, but also an astutenbsp;statesman, who conciliated the conquered and encouragednbsp;them to adopt Roman institutions and acquire the Latinnbsp;language.

The Emperor Hadrian, who came here in 120, appears to have put down an insurrection, and he is representednbsp;building a wall from the Solway to the Tyne. In 139, hisnbsp;successor, Antoninus Pius, sent his general Lollius Urbicusnbsp;here, and he reduced the Brigantes dwelling beyondnbsp;Hadrian�s Wall, and constructed in 143 a turf wall acrossnbsp;the country from the Clyde to the Forth. The reason givennbsp;for the war on the Brigantes was, that the latter had begunnbsp;to invade the territory of certain Roman subjects describednbsp;by Pausanias, the only writer who alludes to them, asnbsp;^ Tevovvca Molpa, � the Division or Cohort called Genunia,�nbsp;a thoroughly non-Brythonic designation, which recallsnbsp;Adamnan�s Geona CoAonf, and such tribal names d.sDdl-Riadanbsp;and Ddl-Cais, the division of Riada and Cas respectively.^nbsp;The Genunians probably occupied some part of Galloway,

* See Reeves�s Adamnan�s � Life of St. Columba,� i. 33 (p. 62), p. 92, note. One is strongly tempted to identify the terms reyowia Noipa and Geona Cohors:nbsp;all that is necessary is to suppose the latter written GJonte, that is Genoncenbsp;instead of Geonce. But what would such an identification mean geographicallynbsp;and historically ?

ROMAN BRITAIN.

possibly that assigned by Ptolemy to a people whom he called Selgovas, a name which meant hunters: they werenbsp;probably a tribe of the Aboriginal Piets, more or lessnbsp;Goidelicised, and depending on the Romans for protectionnbsp;against the aggressiveness of the Brigantic Brythons. Innbsp;162 there were troubles again in the north, and undernbsp;Commodus they became still more serious. The fortressesnbsp;beyond Hadrian�s Wall, and that Wall itself, fell into thenbsp;hands of the northern enemy, who spread devastation innbsp;the province ; it was repelled in 182. In the struggle fornbsp;the purple of empire, following the death of Commodus,nbsp;Albinus, who was in command of the forces in Britain,nbsp;crossed to the Continent, where he met with his death. Henbsp;had, in 197, succeeded in making terms with the tribesnbsp;beyond Hadrian�s Wall; but his successor found himselfnbsp;unable to keep them quiet, so he purchased peace fromnbsp;them at a great price.

It was not long, however, before they attacked the province with such determination, that the Emperor Severusnbsp;resolved to make a great expedition in the north innbsp;208. Severus penetrated, it has been supposed, as far asnbsp;the Moray Firth, and he is credited with having builtnbsp;or rebuilt a wall from the Clyde to the Forth ; in anynbsp;case, as Britain figures little in history from the time ofnbsp;Severus�s death in 211 to the reign of Carausius, his expedition must be pronounced very effective. It is to be noticednbsp;that in the time of Severus one finds the populations of thenbsp;north grouped under the two names of Caledonii andnbsp;Maeatae. By the former, we are to understand the Caledonians, or native Piets of the Highlands, while the latternbsp;name appears to have comprised a mixed people of Pietsnbsp;and Celts occupying approximately the country assignednbsp;by Ptolemy to the northern portion of the Dumnonii,nbsp;together with the tribes of the Lowlands nearest to them.nbsp;Dion Cassius (Ixxvi. 12) locates the MawTai close to the Wall

W.P. nbsp;nbsp;nbsp;H

THE WELSH PEOPLE, (chap, in.)

which divided the island in two. If, as might be expected in his time, he meant Hadrian�s Wall, the Mseatae must benbsp;supposed to have included the Brigantes dwelling beyondnbsp;that Wall and the tribes overshadowed by them, such asnbsp;were probably the Votadini,^ who occupied the coast fromnbsp;the southern Wall to the Firth of Forth. But we find itnbsp;hard to accept this, and would prefer supposing that Dionnbsp;was loosely following some previous author who meant thenbsp;northern Wall. Otherwise one gets into difficulties asnbsp;regards the subsequent history of the tribes involved, andnbsp;runs counter to the later traces of the home of the Maeatae;nbsp;for their name is found to survive beyond the morenbsp;northern Wall, to wit, in Dun-My at, � the fortress of thenbsp;Mceaice or Miatiquot; as the name is once given by Adamnan.�nbsp;Dunmyat is now the name of one of the westernmost pointsnbsp;of the Ochil Hills, over against Stirling, on the other sidenbsp;of the Forth. Other fixed points of the same kind are Maynbsp;Water, near Forteviot, and Maya Insula, well known as thenbsp;Isle of May, off the north-east coast of Fife. So it may benbsp;supposed that the country of the Maeatae comprised atnbsp;least most of the tract covered by Stirlingshire and thenbsp;whole of the Lowlands as far as the Tay, whatever may benbsp;said as to the region beyond that river.�

* The name of these people is given in most of the manuscripts of Ptolemy as �ilToSijvoi or �nraSivoi, but the Weish form, according to Nennius, wasnbsp;Guotoiiin, which should now be Gododin ; and it follows that the ancient formnbsp;should have been more nearly OuotoS�ivoi or OioraSivoi, perhaps better stillnbsp;OioToSiyol. Their country embraced the district around North Berwick, andnbsp;the headland over against Fife is alluded to in Irish literature (Skene�s � Piets andnbsp;Scots,� p. 57) as the promontory of Fothudan, which agrees, except in itsnbsp;termination, with the Guotodin of Nennius. We have a simpler form of thenbsp;same origin in the Irish personal name Fothad, and in the leading elementnbsp;of the genitive Voleparigis, Votecorigas, on the bilingual stone of Casteli D wyran,nbsp;Carmarthenshire. Seethe �Arch. Cambrensis,� 1895, pp. 303-13; 1896,nbsp;pp. 107-10, 138.

2 nbsp;nbsp;nbsp;See Reeves�s Adamnin, i. 9 (p. 36); also i. 8 (p. 33), where the spellingnbsp;is Miathi.

3 nbsp;nbsp;nbsp;See Skene�s �Piets and Scots,� pp. clxi. 423-4.

ROMAN BRITAIN.

Carausius seized the reins of government in Britain in 287, and he is described as of the most plebeian descent,nbsp;as a Menapian citizen, and as an alumnus of Batavia.nbsp;His history was that of a man who had once worked fornbsp;wages as a mariner, and by degrees made his way up tonbsp;the very responsible position of commanding the Romannbsp;fleet charged to keep the sea opposite the Belgic andnbsp;Armoric coasts clear of the Saxons and Franks whonbsp;infested it; among other distinctions he had earnednbsp;a great reputation in the war against the Bagaudae innbsp;Gaul. The headquarters of his fleet were at Bononia ornbsp;Boulogne, and he did his work with great success, butnbsp;at length he fell under the suspicion of conniving at thenbsp;doings of the pirates and of receiving a share of thenbsp;plunder. This resulted at last in an order that henbsp;should be put to death, whereupon Carausius declarednbsp;himself Caesar, and made Britain for a time independentnbsp;of Rome, and not a little prosperous. He died in 294,nbsp;assassinated by a certain Allectus, who is representednbsp;as one of his associates, and Allectus enjoyed power tillnbsp;he was slain in a battle fought in 296 with the armynbsp;of Constantins Chlorus, who then re-united Britain to thenbsp;Roman Empire.^

It will help one to understand the career of Carausius, if it be borne in mind that being a citizen of Menapianbsp;does not necessarily mean that he was a native of thenbsp;Continental Menapia: he may have been born in thenbsp;Manapian town which Ptolemy places somewherenbsp;between Wexford Haven and Avonmore, on the east coastnbsp;of Ireland; and when one comes to look into the namenbsp;Carausius, this becomes probable. For it can hardly be annbsp;accident that Carausius admits of being equated with the

* For the history of Carausius see Aurelius Victor�s Csesars (edited by Picblmayer) xxxix. 20, 39, Eutropius, ix. 21, and other authors cited by Holder,nbsp;s.v, Carausius.

H 2

�-oo THE WELSH PEOPLE, (chap, iir.)

Irish name C�-r�i} which, according to the analogy already discussed (pp. 68, 72), seems to have meant the Hound ofnbsp;the Plain or of the Field, probably of the Battlefield ; and itnbsp;was borne by a personage well known in Irish literature,nbsp;Cu-r�i is represented in Irish literature as a consummatenbsp;magician and a great warrior usually engaged in expeditionsnbsp;to Scythia and other distant lands; but he has unfortunatelynbsp;been inextricably confused with a great ancestral figure innbsp;the West of Ireland, whose name was Cu-ri, and his fortressnbsp;has accordingly been identified with Cathair Con-ri^nbsp;�Caher Conree, or the City of Cu-ri,� in Kerry. The truenbsp;position, however, of C�-r�i�s stronghold is more correctlynbsp;indicated by the circumstantial evidence of the locality ofnbsp;certain foes who made attacks on it at night: some comenbsp;from Breg, the eastern portion of Meath, roughly speaking,nbsp;and some from Sescenn Uairbedil, �Marsh of Uarbel,� somewhere on the east coast of Leinster. On the Welsh side,nbsp;poem xlii. in the Book of Taliessin purports to be the elegynbsp;of Cu-r�i, calls him Corroi, represents him holding a helm onnbsp;the Sea of the South, and connects him seemingly with Dover.^

If this view should prove correct, it is easy to understand that, belonging to a colony which probably traded with

* The first part of C�-r�i consists of cii, � hound,� the vowel of which is long; but in a compound like Cu-r�i the stress accent being on the second elementnbsp;m, the K of C� would lose a part of its length, yielding practically C�-roi ornbsp;CH-rSi. Then as to the spelling of Ca-rausius wnth a, compare Kanovio on anbsp;Roman milestone (now in the British Museum) for Conovio, now Conwy,nbsp;Anglicised Conway, in North Wales. As to the other part of the name Ca-rausius, the intervocalic s, according to rule, disappears, yielding r6i, concerningnbsp;which see Stokes�s � Urk. Sprachschatz,� p. 235, where he traces r�e, r�i,nbsp;�ebenes Feld,� to the same origin as the Latin rtis, genitive r�ris. Corroi innbsp;the Taliessin poem is derived from the Irish genitive Con-r�i.

^ See the story called Fled Bricrenn in Windisch�s � Irische Texte,�� IJp. 294-300. As to Sescenn Uairbeoil, this has been recently identifiednbsp;with the Esgeir Oervel of the Twrch Trwyth story (Oxford Mabinogion,nbsp;pp. I3S~�) by Prof. Meyer, in the �Transactions of the Cymmrodorion Society,�nbsp;1895-6, p. 73. Uarbel seems to mean the � Cold-mouth, � referring to somenbsp;gap or gully where a cold wind usually blows ; compare The Sloe, in the Islenbsp;of Man known in Manx as the Great Mouth of the Wind.

ROMAN BRITAIN.

Britain and with the Belgic coast, including the Menapian mother-state, Carausius was used from his boyhood to thenbsp;sea. If, moreover, the colony was, as we have supposed,nbsp;Belgic or Brythonic, he could presumably pass as a Brython,nbsp;but the fact of his name being Goidelic argues his beingnbsp;partly of Goidelic descent: in other words, he would seemnbsp;to have been favourably situated to become popular withnbsp;both Brythons and Goidels, and to make them considernbsp;him one of themselves, whether in Britain or on the Continent. During his time, at all events, we read of no difficulties with the tribes beyond the Roman Wall. The samenbsp;remark applies to the three years of Allectus�s rule; butnbsp;his name^ points to the North of Britain, whence also hisnbsp;troops may have been largely drawn. No sooner, however,nbsp;had Rome resumed possession of the province than thenbsp;northern tribes began to be troublesome once more, andnbsp;they are now, for the first time, spoken of as Caledonesnbsp;and �other Piets,�^ against whom Constantins Chlorusnbsp;undertook, in 296, an expedition beyond the Wall. Thenbsp;effect of the chastisement which he then inflicted on themnbsp;appears to have lasted some time.

The next serious attack on the province took place in 360, when the Piets from the north were joined by a peoplenbsp;from Ireland, figuring for the first time in history undernbsp;the name of Scotti. They were probably mixed bands ofnbsp;Goidels, Cruithni or Piets of Ireland, and Fir Ulaid or Truenbsp;Ultonians. These last had been crowded into the northeast corner of that island in consequence of the conquestnbsp;of Oriel or southern Ulster some years previously by Celts

* We have it in the name of the Perthshire town of Alyth, in an older form Aleecht, and probably also in the Welsh name Elaeth, borne, according tonbsp;Williams�s �Eminent Welshmen,� by a sixth century saint and poet, who had,nbsp;before he took to a saintly life, been a king in a district in the north of England.nbsp;See also Skene�s � Four Ancient Books,� ii. 344.

2 See Eumenius�s � Panegyricus Constantino, � c. 7 : Non dico Caledonum aliorumqm Pictorum divas etpaludes.

102 THE WELSH PEOPLE, (chap, hi.)

from the direction of Meath. The Scotti presumably crossed at first into Galloway, where we have surmisednbsp;the Genunians to have dwelt; and on this occasion thenbsp;jieople of Galloway, the descendants of the Genunians ofnbsp;a former age, take part themselves in the attack on thenbsp;province; these were probably the Atecotti representednbsp;as located between the southern and the northern Walls,nbsp;and their name would seem to have meant the Old or Ancientnbsp;Race. The Piets from the north are this time described asnbsp;consisting of Dicaledonse and Verturiones, in whom wenbsp;seem for certain to have the Caledonii and Mseatse of Dionnbsp;Cassius respectively; for the term Dicaledonae used bynbsp;Ammianus Marcellinus (xxvii. 8) seems to have meant thenbsp;inhabitants of the Highlands conceived of as forming twonbsp;Caledonias, severed by the waters of Loch Linnhe, Lochnbsp;Lochy, Loch Ness and Inverness Firth, or else as consistingnbsp;of a Lowland and a Highland region. As to Verturione.s,nbsp;that is a name which gave rise ultimately to the designation Fir Fortrenn} as it were VFi Verturionis, � thenbsp;Men of Fortrennquot; a district in which the Picto-Brythonicnbsp;people of the Lowlands had their headquarters at Forteviotnbsp;till the centre of gravity was shifted to the banks of thenbsp;Tay by Kenneth mac Alpin and his dynasty. Thus thenbsp;Verturiones seem to have been the Boresti of Tacitus andnbsp;the Maeate of Dion. Whilst these tribes were attackingnbsp;the province on one side, the Saxons were plundering it onnbsp;another, especially the south-eastern coast. In 369 Theodosius arrived and put a stop to the devastation, which hadnbsp;extended to the heart of the province, and he renewednbsp;the stations on the Wall. Add to this that the Atecotti

� It is possible that we have a survival of the nominative in Fothrev-e or Fothrif, the name of a district embracing Kinross and a part of Fife: seenbsp;Skene�s �Piets and Scots,� pp. Ixxxiv., Ixxxv., 136. This would imply thatnbsp;FothriJ stands for an earlier Forthriu, and that the old Brythonic formsnbsp;were approximately Vorthr^o, genitive Vorlhrionos. See � Pro. Soc. Antiq-Scotland,� xxxii. 396.

ROMAN BRITAIN.

were subdued and their able-bodied men drafted into the Roman army in Gaul, where St. Jerome reports^ havingnbsp;seen some of that race, which was in his time believed tonbsp;be cannibals. The province continued to be attacked fromnbsp;without by land and sea, and to be cleared from time tonbsp;time of the spoilers by the Roman soldiers, until at last thenbsp;exigencies of the empire compelled Rome to withdraw hernbsp;troops from the island altogether early in the fifth century.

A word now as to the administration. At the date to which the Notitia Dig?iitatum or the Table of Dignitiesnbsp;belongs, the military command of Roman Britain was distributed as follows:�(i) There was the Count of Britain,nbsp;Comes rei militaris Britanniariim, Co7nes Britannim'um ornbsp;Comes Britannice, who with his troops was not fixed in anynbsp;particular locality; (2) the General or Duke of Britain,nbsp;Dux BritMiniarum or Dux Britannice, who had commandnbsp;of the troops on the V/all and in the country south of itnbsp;to the Humber; and (3) the Count of the Saxon Shore,nbsp;Comes litoris Saxonici per Britannias, who had chargenbsp;of the south-east of the island, where there were, from thenbsp;fourth century, military posts placed at intervals from thenbsp;Wash to the Isle of Wight, as a defence against invasionsnbsp;by the Saxons and other Teutonic tribes. Comparativelynbsp;little is known concerning the civil administration of Britainnbsp;under the Romans, but the organisation of it, at the headnbsp;of which stood the Vicar of the Britannias, Vicm'ius Bi'itan-niarum, was practically subordinated to the military system,nbsp;owing doubtless in a great measure to the continuous attacksnbsp;to be expected from without. Roman Britain was treatednbsp;as a single province till the year 210, when Severus dividednbsp;it into two, called Lower and Upper Britain, Britannianbsp;Inferior and Bi'itannia Superior; but we have no indication as to their respective positions beyond the fact thatnbsp;Eburacum, �York,� was in Lower Britain; while Deva,

* See � Hieronymus adversus lovinianum,� ii. 7 (p. 50).

THE WELSH PEOPLE, (chap, hi.)

�� Chester,� and Isca Silurum, � Caerleon,� on the Usk, were in Upper Britain. This naturally suggests that Lowernbsp;Britain was the area drained by rivers flowing to the Northnbsp;Sea and the English Channel, and that Upper Britain wasnbsp;so situated as to be reached by travelling up the valleys ofnbsp;the rivers of Lower Britain, in other words the region beyondnbsp;the watershed and draining into the various parts of thenbsp;Irish Sea. But Mr. Haverfield is of opinion that Uppernbsp;Britain was the first portion of the island reached by thenbsp;Romans, that it comprised the whole of the south as far asnbsp;a line drawn from the Mersey to the Humber, and thatnbsp;Lower Britain lay north of that line ; such a division wouldnbsp;seem to have the advantage of fitting in better with thenbsp;military arrangements already suggested. Possibly thenbsp;explanation is, that Severus found the terms Lower andnbsp;Upper Britain already in use in his time in the sense whichnbsp;we have suggested, and that he altered their application sonbsp;far as to make Lower Britain comprehend all south of thenbsp;Mersey and the Humber, and Upper Britain all the Romannbsp;territory beyond those waters; for we are not convincednbsp;that the relative position of Lower and Upper Britain wasnbsp;the reverse.

In 297 Diocletian divided Roman Britain into four provinces called Prima, Secunda, Flavia Caesariensis, andnbsp;Maxima Caesariensis; but all that has been made outnbsp;as to their positions is that Cirencester was in Britannianbsp;Prima. In 369 a fifth province was made, called Valentia,nbsp;after the Emperor Valens, but the position of this also isnbsp;uncertain; it has been supposed to have been the districtnbsp;between the two Walls. In the second and third centurynbsp;the forces here consisted of three legions, one stationed atnbsp;York, one at Chester, and a third at Caerleon, Besidesnbsp;these there were auxiliary cohorts, with their cavalry, mostlynbsp;recruited in Germany, stationed on the Wall and at variousnbsp;points in the district between the Tyne and the Humber,

ROMAN BRITAIN.

which was largely given up to those troops. The Brigantes were not reduced to quietness till the end of the secondnbsp;century, and for another century native troops were seldomnbsp;employed in garrisons in the island : they were drafted awaynbsp;to Germany.

When the Roman legions finally departed the provincials appear to have been on the whole equal to the task ofnbsp;repelling attacks from the north; at any rate, there is nonbsp;evidence that any Caledonians or other Piets were able tonbsp;effect a single settlement south of the Clyde or the Forth.nbsp;The question of invasions from Ireland is a more difficultnbsp;one, as it cannot be severed from the other question, alreadynbsp;touched upon: What Goidelic populations had their homenbsp;in this country before the coming of the Romans, andnbsp;remained here till after the departure of the legionaries ?nbsp;But in spite of the threnody of Gildas over the ravagesnbsp;committed by Piets and Scots, the principal misfortunesnbsp;of the Brythons came from a different quarter, namely,nbsp;from the Continent; we mean the permanent conquestsnbsp;effected here by the Teutonic peoples of the Saxons, thenbsp;Angles, and the Jutes. In the continued effort to holdnbsp;their own the Brythons may naturally be expected to havenbsp;at first endeavoured to maintain the offices to which Romannbsp;administration had accustomed them. Thus they wouldnbsp;probably have somebody filling the office of Count ofnbsp;Britain, or, perhaps more likely, of that and the office ofnbsp;Emperor all in one, now that the Emperor of Rome concerned himself no more with the affairs of the island.nbsp;Welsh literature does not fail to supply us with a personagenbsp;fitted for such a position, and that is Arthur, at any ratenbsp;in so far as Arthur can be treated as a historical man andnbsp;not a myth. He exerted himself, according to Nennius, asnbsp;the Dux Bellorum of the kings of the Brittones, and hisnbsp;activity manifested itself in all parts of the country. Innbsp;Welsh story he is called Amherawdyr, which is the Latin

io6 THE WELSH PEOPLE, (chap, hi.)

word /mJgt;eraior borrowed, and also Penteyme�yr Ynys honn, �Chief of the Rulers of this Island,�^ but not king till late.

With more confidence, however, one detects a post-Roman officer filling the position of the Dux Britannics, in command of the forces on the Wall and in the adjacentnbsp;district, namely, in Cuneda the Gwledig or Ruler, to whomnbsp;reference has already been made (p, 9). His pedigreenbsp;represents him as son of Hitern i^ternus), son of Paterttnbsp;Pesrut (Paternus of the Red tunic, in reference probably tonbsp;the purple of office), son of Tacit [Tacitus)�, while one ofnbsp;his own sons was called Dunawd [Donatus), and one of hisnbsp;grandsons was Meriaun [Marianus), after whom is callednbsp;Meirionyd or Merioneth? So the family of Cunedanbsp;must have been Christian, and perhaps partly of Romannbsp;descent. The most powerful branch of it supplied Venedosnbsp;or Gwyned with kings, and the most powerful of themnbsp;appears to have been Maglocunos or Maelgwn, who wasnbsp;contemporary with Gildas, by whom he is called Insularisnbsp;Draco, meaning probably thereby the Dragon or Leadernbsp;of the Island of Britain. The explanation of the term isnbsp;presumably that the general or leader had as his ensign anbsp;dragon, rvhich had descended to him from the Dux Britannicenbsp;in Roman times. In the seventh century this dragonnbsp;figures, as heraldry teaches, as the Red Dragon of Kingnbsp;Cadwaladr, who was the last of the line of Cuneda andnbsp;Maelgwn to try to wield the power which Maelgwn enjoyed.nbsp;What that power precisely was, is only a matter of inference.nbsp;Maelgwn was King of Gwyned, but he seems also to havenbsp;exercised sway over the whole of the country from thenbsp;Severn Sea to the Firth of Forth. How he obtained thisnbsp;wider power becomes intelligible on the supposition thatnbsp;the office of Dux Britannice had been continued from

' See the Story of Kulhwch and Olwen, in the Oxford Mab., p. loj.

2 See � Y Cymmrodor,� ix. 170, 178, 1S2, and see the pedigree in the note on p. 120 below.

ROMAN BRITAIN.

Roman times, and that the Gwyned branch of the Cuneda family had been able to keep it in their own hands. Thusnbsp;however widely the east and south of the island had beennbsp;wrested from the Brythons by Teutonic tribes, there stillnbsp;was what might be called Roman Britain, though it nonbsp;longer owed allegiance to Rome, and its head wouldnbsp;naturally be hailed by Gildas as the Dragon of the Island.nbsp;It was more congenial to his style to describe him in thatnbsp;way than to call him simply Dux Britannicz. Thoughnbsp;Gildas had grave faults to find with Maelgwn, the latter wasnbsp;of his own race, and that must have counted for a greatnbsp;deal with one who hated Piets and Scots and Saxons withnbsp;the bitter hatred of an irritable saint.

Maelgwn�s son and successor, Rhun, appears to have been less able than his father, but we read of him making war asnbsp;far north as the river Forth,i probably in order to retain hisnbsp;father�s power; and it seems to have been successfully sonbsp;retained, to be lost only to the Angles after a prolongednbsp;struggle. This may be said to have begun with the winningnbsp;of the battle of Chester by ^thelfrith of Deira in the yearnbsp;616, and to have been continued later in a war between Cad-waHon, king of Gwyned, and Eadwine or Edwin, king ofnbsp;the Angles of Bernicia. Bede, in speaking of Cadwalton,nbsp;calls him oftenest Rex Brettonum, � King of the Brythons,�nbsp;but he is also found once using the term Brettonum Dux,nbsp;� general or leader of the Brythons.� Edwin triumphed fornbsp;a time over Cadwafton, and it appears from what Bede saysnbsp;that Edwin was the first of the kings of the English to havenbsp;a banner carried before him when he rode forth and a tuftnbsp;of feathers when he went on foot. This Roman fashionnbsp;was probably also that of the Dux Britannice down to Cad-wailon, from whom Edwin would seem to have adoptednbsp;it as a visible indication that he had taken the position ofnbsp;Cadwalton. It was then also presumably that was firstnbsp;1 See Aneurin Owen�s � Laws and Institutes of Wales,� i. 104, 5.

THE WELSH PEOPLE, (chap, hi.)

heard the title of Bretwalda} Ruler of � Britons,� otherwise Bryten-walda or Bryten-wealda, Breten-anwealda, and Bryten-weald, � Ruler of Britain; � for all these formsnbsp;of it are given in the manuscripts of the Saxon Chronicle,nbsp;when, in speaking of Ecgbryht, it enumerates those of hisnbsp;predecessors supposed to have wielded more power thannbsp;the others. The diversity of form which the title shows innbsp;English documents argues a certain amount of hesitationnbsp;between Dux Brittonum, which would in Welsh be Givledignbsp;Brython, and Dux BritannicB, which Welsh poetry rendersnbsp;Pry dein Wledic? Nay, one would not err perhaps in supposing that the term gwledig (earlier \^g\Lvledic, wletic) hadnbsp;something to do with the choice in Anglo-Saxon of the wordnbsp;walda or an-wealda, � a ruler or sovereign,� to translate thenbsp;title into that language. As it happens, the words are alsonbsp;cognate, for gwlad, [g^wlat, meant the government ornbsp;power of the state, while Anglo-Saxon wealdan was �to rulegt;nbsp;or wield power.�

The battle of Chester made no difference in the claims of the kings of Gwyned to be gwledigs or overlords ; and thisnbsp;is the light in which should be read the epitaph of Kingnbsp;Cadfan, put up by his son Cadwafton or by his grandsonnbsp;Cadwaladr : Catamanus rex sapientisimus opinatisimus

King Cadfan the wisest, the most

renowned of all kings.� It was inspired not so much by a spirit of random flattery, perhaps, as by the ancient pretensions of the family. The final history of the strugglenbsp;between the kings of Gwyned and the princes of the Anglesnbsp;was this: Cadwaiton returned from exile in Ireland andnbsp;was for a time triumphant in the assertion of his ancient

� See the Oxford New English Dictionary, s.v. Bretwalda, where the untenable nature of Kemble�s interpretation of � wide ruler �� is exposed, andnbsp;the equivalence of .lEthelstdn�s Brytenwalda ealles zfyses iglands with rectornbsp;totius huius Britannia insula is advanced.

� See Skene�s � Four Ancient Books of Wales,� ii. 138.

* Rhys�s � Lectures on Welsh Philology,� p. 364.

ROMAN BRITAIN.

right to the office of gwledig or overlord, Edwin having been slain by him in a great battle fought in 633 at Heth-field, somewhere near Doncaster. The Angles continuednbsp;the struggle under ..lEthelfrith�s son Oswald, and Cadwattonnbsp;fell in a battle fought with him in 635 in the neighbourhoodnbsp;of Hexham. Cadwalton�s son Cadwaladr tried for some timenbsp;to recover the position of his ancestors, but his efforts failednbsp;and his personality passed into legend as that of Cadwaladrnbsp;the Blessed. For some time afterwards the bards of thenbsp;Brythons sang of the expected return of Cadwaladr to leadnbsp;his people to victory, and to assert the ancient rights ofnbsp;his family, described in this context as Kessarogyon ornbsp;Caesarians.^

The Dux Brittonum had long been also rex and king, to wit, king of Gwyned, so the title of Dux Brittonum verynbsp;naturally passed into that of Rex Brittomim; in the pagesnbsp;of Bede one finds this latter all but uniformly preferred.nbsp;Henceforth also the domain of the Rex Brittonum, whatnbsp;was left of Roman Britain, dwindled down to the dimensionsnbsp;of Wales. There, however, the title continued in vogue:nbsp;witness the oldest version of the Annales Cambrim, compiled about the middle of the tenth century, which undernbsp;the year 754 record the death of Rotri rex Brittonum, and innbsp;950 that of Higu-el rex Brittonum? The Bruts continued fornbsp;some time to use the same phraseology;� thus under 1056nbsp;is mentioned Grufud vrenhin y brytanyeit, � Griffith, kingnbsp;of the Britons.� In 1091, with the death of Rhys, son ofnbsp;Tewdwr, the kingdom of the Britons {teyrnas y brytanyeit)nbsp;is said to have fallen. Under the year 1113 allusion isnbsp;made to the wish of certain Welshmen to renew the kingdom

' See Skene's �Foui- Ancient Books of Wales,� i. 444-46, 487-90; ii, 25-8, 211-3.

* nbsp;nbsp;nbsp;�Y Cymmrodor,� ix. 161, 169.

* nbsp;nbsp;nbsp;See Rhys and Evans�s �Bruts� (Oxford, 1890), pp. 267, 270, 296, 309,nbsp;341. 3SS. 361, 365, 368, 369, 375, 379; and � Brut y Tywysogion � (Rolls,nbsp;i860), pp. 44, S4f 124, 158. 252, 288, 306, 316, 326, 356.

no THE WELSH PEOPLE, (chap, hi.)

of the Britons {brytana�l teyrnas). Under 1135 two of the sons of Gruffud ab Kynan are represented as jointlynbsp;holding together the whole kingdom of the Britonsnbsp;{hoti deyrnas y brytanyeii). In the subsequent entries thenbsp;phraseology changes, Kymry, �-Wales, Welshmen,� beingnbsp;introduced. In 1198 Gwenwynwyn, prince of Powys,nbsp;purposed an attempt to secure for the Kymry their ancientnbsp;rank, their ancient rights and boundaries. In 1216 ILywelyn,nbsp;son of lorwerth, summoned to him, at Aberdovey, all thenbsp;princes of Wales {liott tywyssogyon kymry') to partition thenbsp;land of Wales ; and in 1220 he summoned most of the Welshnbsp;princes to him in order to join in an attack on the Anglo-Flemings of Roose and Pembroke. In 1228 the Englishnbsp;king, after making an expedition into Wales, makes peacenbsp;with ILywelyn ab lorwerth, and during the latter�s lifetimenbsp;all the princes of Wales swore allegiance to his son Davidnbsp;in 1238, at Strata Florida; lastly, when ILywelyn died innbsp;1240, he is called prince of Wales {tyivyssa�c kymry). Innbsp;1258 we read of an assembly of the princes of the countrynbsp;swearing allegiance to ILywelyn, son of Gruffud, and in 1267nbsp;we find the English king, Henry III., formally and solemnlynbsp;acknowledging the right of the prince of Wales to the .nbsp;homage of the barons who held land in Wales. Thisnbsp;ILywelyn ab Gruffud proved to be the last of the line ofnbsp;Cuneda and Maelgwn to occupy the position of prince ofnbsp;Wales, for as is well known the military successes andnbsp;shrewd policy of Edward I. achieved the substitution ofnbsp;the heir to the English crown for a native prince of thenbsp;race of Maelgwn and Cuneda. Even thus the Prince ofnbsp;Wales of the present day is historically the actual representative of the Dux Britannice of Roman Britain ; but hisnbsp;Roman Britain is Wales, and it differs in one importantnbsp;particular from the Gaulish portions of the Roman Empire,nbsp;namely, in that it has not, like them, adopted the Latinnbsp;language. The number of Latin words, however, in the

ROMAN BRITAIN.

vocabulary of the Celtic language of Wales shows that the latter began to give way to Latin; and this would havenbsp;continued to go on had not the Latin firmly rooted in thenbsp;east and south of Britain been submerged. Strange as itnbsp;may appear, had it not been for the English language, bynbsp;which the existence of Welsh is now threatened, the Welshnbsp;language would have long ago given way to a Latin idiomnbsp;resembling French.

Lastly, another glance at the map of southern Britain and the position on it of some of the Brythonic tribes willnbsp;enable us to infer the relative dates of their advent. Thusnbsp;the Britanni from the opposite coast of the Straits of Dovernbsp;probably took possession first of Cantion or Kent. Passingnbsp;by the Regni of obscure origin and overshadowed by thenbsp;woods of Anderida, we find next in order the Belgse, whonbsp;may have been preceded by the Dobunni; but these lastnbsp;may have made their way round Cornwall and sailed up thenbsp;estuary of the Severn, or they may have drifted westwardsnbsp;from the Midlands. A similar uncertainty attaches to thenbsp;Atrebates ; they may have come up the Thames, but it isnbsp;perhaps more likely that they came about the same timenbsp;with the Belgse and pushed inland from the neighbourhoodnbsp;of the Isle of Wight. We now come to a second group,nbsp;some members of which must have made for the mouthnbsp;of the Thames. The first of these were probably thenbsp;Trinovantes, who posted themselves on the coast of Essexnbsp;and the banks of the Thames as far, at any rate, as the sitenbsp;of London, which the fashionable Romans of a later daynbsp;thought they had re-named Augusta for all time.^ The

1 See Tacitus, Ann. xiv. 33, and Ammianus, xxvii. 8 ; xxviii. 3. Luckily for the historian the ease with which a superior race thinks it consigns tonbsp;oblivion a place-name current among its subjects often proves delusive, as innbsp;this instance of London ; but it is a matter of regret that no Roman inscription discovered in London or elsewhere gives the full name of Londinium,nbsp;or whatever the Latin spelling may have been : we have nothing more thannbsp;the abbreviation LON.

II2 THE WELSH PEOPLE, (chap, in.)

Catuvellauni, coming about the same time, had to proceed higher up the river before landing to conquer the Midlands.nbsp;Near the Trinovantes settled also the Eceni, who landednbsp;on the coast between the Trinovantes and the Wash; wenbsp;offer no guess as to their home on the Continent. Thenbsp;same remark applies to the Brythons beyond the Wash,nbsp;namely the Coritavi, who arrived, presumably, by the waynbsp;of the Humber, as did also, perhaps about the same time,nbsp;the Parisi, the remains of whose iron chariots impart anbsp;special interest to the archseology of the opposite districtnbsp;between the Derwent and the North Sea.

If we look westwards, our attention is challenged by the Ordovices and the Cornavii, both of obscure origin. Thenbsp;former possibly acquired their individuality in the Midlands,nbsp;and thence gradually pushed their way westwards to thenbsp;sea, leaving in the possession of unnamed Goidels what isnbsp;now the north-west corner of Wales, including Mona, ornbsp;Anglesey, where Agricola found Druids. The latter, namelynbsp;the Cornavii, seem to bear a geographical name describingnbsp;them as the inhabitants of the horn or peninsula, as thoughnbsp;they had landed between the estuaries of the Dee and thenbsp;Mersey and thence penetrated inland. The same interpretation fits the Celtic corn (Latin cornu, English horn) in thenbsp;name of Cornwall and in that of the Cornavii, in the extremenbsp;north. It need not be supposed to imply identity of race :nbsp;thus the Cornavii on the Dee were Brythons, while thenbsp;northern Cornavii were as probably a tribe of the Aborigines.nbsp;The last groups of Brythons consisted of the Brigantes andnbsp;the Votadini, that is unless we should include with themnbsp;the Parisi from the banks of the Seine. This could hardlynbsp;be correct if our conjecture (p. 86) fixing the home of thenbsp;Brigantes in the country at the mouth of the Rhine shouldnbsp;prove tenable. The position of the Brigantes in this countrynbsp;would seem to show, that they arrived comparatively latenbsp;and landed probably from the Humber or the Tees, or from

ROMAN BRITAIN.

both. Last of all came the Votadini, who took up their abode on the coast from the Tyne to the Firth of Forth,nbsp;together with the adjacent country as far, perhaps, asnbsp;Bannockburn.

This, however, does not cover the whole Brythonic area towards the west and the north, since Brythonicnbsp;speech is found to have acquired, previous to Kennethnbsp;mac Alpin�s reign, a footing among the Piets of Forteviot,nbsp;to which must be added the fact that Goidelic alsonbsp;appears, in later times, in the valley of the Tay possessed of such a hold there as to be difficult to accountnbsp;for. So we are forced to suppose that a considerablenbsp;mixture of Pictish, Goidelic and Brythonic must havenbsp;existed in the country extending from the Firth of Clydenbsp;to the banks of the Tay ; in other words, the Piets beyondnbsp;the Forth were fairly well protected by the deep mud of thenbsp;Forth 1 on one side and by the Ochil range on another, whitenbsp;their Celtic aggressors took one and the same path towardsnbsp;the Tay,namely, that passing between Stirling and Dunmyat,nbsp;and now sufficiently indicated by the line of railway fromnbsp;Stirling to Perth: it became also the route of the Romannbsp;legions, as indicated by the camp at Ardoch.

But roughly speaking, the inland region from the Firth of Clyde to the basin of the Tay is that assigned bynbsp;Ptolemy to the Dumnonii, and there were Dumnoniinbsp;also in the south-west of the island neither appear to

' It is probably the muddiest river in the kingdom, and its name Foylh may be supposed to refer to this peculiarity of its waters, if we may take thenbsp;word to be Celtic and the etymological equivalent of its Welsh name Gweryd,nbsp;w'hich would seem to be the same word as Welsh gwtryd, � soil, mould, ornbsp;earth.� Similarly, its ancient name of Bodotria seems to have its explanationnbsp;in the Welsh budr, �dirty � : Ptolemy calls it BoSepia, which does not harmonisenbsp;with iudr with its for an older � or d. Skene�s � Piets and Scots � gives Forth anbsp;dative Forciu (to be read probably Forthiu) and a genitive Forthin, pp. 10,43,

^ Holder, s.v. Dumnonii, mixes the two peoples up, and declares them to have been Brythons; but it is right to say that the article seems to havenbsp;accidentally escaped revision.

W.P. nbsp;nbsp;nbsp;I

THE WELSH PEOPLE, (chap, hi.)

have been Brythonic. In fact, the position of the two peoples so designated suggests the hypothesis that theirnbsp;countries are to be regarded as extreme portions of thenbsp;Goidelic area which had escaped conquest by the Brythons,nbsp;and that the word Dumnonii was a collective name ofnbsp;the Goidels of Britain when the Brythons arrived. Thenbsp;adjective discloses the stem Dmnnon, which, treated in anbsp;way not unusual with Irish names, would yield a nominativenbsp;Domnu, genitive Domnann; and we are thus led back bynbsp;easy steps to the Fir Domnann, that is, Viri Dumnonis, ornbsp;Domnu�s Men, already mentioned. We infer that onenbsp;of the Fir Domnann�s landing places in Ireland was thenbsp;river-mouth known as Malahide River, between Howth andnbsp;Balbriggan, in County Dublin ; for in Med. Irish it is callednbsp;Inber Domnann, or the Domnu river-mouth, and from thatnbsp;point the Dronga Domnand} or the Multitudes of Domnu,nbsp;proceeded to the conquest of the fertile soil of Meath andnbsp;adjacent districts. This we should have to regard as annbsp;attack on Ireland in front, but we are reminded that shenbsp;was also assailed from behind, so to say: witness such anbsp;place-name as Irrus Domnann, � the lorrus of Domnu,� nownbsp;the barony of Erris in the north-west of the county of Mayo,nbsp;and witness also the Irish stories of early invasions of thenbsp;north of Connaught from the sea. The explanation isnbsp;probably that some of the Dumnonii, from their home nearnbsp;the Firth of Clyde, sailed round the north of Ireland, andnbsp;landed in the nearest part of Connaught: hence the Firnbsp;Domnann of some of the Irish legends. Nor is this all, fornbsp;they may have coasted further southwards ; and this maynbsp;possibly be the key to the legend which represents Scota, thenbsp;eponym of the Irish Scots of the Milesian group, as buried innbsp;Keriy^, where her grave and that of one of her companions arenbsp;pointed out in the barony of Troughanacmy, in that county.�

� See 0�Curry�s MS, Materials, p. 485, and Stokes�s �Patrick,� p. 34.

- See 0�Donovan�s notes to the Four Masters under the year A.M. 3500.

ROMAN BRITAIN.

However that may be, the Fir Domnann on this side cf the Irish Sea have left the name of the goddess Domnu^nbsp;to the county of Devon ; and in the North, likewise asnbsp;Devon, it has become the name of a river rising within theirnbsp;natural boundary in the Ochil Hills. Even should thesenbsp;conjectures prove tenable, the tribes in the west of Romannbsp;Britain must be pronounced hard to classify, while on thenbsp;eastern side, in spite of the usual scantiness of the data,nbsp;there is little room for error in this respect. There thenbsp;Brythonic settlements were continuous from Dover to thenbsp;Forth, as it has been comprehensively put in the Duannbsp;Albanach, a historic poem concerning Alba, or Scotland,nbsp;which is surmised to have been written in Ireland in thenbsp;eleventh century and to represent the ideas of a still earliernbsp;time. The Duan begins with the Trojan story, and represents Albanus and Brutus, treated as eponymi of Albanbsp;and the territory of the Brittones respectively, takingnbsp;possession of Britain. The third stanza runs thus�;�

This last doubtless meant, as already suggested, a promontory in the country of the Guotodin, somewhere

* nbsp;nbsp;nbsp;In the transition from Domnu or Domnann to Devon it is to benbsp;remembered that Welsh has made mn into vn and Irish into wn. In thenbsp;Anglo-Saxon Chronicle (with f =v) Devonshire is sometimes writtennbsp;Def\e]nasdre and sometimes Def\e'\}iansc{re. See Thorpe�s Rolls ed., i.,nbsp;120, I2I, 146, 147, 166, 167, 246, 247.

* nbsp;nbsp;nbsp;For the text see the Irish Nennius, pp. 272-3; and Skene�s �Piets andnbsp;Scots,� p. 57�the original MS. appears to have been lost (Ibid., pp. xxxvi-iii).nbsp;We have substituted loth for Icht, which makes no sense here, as Muir n-Ichtnbsp;would be the English Channel. As to loth and lodeo, see Rhys�s Rhind Lecturesnbsp;in the � Scottish Review� for 1891, p. loi, where perhaps it might be more correct to say that merin was obtained by analysing into tra merin the Latinnbsp;transmarinus; as, for instance, in Gildas�s duabus primum gentibus trans-marinis vehementer savis, Scotorum a circiom Pictorum ah aquilontnbsp;(Hist. � 14).

I 2

ltd THE WELSH PEOPLE, (chap, hi.) near North Berwick; and the whole of the island up to thatnbsp;point means all the country from the Straits of Dover tonbsp;the Firth of Forth, which is here termed Muirn-Ioth, or Seanbsp;of loth. This is undoubtedly to be identified with the waternbsp;called Merin lodeo in the Book of Aneurin, and lodeonbsp;further equates letter for letter with Nennius�s name of anbsp;town of ludeu, which Bede calls Urbs Giudi} In Scotchnbsp;history the Firth was well known as Scottewatre and Scottisnbsp;See. Lastly, as will have been gathered from our previousnbsp;remarks, the fact of giving the name Alba to Britain,nbsp;when referring expressly to the southern half of the Island,nbsp;from the English Channel to the Firth of Forth, argues anbsp;very respectable antiquity for the tradition set forth in thenbsp;poem.

' As references here may be further mentioned�for Merin lodeo, Skene�s ��Bcur Ancient Books,�ii. 103; Thomas Stephens�s �Gododin� (London, 1888),nbsp;pp. 348-9. For ludeu, see Rhys�s �Celtic Britain,� pp. 134, 152, 271,nbsp;San-Martes �Nennius und Gildas,��65 (p. 74); and for Bede�s Giudi ornbsp;ludi, see Plummer�s Bede�s � Hist. Eccles.� i. 12 (vol. i., p. 25). Here shouldnbsp;be added Muir n-Giudan, quoted (from the book of Lecan) in Reeve�snbsp;Culdees, p. 124, and to be explained probably as having ni modified intonbsp;ngi (rather than influenced by Bede�s Giudi). This sort of change is commonnbsp;enough, for example, in Manx Gaelic; see Rhys�s quot; Manx Phonology,� pp.nbsp;135-6. But more interesting philologically is the identity of the termination ofnbsp;lodeo and ludeu with the ew of such names as Frobbaccennew in .he Aboynenbsp;Ogam: see the �' Proceedings of the Society of Antiquaries of Scotland,�nbsp;xxxii. 396 ; also the suggestion in the note, at p. 102 above, as to the namenbsp;' Fothrif.

CHAPTER IV.

EARLY HISTORY OF THE CYMRY.

In this and the following four chapters we propose to deal briefly with the history of the Cymric nation,^ to givenbsp;some account of the laws and customs of the people duringnbsp;the time of their independence, and to trace the steps bynbsp;which Wales became politically assimilated to England.nbsp;We do not affect to write a history of Wales and thenbsp;Marches, for that is in our opinion a task that cannot benbsp;successfully performed with the aid only of the materialsnbsp;at present at our command.

Many of the sources of information as to the middle and later periods of the Welsh story which have survived to ournbsp;day are to be found at the Record Office and elsewhere,nbsp;but have not been published or even properly examined ;nbsp;and even the well-known authorities have for the mostnbsp;part been only very indifferently edited and printed.^nbsp;Under these circumstances any work dealing with thenbsp;history of Wales must be looked on as merely tentative,

* It may be well to state here that Cymru means the land of the Cyrory, i.e., Wales ; and that Cytnry means the Welsh people. Originally, of course, thenbsp;latter term only included the men of the dominant tribes or clans, and notnbsp;classes or persons subject to them. For the meaning and origin of the termnbsp;see p. 26 above.

� Progress is being made. See the � Public Records relating to Wales,quot; by R. Arthur Roberts, Barrister-at-Law, in � Y Cymmrodor,� x., p. 157 ; and thenbsp;� � Ruthin Court Rolls � (in the Cymmrodorion Record Series), edited, with translation, notes, etc., by the same author (Chas. J. Clark, Lond., 1893), affordsnbsp;a good model for the treatment of the legal materials at the Record Office.

ii8 THE WELSH PEOPLE, (chap, iv.)

and we only present what follows as matter likely to be useful and suggestive to the student of things Welsh.

Most people have either forgotten or never observed that it was only in the century now ending that Wales wasnbsp;completely assimilated to England. Before the Normannbsp;Conquest we may truthfully say that Wales, though itsnbsp;rulers were in some sort of subjection to the kingsnbsp;of England, formed no part of the English realm, andnbsp;it is only by gradual steps that it has been absorbednbsp;into the English body politic. Our principal aim isnbsp;to point out the chief stages in the process by whichnbsp;the present constitutional position has been broughtnbsp;about.

The history of Wales, as distinct from other portions of the Island, commences after the departure of the Romans.nbsp;The scanty and obscure character of the evidence relatingnbsp;to the fifth and sixth centuries does not enable us to speaknbsp;with any confidence as to the commencement of thenbsp;national life of the Cymry. What is now Wales appearsnbsp;to have been during the time of the Roman occupationnbsp;part of the territory extending, roughly speaking, from thenbsp;Bristol Channel to the Clyde and the Forth, under thenbsp;charge of the military official called Dux Britanniarumnbsp;The word Cymro means, according to the best philologicalnbsp;authorities, � compatriot,�^ as we have seen, and it was innbsp;.the contests of Celtic tribes with Teutonic immigrants thatnbsp;it became a national name. It seems perfectly clear thatnbsp;for something over 200 years after the Roman occupationnbsp;had ceased the western part of the Island, from the Bristolnbsp;Channel to the Solway Firth and the Clyde, as well as thenbsp;south-western peninsula, were in the possession of tribesnbsp;who may, subject to what has been said above,*^ benbsp;described as � Celtic,� and who succeeded in maintaining

' Supra, p. 26.

2 Supra, pp. 34-5.

EARLY HISTORY OF THE CYMRY. 119

their predominance in that part of the Island till a few years after the middle of the seventh century. The evidence ofnbsp;the Welsh laws referred to below tends to show that thenbsp;tribal system therein disclosed was similar in its main andnbsp;fundamental particulars to a stage of society through whichnbsp;other Indo-European races have passed. It seems clear,nbsp;too, that these tribes were bound together in some loosenbsp;form of confederation, and that from the time they recognised the term Cymry they looked upon themselves collectively as one nation. They appear to have acknowledgednbsp;the over-lordship or leadership of a king or ruler, who wasnbsp;called the �gwledig,� and whose office or dignity wasnbsp;sooner or later known as the � Crown of Britain.� Thenbsp;authority of the gwledig appears to have been partly basednbsp;on his claim to be the successor of the Roman officer callednbsp;the Dux Britanniarum, and partly on earlier tribal notionsnbsp;of political and military organisation.^

In time the territory over which the confederation spread came to be called Cymru, and the predominant languagenbsp;Cymraeg. The earliest ruler of the Cymry and of Cymrunbsp;of whom there is distinct evidence is Cuneda, whose namenbsp;often occurs in Welsh literature. In an elegy in the �Booknbsp;of Taliesin� he is said to be a man from Coelin, by whichnbsp;was apparently meant the district since called Kyle innbsp;Ayrshire. In Nennius�� Historia Britonum � there occursnbsp;the following passage :�� The great King Mailcun reignednbsp;among the Britons in the district of Guenedota because hisnbsp;great-great-grandfather Cuneda with his twelve sons hadnbsp;come before from the left-hand (or northern) part, i.e., fromnbsp;the country which is called Manau Guotodin, 146 yearsnbsp;before Mailcun reigned, and expelled the Scots with muchnbsp;slaughter from those countries, and they never again

1 One of the ancestors of CnneiJa is called Padam Pesrud (literally, Paternus of the red tunic). See Rhys�s �Celtic Britain,� 2nd ed., p. 118.nbsp;See the pedigree printed in note 2 on p. 138 below.

120 THE WELSH PEOPLE, (chap, iv.)

returned.� 1 Guenedota here is obviously Gwyned (speaking broadly, North Wales); Manau Guotodin is evidently the Manaw of the Gododin of Welsh poems, and appearsnbsp;to have been a district in Scotland situate somewhere southnbsp;of the Forth, but we have no means of determining itsnbsp;boundaries.^ If the conjecture is true, Manaw was south ofnbsp;the northern Roman Wall, in the province of Valentia. Thenbsp;entry in Nennius is confirmed by Welsh tradition andnbsp;by many items of archaeological, philological, and literarynbsp;evidence So that we may take it that the Cymric kingdom was founded upon conquest, and that the aspect whichnbsp;society in what is now Wales presented in the centuriesnbsp;to which we are referring was that of an aggregate ofnbsp;Brythonic clans forming a tribal aristocracy superimposednbsp;upon Goidelic tribes, partly Celtic and Aryan in originnbsp;and partly Aboriginal, who had before occupied the land.nbsp;The conclusion thus arrived at as to the early structure ofnbsp;Welsh society is borne out by the Welsh laws and customsnbsp;of a later time.

It is unnecessary to trace in detail, even were it possible for us to do so with accuracy, the steps by which

* The date of Cuneda�s occupation of North Wales cannot be exactly determined ; but it prob.ab!y took place early in the fifth century, and verynbsp;near to the departure of the Romans. The passage cited does not give thenbsp;date of Maelgwn�s reigning. The � Annales Cambriae� record his death asnbsp;taking place in 547. He was a contemporary of Gildas� (see � Gild. Epist.�nbsp;s. 33). The most probable dates for the birth and death of the latter arenbsp;516 and 570 respectively (Smith�s �Diet. Christ. Biog.,� s.k. Gildas). Thenbsp;� Ann. Cam.� assign his birth to 516, and the � Annales Tigernachi � his deathnbsp;to 570.

2 In the genealogies annexed to the �Annales Cambriae� (as to which see below, p. 132), the number of Cuneda�s sons is put at nine. The entry (whichnbsp;deserves the notice of the student) is as follows : � [Hjec sunt nomina filiorumnbsp;Cuneda quorum numeras erat. ix. Typipaun primogenitus qui mortuus innbsp;regione que vocatur manau guodotin. et non uenit hue cum fratribus suis. prenbsp;[dictis] meriaun. filius ejus. divisit possessiones inter fratres. suos. ii. Osmailnbsp;hi. rumaun. iiii. dunaut. v. Ceretic. vi. abloyc. vii. enniaun. girt. viii. docmail,nbsp;ix. etern.� See �Y Cymmrodor,� ix., p. 182, and below, p. 138,

EARLY HISTORY OF THE CYMRY. 121

invading Teutonic tribes advanced upon the western half of the Island, and by slow steps broke up the Cymricnbsp;federation. Two well-ascertained events mark the process.nbsp;By the loss of the battle of Deorham in 577, the Cymrynbsp;of what is now Wales were severed from the Celtic tribesnbsp;of the south-western peninsula, and afterwards, as a resultnbsp;of the battle of Chester in 616,^ the Cymry of Wales werenbsp;also cut off from their northern allies. The Cymry werenbsp;thus enclosed by Teutonic kingdoms within that part of thenbsp;west of the island which subsequently was called Cymru bynbsp;the inhabitants themselves and Wales by the conqueringnbsp;Saxon.

Notwithstanding these disastrous battles, the Cymry proper maintained a vigorous struggle with very varyingnbsp;fortunes against the Saxon or English kingdoms. But thenbsp;result of continuous warfare, though it did not bring aboutnbsp;incorporation with the English kingdom till long after, wasnbsp;to create a state of complete disorganisation, from a militarynbsp;and a political point of view; and by the defeat ofnbsp;Cadwaladr, shortly after the middle of the seventh century,nbsp;the Cymric kingdom in the older sense came to a melancholy end. This Cadwaladr is deemed by Welsh traditionnbsp;to be the last king of the Cymry who wore the � Crownnbsp;of Britain�; and that the result of the conflict of centuriesnbsp;was adverse to the Cymric nation is admitted by the briefnbsp;but graphic entry in �Brut y Tywysogion,� which says :nbsp;�Cadwaladr died at Rome as Merdyn had previouslynbsp;prophesied to Vortigern of repulsive lips, and thenceforthnbsp;the Britons lost the crown of the kingdom and the Saxonsnbsp;gained it.��

1 The date is uncertain. The Annals of Tighernach (see O�Connor�s � Scriptores rerum hibernicarum � and �Ann. Cam. �) put the battle under 613.nbsp;But the true date seems 616. See Plummer�s � Bsedse Opera Historica,� ii.,nbsp;pp. 76, 77. Tighernach antedates the battle of Daegsastan by three years,nbsp;and probably does the same in regard to the battle of Chester.

� See Murray, �Eng. Dict.,�r.r/. �Bretwalda.�

THE WELSH PEOPLE, (chap, iv.)

With this completion of a series of events the history of Wales in its limited and modern sense commences.

The subsequent history of Wales may be divided into the following periods :�

First.�From the death of Cadwaladr to the Norman Conquest of England.

Second.�From that Norman Conquest to the conquest and settlement of North Wales by Edward I.

Third.�From the settlement of Wales by Edward down to the incorporation of Wales into the English organisation, in the reign of Henry VIII.

Fourth.�From the time of Henry VIH. onward.

CHAPTER V.

HISTORY OF WALES FROM CADWALADR TO THE NORMAN CONQUEST.

(A.D. 664�A.D. 1066.)

Little is known of the history of Wales from the death of Cadwaladr to the death of Gruffyd ab ILewelyn.nbsp;Literary tradition has preserved the names and a baldnbsp;account of the deeds (chiefly inconsiderable battles) of anbsp;line of kings or princes, some of whom are representednbsp;as kings of all Cymru or all the Britons; but the personsnbsp;it hands down to us are for the most part as shadowy asnbsp;the ghosts of Banquo�s issue. The account is colourless,nbsp;and the men it brings to our notice in this period havenbsp;hardly more living interest than the names in a genealogicalnbsp;tree. No relation of the events that happened in Walesnbsp;during this time can be lively or dramatic unless one basesnbsp;it more on plausible efforts of imagination than on crediblenbsp;evidence.^ But though this is the case, in order to understand the subsequent history, it is necessary to see whatnbsp;trustworthy authority has to say about this period, andnbsp;especially to discover as well as we can what were thenbsp;chief political divisions of Cymric territory, or, to put thenbsp;matter perhaps more accurately, how Cymric land wasnbsp;apportioned among the leading royal or princely families.

* We have no assistance from bardic or poetic literature for the period from the sixth century down to about loSo, whenMeilirlameutedTrahaearn (defeatednbsp;and slain by Grufifytf ab Kynan). Stephens�s �Literature of the Kymry�nbsp;(and ed.), pp. lO, li.

THE WELSH PEOPLE, (chap, v.)

There is a considerable number of works dealing with the history of Wales, or with the history of Britain from anbsp;specially Welsh point of view. Passing over Geoffrey ofnbsp;Monmouth�s 1 work, which cannot be treated as seriousnbsp;history, and for the moment Caradog of ILancarvan�snbsp;�Historie of Cambria,� we may refer to the ambitiousnbsp;�Cambria Triumphans � of Enderbie as the principalnbsp;example of an older type.* This writer carries back thenbsp;Cymric story to Troy, and thence to the Tower of Babel.nbsp;His fundamental conception is that all history may benbsp;reduced to a system of events radiating from Troy as anbsp;centre. For the early period dealing with this islandnbsp;he relies on Geoffrey of Monmouth, and for the events ofnbsp;the later period on Caradog of ILancarvan. It is only innbsp;regard to the �modern estate� that the book is of any valuenbsp;to the student of history.

But besides works of this class there are other histories which discard older theories, though they are not adequatelynbsp;critical, of which Warrington�s and Jane Williams�^ are thenbsp;best. Both are in the main founded on a sixteenth centurynbsp;compilation�� The Historie of Cambria, a part of the mostnbsp;famous ylande of Britaine, written in the British language

' For a recent account of Geoffrey and the character and value of his work, see Morley�s �English Writers,� vol. iii., pp. 44-57. See also Professor W.nbsp;Lewis Jones� paper on Geoffrey in � The Transactions of the Hon. Soc. ofnbsp;Cymmrodorion,� session 1898-9, p. 52.

quot; �Cambria Triumphans, or Britain in its perfect lustre, showing the origin and antiquity of the illustrious nation, the succession of their kings and princes,nbsp;the description of the countrey, the history of the modern estate, etc., etc.,�nbsp;by P. Enderbie; folio (Lond., 1661); reprinted 1810. See also Lewis�nbsp;�History of Great Britain, etc., to which is added the Breviary of Britaynenbsp;by HumfreyLwyd, and lately Englished by Thomas Twine,�folio(Lond., 1729I.

^ � History of Wales in nine books, with an Appendix,� by the Rev. William Warrington ; 4to (Lond., 1786).nbsp;nbsp;nbsp;nbsp;� History of Wales,� by Jane Williams,

8vo (Lond., 1869). Reference should also be made to the � History of Wales,� by John Jones (barrister-at-law), Lond., 1824; and to � Hanes Cymru,� by thenbsp;Rev. Thomas Price, commonly called by his bardic name, CarnAuanawc (1842).nbsp;See also his � Literary Remains � (Landovery, 1854-5). See also O. M. Edwards�nbsp;� Hanes Cymru,� part i. (1895), an excellent text-book for Welsh students.

CADWALADR TO NORMAN CONQUEST. 125

about two hundred years past; translated into English by H. Lloyd/ gentleman, corrected, augmented, and continuednbsp;out of records and best approved authors, by David Powelnbsp;Doctor in Divinity.� ^ Caradog of ILancarvan,� the friendnbsp;and contemporary of Geoffrey of Monmouth, was one ofnbsp;the band of men of letters who gathered around Robert,nbsp;Earl of Gloucester, the illegitimate son of Henry I. The datenbsp;of his birth is unknown, but it is supposed he died in 1147.nbsp;It was to him that Geoffrey left the history of the kingsnbsp;who succeeded the Ivor and Ini, who had � fiercely attackednbsp;the nation of the Angles� but to little purpose, just as henbsp;committed the kings of the Saxons to William of Malmesbury and Henry of Huntingdon.^ That Caradog wrote anbsp;chronicle is clearly proved, but in its original form it is notnbsp;extant. Professor Tout thinks (with the probability of thenbsp;case on his side) that it was written in Latin.� Accordingnbsp;*to the address to the reader given by Powel, Caradog collected the successions and acts of the British princes afternbsp;Cadwaladr to 1156 ; several copies of the collection werenbsp;kept in the abbeys of Conway and Strata Florida, whichnbsp;were �yearly augmented as things fell out,� the twonbsp;abbeys comparing the entries every third year. Powelnbsp;says that the entries were continued to the year 1270, and

* Humfrey ILwyd (physician and antiquary) was born in 1527, and died in 1568. The MS. of his translation of Caradog�s translation is preserved in thenbsp;British Museum (Cotton MS. �Caligula,�A. vi.),iv. �Diet. Nat.Biog.,� s. nom.

^ Small quarto, London, 1584; 2nd ed. (Lond., l8it). See also Wynne�s � improved edition� (Lond., 1697); 2nd ed. (Lond., 1774); 3rd ed. (Merthyr,nbsp;1812); 4th ed. (Shrewsbury, 1832). The edition of 1811 is the only exactnbsp;reproduction of Bowel�s work. David Powel was born in 1552 (?), and diednbsp;in 1598. He was vicar of Ruabon and rector of Lanfyitin. The livingnbsp;of the latter parish he exchanged afterwards for Meifod {vide Diet. Nat.nbsp;Biog., sub nom.). He is honourably mentioned in �Strype�s Annals,�ii. 472-3nbsp;(ed. 1824).

� See Diet. Nat. Biog., sub. nom.; also Morley�s � English Writers,� vol. iii., pp. 95. 96. 97-

lt; See Geoffrey�s '� British History,� book xii., ch. 19 and ch. 20.

� Diet. Nat. Vgt;\og., sub nom. �Caradog.�

THE WELSH PEOPLE, (chap. V.)

that transcripts of the work were made by divers, and that over a hundred copies were extant in Humfrey Lwyd�snbsp;time. Lwyd translated the work into English, and a copynbsp;of the translation was bought by Sir Henry Sidney, President of the Court of the Marches, and he, desiring itsnbsp;publication, entrusted the work to Dr. Powel. The editornbsp;collated the copy with three versions of the Welsh work ; henbsp;added matter from other chronicles showing the additionsnbsp;by a change of type, prefixed a description of Wales bynbsp;Sir John Prise, and added brief accounts of the Princesnbsp;of Wales after the Edwardian Conquest. In this form thenbsp;work was published in 1584, under the title .set forth above.

In our judgment the statements made in a work thus compiled, and published so late as 1584, cannot be reliednbsp;on unless confirmed by the Welsh chronicles, to which wenbsp;refer below, or by the authorities accepted by competentnbsp;students of English history as trustworthy�at any rate,nbsp;far as the pre-Norman period is concerned. At the samenbsp;time we cannot deny to the work considerable value, andnbsp;assuming that the main text is down to the middle of thenbsp;twelfth century the work of Caradog, we may look uponnbsp;that part of the history as representing the Welsh traditional view of the general course of Welsh affairs at anbsp;time when the memory of many of the events was comparatively recent.

But really for the period we are now dealing with, the principal Welsh authorities which are entitled to credencenbsp;are �Brut y Tywysogion� and the �Annales Cambriae.�!

' �Brut y Tywysogion� (i.e., history of the princes) and �Annales Cambria� are the names given to two sets of chronicles which specially record affairsnbsp;concerning Wales, and which in MS. seem to have been produced as anbsp;whole within the Cymric limits, though some of the entries in the � Annalesnbsp;Cambrim � appear to have been written in Ireland, or at any rate to have beennbsp;of Irish origin. The former set of MSS. is in Welsh, and the latter in Latin.nbsp;The best critical account of the origin, the date, and the value of the MSS. isnbsp;to be found in Mr. Egerton Phillimore�s able paper entitled �The Publication

CADWALADR TO NORMAN CONQUEST. 127

From these sources, supplemented by the authorities relied on for the early history of the island by competent Englishnbsp;writers, the story of these pre-Norman Welsh kings andnbsp;princes must (if it be possible to do so at all) be constructed.

It is clear from the entry in the Brut that we have quoted above and from other sources that the death ofnbsp;Cadwaladr was regarded by the Cymry as an event ofnbsp;great importance, but as to its exact date we have nonbsp;certain evidence. The Brut puts it as taking place in 681,nbsp;but the writer uses language which shows that for somenbsp;reason he confounded Cadwaladr with Ceadwalla, king ofnbsp;Wessex, who did die in that year. If from the few datanbsp;we have to rely on the matter is traced out there can benbsp;no doubt that the year 681 is too late, and that in all

of Welsh Historical Records � in � Y Cymmrodor,� vol. xi., p. 133 (1892). The earliest known version of the � Annales Cambrix � is printed in � Y Cymmrodor,� vol. ix., pp. 152�169, under Mr. E. Phillimore�s editorship. Thenbsp;portion of the � Annales � dealing with the events up to the Norman Conquest isnbsp;also printed in � Monumenta Historica Britannica,� vol. i. {1848), under thenbsp;editorship of Petrie (really under that of Aneurin Owen, the editor of � Thenbsp;Ancient Laws and Institutes of Wales�). The �Annales Cambrias � werenbsp;also published in the Rolls series in i860 (edited by Ab Ithel). The �Brutnbsp;y Tywysogion � up to 1066 is also printed in � Monumenta Historica Britannica,� and the whole of it was published (ed. Ab Ithel) in the Rolls series innbsp;i860. The versions of the �Annales� and the �Brut� in the Rolls seriesnbsp;are subjected to severe but just criticism by Mr. Phillimore. In 1890, however, the text of the �Brut,� as transcribed in the �Red Book of Hergest,�nbsp;was published in the series of Welsh texts produced at Oxford under thenbsp;editorship of Professor J. Rhys and Mr. Gwenogvryn Evans (Clarendon Press).nbsp;For further Information as to the MSS. see Mr, Evans� preface to the � Rednbsp;Book of Hergest,� vol. ii., and Aneurin Owen�s posthumously printed introduction to the �Gwentian Chronicle� in �Arch. Cambr. � (1864). Thenbsp;so-called �Gwentian� or � Aberpergwm Brut� is printed in � Myv. Arch.�nbsp;(vol. ii., pp. 468�582), and a copy made by Aneurin Owen from the � Myv.nbsp;Arch.� is printed in �Arch. Camb � for 1864. The date of its compilationnbsp;was not earlier than 1550, and it has not the authority of the genuine andnbsp;older � Brut � (see Egerton Phillimore�s paper cited above, � Y Cymmrodor,�nbsp;xi. 163-168). As to the genealogies appended to �Annales Cambrix � innbsp;Plarl. MS. 3,859, and printed in �Y Cymmrodor,� xi., p, 169 et seq., seenbsp;pp. 132, 138, below.

THE WELSH PEOPLE, (chap, v.)

probability it was in or very near to 664 that Cadwaladr died.

If we assume that this date is correct, the period now under consideration comprises 402 years, and the scantinessnbsp;of the direct evidence as to what took place may benbsp;estimated from the fact that there are in the Brut onlynbsp;about 200 entries up to the Norman Conquest, and onlynbsp;431 for the 180 years that elapsed from the time thenbsp;Britons lost the crown of Britain to the accession of Rhodrinbsp;Mawr in 844, and that these entries are always brief andnbsp;often obscure. We have, however, some incidental helpnbsp;for the construction of this direct evidence from othernbsp;sources, amongst which the Laws of Howel Da and othernbsp;legal treatises must be given the first place,� for from themnbsp;we can discover with reasonable certainty the structure ofnbsp;Welsh society in these times from a legal and economicnbsp;point of view; and the � Liber Landavensls � � properly and

� Forty-one in �Annales Cambria: � (Phillimore, ubi supra).

* For an account of the legal treatises see below, p. 176 et seq.

^ Liber Landavensis � is the name given to a work supposed to have been compiled by Galfrid (Jeffrey or Geoffrey), the brother of Urban, the lastnbsp;Bishop of Llandaff mentioned in it. This Galfrid is identified by Mr.nbsp;Gwenogvryn Evans with Geoffrey of Monmouth (see preface to the Oxfordnbsp;text mentioned below). It is a chartulary or collection of documents concerning the Bishops of Llandaff, the endowments of the Church, and eventsnbsp;connected with the history of the diocese. There are several MSS. of thenbsp;work. The oldest and, as it seems, the original one is the Gwysaney MS.nbsp;(belonging to Mr. P. B. Davies-Cooke, of Gwysaney, Flintshire, and Owston,nbsp;Yorkshire). For information as to the MSS. see the prefaces to the printed textsnbsp;by W. J. Rees and Mr. Gwenogvryn Evans. The book was first printed innbsp;1840, under the auspices of the Welsh MSS. Society. The Gwysaney text hasnbsp;been recently published in the Oxford series. See � The Liber Landavensisnbsp;Lyfr Teilo, or the Ancient Register of the Cathedral Church of Llandaff, fromnbsp;MSS. in the Libraries of Hengwrt and of Jesus College, Oxford, with annbsp;English translation and explanatory notes by the Rev. W. J. Rees, M.A.,nbsp;F.S.A., amp;c. � (Landovery, 1840). This has been, however, superseded bynbsp;the Oxford work, in which the Gwysaney text is diplomatically reproducednbsp;under the editorship of Mr. Gwenogvr)'n Evans, � The Text of the Book ofnbsp;Llan Dav, reproduced from the Gwysaney Manuscript by John Gwenogvrynnbsp;Evans, Hon. M.A. Oxon., with the co-operation of John Rhys, M.A.,

CADWALADR TO NORMAN CONQUEST. 129

critically used is not to be ignored. From Welsh literature (other than the Brut and the Laws) no help is to benbsp;obtained for this period.

The task of the historian of Wales has, however, been lightened by the excellent work of those who during thisnbsp;century have devoted themselves to clearing up the earlynbsp;history of England, and it is only by labouring in the lightnbsp;of what they have made known, and according to thenbsp;methods they have adopted, that he can succeed in gettingnbsp;at the truth about the story of the Cymry.

As pointed out above, the break-up of the older Cymric kingdom left Wales in a state of complete political disorganisation. Memories of the old kingship and of thenbsp;old bonds undoubtedly survived in theory and sometimesnbsp;reappeared in fact; but, speaking broadly, the aspect thatnbsp;Wales presents during the succeeding centuries is that ofnbsp;a disunited, or very loosely connected, aggregate of clans,nbsp;or petty kingdoms, or lordships engaged in perpetual warfare both among themselves and with English kingdomsnbsp;and English rulers. It would be untrue to state that therenbsp;was absolutely no conception of a collective nation or of anbsp;united kingdom, but, so far as we can ascertain, on nonbsp;occasion was the whole country effectively under the rulenbsp;of one sovereign. The material is so scanty that it wouldnbsp;be dangerous to make any general assertion in other thannbsp;a tentative fashion.

In an endeavour to clear up the history of a country thus disorganised one of the first questions that must occur to

Professor of Celtic in the University of Oxford � (Oxford, 1893). The � Liher Landavensis � is also called � Hyfr Teilo � (the Book of Teilo). Teilo is one ofnbsp;the principal traditional saints of Wales. He is represented as a cousin of St.nbsp;David�s and as Bishop of Llandaff, but he seems to have advanced archiepiscopalnbsp;claims. For an account of him see Smith�s � Dicty. Christ. Biog.,� subnbsp;nom. The chief authority for his life is a portion of the twelfth centurynbsp;MS. with which this note deals. See � Lib. Land.� (Oxf. ed.), pp. 97 et seq.nbsp;There is no life of, but there are several references to, Teilo in Rees� �Livesnbsp;of the Cambro-British Saints� (Landovery, 1853).

W.P. nbsp;nbsp;nbsp;K

THE WELSH PEOPLE, (chap, v.)

one�s mind is, what were its divisions for the purposes of such government as existed, especially when as in thenbsp;case of Cymru one finds the names attached to variousnbsp;areas ancient. We have in the introduction pointed outnbsp;that Wales was in the times of its practical independencenbsp;of the English monarchy undoubtedly divided into can-trefs and cymwds. In the laws of Howel Da, whichnbsp;are legal treatises once in practical use, the land divisionsnbsp;and measurements are ascribed to Dyfnwal Moelmud, whonbsp;was king � before the crown of London and the supremacynbsp;of this island were seized by the Saxons, and who firstnbsp;established good laws in this island . . . and after thatnbsp;Howel enacted new laws, and abrogated those of Dyfnwal;nbsp;yet Howel did not, however, alter the measurement of thenbsp;lands in this island, but continued them as they were leftnbsp;by Dyfnwal; because he was the best measurer.� ^ Nownbsp;�the cause of his measuring of the island was that he mightnbsp;know the tribute of this island, the number of the miles,nbsp;and its journeys in days.��^ What is expressly ascribednbsp;in the laws to Dyfnwal is the determination of the unitsnbsp;of measurement and the division of the area called thenbsp;cymwd into smaller parts, having some, though to us notnbsp;quite clear, significance in a tribal system. It is notnbsp;said that Dyfnwal marked out the Cymric land into can-trefs and cymwds, but as the cymwd is represented as annbsp;aggregate of smaller divisions, themselves having referencenbsp;to the prescribed units of measurement, it seems to benbsp;implied that he did in fact constitute the division intonbsp;those larger areas.

The matter is not, however, free from difficulty, for if we are to read the text literally as a division ofnbsp;the whole of Cymru, the area of each cymwd ought tonbsp;have been of equal superficial extent. In fact, the

* �Ancient Laws and Institutes of Wales,� vol. i,, pp. 182-185 (1841, Rolls series, ed. A. Owen).

CADWALADR TO NORMAN CONQUEST. 131 cantrefs and cymwds varied very greatly in size. Wenbsp;cannot profess that we have discovered any final solutionnbsp;of the difficulty thus created. It would seem that thenbsp;primary purpose of the division was to facilitate the equitable assessment of the food-rents due to the chieftainsnbsp;as well as in the first instance to secure a fair apportionment of a conquered territory among the new settlers. Itnbsp;may be that in the time of Dyfnwal the different familiesnbsp;of Cymric origin took possession by arrangement amongnbsp;themselves; that they, as was natural, made their firstnbsp;establishments on the more developed and fertile areas;nbsp;that each cenedl on whom the liability for customarynbsp;dues fell became associated with a particular area ; that innbsp;some cases the area came to be called by the name of thenbsp;head of the cenedl at the time of its settlement,^ and innbsp;others that the name of some pre-existing division survived ;nbsp;and that what Dyfnwal the legislator really did was tonbsp;create a system of measurement and division applicablenbsp;roughly to an established order of possession with a viewnbsp;to making the incidence and rendering of the customarynbsp;food-rents fair and easy. However this may be, it isnbsp;certain, that in the tenth century Cymru was dividednbsp;into cantrefs and cymwds, with boundaries ascertainednbsp;well enough for practical purposes, and that the divisionnbsp;was then deemed to be ancient.

Dyfnwal Moelmud is generally supposed to have reigned about 400 years before Christ. This seems due to the placenbsp;given to him by Geoffrey of Monmouth in his line ofnbsp;British kings. According to Geoffrey he gained the sceptrenbsp;of Britain after a civil war that followed the slaying ofnbsp;Porrex, and on his death his sons Belinus and Brenniusnbsp;became kings of Britain.^ If we were to assume this tonbsp;be true, the date of Dyfnwal�s flourishing and legislation

^ Meirionyd. Meirion was a son of CuneSa.

2 � British History,� book ii., cc. l6 and 17; book iii., c. i.

K 2

THE WELSH PEOPLE, (chap, v.)

might be about B.C. 400.1 there is nothing to support Geoffrey�s narrative, while there is credible evidence thatnbsp;about the time of Cuneda (who, as leader of the Cymry,nbsp;conquered Wales) there was a Dyfnwal Moelmud who,nbsp;whether or not he was ever head-king of the Cymry, ornbsp;simply a member of a ruling family, was so closely connected with the line of Cuneda, that his name may have wellnbsp;become associated with legislation over Cymric territory.nbsp;It fortunately happens that there are several pedigreesnbsp;appended to the earliest MS. of the � Annales Cambrim,� andnbsp;as they are undoubtedly old, and came into being at a timenbsp;when every one�s genealogy was most religiously preservednbsp;and remembered as a kind of title-deed to his status in thenbsp;then existing legal and social system, we may with a highnbsp;degree of confidence look upon them as in substancenbsp;accurate.^ The name of Dyfnwal Moelmud occurs innbsp;Pedigree X. He was son of Garbaniaun, and grandson ofnbsp;Coel Hen, whose daughter was Cuneda�s wife, and hisnbsp;pedigree was traced �to Beli et Anna.�� He may thereforenbsp;have been a contemporary of Cuneda�s, and may havenbsp;survived during the lives of one or two generations of hisnbsp;descendants. Save so far as we may infer it from thenbsp;statement in the Welsh laws, that he was king before thenbsp;loss of the crown of Britain, there is nothing to show thatnbsp;he was king of all the Cymry ; and in fact the text does not

' Geoffrey�s Brennius took the city of Rome, and seems to be meant for Brennus, who, according to current computation, in B. c. 390, did capture thenbsp;city and besiege the citadel. Mommsen, �History of Rome,� i., p. 366nbsp;(English translation).

2 nbsp;nbsp;nbsp;� The annales and genealogies in their present form show marks of having

been composed in the last half of the tenth century � (E. Phillimore, �Y Cymmrodor,� xi., p. 144).nbsp;nbsp;nbsp;nbsp;� But the date of the MS. is upwards of a century

later than that of the composition of the Annales and Welsh Genealogies� (Hid., p. 145). See also �The Welsh Pedigrees,� a paper by Henry F. J.nbsp;Vaughan, B.A., S.C.L., printed in �Y Cymmrodor,� x, 72 (1890).

3 nbsp;nbsp;nbsp;� Y Cymmrodor,� ix., p. 174. As the pedigree is not very accessible tonbsp;the student we reproduce it:�

CADWALADR TO NORMAN CONQUEST. 133

necessarily give him the position of a head-king or gwledig ; and there is nothing inconsistent between what is stated innbsp;the laws, and the inference that he was a king of somenbsp;territory subject to Cuneda, and that he was authorisednbsp;by the gwledig, upon the conquest of Wales, to settle thenbsp;affairs of the newly-won lands. This conjecture accordsnbsp;well with the ascertained facts. There were in all probability earlier divisions of Wales, and we need not assumenbsp;that they were entirely superseded by Dyfnwal�s work.nbsp;His main object was evidently, as we have said, to make annbsp;arrangement for the fair imposition of the food-rents of thenbsp;occupiers of Cymric land. Under his system, confirmednbsp;by Howel, the families in each cymwd were liable for thenbsp;same food-rent, that is, the same amount was levied on eachnbsp;cymwd, irrespectively of the number of the families or itsnbsp;size; and as the occupied land varied very greatly innbsp;fertility and productiveness, the operation of practical

THE WELSH PEOPLE, (chap, v.)

causes .may well have led to great difference in the relative extents of the rent-paying units.

In the �Chronicles� and the �Liber Landavensis� we find individuals represented as kings or princes of areas largernbsp;than cantrefs, to which names other than those of thenbsp;cantrefs are given. These may have been, probably were,nbsp;aggregates of the smaller divisions. They representednbsp;lands over the inhabitants of which certain regal or princelynbsp;families descended from, or assumed to be descended from,nbsp;Cuneda, or it may be the descendants of other foundersnbsp;of ancient and pure-blooded tribes of the Cymric racenbsp;exercised a tribal sway, and possessed customary privileges.nbsp;In the � Brut� we find mention of the following areas for thenbsp;most part in terms implying that they were kingdoms:�nbsp;Gwyned, Powys, Ceredigion, Dyfed, Morgannwg, Gwent,nbsp;Brecheiniog, Buaflt, Ystrad Towi,^ Rhuvoniog, Cidweli,nbsp;Gwyr, Mon.

We need not assume the list to be exhaustive. There were very likely other kingdoms or principalities which donbsp;not happen to be mentioned. The names of other districtsnbsp;are certainly to be found {e.g., Gwentlwg, ILeyn, Meironyd),nbsp;but not in terms necessarily suggesting they were undernbsp;separate sovereign families.

It is worthy of notice that before Rhodri�s time there is no mention of a king of Deheubarth (ordinarily used asnbsp;equivalent to South Wales), though the word Deheubarthwyrnbsp;(men of Deheubarth) occurs once.

To the rulers of these larger areas the names � brenin � (king) and � tywysog�^ (prince) are applied. In the earliernbsp;times the former title is liberally accorded. The king isnbsp;usually described as king of a particular district, eg.,

' Ab Ithel in the Rolls edition of the �Brut� translates this as the � Vale of Towi,� but it practically means district of the Towi. Ystrad literally meansnbsp;strand, �strath.�

quot; See s.a. 856 : � Y bu uar6 lonathal tywyssawc Abergeleu.�

CADWALADR TO NORMAN CONQUEST. 135

Gwgaun, son of Meurug, king of Ceredigion ; ^ but some of the chieftains are described as � king of the Britons.� ^ Now-later writers proceed on the theory that Cymru was dividednbsp;into three principal parts, GwyneS, Deheubarth, andnbsp;Powys, with three royal residences�Aberfifraw in the firstnbsp;Dinefwr in the second, and Mathrafal in the third, and annbsp;over-lordship is ascribed to the king of Gwyned. Therenbsp;is nothing in the � Chronicles � absolutely inconsistent withnbsp;this, but on the other hand there is nothing directly tonbsp;support it; but some of the later legal treatises accord anbsp;pre-eminence to the king of Gwyned, and many isolatednbsp;facts tend to support this view, so far as the kingdoms ofnbsp;Deheubarth and Powys are concerned but there seems nonbsp;evidence proving with certainty that the regal families ofnbsp;South-eastern Wales, which was divided into the kingdomsnbsp;of Morgannwg, Gwent, Brecheiniog, and BuaHt, generallynbsp;acknowledged the over-lordship of Gwyned. There is,nbsp;however, no improbability in the view that the chieftainsnbsp;of Cymru at one time regarded themselves as forming anbsp;kind of hierarchy of kings and certainly the organisationnbsp;of each kingdom, as described in the codes, seems to involvenbsp;a gradation of lordships very nearly resembling a feudalnbsp;system. But though the supremacy of Gwyned and thenbsp;subordination of one ruler to another, in some settled mannernbsp;grouping all Wales into a collective nation, may have beennbsp;a legal first principle, the actual facts, as gathered or inferrednbsp;from the �Chronicles,� hardly seem to square with the theory.

* nbsp;nbsp;nbsp;871. �Ann. Cam.,� j-.a. 871.

^ -E.g., �Brut,� s.a. 998 : �Mareduit, son of Owain, the most celebrated King of the Britons.�

� Mr. Seebohm adopts the theory. �Tribal System in Wales,� pp. 134� 139-

�* Cf. ^the case of Ireland, where there seem to have been recognised an Ard-rf Erend and three classes of subordinate kings. O�Sullivan�s Introductionnbsp;to 0�Curry�s �Lectures on the Manners and Customs of the Ancient Irish,�nbsp;pp. ccxxix-xxxi. See also Ginnell�s �The Brehon Laws� (London, 1894),nbsp;pp. 63 et seq.

THE WELSH PEOPLE, (chap, v.)

and what we do find is an aggregate of small states under separate kings or ruling families continually quarrellingnbsp;among themselves.

The period now under consideration�that from the death of Cadwaladr, in 664, to the Norman Conquest�nbsp;may itself be conveniently divided at the accession ofnbsp;Rhodri Mawr. For the one hundred and eighty years thatnbsp;elapsed from the loss of the crown of Britain to 844, whennbsp;Rhodri became chief king of the Britons, we know little ofnbsp;what took place in Wales, and all we can gather fromnbsp;trustworthy sources is the names of certain kings andnbsp;battles, and the general conclusion that the limits of thenbsp;Welsh area were further contracted.

According to Caradog of ILancarvan, Ivor, son of Alan, king of Armorica (who becomes strangely confoundednbsp;with Ine, king of Wessex), succeeded Cadwaladr on hisnbsp;death in 681, and reigned till 720. Then Rhodri Molwynog,nbsp;son of Edwal Ywrch, became � King of the Brytains,� andnbsp;survived till 750, and was followed by his son, Conannbsp;Tindaethwy, who continued chief king till 817, when henbsp;died (after chasing his brother Howel out of Mon in thatnbsp;year), leaving a daughter, Esyllht, married to Merfyn Frychnbsp;ab Gwriad. Merfyn and his wife took possession of thenbsp;kingship, and Merfyn reigned till 841. In that year (�asnbsp;some do write �) he was killed in a battle at Cetteh betweennbsp;the Welsh and the Mercians under Burchred, and then hisnbsp;son Rhodri succeeded. This account is usually followednbsp;by Welsh historians, but it is barely credible. For a periodnbsp;of one hundred and fifty-three years {i.e., 664 to 817), onlynbsp;four chief kings (one following the other immediately) arenbsp;assigned, and Conan is made to reign nearly seventy years.

Caradog appears to have been trying, very likely honestly enough, to represent a continuity in the devolution of thenbsp;Cymric over-lordship that had had no existence except innbsp;the imagination of later rulers and those who were connected

CADWALADR TO NORMAN CONQUEST. 137

with them as bards and genealogists. The � Brut,� while not absolutely inconsistent with the view of Caradog, can hardlynbsp;be described as entirely supporting it. It says that afternbsp;Cadwaladr, Ivor, son of Alan, king of ILydaw, which is callednbsp;little Britain, reigned not as king, but as a chief or prince,^ andnbsp;after him Rhodri Molwynog reigned. The entries down tonbsp;844, when it records the death of Merfyn Frych, are verynbsp;brief. Rhodri Molwynog died in 754. The earliest referencenbsp;to a Kynan is in 812, when it is recorded that a battle tooknbsp;place between Howel and Kynan, and that the latter wasnbsp;beaten, and three years later (815) Howel expelled Kynannbsp;from M�n.^ Then, in 817, Howel was �a second time�nbsp;driven from M�n, and Cynon (who, we may fairly assume,nbsp;was identical with Kynan) died, and the Saxons ravagednbsp;the mountains of Eryri, and took the kingdom of Rhuvoniog.^nbsp;No mention is made of Esyllht, but it is stated thatnbsp;Merfyn Erych died in 844. We may therefore look uponnbsp;the existence of Ivor, Rhodri Molwynog, Kynan, and Merfynnbsp;Frych as proved, and we may believe that they were verynbsp;important chieftains of the Cymry in the time after thenbsp;death of Cadwaladr and before Rhodri Mawr�s accession.nbsp;But there are other kings mentioned, such as Caradog,nbsp;king of Gwyned, Maredud and Rein, kings of Dyfed,nbsp;Arthen, king of Ceredigion, and CadeE, king of Powys.^nbsp;To Rhodri Molwynog, indeed, the title of � brenin ynbsp;Brytanyeit� is accorded, but the deaths of Conan andnbsp;Merfyn are mentioned as if they were simply kings ofnbsp;districts; and we cannot avoid noticing that if Gwyned

* The words in the text of the �Brut� (Oxford edition, p. 257) are � ac nyt megysbrenhin namyn megys pennaeth neu tywyssauc.� They are importantnbsp;as showing clear recognition of the change in the position of the Cymry in thenbsp;island, which had been brought about by the events that led up to the death ofnbsp;Cadwaladr and the loss of �the crown of Britain.�

2 nbsp;nbsp;nbsp;�Brut,� s.a. 815.

3 nbsp;nbsp;nbsp;� Brut,� s.a. 817.

See under the years 798, 796, 807, and 808, �Ann. Cam.,� 798, 807, 808.

THE WELSH PEOPLE, (chap, v.)

had at that time a practical pre-eminence, it is strange that Caradog (a king of that part of Wales) should find nonbsp;place in the succession of kings of all Wales as tracednbsp;by Caradog of ILancarvan.

The sources of information by means of which we may check or correct the traditional or usually adopted accountnbsp;of this period are not limited, however to the Chronicles.nbsp;The first of the pedigrees appended to the MS. of thenbsp;� Annales Cambriae,� edited by Mr. E. Phillimore, is that ofnbsp;Owain, the son of Howel Da, and great-grandson ofnbsp;Rhodri Mawr, and it carries back his genealogy a very longnbsp;way. It is a genuinely old compilation, and however muchnbsp;we may doubt, or rather be in a state of indifference as tonbsp;the more remote stages, yet if we bear in mind the legalnbsp;structure and general complexion of the community innbsp;which it was produced, it would be an excessive display ofnbsp;the sceptical spirit to deny its accuracy for many generations, especially as there is evidence from many sourcesnbsp;that most of the nearer ancestors of Owain whom itnbsp;discloses really lived and played their parts among thenbsp;Cymry in a sequence of events that is not inconsistent withnbsp;the order of the names in the pedigree in question.�-

According to this pedigree, the names of the successors of Cadwaladr were :�lutgual, Rotri, Cinnan, Etthil, Mer-min, Rotri (Mawr).^ lutgual is probably the Idwal Ywrchnbsp;of Caradog ; the first Rotri is evidently Rhodri Molwynog,nbsp;king of the Britons, who, according to the ��Brut,� died innbsp;754; Cinnan seems to be the Kynan or Cynon of the �Brut,�nbsp;who fought with Howel in 812 and 815, and died in 817,nbsp;and Caradog�s Conan Tindaethwy; Etthil, daughter of

* In Mr. Phillimore�s opinion, �up to the date when all Welsh records necessarily became more or less fabulous, these genealogies have every claimnbsp;to rank beside the � Annales � and the � Saxon Genealogies � as a valuablenbsp;historical authority.� � Y Cymmrodor,� ix., p. 149 (1888).

2 This important pedigree deserves the most careful study. It is printed in ihe preface to Aneurin Owen�s �Welsh Laws,� etc. (vol, i., Preface, p. xiv.,

CADWALADR TO NORMAN CONQUEST. 139

Cinnan, is the Esyllht who, in his account, was married to Merfyn Frych; Mermin, son of Etthil, is probablynbsp;identical with the Merfyn Frych whose death is assignednbsp;by the Brut to the year 844, and the Mermin whose deathnbsp;is mentioned in the Annales as taking place in the samenbsp;year. The second Rotri is Rhodri Mawr.

We may, then, take it that the existence of Idwal Ywrch, Rhodri Molwynog, Conan Tindaethwy, Etthil, Merfynnbsp;or Mermin Frych is confirmed by the pedigree, and thatnbsp;they were descendants of Cadwaladr; but the intervalnbsp;between 754, when Rhodri Molwynog died, and 817 ornbsp;816, when Conan died, is very long. A Caradog, king ofnbsp;GwyneS, is stated to have been killed by the Saxons in 798.^nbsp;He may have been a son of Rhodri Molwynog, and have

140 THE WELSH PEOPLE, (chap, v.)

been omitted by the compiler of the genealogy, but there can be no certainty about the question. Nor can we speaknbsp;positively as to the district over which this line held sway,nbsp;though the mention in the Brut and Annales of kingsnbsp;of various districts,^ who do not appear to have been descendants of Cadwaladr, leads us to the opinion that lutgualnbsp;and his immediate successors were rulers of Gwyned.

In addition to the scanty information we have as to the names of these kings, we know, from the Welsh and othernbsp;sources, that there was almost continual warfare between thenbsp;Cymry and their English neighbours, and very frequentlynbsp;among themselves, and that, as a result, the Cymric areanbsp;was again diminished.

It is with the name of Offa of Mercia that the further and definite lessening of the Cymric land is chieflynbsp;associated. He began his reign in 757.� Of his deedsnbsp;during its earlier years little is recorded, but later on henbsp;engaged in seemingly fierce contests with the Welsh. Innbsp;776 the destruction of the South Wales men took place,�nbsp;and some years after he pushed over the Severn, � andnbsp;spoiled the Britons in summer time.�* The king of Powysnbsp;was driven from Pengwern (Shrewsbury), till then thenbsp;capital of his realm, and the boundaries of Mercia werenbsp;practically carried to the Wye. It was probably aboutnbsp;this time that the Mercian king caused the earthworknbsp;known as Clawd Offa, or Offa�s dyke, to be constructed.nbsp;Speaking roughly, this work extended from the estuarynbsp;of the Dee to the mouth of the Wye. Whether it wasnbsp;intended for military purposes or simply as a visible mere

' nbsp;nbsp;nbsp;�Arthgen rex cereticiaun,� �Regin rex demetoram; et catel,

pouis moriuntur,� � Ann.,� s.a. 807, 808.

* nbsp;nbsp;nbsp;�Dictionary of National Biography,� suh nom. Offa.

� �Brut,� s.a. 776. �Vastatio Brittonum dexteralium apudOffa,� �Ann.,�� s.a. 778.

* nbsp;nbsp;nbsp;�Biut,�j.a. 779. (^�Vastatio Brittonum cum Offa in estate,� �Ann.,�nbsp;s.a. 784.

CADWALADR TO NORMAN CONQUEST. 141

between England and Wales, it became recognised as the boundary line of Cymru. The dyke is not mentioned innbsp;the Oxford text of the �Brut,� though in one of the MSS.nbsp;on which the Rolls edition is based, it is stated that Offanbsp;caused the dyke to be made to enable him more easily tonbsp;withstand the attacks of his enemies. Probably this is anbsp;late addition to the Brut, but the making of the dyke isnbsp;mentioned by Asser,^ and its existence is an indisputablenbsp;fact. Portions of it were a few years ago noticeable. Innbsp;Radnorshire at the beginning of this century its remainsnbsp;were � almost as fresh as if cut yesterday, excepting thenbsp;edges, which are clothed with a fine verdure.��

The consolidation of the Teutonic kingdoms in England under Ecgbryht (the first Saxon king who called himselfnbsp;Rex Anglorum) in the early years of the ninth century hadnbsp;an immediate effect upon the fortunes of the remnantnbsp;of the Britons. In his reign the Cornish people werenbsp;subdued, and henceforth, though they maintained somenbsp;kind of separate organisation, they never successfully threwnbsp;off the yoke of Wessex. After the reduction of Mercianbsp;and Northumbria, it seems clear that he extorted thenbsp;submission of the Cymric princes^�the English certainlynbsp;made temporarily successful invasions into the heart ofnbsp;the Cymric land.^

� Asser, M. H. B. 471.

2 Williams�s � History of Radnorshire,��58. See also Pryse�s � Descriptio � prefixed to �Caradocof]Lancarvan,�ed. 1584; also Guest�s �OriginesCelticse,�nbsp;vol. ii., p. 273 ; and Longucville Jones�s article in ArcA. Cambr. (3rd series,nbsp;vol. ii., pp. 1-3, and pp. 151-4); also Earle�s paper, Arch. Cambr., 3rdnbsp;series, vol. iii., pp. 196-209. Most of what is known about the dyke is wellnbsp;stated by Mr. A. N. Palmer in his paper, � Offa�s and Wat�s Dykes,� innbsp;�Y Cymmrodor,� vol. xii., p. 65(1897).nbsp;nbsp;nbsp;nbsp;� Freeman, N. C. i., p, 42.

* In 817, � The Saxons ravaged the mountains of Eryri and took the kingdom of Rhuvoniog,� �Brut,� r.a.

In 818, �A fight took place in Mona, called the action of Llanvaes.�

In 819, � Kenulf ravaged the kingdoms of Dyfed.�

In 823, � the Castle of Deganwy was destroyed by the Saxons, and then the Saxons took the kingdom of Powys into their possession.quot; � Brut,� s, a. 823.

THE WELSH PEOPLE, (chap, v.)

The reign of Ecgbryht marks a distinctstep in the development of the English monarchy ; but just as the English state was attaining to considerable power, its own existence,nbsp;as well as that of all the kingdoms of the island, wasnbsp;threatened by the formidable invasions of the Northmennbsp;or Danes, who were still outside the pale of Christianity.nbsp;Sporadic incursions had taken place before the accessionnbsp;of Ecgbryht. The Welsh Chronicle says that 790 was thenbsp;year of Christ when the Pagans first came to Ireland.^ Atnbsp;the beginning of the ninth century their raids becamenbsp;more frequent and more effective ; but shortly before hisnbsp;death Ecgbryht defeated the invaders and the Cornish Britonsnbsp;(who had joined them) in a great battle at Hengestendunnbsp;in Cornwall.^

Freeman makes the Danish invasions of England to fall into three periods�one of mere plunder, one of settlement,nbsp;and one of political conquest. The first extended from thenbsp;first appearance of the Scandinavians in the later years ofnbsp;the eighth century to 855 ; the second, from that year tonbsp;897 ; and the third from 980 to 1016, when Cnut commenced to reign as king of the English.* In the firstnbsp;period the Welsh seem to have suffered much as thenbsp;English did, though to a less extent. As to the secondnbsp;period, there was no large settlement of Northmen innbsp;Cymru.^ As to the third, the Danish Conquest did notnbsp;materially alter the relations of the Welsh princes to thenbsp;government of England. However much thfe people settlednbsp;on the coast of Wales may have suffered from the Danishnbsp;raids, it seems clear enough that for the Cymry, as a whole,nbsp;the arrest of the growth of the English monarchy andnbsp;the incoming of fresh settlers was an advantage. The

� Cf. � Ann.,� s.a. 796.

2 In 836, Freeman, � Norman Conquest,� i. 43.

� Freeman, tibi supra,

* See above, pp. 27, 35.

CADWALADR TO NORMAN CONQUEST, 143 incursions of the Northmen assisted the preservation of practical independence by the Welsh nation for a long time, andnbsp;by distracting the English gave an opportunity for the operation of forces which were slowly tending to the consolidationnbsp;of the little British kingdoms beyond the dyke. Probablynbsp;it was due to the absorption of the English in the conflictsnbsp;of the ninth century that Rhodri Mawr found it possible tonbsp;extend his dominion over a very large part of Wales, andnbsp;make his house the really predominant power there.

While Merfyn Frych was still reigning, some cessation of Danish attack enabled Burchred, king of Mercia, to turnnbsp;his attention to the Welsh. In 844 he engaged and defeatednbsp;them at a place called CetyH in the Brut; Merfyn wasnbsp;slain, and was succeeded by his son Rhodri, who came tonbsp;be known as � the Great � (Mawr). With his accession wenbsp;reach ground somewhat surer.

By the death of Merfyn he had become head of the line of Gwyned. Afterwards, by his marriage with anbsp;daughter of Meurig ab Dyfnwalion, he became lord ofnbsp;Ceredigion and Ystrad Towi on the death of her brothernbsp;Gwgan;^ and he is said to have become possessed of Powysnbsp;through his grandmother. Nest, sister and heiress of Congennbsp;ab CadeR, king of Powys.� Whether Rhodri ever directlynbsp;ruled over Powys is not clear,� but it is certain that hisnbsp;dominions included the rest of Wales except Dyfed,nbsp;Morgannwg, Gwent, and the principalities roughly corresponding to the modern Brecknockshire and Radnor. Itnbsp;is, of course, possible that he may have exercised some kindnbsp;of over-lordship even over these territories. We know sonbsp;little of Rhodri that it is not very plain why he came to be

� See Jesus Coll. MS. 20: � CymmrocJor, � viii. 87; Harleian MS. 3859; � Cymmrodor,� ix., p. 180 : Pedigree xxvi.

2 The death of a CadeH, king of Powys, is recorded in the �Brut,� s.a. 808. See Pedigree xxvii. in � Y Cymmrodor,� ix., p. 181.

� In 823 the Saxons took possession of Powys. �Brut,�r.a. The quot;Brut� is silent as to Powys from this time to the Norman Conquest.

THE WELSH PEOPLE, (chap, v.)

called the Great,^ unless it be from the fact that he ruled over an area much larger than any then recent predecessor,nbsp;and that this, coupled with military successes of which wenbsp;have no sure evidence, made him an exceptionally powerfulnbsp;king among his contemporaries. He had continual conflicts with the Mercians and the Danes. According to Irishnbsp;authorities, he slew a Danish chief called Horm in 855.nbsp;The end of his reign was clouded in misfortune; for in 876nbsp;he sustained a great defeat at the hands of the English, andnbsp;was obliged to flee to Ireland. Returning in the followingnbsp;year, he and his brother Gwriad were slain by the Saxons.�

Great importance is attached by the later writers on Welsh history to Rhodri�s reign, for with it, or its conclusion, is associated the division of Wales into the threenbsp;kingdoms we have mentioned � Gwyned, Powys, andnbsp;Deheubarth. Some say that Rhodri made the divisionnbsp;during his lifetime, but the earlier authorities attribute thenbsp;division to his sons after his death. The text of Caradognbsp;of ILancarvan says that Rhodri had divers sons, as Anarawd,nbsp;to whom he gave Aberffraw with North Wales ; Cadeft, thenbsp;second son, to whom he gave Dinevwr with South Wales, alsonbsp;took Powys land by force from his brethren after the death ofnbsp;Merfyn, the third son, to whom his father had given the same.

Powel, in his note, amplifies this statement in substance thus; � Rhodri was the undoubted owner and possessor ofnbsp;all Wales; * Gwyned he had through EsyHt,� his mother;

' This is a convenient place for our calling attention to the excellent biographies of Welsh princes in the �Dictionary of National Biography.�quot;nbsp;They deserve the attention of all students of Welsh history. Most of themnbsp;are written by Professor Tout, Professor J. E. Lloyd, or Mr. Lleufernbsp;Thomas.

^ �Brut,� s.a. 877. �Ann. Camb.� 877. According to the latter, Gwriad was Rhodri�s son.

� �Car. ofLan.,� p. 35.

* This is certainly not true if the term Wales is used to cover the present thirteen counties.

^ Seemingly the Etthil of the pedigree cited above is meant.

CADWALADR TO NORMAN CONQUEST. 145

Demetia, or South Wales, came to him by his wife, daughter of Meurig ab Dyfnwal, king of Ceredigion ; Powysnbsp;he had by Nest, his grandmother. These three dominionsnbsp;he appointed under their meares and bounds, withnbsp;a princelie house in every of them, which he namednbsp;�y tair talaeth,� and left the same unto three of hisnbsp;sons, Anarawd, Cadeit, and Merfyn, which were callednbsp;� y tri thywysog taleithiog � (the three diademed princes).^nbsp;The historians of Wales generally accept this account, oftennbsp;speak as if the division amounted to a splitting up of allnbsp;Cymru, and deplore it as an impolitic act.

There is, however, something wrong in the aspect it gives to the division of Rhodri�s dominions, whether it took placenbsp;during his life or after his death. Gwyned and Powys (asnbsp;we have seen) were separate kingdoms before Rhodri�s time.nbsp;So also were Dyfed and Ceredigion, as we know from thenbsp;Welsh chronicles, which are, however, silent as to any division by Rhodri.^ The principalities or kingdoms of the southeastern part of Wales�Morgannwg, Gwent, Brecheiniog,nbsp;BuaHt, and smaller areas�clearly preserved separate organisations. The evidence of Asser confirms this.� In the � Lifenbsp;of Allfred� he says that King Hemeid,with all the inhabitantsnbsp;of Demetia, compelled by the violence of Rhodri�s six sons,*nbsp;submitted to ALlfred, Howel also, son of Rhys king of

' Giraldus, writing more than 300 years after Rhodri�s death, gives the tripartite division as ancient, and says that Rhodri was the cause of thenbsp;division. � Descriptio, � i. c. 2.

^ S. a. 796, we hear of Maredud, king of Dyfed; s.a. 798, of Caradog, king of Gwyned; s.a. 808, of Rein, king of Dyfed, and Cadett, king of Powys; s.a.nbsp;819, of kingdoms of Dyfed ; s.a. 817, of the kingdom of Rhuvoniog ; s.a. 823,nbsp;of the kingdom of Powys; s.a. 848, of Ithel, king of Gwent; s.a. 856, ofnbsp;lonathal, prince of Abergeleu ; s.a. 871, of Gwgawn ab Meurug, king of Ceredigion and the Vale of Towi.

^ We are aware that the worth and genuineness of Asser�s � Life� have been seriously attacked, and give the extract with this caution.

* We can only find four mentioned in the � Brut.� See the genealogical table at the end of this chapter.

W.P. nbsp;nbsp;nbsp;L

THE WELSH PEOPLE, (chap, v.)

Gleguising, and Brocmail and Fernmail, sons of Meurig, kings of Gwent, compelled by the violence and tyranny ofnbsp;Earl Ethered of the Mercians, sought his protection ; whilenbsp;Elised, son of Teudyr, king of Brecknock, compelled by thenbsp;force of the same sons of Rhodri, sought the government ofnbsp;Alfred. Now, of course, it is quite possible that Rhodri,nbsp;after adding Ceredigion and Ystrad Towi to the parcels ofnbsp;his immediate dominions, had obtained some kind of submission from the kings or princes of the smaller areas herenbsp;mentioned, and that he was recognised as king of all thenbsp;Britons or all Cymru. But it is certain that he was notnbsp;possessed of all the Cymric land. He was king of Gwyned,nbsp;with seemingly Aberffraw as his home demesne, and ofnbsp;Deheubarth, with Dinevwr as its chief seat; but for thenbsp;notion that Deheubarth was equivalent to what we now callnbsp;South Wales there is no warrant at all, and no kingdom ofnbsp;Deheubarth is referred to in the chronicles, though the wordnbsp;Deheubarthwyr is used. Dinevwr seems to have been thenbsp;palace of the king of Ceredigion and Ystrad Towi, whichnbsp;roughly corresponded to the present Cardiganshire andnbsp;the greater part of the modern Carmarthenshire. Neithernbsp;Cidweli nor Gwyr (answering to Gower in Glamorganshire)nbsp;which are both mentioned in the Brut, are shown with anynbsp;certainty to have been part of Rhodri�s dominions as matternbsp;of right; while we find that Dyfed, Morgannwg, Gwent, andnbsp;Brecheiniog, and probably other smaller areas, were undernbsp;other rulers. Some time afterwards the line of Dyfednbsp;came to an end, and the district was incorporated innbsp;some fashion, at any rate temporarily, into the kingdom ofnbsp;Dinevwr or Deheubarth; but its subsequent history�itsnbsp;rapid development into a county palatine, without anynbsp;apparent violent breach of the continuity of its story�nbsp;seems to show the survival of a separate organisation.nbsp;Morgannwg, Brecheiniog, and Gwent remained as � separatenbsp;entities,� if we may use a modern phrase, and by a gradual

CADWALADR TO NORMAN CONQUEST. 147 process became counties under the English or Normannbsp;system.

Though there is no early authority on the question, and though the complexion given to the transaction by Caradognbsp;is probably wrong, yet there is ground for thinking thatnbsp;something unusual did take place in regard to the devolutionnbsp;of the regal rights upon the death of Rhodri. The Cymricnbsp;kingship was originally, and probably down to a late time,nbsp;not a personal monarchy, but a tribal or family chieftaincy.^nbsp;The so-called king was the chief of the royal familynbsp;(penkenedl) in whom the tribal sovereignty was vested.nbsp;If we may assume that the laws of Howel Da apply tonbsp;earlier times, then we should expect that on Rhodri�s deathnbsp;Anarawd, the eldest son, would become chief (if he had thenbsp;necessary legal qualifications of a penkenedl) without anynbsp;division of the family dominions of a permanent character,nbsp;though Cadett (e.g-.) might be made arglwyd (lord) of anbsp;particular district of the family lands for reasons of convenience. Then on the death of Anarawd one wouldnbsp;expect his eldest son or some other member of thenbsp;cenedl (kindred) to become chief and rule over the wholenbsp;dominion. In fact, however, the devolution of Rhodri�snbsp;possessions was different. Anarawd became king ofnbsp;Gwyned, and handed it on to his son Idwal; and Cade�nbsp;became king of Deheubarth, and was succeeded by hisnbsp;son Howel. Merfyn does not seem to have transmittednbsp;any claims to Powys. The two former undoubtedly foundednbsp;the princely lines of Gwyned and Deheubarth. The facts,nbsp;therefore, seem to show that the succession to Rhodri�snbsp;dominions did not proceed in the ordinary way; and,nbsp;perhaps, what took place may have marked a stage in thenbsp;change from a tribal chieftaincy to a territorial sovereignty.

Upon the death of Rhodri his eldest son Anarawd, as we have said, succeeded to Gwyned; CadeH to Deheubarth,

* See below, pp. 202-3.

L 2

THE WELSH PEOPLE, (chap, v.)

and perhaps Merfyn to Powys. Of the south-eastern principalities we learn practically nothing on trustworthynbsp;authority. No grave internal troubles seem to have occurrednbsp;during Anarawd�s long reign of thirty-eight years. Innbsp;880, three years after his accession, there was an Englishnbsp;invasion, and Anarawd defeated the Saxon enemy �fornbsp;God to avenge Rhodri.� ^ The battle was fought nearnbsp;Conway, and came to be referred to as � Dial Rhodri �nbsp;(Rhodri�s revenge).

During the later years of the ninth century the house of Rhodri was undoubtedly the predominant powernbsp;in Wales. We have seen how the kings outside the palenbsp;of Rhodri�s possessions were compelled by the oppressionnbsp;of his �six sons� to seek the protection of .Alfred thenbsp;Great; but now Anarawd himself, with a brother (seeminglynbsp;Cadeit of Deheubarth), abandoned close relations withnbsp;the Northumbrians and came into the great king�s presence and sought his friendship. He was received by thenbsp;king with honour as his son by the Bishop�s confirmation,nbsp;and was presented with many gifts.^ Probably Anarawdnbsp;at first pursued a policy of friendship and alliance withnbsp;the people of Northern Britain as against the Merciansnbsp;and West-Saxons, as he did with the Danes for a time.nbsp;His submission to ./Elfred, and that of his brother, nonbsp;doubt paved the way to that usually friendly relationnbsp;which existed between the chief rulers of Wales andnbsp;the kings of the house of Ailfred during the greaternbsp;part of the tenth century. We know not when Anarawdnbsp;and .Alfred met; not even whether it was before or afternbsp;a temporary quarrel with CadeH which led, in 893, to annbsp;inroad into Ceredigion and the Vale of Towi by the North-Welsh prince. Probably, however, the meeting took place

' �Brut,� s.a. 877; �Ann. Cam.,� 877. � Asser, M. H. B., p. 488.

CADWALADR TO NORMAN CONQUEST. 149

after the defeat of the Danes under Hasting by Alfred in 897.^

The remaining years of Anarawd�s reign were quite barren of any events of importance. He died in 915, andnbsp;was succeeded by his son Idwal Voel. Cade� had predeceased his brother, and Howel afterwards called Danbsp;(the good) became king of Deheubarth, and later on ofnbsp;Gwyned as well. Under the rule of Idwal and Howel thenbsp;Cymry enjoyed unwonted peace.^ They were contemporaries of Eadward the Elder and Aithelstan, and thusnbsp;lived when the power of the house of HLlfred was at itsnbsp;greatest height. Howel throughout his whole careernbsp;remained � on peaceful terms with the English court, andnbsp;so far as we know (except in regard to a dispute withnbsp;Morgan Hen of Morgannwg) with the other Welsh princes.nbsp;Both Idwal and Howel did homage to the English kings,nbsp;and seem to have behaved as faithful vassals.

To appreciate the significance of the isolated facts which we can gather concerning Idwal and Howel it is necessarynbsp;to bear in mind the change that had been wrought innbsp;England by the settlement of the Danes. By the peace ofnbsp;Wedmore England north of the Thames had been dividednbsp;by a line roughly drawn from north to south from thenbsp;Kibble to the upper valley of the Thames. This involvednbsp;the division of the ancient and important kingdom ofnbsp;Mercia into an English and Danish Mercia. The former

' Green, � Conquest of England,� pp. 172-3,183; � Eng. Cliron.,�r.3. 897. The fact, however, that Gwyneit and Deheubarth escaped the ravages of thenbsp;Northmen in 894 suggests that the house of Rhodri was then in alliance withnbsp;them.

^ In the Brut� only four battles are mentioned between 9^4 ^ttd 948 :��

914. The people of Dublin {i.e., Norsemen) made a descent on M�n.

919. A battle took place at Dinas Newyd. (Ann. Cam., 921.)

935. The battle of Brun took place. (Ann. Cam., 938.)

944. Ystrat Clut (Strath Clyde) was devastated by the Saxons. (Ann.

Cam., 946.)

� At any rate, after the first five years of his reign.

150 THE WELSH PEOPLE, (chap, v.)

was a stretch of country from the Ribble to the Thames and the Avon, and it seems clear that it was one otnbsp;.Alfred�s objects to separate the Danes from the Welshnbsp;by this English Mercia. At the beginning of the tenthnbsp;century this remnant of the older Mercia was ruled by thenbsp;Ealdorman ALthelred and his wife Aithelflaed, and afternbsp;the death of the former, by the latter, the celebrated � Ladynbsp;of Mercia.�

No marked change took place in regard to the Welsh principalities during the latter part of the ninth or earliernbsp;part of the tenth century. There is nothing to shownbsp;that either Idwal Voel or Howel extended his dominionsnbsp;over Morgannwg, Brecheiniog, Buaflt, or Gwent. Of Powysnbsp;we hear nothing, but it may be presumed with probabilitynbsp;that the lordships into which, as we gather from laternbsp;sources, it was divided were in some sort of subjection tonbsp;Mercia. As to Deheubarth, it is likely that Howel becamenbsp;possessed of Dyfed, for he married Elen, daughter ofnbsp;Loumarc ab Hymeid, king of Dyfed. This Hymeid wasnbsp;seemingly the king of Dyfed who sought the protectionnbsp;of .^Ifred.i We hear no more of kings of Dyfed, thoughnbsp;it seems to have kept a separate character. On the eastnbsp;Howel extended his rule peaceably over Kidweli and Gwyr,nbsp;and thus became the immediate neighbour of Morgannbsp;Hen (king of Morgannwg), for we find that the undoubtednbsp;possessions of Maredud ab Owain�the grandson of Howelnbsp;Da�included those two districts in addition to Ceredigionnbsp;and Ystrad Towi.^ Gwyned certainly included Mon, thenbsp;present shire of Carnarvon, and part of Merionethshire, and

� See above, p. 14$. In pedigree ii. appended to the � Ann.� (� Y Cymm-vodor,� ix. p. 171), Elen is daughter of loumarc (a mistake for Loumarc), son of Himeyt. Loumarc is the Welsh Lywarch.

^ The �Brut,� s.a. 991, with a videlicet, describes the kingdoms of Maredud as�Dyfed and Keredigion, and Gower and Kydweli. Ystradnbsp;Towi is not specifically mentioned, but probably it was covered by the termnbsp;� kingdom of Keredigion.�

CADWALADR TO NORMAN CONQUEST. 151

probably North Wales up to the Dee, but at this time it did not comprise Chester.^

Pursuing the policy adopted by Alfred, of completely interposing an English kingdom between the Welshnbsp;and the Danes, the rulers of Mercia early in the tenthnbsp;century proceeded to re-fortify that city. Since its surrender to iEthelfrith in 616 the city had lain waste andnbsp;desolate. Situate on the Dee, being the point of junctionnbsp;of ancient ways, and commanding the old coast routenbsp;from England to Anglesey, it was a place of militarynbsp;importance, and its effectual occupation cut off the shortestnbsp;communication between the Welsh of Gwyned and thenbsp;Danish Mercians. Its Roman walls still existed in anbsp;damaged condition, and little exertion must have beennbsp;necessary to make it a comparatively strongly-defendednbsp;centre of operations. In 907 the Ealdorman of Mercianbsp;� renewed � Chester, though we are not informed as to thenbsp;extent of the new fortifications he erected. A smallnbsp;settlement was made, and a secular house of St. Werburghnbsp;was founded in the city. The event was of consequencenbsp;in Welsh history, and henceforth Chester played a considerable part in the military and the economic fortunes ofnbsp;the men of Gwyned and Powys.

Bearing these general considerations in mind, we now return to the personal history of Idwal and Howel.

Few facts are known concerning Idwal's reign over Gwyned. In 922, when Eadward the Elder had subduednbsp;all Mercia (Danish as well as English), Idwal, together withnbsp;Howel Da and a Welsh king called Clydawc,^ received him asnbsp;their lord,^ and the two former did homage to .(Ethelstannbsp;in 926* at Hereford and, it is said, rendered tribute to him.

' We can find no evidence of a re-peopling of Chester by the Welsh.

�We know not where he reigned. The death of a Clydog is recorded in the �Brut,� s.a. 917, and in �Ann. Cam.,� s.a. 919.

� �Eng. Chron.,�j.a. 922.

^ �Eng. Chron.,� s.a. 926.

THE WELSH PEOPLE, (chap, v.)

A word or two ought here to be said as to the submission of the Welsh kings to these kings of the English. There cannbsp;be no reasonable doubt that the entries in the Englishnbsp;chronicles, confirmed as they are by authentic chartersnbsp;and by the documents relating to the litigation betweennbsp;Howel Da and Morgan Hen referred to below, recount realnbsp;historical events. But there is danger of misunderstandingnbsp;what took place. The kings of the English did not bynbsp;reason of the commendation of the Welsh kings obtain thenbsp;right of directly interfering in the affairs of the Welshnbsp;kingdoms. Though Idwal and Howel became his men, theirnbsp;subjects�under-kings or lords or uchelwyr�stood in nonbsp;legal relation to the English king. The effect of the commendation was that the over-lord took upon himself thenbsp;duty of protecting his vassals from their enemies, while onnbsp;their side they incurred the obligation of fighting againstnbsp;their lord�s enemies. The tie was necessarily, in those times,nbsp;a loose one, and was often broken.^ The reality of thenbsp;relationship in the first half of the tenth century is shownnbsp;by the attendance of the Welsh princes at the meetings ofnbsp;the Witenagemot. If we can trust a charter (which is, however, of doubtful authenticity), Idwal was taking part in thenbsp;proceedings of the English assembly at Exeter in 928.^nbsp;That Howel Da attended these meetings on severalnbsp;occasions is certain. The silence of the Brut about Idwalnbsp;till 943 affords some indication that nothing of any importance took place in Gwyneff between 926 and that time;nbsp;but as in that year Idwal and his brother Elised were killednbsp;by the English,� we may presume that Idwal had revolted,nbsp;or perhaps had refused to pay tribute, but there is nonbsp;certainty about the matter.

We have somewhat fuller information about Howel Da,

' See as to the effect of �commendation � Freeman, N. C. i., pp. 131-3.

2 See charter, �Cod. Dipl.,� Iioi.

5 �Brut,� s.a. 941 ; �Ann. Cam.,�r.i�. 943.

CADWALADR TO NORMAN CONQUEST. 153

though of the earlier years�that is, the years before the submission to Eadward the Elder�we know nothing.nbsp;Whether or not he took part in the actions of Dinas Newydnbsp;or of Brun we cannot determine. After doing homagenbsp;to ^thelstan he is said to have made a pilgrimage tonbsp;Rome.1

There seems no reason to doubt this, but if he did so we may infer that there was an unusually stable position of thingsnbsp;in Wales, and no internal events seem to have made suchnbsp;a journey necessary. Upon his return he resumed peaceable exercise of his regal rights. From that time to the yearnbsp;before his death he was a frequent attendant at the meetingsnbsp;of the Witenagemot of his English over-lords. We knownbsp;from his subscription of charters that he did so in 931,nbsp;932, 933, 934, 937, 946, and 949.� In the earlier chartersnbsp;he attests as � sub-regulus � ; in the later ones he subscribesnbsp;as � regulus � and � rex.� Perhaps the difference is due tonbsp;his having on Idwal�s death succeeded to Gwyned andnbsp;becoming recognised as king of the Britons or of thenbsp;Cymry.

The nature of Howel�s relations with the English kings is made clearer by the account, preserved in the Book ofnbsp;Llandaff, of a dispute between him and Morgan Hen,

* �Brat,� s.a. 926; �Ann. Cam.,� s.a. 928. The date is uncertain; see below, p. 182-3.

^ The following are the dates of thequot;charters, and the references to them in Kemble�s �Cod. Diplom.� :�

21 July, 931 (�Cod. Dipl.� v. 199).

12 Nov., 931 (rA, ii. 173).

30 Aug., 932 {id., V. 208).

15 nbsp;nbsp;nbsp;Dec., 933 {id., ii. 194).

28 May, 934 {id., ii. 196).

16 nbsp;nbsp;nbsp;Dec., 934 {id., V. 217).

937 {id., ii. 203).

946 and 949 {id., ii. 269, 292, 296).

There are also subscriptions of Howel�s to doubtful charters of the 17th June, 930, and the ist Jan. and 21st Dec., 935 (�Cod. Dipl.� ii, 170; v. 222;nbsp;ii. 203).

THE WELSH PEOPLE, (chap, v.)

king of Morgannwg. The entry in the Book is not, however, free from ambiguity. It says Eadgar and Howel Da andnbsp;Morgan Hen were kings of all Britain, and those two werenbsp;subject to King Eadgar, and Morgan enjoyed the whole ofnbsp;Glamorgan in peace and quietness, but Howel would takenbsp;from him Ystradyew and Ewyas if he could. Then Eadgarnbsp;summoned his under-kings and Morgan�s son, Owain, beforenbsp;him, and having examined the matter in dispute, gavenbsp;judgment in favour of Morgan, and with the common assentnbsp;and testimony of all the archbishops, bishops, abbots, earls,nbsp;and barons of all England and Wales granted to Owain, thenbsp;son of Morgan Hen, �the said two districts of Ystradyewnbsp;and Ewyas, declared by name to be in the diocese of Llan-daff, as his own proper inheritance.�^ There are severalnbsp;points of difficulty connected with this document�pointsnbsp;which we do not affect to solve. Eadgar, though he hadnbsp;ruled in Mercia before, was not king of England till 958,nbsp;some eight years after Howel�s death. Morgan Hen survivednbsp;Howel, for he was a witness to a charter of Eadwig�s in 956,nbsp;together with Eadgar, sub-regulus of Mercia. There is anbsp;Welsh version of the same Latin document to be found innbsp;the Myvyrian Archaiology, among the collection callednbsp;� Y cwta cyfarwyd o Forganwg.� It is clearly impossiblenbsp;that Howel could have appeared before Eadgar, king ofnbsp;England. It is, of course, possible that the dispute maynbsp;have arisen during the time of Howel, and, lingering on fornbsp;years, may have been decided by Eadgar. But, on thenbsp;other hand, it is quite possible that the dispute may havenbsp;been decided, not by Eadgar, but by Eadward the Elder,nbsp;and that the mistake may have occurred in the transcriptionnbsp;of the account of the dispute and of the grant, especially asnbsp;we find it expressly stated that it was inserted in the Book

* � Book of Llan Dav � (Oxford, 1893), P- ^4^ gt; � Liber Landavensis �� (Llandovery, 1870), p. 237; Palgrave, � English Commonwealth,� v. 2,nbsp;p. ccxliv.

CADWALADR TO NORMAN CONQUEST. 155

of l.andaff because the paper on which it had been written had nearly perished from its great age.^

It is, however, on his legislation that the fame of Howel Da chiefly depends, for to him is attributed the settingnbsp;down in writing of the laws and customs of the Cymry.nbsp;The preamble prefixed to each of the codes that havenbsp;been handed down to us in substance (though in varyingnbsp;language) records that Howel summoned four men fromnbsp;each cantref in his dominions to the Ty Gwyn, which isnbsp;identified by far-reaching tradition with Whitland in Carmarthenshire. We deal with these laws in the followingnbsp;chapter. Though there is no mention in the Brut of thenbsp;summoning of this meeting at Ty Gwyn, there is no reasonnbsp;to doubt that the preambles of the Code preserve anbsp;historical transaction.^ It is probable that the compilationnbsp;of the work took place after Howel had become king ofnbsp;Gwyned upon Idwal�s death, and therefore some timenbsp;between 943 and 950. These facts are all that we cannbsp;glean upon trustworthy evidence concerning a king whonbsp;was, to use the words of a later writer, �for his godlienbsp;behaviour, discreet and just rule, beloved of men.� Theynbsp;are too few to enable us to draw a vivid picture of hisnbsp;character or personality, but they corroborate the view ofnbsp;him popularly entertained among the Welsh people, andnbsp;justify us in inferring that he was an able and politic prince,nbsp;under whom Wales enjoyed a period of unusual repose andnbsp;prosperity.

The peace that Howel had kept disappeared at his death. There was war at once between his sons (Owain, Dyfnwal,nbsp;Rhodri, and Edwyn) and the sons of Idwal Voel (leuafnbsp;and lago), and the eighty-nine years that elapsed from the

gt; Palgrave, uii supra, says that according to usual custom the Welsh scribe omitted the final d and substituted a for a w, relying, he observes, on thenbsp;authority of a genealogical MS. (� Bibl. Hart,� 4181).

2 See the next chapter.

THE WELSH PEOPLE, (chap, v.)

commencement of this conflict to the accession of Gruflyd ab ILewelyn in 1039 a time of almost inextricablenbsp;confusion. In a battle fought at Carno in the very yearnbsp;that Howel died his sons sustained a defeat at the handsnbsp;of leuaf and lago, who, setting aside their elder brother,nbsp;obtained joint possession of Gwyned.^ In 952 theynbsp;ravaged Dyfed twice, and slew DyfnwaHon, who wasnbsp;probably the prince of that region.^ The sons of Howelnbsp;in 954 invaded North Wales, but were again apparentlynbsp;unsuccessful, being beaten in an engagement at ILanrwst,�nbsp;by the sons of Idwal, who thereupon devastated Ceredigion�nbsp;whence, however, it is said they were driven back withnbsp;great slaughter.* After this there was an interval of peacenbsp;between Gwyned and Deheubarth, but raids of the Danesnbsp;gave some trouble.

There was a quarrel with the English in 965, and Alvryd invaded and ravaged Gwyned; while in 970 Godfrey sonnbsp;of Harold subdued and for a time held M�n.� Beforenbsp;this latter event, however, the brothers of Gwyned hadnbsp;quarrelled. lago seized leuaf and caused him to be blindednbsp;and then hanged. The relations between the Welshnbsp;princes and Eadgar (9S8�97S) were fairly peaceable,nbsp;though there seems to have been an invasion of Gwynednbsp;in 968,� but the English hold on Wales was graduallynbsp;relaxing. No fewer than four Welsh princes attended anbsp;Witenagemot held by Eadred. During Eadgar�s time,nbsp;so far as we can tell, the Welsh no longer attended thenbsp;English Court, and their dependence on the English Crown,

gt; � Brut,� s.a. 948, 950, 951, 952.

^ He may have been Dyfnwal ab Howel Ba; but the � Brut � places the death of this Dyfnwal in the following year.

� � Brut,� s.a. 952.

* Gwentian � Brut � (Myv. Arch, ii., 468 et seq). s �Brut,� s.a. 965 and s.a. 970.

� �Ann. Cam.,� s.a. 968. It is possible this is the same invasion as that by Alvryd noted in the � Brut,� s.a. 965.

CADWALADR TO NORMAN CONQUEST. 157 which had been real enough, as we have seen, earlier in thenbsp;century, was now becoming nominal as the power of thenbsp;house of iElfred was waningd

In later chronicles it is said that Eadgar went to Chester and summoned eight under-kings�including five Welshnbsp;princes�to his presence, that they did homage and sworenbsp;fealty, and that as a mark of their subordination he causednbsp;his vassals to row him in his barge on the Dee from thenbsp;palace to the monastery of John the Baptist, and afternbsp;divine service there back to the palace.^ Possibly thenbsp;invasion of Gwyned mentioned above may have led tonbsp;a renewal of oaths of fealty and payment of tribute; butnbsp;this twelfth-century story cannot be accepted as certainnbsp;history, and a similar observation must be made as to thenbsp;imposition by Eadgar of a tribute of three hundred wolves *

Returning to the affairs of Gwyned, the murdered leuaf had left a son called Howel, who was not long innbsp;avenging his father�s death, for in 972 he succeeded innbsp;expelling lago and taking possession of Gwyned in hisnbsp;stead. lago was captured by Danes in 978, and we hearnbsp;of him no more.^

Howel�s rule was soon challenged, for the cause of the defeated chieftain was espoused by his son Kystenin, andnbsp;in the year after the capture of his father he, with the helpnbsp;of Godfrey son of Harold (of whom we have alreadynbsp;heard), made a raid on ILeyn (in the modern South

* nbsp;nbsp;nbsp;Green, � Conqu. of Eng.,� p. 323. There is one charter to which lago�snbsp;name (as lacob) appears dated at Bath, Whitsuntide, 966, but the documentnbsp;is suspect. Kemble, �Cod. Dipl.,� 519.

=gt; �Will. Malm. Gest. Reg.� (Hardy) i., p. 251 ; �Flor. Wore.� (Thorpe) i. 142.

5 �Will. Malm. Gest. Reg.� (Hardy) i. 251. Palgrave prints a charter (the date of which must have been before 971, since one of the subscribers, Oscytel,nbsp;died in that year) which lukill and lacobus attest. �Eng. Com.� ii.,nbsp;p. ccliii. They may have been leuaf and lago ; but lukill does not look likenbsp;leuaf at all.

* nbsp;nbsp;nbsp;The dates are quite uncertain. See � Brut,� s.a, 972 and 978.

158 THE WELSH PEOPLE, (chap, v.)

Carnarvonshire) and M�n, but he was met by Howel ab leuaf at Hirbarth, and fell in the battle.^ Howel, whonbsp;acquired (in curious antithesis to the case of Howel Da)nbsp;the epithet of Drwg, or the Bad, does not seem to havenbsp;been attacked by other Welsh claimants to Gwyned; butnbsp;the Godfrey son of Harold (who had taken the side ofnbsp;Cystenin, as we have seen) still vexed the Welsh kingdom.nbsp;In 981 he ravaged Dyfed, and in the next year Brecheiniog,nbsp;and the territories of Einion ab Owain ab Howel Da werenbsp;overrun by the Saxons under Alvryd. In 984 Howel wasnbsp;killed by the � Saxons through treachery.� ^ He left two sonsnbsp;�at least�Maig, who was killed in the following year, andnbsp;Cadwallon, who took possession of Gwyned, but who wasnbsp;almost immediately defeated and slain� by Maredud abnbsp;Owain, king of Deheubarth (a grandson of Howel Da),nbsp;who played a considerable part for the next few years.

We must now turn for a moment to the affairs of Deheubarth. Upon Howel Da�s death Owain and his three brothers succeeded to that kingdom, but failed to makenbsp;good whatever claim they may have had to Gwyned asnbsp;against the sons of Idwal. Dyfnwal, Rhodri, and Edwynnbsp;died very soon,^ and Owain reigned alone till he died innbsp;987 or 989,� and was succeeded by the Maredud justnbsp;mentioned.

From an incidental statement in the Brut� we know that Maredud�s possessions included Dyfed, Ceredigion, Gwyr,nbsp;and Kydweli, and no doubt Ystrad Towi, which had long

� � Brut,� s.a. 979.

2 �Brut,� s.a. 984. Probably, however, a year or two later.

� � Brut,� s.a. 984.

* In 951 and 952, according to the � Brut.� Probably in either case later.

^ � Brut,� s.a. 987. In � Ann. Cam. � Owain�s death is placed in the next entry after 987.

� �Brut,� s.a. 991. It is curious that this is pretty nearly the kingdom of Pryderi ab Pwyft, as described at the end of the story of Pwytl, prince of Dyfed,nbsp;in the � Mabinogion.� See Oxford edition of the Red Book, i. 25.

CADWALADR TO NORMAN CONQUEST. 159

been connected with Ceredigion. Gwyr may have passed from Howel�s sons on his death, for we have it recordednbsp;that Einion ab Owain, his grandson, devastated that districtnbsp;twice�the first time, according to the Brut, in 968, and thenbsp;second time in 976. Probably Einion was extending thenbsp;family territory to the east, and under his father Owain wasnbsp;able to annex some of the smaller lordships or areas tonbsp;Deheubarth. We hear that Brecheiniog and all the territorynbsp;of Einion were devastated by the Saxons in 982 ^�an entrynbsp;which seems to show that Einion�s territory extended tonbsp;south-eastern Wales. The year after this raid Einion wasnbsp;killed through �the treachery of the nobles of Gwent.� ^

Apart from such exploits as may have been performed by Einion and Maredud, the only events of Owain�s long reignnbsp;were the usual raids of Danish leaders, and some conflictsnbsp;with the English. Of his relation to Powys we knownbsp;nothing certain. We may conjecture with some probabilitynbsp;that he shared some of the qualities of his father, for uponnbsp;his death he handed on to Maredud the kingdom ofnbsp;Deheubarth with its area undiminished, and before thatnbsp;event, as we have recounted, his son Maredud, takingnbsp;advantage of the fall of Howel Drwg, had founded a claimnbsp;to Gwyned by attacking and killing that king�s brother,nbsp;Cadwatton ab leuaf.

Maredud does not seem to have been able to obtain real possession of Gwyned, though Caradog places himnbsp;in the line of kings or princes of all Wales.� He was

' �Brut,� s.a. 982. Cf. Ann. Cam. (Rolls Series), p. 20.

^ � Brut,� s.a. 983. Gwent was still farther to the east, and beyond Morgannwg.

2 We may conjecture, too, that the real effect of Maredud�s victory over Cadwatton ab leuaf was to create a kind of interregnum in Gwyned. Thenbsp;student must bear in mind that the fact that there was no king of Gwyned atnbsp;any particular moment did not disorganise the life of the territory as the wantnbsp;of a head in the highly-centralised systems of to-day generally does now.nbsp;Except in time of attack from without, there was some advantage to the ordinarynbsp;head of a family, for the king�s progresses, etc., were for the time discontinued.

THE WELSH PEOPLE, (chap, v.)

chiefly occupied in repelling the Danes (to whom he paid tribute on at least one occasion)/ and in attacks uponnbsp;Gwyned and Morgannwg.** It is said he ruled Powys innbsp;right of his mother, but there is no sufficient authority fornbsp;this.^ However this may be, he fairly maintained thenbsp;prestige of the house of Howel Da in a very disturbednbsp;period, and died a natural death in 998 or 999,'* leavingnbsp;only one child�a daughter, married to ILewelyn ab Seisyllt,nbsp;who, apparently in right of his wife, assumed the governmentnbsp;of Deheubarth.

Ever since the retreat of Maredud from Gwyned, after his victory over Cadwallon, that kingdom had been in anbsp;condition of extreme confusion ; and there was probablynbsp;from the death of Cadwallon ab leuaf to lago ab Idwal�s^nbsp;accession a kind of interregnum.� Meurig ab Idwal Voelnbsp;(apparently he who was ousted by his brothers leuaf andnbsp;lago) � fell sick � and died,^ but he left issue. Among hisnbsp;sons was an Idwal, who fled to ILancarvan in Morganwg,nbsp;in the lifetime of Maredud, who made an attempt to seizenbsp;him. Idwal�s claims on Gwyned seem to have been just,nbsp;according to the legal rules of succession, and Maredud�snbsp;attempt on his person having failed, he secured somenbsp;adherents, and in 992 returned to Gwyned.� In thenbsp;following year, 993, a battle took place between him and

� �Brut,�r.a. 988.

* nbsp;nbsp;nbsp;In 990 he devastated also Maes Hyfeid (roughly equivalent to the presentnbsp;Radnorshire). See �Brut,�r.o. 990.

^ Warrington, p. 20, cites � Brit. Antiq. Revived,� by G. Vaughan of Hengwrt, pp. 5, 14.

�* �Brut,� s.a. 998.

� /.If., 984 to 1021.

� Caradog, however, fills the time up with Idwal ab Meurig, Conan ab Howel, and Aedan ab Blegored.

^ �Brut,�j.a. 972; but cj. �Ann. Cam.,� � Meuric filiusIdwalcsecatusest � (974)-

* nbsp;nbsp;nbsp;�Brut,� s.a. 992: �the sons of Meurig made an inroad intonbsp;Gwyned.�

CADWALADR TO NORMAN CONQUEST. i6i his brothers^ and the sons of Maredud, in which the formernbsp;triumphed, with the result that Idwal ab Meurig becamenbsp;king of Gwyned;* but he did not enjoy his success long,nbsp;for two years after he was killed, probably by the Danes.�

He left a son of tender years, I ago, who, though passed over for the moment, many years after obtained Gwyned.nbsp;Upon Idwal�s death Cynan ab Howel, Aedan ab Blegored,nbsp;and others, with little or no show of right, � did aspire tonbsp;the government,� and sought the rule of the land.^ Therenbsp;was again a contested succession, but Kynan was killednbsp;(presumably in a battle with Aedan) in 1003,� and Aedannbsp;seems to have usurped the throne. We know nothing morenbsp;of him except that he and his four sons were killed in 1016nbsp;in a fight with ILewelyn ab Seisyllt, who once more joinednbsp;Gwyned to Deheubarth.� The troubles of England undernbsp;ALthelred the Unready, culminating in the fall of the housenbsp;of Ailfred and the accession of Cnut, seem to have affordednbsp;some relief to Wales from attacks from the English bordernbsp;as well as by Danish forces, and with the reign of ILewelynnbsp;begins a fresh growth of Cymric power that attained itsnbsp;greatest development in the reign of his son, Gruffyd abnbsp;ILewelyn. It is assumed that ILewelyn ruled over Powys,^nbsp;but for this there is no certain warrant. Though Deheu-barth seems to have been fairly quiet under his rule, henbsp;had one rising at least to contend with. The uchelwyrnbsp;of South Wales �loved not ILewelyn,�� and, led by one

' In the �Brut� Idwal�s party is described as the sons of Meurig. The only one of the sons of Meurig besides Idwal whose name we know seems tonbsp;be the Jonaval who was killed by Cadwatton ab leuaf in 984.nbsp;nbsp;nbsp;nbsp;� Brut,�� s.a.

984. Cf. �Ann. Cam.,� p. 24.

^ � Brut,� s.a. 993.

3 � Brut,� s.a. 995. Three years after, according to �Ann. Cam.,� p, 21.

? nbsp;nbsp;nbsp;Caradog, p. 74.

5 �Brut,�r.a. 1003.

� nbsp;nbsp;nbsp;� Brut,� s.a. 1016.

I Warrington, p. 205.

* nbsp;nbsp;nbsp;Caradog, p. 85.

W.P. nbsp;nbsp;nbsp;M

THE WELSH PEOPLE, (chap, v.)

Meurig ab Arthvael, revolted in 1019, but they were at once subdued.^

In the following year there was a more formidable rebellion, when a pretender called Rein Yscot, affecting to be a son of Maredud ab Owain, got together the strength of Deheu-barth at Abergwili, and waited the coming of the king.nbsp;ILewelyn, � daring and fearless,� with the host of Gwynednbsp;engaged in battle and conquered Rein, who, though bravenbsp;and confident in assault, � retreated shamefully in a fox-likenbsp;manner, and never thenceforward made his appearance.�^nbsp;The men of Gwyned wrathfully pursued the enemy, andnbsp;devastated the country to the Mercian border. In 1023nbsp;ILewelyn at the height of his power died.� It is said bynbsp;Caradog he was slain by two descendants of Howel Da, butnbsp;the � Brut � simply records that he died. His son Gruffyd,nbsp;who was destined to play a great part in the years justnbsp;before the Norman Conquest, must have been at thisnbsp;time very young, and did not succeed to either kingdom. Gwyned fell to lago, the son of that Idwal whonbsp;had possessed it in defiance of Maredud ab Owain; butnbsp;possibly he did not make good his claims till some timenbsp;after ILewelyn�s death.'^ But it was only to Gwyned thatnbsp;he succeeded, for Deheubarth was seized by Rhyderchnbsp;ab lestyn.^ Though possessing neither of the two principalnbsp;divisions of his dominions by a legal title, ILewelyn�s rulenbsp;left a marked impression upon the Welsh people. According to the � Brut,� � In his time it was usual for the elders ofnbsp;his kingdom to say that his dominion was from one sea tonbsp;the other complete in abundance of wealth and inhabitants;

' Meurig was killed. �Brut,�i.a. 1019. � Book of ILan Dav � (Oxf.), p. 2cxx � �Brut,��r.�. 1020.

� �Brut,�j.a!. 1021. �Ann. Cam.,� s.a. 1023. Caradog, p. 86.

�* � Brut,� s.a. 1031, says : �And then lago ab Idwal held the government of Gwyned after ILewelyn.�

^ We infer this from later events recorded in the � Brut.�

CADWALADR TO NORMAN CONQUEST. 163

so that it was supposed there was neither poor nor destitute in all his territories ; nor empty trev nor any deficiency.� ^nbsp;Though we cannot fix the exact year in which lago abnbsp;Idwal obtained a real possession of Gwyned, it seemsnbsp;clear that his reign was uneventful. Nor did anything ofnbsp;special consequence happen in the south. Great events,nbsp;however, had been taking place in England, for the Danishnbsp;king Cnut had become in 1016 ruler of all England,� whichnbsp;he divided into the great ealdormanries of Wessex, Mercia,nbsp;East Anglia, and Northumbria. Cnut�s accession broughtnbsp;peace to the whole country. The Danish pirate fleetsnbsp;(speaking broadly) ceased to ravage the coast of the island,nbsp;while indirectly the vigorous government of Cnut benefitednbsp;Wales as well as the territories under his more direct rule.nbsp;English manufactures and trade began to make somenbsp;progress. Worcester was growing to be a place of importance, and Gloucester was rapidly rising to a position whichnbsp;enabled it in the years after the Norman Conquest tonbsp;exercise a marked influence on the development ofnbsp;South Wales. In the north, Chester, restored as we havenbsp;seen some one hundred years before by .^thelflaeda,nbsp;was now a centre of commerce and the common meetingnbsp;ground of Irish, Welsh, Cumbrians, English, and Danes.nbsp;The quiet of lago�s reign is probably largely explained bynbsp;these and other more general circumstances. Deheubarthnbsp;and the small principalities to the south-east enjoyed annbsp;unwonted immunity from external attack, but there wasnbsp;as usual internal trouble. Rhyderch of Deheubarth wasnbsp;slain, it is said by Irish-Scots, in 1031 or 1033.� Howelnbsp;and Maredud, sons of Edwin,^{1 2} took his place, but a year

1

S.a. 1020.

2 nbsp;nbsp;nbsp;Green, � Conqu. of Eng.,� pp. 411 et seq.

3 nbsp;nbsp;nbsp;�Brut,�r.a!. 1031. �Ann. Cam.,� r.a. 1033.

2

Apparently this Edwin was son of Einion, one of the grandsons of Howel Da ; see the genealogical table at end of this chapter.

M 2

THE WELSH PEOPLE, (chap, v.)

afterwards the sons of Rhyderch revolted, and a battle was fought at Hiraethwy, in which the latter were probablynbsp;defeated. Maredud ab Edwin was soon after killed in annbsp;obscure conflict,^ and Howel, his brother, was left in solenbsp;possession of Deheubarth, though Gruffyd ab Rhyderchnbsp;survived to create further disturbances in after years.

The peace of Gwyned was some six years after these events broken by the assertion of his claims by Gruffydnbsp;ab ILewelyn ab Seisyftt, who had, though still young, bynbsp;that time reached manhood. Of his early years nothingnbsp;is known. The immediate occasion of his attack uponnbsp;Gwyned appears to have been that one lestyn ab Gwrgant,nbsp;having ravished Gruffyd�s cousin Arden, the daughter ofnbsp;Robert ab Seisyftt, fled to lago, who gave him his protection. Gruffyd thereupon raised a force, engaged the armynbsp;of lago, slew the king, and seized his kingdom.^

It was during the reign of Gruffyd ab ILewelyn�extending from 1039 to 1063�that the Cymry reached the point of greatest strength since the death of Cadwaladr, and that fornbsp;the first time for many years their leader was able to exercisenbsp;an appreciable influence on affairs beyond the border. Innbsp;103 s Cnut had died. At the meeting of the Witan atnbsp;Oxford held after his death, notwithstanding the resistancenbsp;of the powerful Godwine, Earl of Wessex, who endeavourednbsp;to enforce the will of Cnut in favour of Harthacnut,nbsp;Harold Harefoot was chosen king, with the aid of Leofric,nbsp;Earl of Mercia. Godwine�s influence was, however, strongnbsp;enough to secure part of the late king�s dominions for thenbsp;younger son of Cnut, who was recognised as king ofnbsp;Wessex. It was after England had been once morenbsp;divided, and the house of Godwine had received a temporarynbsp;check in its path of aggrandisement, that Gruffyd becamenbsp;king of Gwyned. Nearly ninety years had passed since

' �Brut,� 1033. �Ann. Cam.,� j.a. 1035.

quot; � Brut,� j.a. 1037. �Ann. Cam.,� j.n. 1039.

CADWALADR TO NORMAN CONQUEST. 165

the death of Howel Da. They had been for the most part, as the bald narrative we have been able to give shows,nbsp;years of almost continued internal confusion, of bordernbsp;troubles, and of vexatious invasions from beyond the sea.nbsp;But the career of Gruffyd ab ILewelyn seems to show thatnbsp;the conflicts that had been waged and the events that hadnbsp;taken place had not sensibly affected the power of thenbsp;Cymric clans as a whole. It is difficult to avoid the inference that the wars during this disturbed period, of whichnbsp;the recollection is preserved in the � Brut,� were on a merelynbsp;petty scale�an inference strengthened, of course, by ournbsp;knowledge that the population was very small.^

However this may be, there can be no doubt that under the leadership of Gruffyd the Cymry suddenly developednbsp;an amount of military capacity and activity which had notnbsp;been displayed for centuries, and which resulted in theirnbsp;becoming a factor of some considerable importance in thenbsp;affairs of the whole island. The divisions of race in England,nbsp;the rivalries of the great Earls, and other circumstances,nbsp;combined to assist Gruffy^ in uniting the forces of Wales,nbsp;consolidating his own position, and making himself notnbsp;only the predominant chieftain in Wales, but a dangerousnbsp;and powerful foe to the English king, or at any rate to thenbsp;house of Godwine. One reason no doubt was that afternbsp;the first year of his reign his policy�one consistentlynbsp;pursued�was to remain on friendly terms with the Earlnbsp;and people of Mercia, or rather the English part of thenbsp;old kingdom of Mercia, then forming the ealdormanry ofnbsp;Leofric. Of the personal characteristics of this the greatestnbsp;military chief of the Cymry (except, perhaps, ILewelyn abnbsp;lorwerth, who was to exhibit similar qualities two hundrednbsp;years later), we know nothing except what may benbsp;inferred from his deeds. The burst of literary activitynbsp;which commenced among the Cymry shortly after thenbsp;* See the Introduction, above.

THE WELSH PEOPLE, (chap, v.)

Norman Conquest has preserved for us much information about later and far less important princes and warriors;nbsp;but though Gruffyd, like them, had bards in his trainnbsp;and at his court, not even fragments of their poems havenbsp;come down to enlighten us concerning the lord whom theynbsp;doubtless delighted to honour. For Gruffyd�s life, as innbsp;regard to those of his predecessors, we must rely simply onnbsp;the short entries in the chronicles.

Gruffyd�s character showed itself at once. In the very year of his accession he led a raid over the border intonbsp;Mercia, and beat the English forces in a battle at Rhyd-y-Groes,i on the Severn, in which Eadwine, brother of Earlnbsp;Leofric, was slain. This event does not seem to havenbsp;embroiled him with the Mercian Earl, for henceforth we findnbsp;him in alliance with the house of Leofric, and many yearsnbsp;elapsed before he again invaded England. He immediatelynbsp;turned his attention to Deheubarth. Howel ab Edwin, whonbsp;was now possessed of South Wales as a result of the defeatnbsp;of the sons of Rhyderch, was, without any considerablenbsp;pause, attacked by Gruffyd, and defeated in an encounternbsp;at ILanbadarn. Howel was forced to fly to the Irish Norsemen for assistance. Two years afterwards, with theirnbsp;support, he returned to Wales, and penetrated into Kere-digion, but was again beaten by Gruffyd in an engagementnbsp;at Pencader, which was of a decisive character, and innbsp;which the victor captured Howel�s wife, whom he tooknbsp;as his mistress. Howel�s resources were not, however,nbsp;exhausted, and, by one of the sudden changes of fortunenbsp;characteristic of the period, in the following year (1042)nbsp;Gruffyd was himself beaten, with the aid of the �blacknbsp;Pagans,� and taken prisoner, at Pw� Dyvach. Somehow�nbsp;probably by payment of a ransom�Gruffyd regained his

' Literally �the ford of the Cross.� The chief authorities for the life of Gruffyd are the � Brut.� and �Ann. Cam.,� with the English Chronicle. Seenbsp;his life in �Diet. Nat. Biog.�

CADWALADR TO NORMAN CONQUEST. 167

liberty, and returned to his kingdom. Two years after Howel, who seems to have once more gone to Ireland, camenbsp;back with a fleet, and, sailing round Dyfed, proceeded tonbsp;the Towy, but was defeated and lost his life in a battle withnbsp;the army of Gruffyd at Abertowy.'^

This victory secured Deheubarth for the king of Gwyned, though his troubles in that part of Cymrunbsp;were not over. Howel, as we have seen, had himselfnbsp;violently usurped the crown of Deheubarth, after expelling Rhyderch. Two sons of the latter, Gruffyd andnbsp;Rhys, saw in Howel�s overthrow an opportunity ofnbsp;asserting the claims of their house. How far they werenbsp;able to obtain actual possession of the whole or any part ofnbsp;South Wales is not clear; but probably it was Gruffyd abnbsp;ILewelyn who was the actual ruler, while the sons of Rhyd-erch from time to time attacked him or his subordinatenbsp;lords. About ten years, however, elapsed from the defeatnbsp;of Howel before Gruffyd was able finally to suppress thenbsp;house of Rhyderch. That he was strengthening himselfnbsp;with prudence is shown by his peaceful attitude towardsnbsp;Edward the Confessor�s government, and his close relationnbsp;to the Mercian Earl; and when Swein son of Godwine, in ornbsp;about 1045, was Earl of the south-western part of the oldnbsp;kingdom of Mercia, he joined Gruffyd ab ILewelyn innbsp;an expedition against the sons of Rhyderch. His friendship with the house of Mercia was cemented by his marriagenbsp;with Ealdgyth, daughter of JElfgar, the son of Leofric, whonbsp;afterwards became the wife of Harold H. Gruffyd also

� This place is not to be confounded with Aberteivi in Keredigion, nor with Abertawe. Abertowy occurs in the Twrch Trwyth hunt. At �Aber Tywi�nbsp;it was that Twrch Trwyth turned to bay and killed Kynlas son of Kynan, andnbsp;Gwilenhin, king of France. See Rhys�s paper in �Transactions of Cymmro-dorion Society, � 1894-5, P- l�- Abertowy was on the peninsula between thenbsp;Tow)� and the Gwendraeth. About three years ago a storm, carrying aw'aynbsp;parts of the sandbanks there, exposed the foundations of a row of houses.nbsp;(Account furnished to Professor Rhys by Mr. Drummond, agent for Lordnbsp;Cawdor.)

THE WELSH PEOPLE, (chap, v.)

succeeded in obtaining a grant from the English king of all the lands west of the Dee that had theretofore beennbsp;taken possession of by the English.

In 1047 he had to deal with a more than usually serious revolt in Deheubarth. The uchelwyr of Ystrad Towinbsp;suddenly rose and slew one hundred and forty of his men.nbsp;Gruffyd thereupon laid waste that district, as well as Dyfed.nbsp;This rising had probably some connection with the claimsnbsp;of Gruffyd ab Rhyderch and his brother. Two yearsnbsp;later there was further trouble in the south. The Irish alliesnbsp;of Gruffyd ab Rhyderch are said to have ravaged Deheubarth. We hear nothing further of this Gruffyd, exceptnbsp;that he was slain by Gruffyd ab ILewelyn in 1055,nbsp;his brother Rhys, too, disappears from the story.

Even before the death of Gruffyd ab Rhyderch the power of the North-Welsh king had become very considerable. He felt himself strong enough once more to invadenbsp;England. Of the circumstances that led to this course ofnbsp;action we have no information. Whatever the reason fornbsp;the raid, Gruffyd in 1052 penetrated into the land ofnbsp;Hereford, very nigh to Leominster, and fought the � landsmen as well as the Frenchmen of the Castle � on the samenbsp;day on which, thirteen years before, Eadwine had beennbsp;slain.i H�lfgar was outlawed in 1055, without, as thenbsp;English chronicler says, any guilt.� He fled to Irelandnbsp;and collected a fleet of eighteen ships, and with that forcenbsp;proceeded to Wales to Gruffyd, who received him intonbsp;his protection. Gruffyd and his father-in-law, havingnbsp;gathered together a great force, invaded England, andnbsp;defeated the English under Ralph the Earl near Hereford.nbsp;� Before there was any spear thrown the English peoplenbsp;fled because they were on horses, and there great slaughter

' �Eng. Chron,,� s.a, 1052. �And there were slain of the English very many good men, and also of the Frenchmen.�

quot; �Eng. Chron.,� J.a. 1055.

CADWALADR TO NORMAN CONQUEST. 169

was made about four hundred or five, and they made none on the other side.�^ The Welsh then took Hereford, burntnbsp;the town and the minster that the venerable Bishopnbsp;Hithelstan had built, and even slew the priests that werenbsp;within it and many others, and retired carrying awaynbsp;much booty.^

The incapable Ralph was replaced by Harold son of Godwine, and the importance of the Welsh victory is shownnbsp;by the fact that a great force was gathered � from wellnbsp;nigh all England.�� The English army met at Gloucester,nbsp;and started against the Welsh. Gruffyd appears to havenbsp;retreated, or at any rate avoided an engagement, andnbsp;Harold either did not desire or was not able to bring aboutnbsp;a battle. Some obscure negotiations took place betweennbsp;.dElfgar and Gruffyd on the one side, and Harold on thenbsp;other. The result was that peace was restored ; .TLlfgarnbsp;was in-lawed, but Gruffyd lost the lands beyond the Deenbsp;that had been granted to him by the king. One of thenbsp;copies of the English Chronicle says that when the Welshnbsp;� had done the utmost evil this counsel was counselled:nbsp;that Elgar (.^Elfgar) the Earl should be in-lawed and benbsp;given his earldom and all that had been taken from him.�nbsp;The fact that peace in accordance with this counsel wasnbsp;made is the strongest evidence of the formidable influencenbsp;of the Welsh king. Harold forthwith rebuilt Hereford, andnbsp;Bishop .iEthelstan having died, he appointed Leofgar, hisnbsp;mass-priest, to be Bishop on February 7, 1056.

The peace between Gruffyd and Harold was not long kept. In the summer of 1056 Gruffyd (who was probablynbsp;dissatisfied with the arrangements of the year before) again

' �Eng. Chron.,� s.a. 1055.

� �Eng. Chron.,� f.a. 1055. Cf. �Brut.,� j.a. 1054. This entry describes the engagement as a �severely hard battle,� and says the Saxons took to flightnbsp;unable to bear the assault of the Britons, and fell with a very great slaughter.

^ �Eng. Chron.,� r.a. 1055.

170 THE WELSH PEOPLE, (chap, v.)

invaded south-west Mercia. He was met by the new Bishop�who had worn � his knapsack during his priesthood till he was a bishop,� and then �forsook his chrismnbsp;and his rood, his ghostly weapons, and took to hisnbsp;spear and sword after his bishophood��and ^Ifnoth,nbsp;the Sheriff, at the head of the Mercian forces, eight daysnbsp;before Midsummer.^ Gruffyd was again victorious; bothnbsp;the Bishop and the Sheriff were slain, and with themnbsp;�many good men.� Gruffyd seems to have followed upnbsp;his success, for the English chronicler found it �difficultnbsp;to tell the distress, and all the marching and all thenbsp;camping, and the travail and destruction of men andnbsp;also of horses which all the English army sustained.�nbsp;But Leofric, Harold, and Bishop Ealdred of Worcesternbsp;came to Gruffyd and succeeded, though we know not onnbsp;what terms, in quieting him. A reconciliation was effected,nbsp;and Gruffyd swore oaths that he would be to Kingnbsp;Edward � a faithful and unbetraying under-king.� ^

Two years after this event, however, it is recorded in the � Brut � that Magnus, son of Harold, described as �king ofnbsp;Germany,� came to England and ravaged the dominions ofnbsp;the Saxons, and that Gruffyd was his �conductor andnbsp;auxiliary.�� The English Chronicle says that a fleet camenbsp;from Norway in 1058, but does not connect this event withnbsp;Gruffyd; but that there were hostilities between him andnbsp;the English in that year is clear, for vElfgar the Earl, whonbsp;succeeded his father on his death in 1057, was banished, butnbsp;soon returned, with the aid of his son-in-law.^ The Welshnbsp;king was now at the height of his power. So long as .^Elfgarnbsp;lived,however, he seems to have kept the peace. Harold himself had taken possession of the earldom of the Magesaetas

3 �Brat,� s.a. 1056. In �Ann. Cam.seemingly 1058. * �Eng. Chron.,� j.a. 1058

CADWALADR TO NORMAN CONQUEST. 171

and the course of the Severn (i.e., south-west Mercia), no doubt with the intention of holding ^Elfgar and Gruffydnbsp;in check. The exact date of the former�s death is notnbsp;known, but it is probable that he died in 1062. Very likelynbsp;border raids into England were made by the Welsh, ornbsp;some other provocation was given by Gruffyd, for in 1063nbsp;Harold determined to make a strong attempt to crush hisnbsp;dangerous and now too formidable neighbour. The chiefnbsp;palace of Gruffyd was at Rhudlan, which was a site ofnbsp;military value, since it dominated the Vale of Clwyd, and wasnbsp;then a seaport. Itwas against Rhudlan that Harold directednbsp;his first blow. With a small band (probably his own house-carls) he hastened there at the end of 1062, and surprisednbsp;Gruffyd, who, however, escaped by sea. Unable to follow,nbsp;and not strong enough to winter in North Wales, Haroldnbsp;contented himself with burning the house and the remainingnbsp;ships, and set out back for Gloucester on the same day.^nbsp;It may be conjectured from the subsequent course of affairsnbsp;that this event did much to damage Gruffyd�s prestigenbsp;among the loosely united Welsh clans, and especially amongnbsp;the uchelwyr of South Wales. Such a career as his,nbsp;in the circumstances of his time, must have creatednbsp;jealousies and involved the enmity of many families amongnbsp;the Cymry. Only six years had elapsed since the finalnbsp;overthrow of the house of Rhyderch. Many men mustnbsp;have been waiting for the time when a reverse gave anbsp;favourable opportunity for revolt.^

It was immediately after his sudden raid on Rhudlan that Harold, now the most powerful subject of England�indeed,nbsp;its real ruler�planned a systematic invasion of Wales. Henbsp;collected a fleet at Bristol with a view to coasting round thenbsp;country, while he arranged that his brother Tostig� should

1 nbsp;nbsp;nbsp;FI. Wigorn, 1063.

2 nbsp;nbsp;nbsp;Note Giraldus� reference to Gruffyd as one �who by his tyranny for anbsp;long time had oppressed Wales � : � Itin. Cam.,� book i., ch. 2.

� Tostig had become Earl of Northumbria.

THE WELSH PEOPLE, (chap, v.)

cross the border from Northumbria with a land force. At Rogation-tide (in 1063) Harold left Bristol with his fleet andnbsp;sailed along the coast, presumably landing at points wherenbsp;damage could be inflicted on the Welsh.^ Tostig, actingnbsp;in conjunction with his brother, crossed the border. Fornbsp;several weeks the English carried on a vigorous and far-reaching harrying of Gruffyd�s dominions. There is nonbsp;record of any pitched battles, and the warfare was clearlynbsp;of the guerilla kind. Taught by experience, the Englishnbsp;leaders changed their method of fighting. They madenbsp;their men discard armour and give up the close array.^nbsp;Lightly armed, they fought on the same terms as theirnbsp;active enemies. The Cymry defended themselves withnbsp;stubbornness, but the English won many skirmishes andnbsp;gave no quarter. The former suffered more severely thannbsp;at any time since the death of Cadwaladr.� The campaignnbsp;seems to have been carried on over a large part of Wales.nbsp;The result was that � the people � (so says the Englishnbsp;Chronicle) made a truce with Harold and deliverednbsp;hostages.^ Of Gruffyd himself during these weeks in thenbsp;summer of 1063 we hear nothing. Later authority saysnbsp;the Welsh sentenced him to deposition. What is certainnbsp;is that he was slain in August by Welshmen�slain,nbsp;according to the English chronicler, because �of the warnbsp;he waged with Harold the Earl��slain, according to thenbsp;Brut, by the treachery of his own men. � The shield andnbsp;defender of the Britons . . . the man who had been hithertonbsp;invincible, was now left in the glens of desolation, afternbsp;taking immense spoils and after innumerable victories andnbsp;countless treasures of gold and silver, and jewels and purple

1 nbsp;nbsp;nbsp;�Eng. Chron.,�G.3. 1063.

2 nbsp;nbsp;nbsp;See Freeman, �N.C.,�ii. 480.

� Giraldus (writing about 140 years later) says that Harold left scarcely a man alive in Wales.

�Eng. Chron.,� s.a. 1063.

CADWALADR TO NORMAN CONQUEST. 173

vestures.�^ The effects of Harold�s merciless ravaging were long felt. His victory plunged Wales once more intonbsp;confusion, and no doubt contributed to the comparativelynbsp;swift conquest of a great part of the South by the Normansnbsp;a few years afterwards.

However ruthless Harold had shown himself in the campaign, so soon as Gruffyd had been got rid of henbsp;proceeded to arrange a new settlement of Welsh affairs.nbsp;The kingdom of the dead chieftain was divided betweennbsp;Bledyn ab Cynfyn and his brother Rhiwallon,¹� but furthernbsp;considerable portions of Cymric land were added to thenbsp;shires or earldoms on the border. The Vale of Clwyd, ornbsp;the greater part of it, was added to the shire of Chester,nbsp;and seemingly passed under the rule of Eadwine, son ofnbsp;^Ifgar, Earl of Mercia. The whole or a large part of whatnbsp;is now Radnorshire became an English possession. Partnbsp;of Gwent, though we cannot define what part (probably thenbsp;land between the Wye and the Usk), was united to thenbsp;earldom of Harold.� How far, or in what sense, these lands

1

�Brut.,� s.a. io6i. �Ann. Cam.,� s.a. 1063.

^ The �Brut quot;and �Ann. Cam.� are silent as to this, but the subsequent entries confirm the transaction. The �Worcester Chronicle� (1063) records it�makingnbsp;Bledyn and Rhiwafton brothers of Gruffyd; and in the �Brut,� s.a, 1068,nbsp;they are referred to as still reigning, and are mentioned as sons of Cynfyn.nbsp;They were really half-brothers of Gruffyd. Their mother was Angharadnbsp;(daughter of Maredud), who married ILowelyn ab Seisyift, and also Cynfyn.nbsp;Probably the latter was her second husband. The � Brut,� s.a. 1112, explainsnbsp;the relationship. See Freeman, � Norman Conquest,� ii.,p. 483, n. i. In thenbsp;�Brut,�r.a. 1073, Bledyn ab Cynfyn is referred to as �the man who afternbsp;Gruffyd his brother nobly supported the whole kingdom of the Britons.�nbsp;The Gruffyd referred to is Gruffyd ab Hcwelyn.

� See Freeman, quot; N. C.,� v. ii., p. 483-6 ; and note (��) in App., p. 707 (third edition, London, 1877). Harold, it seems, began to build a huntingnbsp;seat at Forth Iscoed. Caradog, son of that Gruffyd who claimed Deheubarthnbsp;and who was slain by Gruffyd ab Lewelyn, made a raid upon the workmennbsp;engaged in the building, slew nearly all of them, and carried away thenbsp;provisions and other things that Harold had collected. � Chronn. Ab. etnbsp;Wig.,� 1065; �Domesday,� 162: �Sub iisdem prsepositis sunt iiii, villcenbsp;wastatEe per regem Caradnech.�

THE WELSH PEOPLE, (chap, v.)

(except the Vale of Clwyd) had been part of Gruffyd's possessions we have no means of determining ; but it isnbsp;clear that one of the results of the Earl�s successful war wasnbsp;that the English made a further advance on Welsh territory.

There can be no doubt that this comparatively great Welsh campaign largely increased the prestige ofnbsp;Harold and his house, and was one of the circumstancesnbsp;which led to his being elected king on the death ofnbsp;Eadward the Confessor early in 1066. He did not enjoynbsp;his high office long, for on the 14th of October in thenbsp;same year he fell in the battle of Hastings or Senlac,nbsp;resisting the Norman invasion, and shortly afterwardsnbsp;William, Duke of Normandy, became king of England.nbsp;We need say nothing for our purpose as to the generalnbsp;circumstances and effect of this Norman conquest, but henbsp;who looks at it from a Cymric standpoint will note withnbsp;curious interest the satisfaction of the Welsh chronicler whonbsp;records that Harold, who had been previously � vauntinglynbsp;victorious,� was despoiled of his life and kingdom bynbsp;William the Bastard �Tywysog� of Normandy, and thatnbsp;� that William � defended the kingdom of ILoegr in a greatnbsp;battle �with an invincible hand and his most noble army.�^

As appendices to this chapter we insert in face of this page

(A) A Chronological Table of the Kings of England and the Kings or Princes of Gwyned and Deheubarth.nbsp;The dates of the accession of the latter are taken

* �Brut,�j.a. 1066. The last sentence is, � Ar G6ilim h6nn6 dr6y dirua�r ur�ydyr a ymdiffynna�d teyrnas Loegr o an orchyfegedic Ia6 a uonhedickaf lu �nbsp;(see � Red Book of Hergest,� vol. ii., p. 268 : Oxford edition). Though atnbsp;the time ILoegr denoted much the same area of the island as England at present,nbsp;yet it did not connote, when the chronicler wrote, all that the word England doesnbsp;now to us. The notion that William was defending ILoegr should be observed.nbsp;The student should also notice that the application of the term tywysog tonbsp;William in regard to Normandy shows it was then used in a very generalnbsp;sense.

TABLE B.

THE HOUSE OF RHODRI. Rhodri Mawr

(rf. 877)

I . I.

Rhodri Edwin [d. 951) (d. 952)

Idwal (d. 960)

A daughter = Lewelyn ab SeisyHt (d. 1021)

Gruffyd (d. 1063)