Skip to main content

Full text of "A history of Wales from the earliest times to the Edwardian conquest"

See other formats












5 1959 


IN this work it has been my endeavour to bring to- 
gether and to weave into a continuous narrative what 
may be fairly regarded as the ascertained facts of the 
history of Wales up to the fall of Llywelyn ap Gruffydd 
in 1282. In a field where so much is matter of conjec- 
ture, it has not been possible altogether to avoid specu- 
lation and hypothesis, but I can honestly say that I have 
not written in support of any special theory or to urge 
any preconceived opinion upon the reader. My purpose 
has been to map out, in this difficult region of study, 
what is already known and established, and thus to 
define more clearly the limits of that "terra incognita" 
which still awaits discovery. The task has not been 
attempted in English since Miss Jane Williams (Ysga- 
fell) published her History of Wales in 1869, and it 
cannot be doubted, therefore, that it was time to under- 
take it anew. 

The enterprise, it need scarcely be said, has been a 
laborious one, and, as the occupation of somewhat limited 
hours of leisure, has been spread over a considerable 
number of years. In some respects this may have been 
an advantage, but it has entailed certain drawbacks also. 
Had the earlier chapters been written more recently, 
they might have owed more than they do to the study 
of such works as Dr. Holmes' Ancient Britain and 
Professor Bury's Life of St. Patrick. For this and 


many other shortcomings I can but crave the indulgence 
of the reader. 

It has been my endeavour to indicate, in the foot- 
notes and elsewhere, my innumerable obligations to 
other workers in this field of study. But I should wish 
here to express my general indebtedness to Sir John 
Rhys, Mr. Egerton Phillimore, Mr. Alfred N. Palmer, 
and Dr. Hugh Williams for the pioneer work which has 
so greatly facilitated the scientific study of Welsh history. 
I owe to them what cannot be expressed in the debit of 
citation and reference, namely, outlook and method and 

For assistance given to me ungrudgingly during the 
progress of the work, I desire to thank Principal J. R. 
Ains worth Davis, M.A., Professor T. F. Tout, M.A., 
Professor J. Morris Jones, M.A., Professor W. Lewis 
Jones, M.A., Mr. Edward Greenly, F.G.S., Mr. Percy 
G. Thomas, M.A., Mr. O. T. Williams, M.A., and the 
Rev. T. Shankland. 

Most of the primary authorities used are discussed in 
some part or other of the book. The reader may notice, 
however, that nowhere is there any full and systematic 
discussion of the chronicles included in Annales Cam- 
bria and Brut y Tywysogion. I had originally in- 
tended to include a critical account of these authorities 
in the work, but afterwards came to the conclusion that 
the task was too ambitious for the present occasion and 
must be separately undertaken. Let it suffice here to 
say that I have throughout treated Brut y Tywysogion 
and Brut (or Brenhinedd] y Saeson as two independent 
translations of a Latin original partially (but by no 
means fully) represented in MSS. B. and C. of Annales 

The Map is intended to be of general service to 


those who may use the book, and does not reproduce 
the political divisions of Wales at any definite point in 
its history. For North Wales, however, it is approxi- 
mately correct for the year 1 200. Cantrefs are usually 
indicated, but in Anglesey, Powys, Ceredigion and 
Morgannwg, commotes are shown as there the more 

In the spelling of Welsh names, I have sought to 
observe the rules laid down in 1893 by the Orthogra- 
phical Committee of the Society for Utilising the Welsh 

My thanks are due to Miss E. M. Samson for the 
compilation of the Index. 


BANGOR, ist^November, 1910. 














3. WALES AT THE CHRISTIAN ERA . ... i . . . -37 






NOTE TO i. CARAT Acus 89 


1. BRITAIN LOST TO THE EMPIRE . . .... . . .91 











2. GlLDAS 134 


NOTE TO i. HARLBIAN MS. 3859 159 

























1. THE CENBDL 283 

2. THE TREF agx 

3. THE CANTREF ..." 300 

4. THE KING 308 











Page 43, line 8. For understand, read understood. 
129, ,, 17. ,, form, ,, from. 

132, ,, 6. Aurelius, ,, Aurelianus. 


A. S. Chr. The Anglo-Saxon or English Chronicles are quoted by reference to 
the annal and sometimes to the MS. also. The text used is that of 
Earle and Plummer (" Two of the Saxon Chronicles parallel," revised 
text, Oxford, 1892) ; references to the introduction and notes of this edi- 
tion are given as Plummer, ii. 

Ang. Sac. Anglia Sacra [ed. Hen. Wharton]. Two vols. London, 1691. 

Ann. ad 1298. Annals from A.D. 600-1298 written in Breviate of Domesday 
(K. R. miscellaneous books i.), if. 29-35, an ^ printed in Arch. Camb. III. 
viii.'(i862), 273-283. 

Ann. C. Annales Cambriae. Ed. J. Williams Ab Ithel (Rolls Series). 1860. 

Ann. Cest. Annales Cestrienses [Mostyn MS. 157]. Ed. R. C. Christie (Lanca- 
shire and Cheshire Record Society, vol. xiv.). London, 1887. 

Ann. Dunst. Annales Prioratus de Dunstaplia in vol. iii. of Annales Monastic! 
(Rolls Series). London, 1866. 

Ann. Marg. Annales de Margan in vol. i. of Annales Monastic! (Rolls Series). 
London, 1864. 

Ann. Osen. Annales Monasterii de Oseneia in vol. iv. of Annales Monastic! 
(Rolls Series). London, 1869. 

Ann. S. Edm. Annales S. Edmund! [Bury St. Edmund's] printed by Lieber- 
mann in Ungedruckte Anglo - Normanische Geschichtsquellen from 
Harl. MS. 447. Strassburg, 1879. 

Ann. Theokesb. Annales de Theokesberia in vol. i. of Annales Monastic! 
(Rolls Series). London, 1864. 

Ann. Ult. The Annals of Ulster or Annals of Senat. Issued by the Royal 
Irish Academy (Dublin). 

Vol. i. 431-1056 (ed. W. M. Hennessy). 1887. 
ii. 1057-1378 (ed. B. MacCarthy). 1893. 
iii. 1379-1541 ( ). 1895. 

Ann. Waved. Annales Monasterii de Waverleia in vol. ii. of Annales Monastici 
(Rolls Series). London, 1865. 

Ann. Wigorn. Annales Prioratus de Wigornia in vol. iv. of Annales Monastici 
(Rolls Series). London, 1869. 

Ann. Wint. Annales Monasterii de Wintonia in vol. ii. of Annales Monastici 
(Rolls Series). London, 1865. 

Antt. Legg. De Antiquis Legibus Liber; Cronica Maiorum et Vicecomitum 
Londoniarum. Ed. T. Stapleton (Camden Society). London, 1846. 

App. Land Com. Bibliographical, Statistical, and other miscellaneous Memor- 
anda, being appendices to the Report of the Royal Commission on Land 
in Wales and Monmouthshire [Compiled chiefly by the Secretary, D. 
Lleufer Thomas]. London, 1896. 



Arch. Camb. Archaeologia Cambrensis, the journal of the Cambrian Archseo- 

logical Association. The capital Roman numeral denotes the series (I. 

1846-9 ; II. 1850-4 ; III. 1855-69 ; IV. 1870-83 ; V. 1884-1900 ; VI. to date), 

the uncial letter the volume (to which the year is added in brackets), 

and the Arabic numeral the page. 
Arth. Legend. Studies in the Arthurian Legend. By John Rhys. Oxford, 

Asser. The Life of Alfred is quoted, by reference to the chapter, from the 

edition of W. H. Stevenson (Oxford, 1904). The notes, etc., of this 

edition are cited as Stev. 

B. Saes. The Welsh chronicle in Cottonian MS. Cleopatra B. v. ff. 111-1646, 
there styled " Brenhined y Saesson," but by the Myvyrian editors " Brut 
y Saesson " (Myv. Arch. II. 468-582 [652-684]), is cited by reference to the 

B. T. Brut y Tywysogion. Ed. J. Williams Ab Ithel (Rolls Series). 1860. 

B. Willis, Bangor. Survey of the Cathedral Church of Bangor, with an ap- 
pendix of records. Collected by Browne Willis. London, 1721. 

Bede, H. E. The " Historia Ecclesiastica Gentis Anglorum " is cited, by refer- 
ence to book and chapter, from vol. i. of " Baedae Opera Historica," 
edited by C. Plummer (Oxford, 1896). References to the introduction, 
notes, etc., of this edition are given as Plummer's Bede, with no. of 
volume and page. 

Be"mont. Simon de Montfort, comte de Leicester. Par Ch. Be"mont. Paris, 

Ben. Abb. Gesta Regis Henrici Secundi Benedict! Abbatis. Two vols. Ed. 
W. Stubbs (Rolls Series). 1867. 

Blk. Bk. The Black Book of Carmarthen (Peniarth MS. i = Hengwrt MS. n) 
is cited from the facsimile edition issued by J. Gwenogvryn Evans 
(Oxford, 1888) and reproduced by the same editor in print (Pwllheli, 
1906). Another printed text will be found in IV. Anc. Bks. ii. 3-61. 

Blk. Bk of St. David's. The Black Book of St. David's [an extent of the lands 
and rents of the bishop in 1326 = Br. Mus. Add. MS. 34,125]. Ed. J. 
W. Willis-Bund (Cymmrodorion Society). London, 1902. 

Breconsh. (2). A History of the County of Brecknock. By Theophilus Jones. 
Originally issued in two vols. (1805, 1809) ; reissued in one by Edwin 
Davies. Brecknock, 1898. 

Britannia. By W. Camden. Cited from the edition of 1600. (George Bishop, 

Bruts. The text of the Bruts [Brut y Brenhinoedd, Brut y Tywysogion, etc.] 
from the Red Book of Hergest. Edited by John Rhys and J. Gwenogvryn 
Evans. Oxford, 1890. 

Buch. GT. ap C. " Buchedd Gruffydd ap Cynan " is cited (by page) from Arch. 
Camb. III. xii. (1866), 30-45, 112-128, with a further reference in brackets 
to the page of Myv. Arch., second edition. 

Bye-Gones. Notes contributed to the " Bye-Gones " column of the (weekly) 
Oswestry Advertiser and separately published in yearly half-volumes. 

C.I.L. Corpus Inscriptionum Latinarum. Vol. vii. (ed. Hu'bner, Berlin, 1873) 
deals with the Latin inscriptions of Britain of older date than 500 A.D. 

Cal. Close R. Calendar of the Close Rolls, prepared under the superintendence 
of the Deputy Keeper of the Records. 

Edward I. vol. i. 1272-9. London, 1900. 

Cal. Doc. Fr. Calendar of Documents preserved in France illustrative of the 


History of Great Britain and Ireland [from 918 to 1206]. Ed. J. H. Round. 
London, 1899. 

Cal. Pat. R. Calendar of the Patent Rolls, prepared under the superintendence 
of the Deputy Keeper of the Records. 

Henry III. vol. i. 1232-47. London, 1906. 
ii. 1247-58. 1908. 
Edward I. vol. i. 1272-81. xgoi. 
Camb. Biog. The Cambrian Biography. By William Owen [W. O. Pughe]. 

London, 1803. 

Camb. Qu. Mag. The Cambrian Quarterly Magazine. London, 1829-33. 
Camb. Reg. The Cambrian Register [edited by W. O. Pughe]. London. 
Vol. i. 1796. 
ii. 1799. 
Hi. 1818. 

Cambro-Br. SS. Lives of the Cambro-British Saints. Ed. W. J. Rees (Welsh 
MSS. Society.) Llandovery, 1853. 

[For a long list of corrections to be made in Rees* text, see Cymr. 

xiii. 76-96 (Kuno Meyer).] 
Card. Priory. Cardigan Priory in the Olden Days. By E. M. Pritchard. 

London, 1904. 
Carlisle, Top. Diet. A Topographical Dictionary of the Dominion of Wales. 

By Nicholas Carlisle. London, 1811. 

Carm. Cart. Cartularium S. Johannis Bapt. [corrige Evang.] de Caermarthen 
[Hengwrt MS. 440]. Privately printed for [Sir] T[homas] P[billips]. 
Cheltenham, 1865. 

Carnh. Hanes Cymru. Gan T. Price (Carnhuanawc). Crughywel, 1842. 
Cart. Brec. See chapter xii. note 128. 

Cart. Glouc. Historia et Cartularium Monasterii Sancti Petri Gloucestriae 
Ed. W. H. Hart (Rolls Series). London. 
Vol. i. 1863. 
ii. 1865. 
iii. 1867. 

Cart. Sax. Cartularium Saxonicum. Ed. W. de Gray Birch. London. 

Vol. i. 1885. 

ii. 1887. 

iii. 1893. 

index, 1899. 

Cartae Gfam. Cartae et alia munimenta quae ad Dominium de Glamorgan 
pertinent. Curante Geo. T. Clark. 
Vol. i. Dowlais, 1885. 
ii. Cardiff, 1890. 
iii. 1891. 
iv. 1893. 
Celt Br. Celtic Britain. By John Rhys. London. Usually cited from the 

(second) edition of 1884 (2), but sometimes from that of 1904 (3). 
Celt. Folklore. Celtic Folklore. By John Rhys. Two vols. Oxford, 1901. 
Celt. Heath. The Origin and Growth of Religion as illustrated by Celtic 
Heathendom (Hibbert Lectures for 1886). By John Rhys. London, 
Celt. Remains. Celtic Remains. By Lewis Morris [1700-65], Ed. D. Silvan 

Evans (Cambrian Archaeological Association). London, 1878. 
Charter Rolls. Calendar of the Charter Rolls, printed under the superinten- 
dence of the Deputy Keeper of the Records. London. 
Vol. i. (1226-1257) 1903. 
ii. (1257-1300) 1906. 


Chron. Scot. Chronicum Scotorum [Trin. Coll. Dubl. MS. H. I. 18]. Ed. W. 

M. Hennessy (Rolls Series). London, 1866. 
Close Rolls. The Close Rolls, printed under the superintendence of the Deputy 

Keeper of the Records (full text). 

Henry III. vol. i. 1227-31. London, 1902. 
ii. 1231-34. 1905. 

iii. 1234-37. 1908. 

Cod. Dipl. Codex Diplomaticus Mvi Saxonici. Ed. J. M. Kemble (English 

Historical Society). Six vols. London, 1839-48. 
Cole, Docts. Documents illustrative of English History in the Thirteenth and 

Fourteenth Centuries. Ed. H. Cole. London, 1844. 
Comment. (2). Humfredi Llwyd Britannicae Descriptionis Commentariolum 

[first published, Cologne, 1572]. Accedunt Aerae Cambro- Britannicae. 

Accurante Mose Gulielmio [Moses Williams]. Londini, 1731. 
Conq. Eng. (2). The Conquest of England. By J. R. Green. Second edition. 

London, 1884. 
Cont. Fl. Wig. The continuation of the Chronicle of Florence of Worcester 

from 1118-1141, for which see Fl. Wig. ii. 71-136. 
Coxe (2). A Historical Tour Through Monmouthshire. By William Coxe [first 

published 1801]. Reissued by Davies & Co. Brecon, 1904. 
Cyff Beuno. Gan Eben Fardd. Tremadog, 1863. 
Cymr. The Cymmrodor, the magazine of the honourable society of Cym- 

mrodorion. London, 1877 to date. 

Davies, Diet. Antiquae Linguae Britannicae Dictionarium Duplex [Welsh- 
Latin and Latin- Welsh : by John Davies of Mallwyd]. Londini, 1632. 

De Nugis. Gualteri Mapes de Nugis Curialium [Bodleian MS. 851]. Ed. T. 
Wright (Camden Society). London, 1850. 

Diceto. Radulfi de Diceto Opera Historica. Ed. W. Stubbs (Rolls Series). 
Two vols. London, 1876. 

Diet. Nat. Biog. The Dictionary of National Biography. London, 1884-1904. 

Domesd. The Domesday Survey is cited from the Record edition [1783], by 
reference to the volume, the folio, the page (a or b) and the column ((i) 
or (2)). 

Dwnn. Heraldic Visitations of Wales. By Lewis Dwnn [fl. 1580]. Ed. S. R. 
Meyrick (Welsh MSS. Society). Two vols. Llandovery, 1846. 

Eadmer. The " Historia Novorum " and the " De Vita S. Anselmi " are 

cited from the edition of Martin Rule (Rolls Series). London, 1884. 
Eng. Hist. Rev. The English Historical Review. London, 1886 to date. 
Evans, Diet. A Dictionary of the Welsh Language. By D. Silvan Evans. 
Five parts. Carmarthen. A. 1887. B. 1888. C. 1893. CH. and D. 
1896. E-Enyd, 1906. 

Evans, Proverbs. Casgliad o Ddiarhebion Cymreig (A collection of Welsh 
Proverbs). By J. Gwenogvryn Evans. Printed in the Transactions of 
the Liverpool National Eisteddfod of 1884, pp. 528-584. Liverpool, 1885. 
Evans, Rep. Report on Manuscripts in the Welsh language, issued by the 
Historical Manuscripts Commission [The special commissioner for this 
purpose was J. Gwenogvryn Evans]. 

Vol. i. pt. i Mostyn MSS. London, 1898. 

,, ,, 2 Peniarth MSS. (first portion). 1899. 

3 Peniarth MSS. (second portion). ,, 1905. 
Vol. ii. pt. i MSS. at Oxford, Cardiff, etc. 1902. 

2 MSS. at Llanstephan. 1903. 

,, 3 Panton and Cwrtmawr MSS. I95- 


Eyton, Itin. The Court, Household, and Itinerary of King Henry II. By R. W. 

Eyton. London, 1878. 
Eyton, Shrops. Antiquities of Shropshire. By R. W. Eyton. Twelve vols. 

London, 1854-60. 

Fenton (2). A Historical Tour Through Pembrokeshire. By Richard Fenton 

[first published 1811]. Reissued by Davies & Co. Brecknock, 1903. 
Feudal Eng. Feudal England. By J. H. Round. London, 1895. 
Fl. Wig. Florentii Wigornensis Chronicon ex Chronicis. Ed. B. Thorpe 

(English Historical Society). Two vols. London, 1848. 

Flores. Flores Historiarum. Ed. H. R. Luard (Rolls Series). Vol. ii. (1067- 
1264). London, 1890. [This is the chronicle formerly known as that of 
" Matthew of Westminster ".] 
IV. Anc. Bks. The Four Ancient Books of Wales. By W. F. Skene. 

Vol. i. Introduction : translations ; vol. ii. Text : notes. Edin- 
burgh, 1868. 
This work includes the following texts : 

1. Black Book of Carmarthen ii. 3-61. (See Blk. Bk. above.) 

2. Book of Aneurin (Cardiff Publ. Libr. MS. i) ii. 62-107. 

3. Book of Taliesin (Peniarth MS. 2) ii. 108-217. 

4. Red Book of Hergest portions only (Jesus Coll. MS. i) ii. 


The translations are by D. Silvan Evans, except the Book of Taliesin. 
translated by Robert Williams of Rhydycroesau. 

Gaimar. Lestorie des Engles. By Geffrei Gaimar. Ed. T. D. Hardy and 
C. T. Martin (Rolls Series). London. 
Vol. i. (text) 1888. 

ii. (translation) 1889. 
Geoff. Mand. Geoffrey de Mandeville. A Study of the Anarchy. By J. H. 

Round. London, 1892. 
Gervase. The Historical Works of Gervase of Canterbury. Ed. W. Stubbs 

(Rolls Series). Two vols. London, 1879-80. 

Gesta St. Gesta Stephani. Ed. R. Howlett, in ' Chronicles of the reigns of 

Stephen, Henry II., and Richard I." (Rolls Series), vol. iii. London, 1886. 

References in brackets are to the edition of R. C. Sewell (English 

Historical Society), London, 1846. 

Gibson. Camden's Britannia, translated into English, with additions. Ed. 
Edmund Gibson. London, 1695. 

"The whole business of Wales was committed to the care of Mr. 
Edward Lhwyd, Keeper of the Musaeum in Oxford" (pref. to the 
Reader, p. 3). 

Gildas. Cited from " Monumenta Germaniae Historica," quarto series (Auctores 
Antiquissimi), tomus xiii. ed. T. Mommsen (Berlin, 1898). 

Gildas and Nennius were issued separately in 1894 as vol. iii. fasc. i, 

of " Chronica Minora saec. iv. v. vi. vii." 
" Williams, Gildas " is the edition by Hugh Williams (Cymmrodorion 

Society), London, 1899. 

For note on MSS. and editions of Gildas, see end of chap. v. 
Gir. Camb. Giraldi Cambrensis Opera. Ed. J. S. Brewer, J. F. Dimock, and 
G. F. Warner (Rolls Series). Eight vols. London, 1861-1891. 
Particular works are referred to as follows : 
Itin. Itinerarium Kambriae. 
Descr. Descriptio Kambriae. 


Gir. Camb. Giraldi Cambrensis Opera (continued). 

Particular works are referred to as follows : 
De Rebus. De Rebus a se Gestis. 
Top. Hib. Topographica Hibernica. 
Exp. Hib. Expugnatio Hibernica. 
Invect. De Invectionibus. 
Sym. El. Symbolum Electorum. 
Gemma. Gemma Ecclesiastica. 
Men. Eccl. De Jure et Statu Menevensis Ecclesiae. 
Spec. Speculum Ecclesiae. 
Princ. Instr. De Principis Instructione. 
Godwin (2). De Praesulibus Angliae. By F. Godwin. Ed. W. Richardson. 

Cambridge, 1743. 

Goss. Guide. The Gossiping Guide to Wales (North Wales and Aberystwyth). 

Traveller's edition, issued annually. London, Oswestry, and Wrexham. 

[Recent editions have been revised and amplified by Egerton Phillimore 

see pref.] 
Gr. Celt. (2). Grammatica Celtica. By I. C. Zeuss. Ed. H. Ebel. Berlin, 


Gw. ap Rhys. Hanes y Brytaniaid a'r Cymry. Gan Gweirydd ap Rhys [ac 
eraill]. Llundain. 
Cyf. i. 1872. 
Cyf. ii. 1874. 
Gw. Brut. The Gwentian Brut or " Brut Aberpergwm " is printed in Myv. 

Arch. II. 468-582 [685-715] from a MS. dated 1764. 

Gwydir Fam. The History of the Gwydir Family. By Sir John Wynne [1553- 
1626]. Oswestry, 1878. 

H. and St. Councils and Ecclesiastical Documents relating to Great Britain and 
Ireland. Ed. A. W. Haddan and W. Stubbs. Oxford. 
Vol. i. 1869. 

ii. pt. i 1873 ; pt. 2 1878. 

[For these two volumes, dealing with the Celtic churches, Mr. Haddan 
was responsible (see pref. to vol. i.), though the second appeared after 
his death.] 

Harl. MS. 3859. See note appended to chap. v. 
Hemingb. Chronicon Walteri de Hemingburgh. Ed. H. C. Hamilton (English 

Historical Society). Two vols. London, 1848-9. 
Hen. Hunt. Henrici Archidiaconi Huntendunensis Historia Anglorum. Ed. T. 

Arnold (Rolls Series). London, 1879. 

Hist. Britt. The " Historia Brittonum " usually coupled with the name of 
Nennius is cited from the same edition as Gildas (see above). For a 
brief account of the " Historia " and of " Nennius," see chap. vii. 3. 
Hist. Ch. York. The Historians of the Church of York and its Archbishops. 
Ed. James Raine (Rolls Series). London. 
Vol. {.1879. 
ii. 1886. 
iii. 1894. 

. Reg. The " Historia Regum Britanniae " of Geoffrey of Monmouth is 
usually cited from the edition of J. A. Giles (London, 1844), but references 
are occasionally made to the (unpublished) Berne MS. The readings 
of this MS. I give on the authority of Mr. G. B. Matthews, who col- 
lated it in 1898 and has kindly allowed me to make use of his notes. 
Hoare, Itin. The Itinerary of Archbishop Baldwin through Wales. Trans- 
lated by R. C. Hoare. Two vols. London, 1806. 


Hoveden. Chronica Magistri Rogeri de Houcdcne. Ed. W. Stubbe (Rolls 
Series). Four vols. London, 1868-71. 

Inq. p. mortem. Calendar of Inquisitions post mortem, prepared under the 
superintendence of the Deputy Keeper of the Records. 
Vol. i. (Henry III.), London, 1904. 
ii. (Edward I.), 1906. 
Inscr. Chr. Inscriptiones Britanniae Christianae. Ed. Hubner. Berlin and 

London, 1876. 
loto MSS. lolo Manuscripts. Ed. Taliesin Williams (Welsh MSS. Society). 

Llandovery, 1848. Reprinted by I. Foulkes, Liverpool, 1888. 
Itin. Ant. Itinerarium Antonii Augusti. Ed. Parthey and Finder. Berlin, 1848. 

Jafife (2). Regesta Pontificum Romanorum ad annum 1198. Ed. Ph. Jaffa-. 
Second edition. Leipzig. 
Vol. i. 1885. 
ii. 1888. 

Jones and Freem. The History and Antiquities of Saint David's. By W. 
Basil Jones and E. A. Freeman. London, 1856. 

L. G. Cothi. Poetical Works of Lewis Glyn Cothi. Ed. John Jones (Tegid) 

and Walter Davies (Gwallter Mechain) (Cymmrodorion Society). Two 

vols. Oxford, 1837. 
Land of Morgan. By G. T. Clark (Cambrian Archaeological Association). 

London, 1883. 

Lap. W. Lapidarium Walliae. By J. O. Westwood. Oxford, 1876-9. 
Leland, Wales. The Itinerary in Wales of John Leland in or about 1536-9. 

Arranged and edited by Lucy Toulmin Smith. London, 1906. 
Letters, Hen. III. Royal and other Letters illustrative of the reign of Henry 

III. Ed. W. W. Shirley (Rolls Series). London. 
Vol. i. 1862. 
ii. 1866. 
Lewis, Top. Diet. A Topographical Dictionary of Wales. By Samuel Lewis. 

Two vols. Cited from the first edition (London, 1833). 
Lib. Land. The Text of the Book of Llan Dav, reproduced from the Gwysaney 

MS. by J. Gwenogvryn Evans. Oxford, 1893. 
Lib. Nig. Liber Niger Scaccarii. Ed. T. Hearne. Second edition. London, 

Lit. Eng. The History of Little England beyond Wales [ = Pembrokeshire]. 

By Edward Laws. London, 1888. 
Lit. Kym. (2). The Literature of the Kymry [during the period 1080-1322]. 

By Thomas Stephens. Ed. D. Silvan Evans. London, 1876. 
LL. Ancient Laws and Institutes of Wales. Ed. Aneurin Owen (Record Com- 
mission). Two vols. London, 1841. 
Ven. Dim. Gw. = Venedotian, Dimetian, and Gwentian codes of this 

edition (vol. i.). 
Lat. A., B., C. = Peniarth MS. 28, Cott. MS. Vespasian E. xi., Harl. 

MS. 1796, as printed in vol. ii. 749-907. 

Llyfr yr Ancr. The Elucidarium and other Tracts in Welsh, from Llyvyr 
Agkyr Llandewivrevi [Jesus Coll. MS. 119]. Ed. J. Morris Jones and 
John Rhys (Anecdota Oxoniensia). Oxford, 1894. 

Lpool. W. Nat. Trans. Transactions of the Liverpool Welsh National Society. 
Issued annually since 1886. 



M. Paris, Chron. Matthaei Parisiensis Chronica Majora. Ed. H. R. Luard 
(Rolls Series). London. 
Vol. iii. 1876. 
iv. 1877. 
v. 1880. 

Mab. The Text of the Mabinogion and other Welsh tales from the Red Book 

of Hergest. Ed. John Rhys and J. Gwenogvryn Evans. Oxford, 1887. 

Wht. Bk. indicates the readings of " Llyfr Gwyn Rhydderch," as 

edited by J. Gwenogvyrn Evans (Pwllheli, 1907 : pref. dated 1909). 

Mak. Eng. The Making of England. By John Richard Green. [First issued 

in 1881.] London, 1885. 
Marchegay. Chartes Anciennes du Prieurg de Monmouth. Publie'es par P. 

Marchegay. Les Roches-Baritaud, 1879. 
Margam Abb. A History of Margam Abbey. By W. de Gray Birch. London , 


Mat. Hist. Becket. Materials for the History of Thomas Becket. Ed. J. C. 
Robertson and J. B. Sheppard (Rolls Series). Seven vols. London, 
Med. Mil. Arch. Mediaeval Military Architecture in England. By George T. 

Clark. Two vols. London, 1884. 

Meyrick, Card. The History and Antiquities of the County of Cardigan. By 
Samuel R. Meyrick. 

(1) = original edition, London, 1808. 

(2) = reissue by Davies & Co. , Brecon, 1907. 

Migne, Patrologiae Cursus Completus. Series Latina. Accurante J.-P. Migne. 

Mon. Angl. Monasticon Anglicanum. By Sir William Dugdale. Ed. Caley, 

Ellis and Bandinel. Six vols. London, 1817-30. 
Mon. Ant. Mona Antiqua Restaurata. By Henry Rowlands. Dublin, 1723. 

[References in brackets to the second edition, London, 1766.] 
Mon. Hist. Br. Monumenta Historica Brittanica materials for the history of 
Britain [to 1066]. Prepared by Henry Petrie and John Sharpe. [Lon- 
don], 1848. 

Mont. Coll. Montgomeryshire Collections, historical and archaeological. Is- 
sued by the Powysland Club from 1867 to date. 

Morris. The Welsh Wars of Edward I. By J. E. Morris. Oxford, 1901. 
Mots Latins. Les Mots Latins dans les Langues Brittoniques. Par J. Loth. 

Paris, 1892. 

Myv. Arch. The Myvyrian Archaiology of Wales. Ed. Owen Jones (Myfyr), 
Edward Williams (lolo Morgannwg), and W. O. Pughe (Idrison). Lon- 

Vol. i. (poetry) 1801. 
ii. (prose) 1801. 
iii. (prose) 1807. 

[References in brackets to the second (one volume) edition, Denbigh, 

Nennius. See Hist. Britt. 

Nenn. V. Nennius Vindicatus. Von H. Zimmer. Berlin, 1893. 

Norm. Conq. (3). History of the Norman Conquest of England. By E. A. 

Freeman. Third edition (Oxford, 1877) of vols. i. and ii. 
Not. Dig. Notitia Dignitatum. Ed. O. Seeck. Berlin, 1876. 


Ord. Vit. Orderici Vitalis Historia Ecclesiastica. Ed. A. Le Prevost Paris, 

[References are to book and chapter, but those in brackets to volume 

and page of this edition.] 
Orig. Cist. Originum Cisterciensium tomus i. Descripsit L. Janauschek 

Vindobonae [Vienna], 1877. 

Owen, Catalogue. A Catalogue of the MSS. relating to Wales in the British 
Museum. By Edward Owen (Cymmrodorion Society). London. 
Pt. i. (Cottonian and Lansdowne MSS.) 1900. 
ii. (Harleian MSS.) 1903. 
iii. (Charters and Rolls) 1908. 

Owen, Pemb. The Description of Pembrokeshire. By George Owen of Henllys 
[1552-1613]. Ed. Henry Owen (Cymmrodorion Society). London. 
Pt. i. 1892. 
ii. 1897. 
iii. 1906. 
[Many of the notes are by Egerton Phillimore.] 

Papal Letters. Calendar of Entries in the Papal Registers relating to Great 
Britain and Ireland Papal Letters, ed. W. H. Bliss. London. 

Vol. i. (1198-1304) 1893. 

Pat. Rolls. The Patent Rolls, printed under the superintendence of the Deputy 
Keeper of the Records (full text). London. 
Vol. i. (1216-1225) 1901. 
ii. (1225-1232) 1903. 

Pees. S. Wales. History of the Princes of South Wales. By G. T. O. Bridge- 
man. Wigan, 1876. 

Peckham. Registrum Epistolarum Johannis Peckham. Ed C. T. Martin (Rolls 
Series). London. 
Vol. i. 1882. 
ii. 1884. 

Pen. MS. 147. See p. 281. 
Penn. Tours in Wales. By Thomas Pennant [1726-1798], Second edition, in 

three vols. London, 1810. 

Pipe Roll, 31 Hen. I. Magnum Rotulum Scaccarii de anno tricesimo primo 

regni Henrici Primi. Ed. J. Hunter (Record Commission). [London] 1883. 

Pipe Rolls, 2, 3, 4 Hen. II. The Great Rolls of the Pipe for the second, third, 

and fourth year of Henry II. (1155-8). Ed. J. Hunter. London, 1844. 
Pipe Rolls, 5, 6, 7, etc., Hen. II. Printed by the Pipe Roll Society (1884 to 

Polychr. Polychronicon Ranulphii Higden Monachi Cestrensis. Ed. C. Babing- 

ton and J. R. Lumby (Rolls Series). Nine vols. London, 1865-86. 
Powel. The Historic of Cambria. By HfumphreyJ Lloyd. Ed. David Powel 
[ob. 1598]. London [1584]. 

This is cited from the reprint of 1811 (London). Humphrey Lloyd of 
Denbigh (ob. 1568) originally compiled the work, which is based 
upon " Brut y Tywysogion". After his death, the MS. was en- 
trusted to Powel, who published it with large additions, indicated 
thus Ht- 
Powys Fadog. The History of Powys Fadog. By J. Y. W. Lloyd. Six vols. 

London, 1881-7. 

Ptol. Claudii Ptolemaei Geographia. Ed. C. Mailer (Firmin Didot). Paris, 
Vol. i. 1883. 


R. de Torigini. The Chronicle of Robert of Torigni. Ed. R. Hewlett in vol. 
iv. (London, 1889) of " Chronicles of the Reigns of Stephen, Henry II. and 
Richard II. (Rolls Series). 

Radnorsh. (2). A General History of the County of Radnor. From the MS. of 
Jonathan Williams. Compiled by Edwin Davies. Brecknock, 1905. 
[Williams's MS. was first printed in Arch. Camb. and was separately 

issued in 1859 (Tenby).] 

Rec. Carn. Registrum vulgariter nuncupatum " The Record of Caernarvon " 
[Harleian MS. 696]. Ed. Henry Ellis (Record Commission). [London], 
Reg. Conway. Register and Chronicle of Aberconway [Harl. MS. 3725]. Ed. 

Henry Ellis (Camden Miscellany, vol. i.). [London], 1847. 
Reg. Sacr. (2). Registrum Sacrum Anglicanum. By W. Stubbs. Second 

edition. Oxford, 1897. 

Rev. Celt. La Revue Celtique. Paris, 1870 to date. 

Rot. Chart. Rotuli Chartarum in Turri Londinensi asservati. Ed. T. D. Hardy 
(Record Commission). [London], 1837. 

Vol. i. pt. i (1199-1216). 

Rot. Glaus. Rotuli Litterarum Clausarum in Turri Londinensi asservati. Ed. 
T. D. Hardy (Record Commission). London. 
Vol. i. (1204-1224) 1833. 
ii. (1224-1227) 1844. 

Rot. Fin. Excerpta e Rotulis Finium. . . Henrico Tertio rege. Ed. C. Roberts 
(Record Commission). London. 
Vol. i. (1216-1246) 1835. 
ii. (1246-1272) 1836. 

Rot. Norm. Magni Rotuli Scaccarii Normanniae . . . ed. T. Stapleton (Society 
of Antiquaries). London. 
Vol. {.1840. 
ii. 1844. 

Rot. Pat. Rotuli Litterarum Patentium in Turri Londinensi asservati. Ed. 
T. D. Hardy (Record Commission). [London.] 

Vol. i. pt. i. (1201-1216) 1835. 

[An Itinerary of King John is appended to the introduction.] 
Rot. regn. Joh. Rotuli de Liberate ac de Misis et Praestitis regnante Johanne. 

Ed. T. D. Hardy. London, 1844. 
Round, Anc. Charters. Ancient Charters prior to A.D. 1200. Ed. J. H. Round. 

(Pipe Roll Society's vol. x.) London, 1888. 

Royal Charters. Royal Charters and other documents relating to Carmarthen, 
Talley, and Ty Gwyn ar Daf. By J. R. Daniel-Tyssen and Alcwyn C. 
Evans. Carmarthen, 1878. 

Rymer. Foedera, Conventiones, Litterae, etc. Ed. T. Rymer and R. Sander- 
son. Fourth edition, in four vols., by Clarke, Holbrooke, and Caley. 
London, 1816-69. 

Sax. Gen. The " Saxon Genealogies," or chaps. 57-65 of Hist. Britt. See 

p. 1 1 6. 

Sim. Dun. Symeonis Monachi Opera Omnia. Ed. T. Arnold (Rolls Series). 

Vol. i. 1882. 
ii. 1885. 

Song of Dermot. The Song of Dermot and the Earl from Carew MS. 596 
(Lambeth). Ed. G. H. Orpen. Oxford, 1892. 
V. refers to the lines of the poem, p. to the pages of this work. 


Spurrell, Carm. Carmarthen and its Neighbourhood. By William Spurrell. 

Second edition. Carmarthen, 1879. 
Stev. See Asser. 
Str. Flor. The Cistercian Abbey of Strata Florida. By Stephen W. Williams. 

London, 1889. 

Stubbs, Const. Hist. The Constitutional History of England. By William 
Stubbs. Third edition. Oxford. 
Vol. i. 1880. 
ii. 1883. 

Sweetman. Calendar of Documents relating to Ireland in the Public Record 
Office. Ed. H. S. Sweetman. London. 
Vol. i. (1171-1251) 1875. 
ii. (1252-1284) 1887. 

Tanner. Notitia Monastica. By Bishop Thomas Tanner of St. Asaph. Ed. J. 

Tanner. London, 1744. 
Tax. Nich. Taxatio Ecclesiastica Angliae et Walliae auctoritate P. Nicholai 

IV., circa A.D. 1291. Record Commission, 1802. 

References are to the page and column (a or b). 
Testa de Nevill. Sive Liber Feodorum temp. Hen. III. et Edw. I. [London] 

Thomas, S. Asaph. A History of the Diocese of St. Asaph. By D. R. Thomas. 

London, 1874. 
Trans. Cymr. Transactions of the Honourable Society of Cymmrodorion. 

Issued sessionally from 1892-3 to date. London. 

Trans. Roy. Hist. Soc. Transactions of the Royal Historical Society. London. 
Trevet. Triveti Annales. Ed. T. Hog (English Historical Society). London, 

Triads. Cited by reference to the three series (i. ii. iii.) printed in Myv. Arch. 

II. 1-22, 57-75 (388-411), the Arabic numeral giving the number of the 

triad. For some account of the Triads, see note to chap. iv. 
Trib. System. The Tribal System in Wales. By Frederic Seebohm. London, 

[App. refers to the documents printed as Appendices.] 

Urk. Spr. Urkeltischer Sprachschatz. By Whitley Stokes and A. Bezzen- 
berger. GOttingen, 1894. 

This is pt. ii. of the fourth edition of A. Pick, Vergleichendes 
Worterbuch der Indogermanischen Sprachen.J 

V. S. Columb. Adamnan's Life of St. Columba is cited from the edition of J. T. 
Fowler. (Oxford, 1894.) 

Valor Eccl. Valor Ecclesiasticus temp. Henr. VIII. Six vols. (Record Com- 
mission.) London, 1810-34. 

W. People. The Welsh People. By John Rhys and David Brynmor-Jones. 

London, 1900. 
W. Phil. (2). Lectures on Welsh Philology. By John Rhys. Second edition. 

London, 1879. 
Walt. Cov. The Historical Collections of Walter of Coventry. Ed. W. Stubbs 

(Rolls Series). Two vols. London, 1872-3. 
War of G. and G. The War of the Gaedhil with the Gaill. Ed. J. H. Todd 

(Rolls Series). London, 1867. 


Welsh SS. Essay on the Welsh Saints. By Rice Rees. London, 1836. 
Wendover. Rogeri de Wendover Chronica. Ed. H. O. Coxe (English Histori- 
cal Society). Four vols. London, 1841-4. 
Wht. Bk. See Mab. 
Williams, Aberconwy. The History and Antiquities of the Town of Aberconwy. 

By Robert Williams. Denbigh, 1835. 
Wm. Malm. The works of William of Malmesbury are cited as follows : 

G. R. = Gesta Regum \ Ed. W. Stubbs (Rolls Series). London, 
H. N. = Historia Novella / 1887-9. 

References in brackets are to the edition of T. D. Hardy (English His- 
torical Society, London, 1840). 
G. P. = Gesta Pontificum. Ed. N. E. S. A. Hamilton (Rolls Series). 

London, 1870. 

Wm. Newb. Historia Rerum Anglicarum. Ed. Richard Howlett in " Chron- 
icles of the Reigns of Stephen, Henry II., and Richard I. " (Rolls Series), 
vol. i. (bks. i.-iv. 1884) and vol. ii. (bk. v. 1885). 

Woodward. The History of Wales. By B. B. Woodward. Two vols. Lon- 
don [1852]. 

Wotton. Cyfreithjeu Hywel Dda ac Eraill. Seu Leges Wallicae. Ed. Gul. 
Wottonus, adiuvante Mose Gulielmio [Moses Williams]. London, 
Wykes. Chronicon Thomae Wykes. Ed. H. R. Luard in vol. iv. of Annales 

Monastici (Rolls Series). London, 1869. 

Wynne. The History of Wales. By Dr. Powel. Newly arranged and im- 
proved by W. Wynne [ob. 1704]. London, 1697. 

[A new edition of the " Historic " of Powel, with additional matter, 
taken mainly from the notes of Robert Vaughan of Hengwrt. See 
pref. 46.] 

Yorke (2). The Royal Tribes of Wales. By Philip Yorke of Erddig [1734-1804]. 
[Original edition, Wrexham, 1799.] Ed. R. Williams. Liverpool, 1887. 

Zeit. Celt. Ph. Zeitschrift fur Celtische Philologie. Issued at Halle from 1897 
to date. 




THE region now known as Wales was inhabited by man in the CHAP. 
earliest period during which science has clearly shown him to 
have dwelt in the British Isles. 2 In the Pleistocene Age of 
geology, 3 separated from our own by an interval which must 
be measured by tens of thousands of years, a rude race of 
hunters and fishers is proved by the discovery of its implements 
roughly chipped flints, carved fragments of bone and horn 
to have ranged the hills and valleys of Southern Britain and 
waged a not unequal struggle with great beasts of prey, of 
which many belonged to species now extinct. Our islands 
had then no separate existence ; in the beds of what are now 
the North Sea and the English Channel mighty rivers flowed 
north and west to a coast-line far out in the Atlantic, which 
lay where the 100 fathom line now marks the beginning of the 
dip toward oceanic depths. The relics of pleistocene, or, as 
he is more commonly termed (from the primitive fashion of his 
stone weapons), palaeolithic man are found both in the beds of 
ancient rivers, left high and dry as the stream has cut its way 
down, and in caves, those natural houses cool in summer and 

1 In writing this chapter I have chiefly used the following : Boyd Dawkins, 
Early Man in Britain; Evans, Ancient Stone Implements; Beddoe, Races of 
Britain; Taylor, Origin of the Aryans ; Munro, Prehistoric Scotland; Green- 
well and Rolleston, British Barrows; Schrader, Prehistoric Antiquities of tht 
Aryan Peoples ; Sergi, The Mediterranean Race ; Ripley, The Races of Europe ; 
British Museum Guides (i) The Stone Age, (ii) The Bronze Age, compiled by 
C. H. Read. 

8 It is assumed that Eolithic man belongs as yet to the region of hypothesis. 

8 I follow the terminology of Prof. Boyd Dawkins, who makes the Pleis- 
tocene Age end with the beginning of the Neolithic. 
VOL. I. I 


CHAP, warm in winter of which man availed himself from the very 
first. In South-eastern Britain it is the river drift which sup- 
plies evidence of the conditions of life in palaeolithic times ; 
in Wales, on the other hand, our knowledge of the period is 
entirely derived from the caves which abound in the carbonifer- 
ous limestone of the district. 

The exploration of caves has yielded traces, indubitable 
however slight, of the presence of palaeolithic man in four Welsh 
regions, namely, the Vale of Clwyd, South Pembrokeshire, 
Gower and the neighbourhood of Monmouth. In 1861, during 
the excavation of the Long Hole in Gower, flakes of flint which 
had been used for cutting were found amid the bones of extinct 
mammals of the Pleistocene Age. It was about the same time 
that the bone-bearing caves of the Tenby district were being 
opened up ; here Mr. Laws discovered in the Coygan cave, not 
far from Laugharne, a bone awl and a worked flint lying under 
bones of the rhinoceros. In King Arthur's Cave, on Great Do- 
ward Hill, near Monmouth, Mr. Symonds found flint flakes 
among abundant remains of pleistocene mammals. Similar dis- 
coveries have also been made in the Denbighshire caves : in that 
of Pont Newydd, one of the famous caves of Cefn Meiriadog, 
palaeolithic implements came to light in association with bones 
of the hippopotamus and the straight-tusked elephant, and the 
presence of man was made doubly certain by the finding of a 
human tooth. On the other side of the valley, Dr. Hicks ex- 
plored the Ffynnon Feuno and the Cae Gwyn caves, near 
Tremeirchion, and in both discovered flint implements, with 
bones of the mammoth and the rhinoceros. There would 
seem to be no doubt, therefore, that palaeolithic man ranged 
over the whole of Wales. 4 

The list of great mammals who disputed with him the pos- 
session of the country is an impressive one. It includes, in 
addition to the mammoth, the rhinoceros (of two varieties), 
the hippopotamus and the straight-tusked elephant, already 
mentioned, the cave lion, the cave bear, the hyaena, the 
bison, the reindeer, the Irish elk and the wild horse. When 
the author of the third series of the Triads described the first 
settlers of Britain as finding it full of " bears and wolves 

4 Evans, Ancient Stone Implements, second edition (1897), pp. 520-1. 


and dragons and long-horned oxen" (eirth a bleiddiau ac CHAP. 
efeinc ac ychain bannog), 6 he was no doubt giving rein to a 
lively imagination, but the truth revealed by science is a hun- 
dredfold stranger than his ingenuous fiction. 

The palaeolithic remains found in England and France have 
enabled students of the subject to trace a gradual improve- 
ment in civilisation, showing itself at last in a quite surprising 
degree of skill in carving and drawing. But the Welsh relics 
are too few to furnish much evidence of the kind, and, so far 
as they have been classified at all, appear to belong to a very 
primitive type. Nor can it be said that their relation has yet 
been finally determined to the glacial epoch of British geology, 
an epoch falling within the limits of the Pleistocene Age, when 
the greater part of Wales was wrapped in a curtain of ice and 
snow as inhospitable as that which now envelops Greenland. 
The high authority of Sir John Evans may be cited in support 
of the view that the palaeolithic climate was not much colder 
than that of our own day and that palaeolithic man was 
post-glacial ; 6 Dr. Hicks, on the other hand, argues from the 
evidence furnished by the Welsh caves that the characteristic 
pleistocene mammals, with whose remains those of man are 
so often associated, lived before the age of the glaciers, and, 
with the palaeolithic race, disappeared as it laid its icy hand 
on the soil. 7 

So far as Britain is concerned, the story of palaeolithic man 
certainly ends abruptly. On the Continent, archaeologists 
have met with some success in the effort to bridge over the 
gap which severs this from the succeeding, the Neolithic Age. 
In our own land this has not yet been done, and it must re- 
main highly probable that, as the strange beasts around him 
disappeared from our valleys, so also did the cunning savage 
who watched and entrapped them, and this without leaving 
any of his posterity behind. Welshmen have inherited neo- 
lithic blood and the neolithic civilisation ; in the palaeolithic 
man of Britain it is probable they have no part. 

5 Myv. Arch. ii. 57 (400). 

8 Presidential Address at the Toronto Meeting (1897) f tnc British Associa- 
tion (Report of the Association for the year, p. 13). 

1 Proceedings of the Geological Society, May, 1898, ci. 




With the opening of the Neolithic or New Stone Age begins, 
so far as is now known, the continuous history of man in 
Wales. That period, in which the use of metal for the practi- 
cal purposes of life was unknown, was in time succeeded by the 
Age of Bronze, when the newly discovered alloy of copper and 
tin gave the hard cutting edge which folk needed for tools and 
weapons, and this in turn by the Age of Iron, which may be 
regarded as lasting to our own day. But no break separates 
these periods from each other ; the arts introduced by neolithic 
man into Britain the management of domestic animals, the 
making of pottery, the grinding of stone implements have 
never been forgotten ; the men who first practised them here 
are still, there is no reason to doubt, plentifully represented in 
the population of these islands. The beginning of each new 
period marks an advance in culture and, probably, the arrival 
of a new race, but the past has not been obliterated ; its 
influence is still potent in the new era. 

In the neolithic period the contour of the British Isles 
presented the same general appearance as at present. Valleys 
and plains had sunk so as to form encircling seas, whose 
billows swept through the Straits of Dover and St. George's 
Channel. The process of depression was, however, not yet 
complete, for it has been shown by discoveries made in many 
parts of our coast-line, notably in Wales, that tracts of mud and 
sand now regularly washed by each tide were in neolithic times 
covered with a luxuriant forest growth, giving no hint of the 
neighbourhood of the sea. The blackened stumps of such a 
forest were laid bare in the winter of 1171-1 172, to the no small 
perplexity of the wise men of Dyfed, by a great gale which 
swept over the sands of St. Bride's Bay, 8 and in recent years 
submerged areas of the kind have been examined at Whitesand 
Bay, near St. David's, at Barry, 9 and at Borth, near Aber- 
ystwyth, and have proved to be land surfaces of the Neolithic 
Age. It is clear, therefore, that during this period the Welsh 
coast-line was, speaking generally, much further out to sea 
than at present, and there may well have been a time within 

8 Gir. Camb. vi. 100 (I tin. i. 13) ; v. 284 (Exp. Hib. i. 36). 

9 Quarterly Journal of the Geological Society, Aug. 1896, pp. 474-89. 


the memory of neolithic man when Anglesey was not an island CHAP. 
and Cardigan Bay, which hardly anywhere exceeds twenty l ' 
fathoms in depth, was dry land. One is tempted to inquire 
whether we may not have in the well-known Cantref y Gwaelod 
and Traeth Lafan legends, which are stories of the submergence 
of flourishing realms beneath the pitiless sea, reminiscences 
handed down through many generations of the effects at times, 
perhaps, startling of this gradual subsidence attested by geo- 
logy. In the original story of the Lowland Hundred, as told in 
a poem in the Black Book of Carmarthen, there is no mention 
of the embankment, invested with such interest for lovers of 
literature by the sardonic humour of Peacock, 10 nor is it a 
drunken " Lord High Commissioner," but a mysterious maid, 
servant of a magic well, who is blamed for the catastrophe. 
Hence there can be no doubt that the tale is thoroughly primi- 
tive ; it remains, however, for students of folk-lore to say whether 
this and kindred legends known to the Welsh people have any 
special features which show them to embody genuine traditional 
history, or whether they are merely specimens of a class of 
story known in all parts of the world. 

At what time neolithic man first crossed the channel and 
the precise degree of civilisation he had attained at the time of 
his arrival here are matters of conjecture, hardly as yet of 
definite knowledge. The neolithic culture of Europe is be- 
lieved to have slowly arisen within the continent itself, perhaps 
out of palaeolithic germs, during a period which began not less 
than ten thousand years ago. Much progress had no doubt 
been made before the long canoes, fashioned out of single tree 
trunks, grated upon our shores and discharged their human 
burthen, the first ancestors of the British people ; the Neolithic 
Age must have been far advanced ere such an immigration 
could have taken place. Probably the chief domestic animals 
the dog, the ox, the sheep, the goat, the pig had been tamed ; 
a rude kind of pottery was made ; skill had been acquired in 
the grinding of smooth stone axes, which were highly polished, 

10 " That the embankment is old, I am free to confess ; that it is somewhat 
rotten in parts, I will not altogether deny ; that it is any the worse for that, I do 
most sturdily gainsay. It does its business ; it works well ; it keeps out the 
water from the land, and it lets in the wine upon the High Commission of Em- 
bankment" (M is/or tunes of Elf hin, chap. ii.). 


CHAP, furnished with sharp edges and hafted in wooden handles. 
Besides these weapons of polished stone, others were in use 
which had been chipped into shape with the utmost care. For 
shelter men betook themselves to caves, where the neolithic 
level of occupation is often to be found above that of palaeolithic 
times, or they dug themselves pits in the ground, or with their 
stone axes cut timber and made square log huts. That they 
were given to the regular tillage of the soil there is nothing to 
show ; primitive folk would seem to have had little taste for 
the toilsome pursuit of agriculture as long as an easier method 
of making a living was open to them. They clothed themselves 
in leathern garments, for the most part, though some use was 
also made of a coarse linen, woven in a primitive kind of loom. 
Their dead they laid to rest, in the crouching attitude of sleep, 
in a cave or a stone chamber which, though encased in a 
mound or a cairn, had an entrance which might be used for 
further burials. By their side they placed the implements 
which they took to be needful for the use of the departed in 
the spirit life just begun. 

Many generations, no doubt, went by ere the civilisation 
just described reached the highlands of Wales. But there is 
no lack of evidence that it rooted itself thoroughly in Welsh 
soil. The limestone caves of Pembrokeshire and Denbighshire 
were occupied by neolithic as well as by palaeolithic man, and 
show us the former race in the early stages of its development. 
In the Little Hoyle cave in Longbury Bank, near Tenby, bones 
of the ox, goat (or sheep), pig and dog were found with charcoal 
(the ashes of the primitive hearth), pottery, and implements of 
bone and flint. The greater pleistocene beasts had by this time 
entirely disappeared, but the brown bear, the wild horse and 
the red deer still provided big game for the neolithic hunter. 
In the exploration of this cave it was a matter of surprise to 
find human bones, representing no less than nine individuals, 
mingled in great confusion with the house rubbish above 
described ; this, however, appears to have been due to an ac- 
cident which had let fall through the roof of the cave the 
contents of the neolithic burying place above. 11 A group of four 
caves in the neighbourhood of Llanarmon in Yale, used partly 

11 Lit. Eng. pp. 15-16. 


as dwelling-places and partly as tombs, illustrates the neolithic CHAP, 
civilisation as found in North Wales. Remains of the same 
domestic and wild animals were found as in the Little Hoyle ; 
there were also " fragments of rude black pottery, hand made, 
composed of clay worked up with small pieces of stone, to 
prevent fracture while it was being subjected to the fire ". A 
few flint implements were discovered, but the most interesting 
find was " a beautiful polished axe made of greenstone, with the 
edge uninjured by use ". The caves of the Cefn Meiriadog 
district were also inhabited in this age and have yielded remains 
of the same type. 12 

On the Pembrokeshire coast a neolithic settlement of a 
different kind, though probably of the same early stage, was 
recently explored by Mr. Laws. Under the blown sand on 
Giltar Head, near Tenby, a clay floor, hardened by fire and 
covered with charcoal, came to light ; upon it lay flint and 
bone implements, but the most valuable evidence as to the life 
led by the men who had used it was derived from a large 
refuse heap or "kitchen midden," of the kind which has 
furnished Danish archaeologists with data of the highest value 
for the early Neolithic Age. This contained the bones of 
domestic animals, but greatly in excess of these in quantity 
were the tons of shells of twenty-four varieties, which showed 
that the dwellers in this seaside camp a place, perhaps, of 
summer resort were great eaters of shell fish and by no means 
relied entirely upon their cattle and upon the chase for food. 18 

In other countries much information about the Neolithic 
Age has been gleaned from the vestiges of lake dwellings, 
built for safety on piles or artificial islands at some little dis- 
tance from the shore. One such " crannog," to use the Irish 
name, has been discovered in Wales, namely, in Llangorse 
Lake, but the remains were too scanty to make it possible to 
fix the age of its construction and occupation, since the lake 
dwelling as such is not specially characteristic of any one of the 
great prehistoric periods. 14 There may also have been a lake 
dwelling in Llyn Llydaw at the foot of Snowdon, for a lower- 
ing of the water level of this lake some years ago exposed to 

u Cave Hunting (1874), by Prof. Boyd Dawkins, chap. v. 

Arch. Camb. IV. xi. (1880), 244-5. 

14 Arch. Camb. IV. i. (1870), 192-8 ; iii. (1872), 146-8. 


CHAP, view a canoe of the early type known in Welsh as " cafn un- 
pren," the hulk fashioned from a single tree, which is the most 
primitive known form of boat. 15 

Stone implements of neolithic pattern have been picked up 
in many parts of Wales, 16 but, while they testify to the occupa- 
tion of the country for many centuries by neolithic man, they 
tell us little else when they are not found in company with 
other remains, domestic or sepulchral. 

The neolithic burying places of Wales next claim attention. 
In the new Stone Age the practice of burning the dead was 
rarely followed, and it was also unusual to provide each corpse 
with a separate grave ; instead, a vault or chamber was pre- 
pared, which was regarded as a house of the dead and opened 
from time to time to receive fresh burials. Sometimes this 
chamber was natural, a cave (as at Rhos Ddigre) 17 or a cleft 
in the rock ; sometimes it was excavated ; but, especially in the 
later Neolithic Age, as greater skill was acquired in the hand- 
ling of large masses of stone, it was very commonly built of 
boulders and loose stones, with a great slab for roof and an- 
other to keep the entrance. The whole was covered with a 
mound of earth, for protection from the weather and from 
beasts of prey, and then formed a " barrow," the Welsh " tom- 
en " ; 18 or, if the ribs of the earth were thereabouts but thinly 
coated with soil, small stones were used instead to form a 
"cairn" or " carnedd ". Owing to the need of providing an 
entrance passage, the structure was usually rather long in pro- 
portion to its breadth, but this was an accidental, not an 
essential feature of it, and for this reason the term " chambered 
barrow" is to be preferred, as suggesting a truer classification, 
to the more familiar " long barrow". The chambered barrows 
and cairns of Wales have not yielded the abundant evidence 
with regard to the civilisation and physique of neolithic man 
which has been furnished by those of the South of England. 

18 Arch. Camb. IV. v. (1874), 147-51. 

16 See index to Evans, Ancient Stone Implements. 

17 Cave Hunting, chap. v. 

18 The simple form "torn" also occurs, as in Cwm Dom (where there is a 
tumulus), near Llanfyllin, Bon y Dom, the original name of Moel y don, Angle- 
sey (Penn. iii. 16), and the line, attributed to Llywarch Hen, 

Tom elwithan ( = Elwyddan) neus gwlych glaw, 
for which see IV. Anc. Bks. ii. p. 291. 


Few of them remain in the cairn or barrow form, and of these CHAP. 
many have been rifled of their contents and are only interest- 
ing to-day as structures. Such has been the fate of Carneddau 
Hengwm, near Barmouth, two huge cairns, of which the larger 
is 150 feet long ; it is possible, however, that all the chambers 
have not yet been opened. 19 The mound near Plas Newydd, 
Anglesey, measuring about 150 by 100 feet and locally known 
as " Bryn yr hen bobl " (The old folks' hill) was opened about 
1730, but there is no record of the exploration. 20 Not far off is 
another carnedd, that of Bryn Celli Ddu, which appears in the 
old one-inch Ordnance map as " Yr Ogof " (The Cave) ; here 
the chamber has been almost entirely stripped of its covering, 
but the entrance passage is still intact. Pennant records the 
discovery of human bones in a very friable condition in this 
chamber when it was first opened up ; a broken flint knife was 
recently found there also, so that it was beyond any doubt a 
neolithic burying place. 21 Two of the chambered carneddi of 
Wales have been explored since the scientific study of such 
monuments began and accordingly have yielded some evi- 
dence, scanty though it be, as to the neolithic population of 
the country. That of Park Cwm in the parish of Penmaen, 
Gower, which measured about 60 by 50 feet in area, contained 
the remains of some twenty or thirty skeletons in such disorder 
as to show that the bodies must have been placed there in 
succession and at long intervals. Some pottery came to light, 
but there were no implements to indicate more precisely the 
stage of civilisation which had been reached by the builders of 
the carnedd. It may, nevertheless, be regarded as suggestive 
of some progress in ideas and in manual skill that the four 
chambers opened into a single corridor, instead of having each 
its own means of access from the open air ; the feature is one 
which appears in some of the " long barrows " of the neigh- 
bouring English counties. 22 The carnedd at Tyddyn Bleidd- 
yn, near St. Asaph, was of a simpler pattern ; in its two separate 

19 Arch. Camb. IV. iv. (1873), 91-5. 

20 Perm. iii. 21 ; Arch. Camb. IV. i. (1870), 51-8 (W. 0. Stanley) ; xi. 
(1880), 81-96 (E. L. Barnwell). For the name see vol. ii. of the Cambrian Register 
(1799), p. 289. 

81 Penn. iii. 53-4; Arch. Camb. I. ii. (1847), I ' 6 > In - ** ( l86 9). T 4<>-7 
(E. L. Barnwell); Griffith, Cromlechs of Anglesey and Carnarvonshire (1900). 

*Arch. Camb. IV. ii. (1871), 168-72. 


CHAP, chambers human remains were found, so huddled together 
that the burials could not have been simultaneous ; there were 
also bones of the domestic animals. 23 

Stripped of its casing of earth and stones, the neolithic 
grave chamber became a cromlech. The bleak Atlantic winds 
and rain storms of many a century and the misguided industry 
of farmers and road menders in recent times have reduced a 
great number of the chambered cairns and barrows of Wales to 
the bare stone skeleton, so that cromlechs are far commoner in 
the country than the structures just described. Their massive 
outlines, telling unmistakably of immense toil lavished upon 
some great purpose and yet giving scarcely a hint as to what 
this might be, have for ages whetted men's curiosity. Nothing, 
indeed, can serve better to show how completely their original 
use has been forgotten than the names which have been given 
to them by the country folk in historic times. The cromlech 
at Ashbury, in Berkshire, was known as early as the tenth 
century as " Wayland's smithy," where the great smith god of 
the Teutonic peoples plied his demon art. 24 So in Wales very 
many cromlechs bear the name of " Coeten Arthur " (Arthur's 
Quoit) ; 25 one is " Bwrdd Arthur " (Arthur's Table), also known 
as " Gwal y Filast " (The Lair of the Greyhound Bitch) ; 26 an- 
other is " Llety 'r Filiast " (The Greyhound Bitch's Lodging) ; 27 
another, "Barclodaid y Gawres" (The Giantess's Apronful); 28 
another, " Llech y Drybedd " (The Tripod Stone), 29 another, 
" Ty Illtud " (Illtud's House). 30 Cromlech itself appears, in 
point of fact, to be a name of this description ; on the analogy 
of " cromglwyd," 31 a wattle roof, it should mean a covering or 

23 Cave Hunting, pp. 161-4. 

24 Kemble, Codex Diplomatics, v. 332. For the legend see Elton, Origins 
of English History, second edition (1890), p. 127. 

w E.g., those at Llugwy, Anglesey (Penn. iii. 56), at Llanfair by Harlech 
(Gibson, 661), and at Cefn Amwlch, Lleyn (Arch. Camb. I. ii. 97). Inquiring 
for a "cromlech" at Rhoslan, near Cricieth, I was told that nothing of the kind 
was known there, but upon its appearing that what I wanted was to see " Coeten 
Arthur," I was taken at once to the spot. Mr. J. E. Griffith says the name is 
frequently given to "blocs perche's". 

26 At Dol Wilym in Carmarthenshire (Gibson, 628). 

27 On the Great Orme (Griffith's Cromlechs). 

28 At Trecastell, Anglesey (ibid.). 

29 Near Nevern, Pembrokeshire (Gibson, 638). 

30 At Llanamwlch, Brecknockshire (ibid. 593). 
S1 For this form see Mob. 76 (cromglwyt). 


sheltering stone, and in the Welsh version of the Bible this is CHAP. 
the sole use : " cromlechydd y creigiau " stands for " clefts " or 
" holes of the rocks ", 32 It is George Owen of Henllys who, 
writing in the same age as the translators of the Welsh Bible, 
first applies the term to a megalithic structure ; the terms of his 
references, however, to " the stone called Maen y gromlegh vpon 
Pentre Jevan lande " 33 clearly show that he is quoting a popu- 
lar description (The Shelter Stone), without the least intention 
of applying the name to this class of monument as a whole. 
But, just as in the region round Harlech " Coeten Arthur" was 
so common a name for these structures as to have almost 
become at one time their usual designation, 34 so in Anglesey at 
the beginning of the eighteenth century they were " vulgarly 
called by the name of cromlech," 35 and, through the writings of 
Henry Rowlands and other Anglesey antiquaries, this came to 
be the term by which they were generally known among the 
English and Welsh writers who speculated as to their origin. 
Yet it would be a mistake to suppose that the name is of 
greater historical significance than the others which we dismiss 
at once as mere freaks of popular fancy : no light, in short, is 
cast upon the origin of the cromlech by any of the names it 

In the eighteenth century the theory was broached that 
cromlechs were Druidical altars, probably because they are 
especially common in Anglesey, which is known to have been 
a centre of Druidical worship. Beginning at the end of the 
seventeenth century with John Aubrey, 36 the view that these and 
other rude stone monuments were the handiwork of the Druids 

32 Isaiah vii. 19; Ivii. 5. Cf. also Jeremiah xlix. 16. 

33 Descr. Pemb. i. 251. 

34 " Nid oedd yn eich amser chwi, Syr, ond bargeinion bol clawdd, a lied 
Haw o scrifen am dyddyn canpunt, a chodi carnedd neu gotten Arthur yn 
goffadwriaeth o'r pryniant a'r terfyneu " (Y Bardd Cwsg, p. 62 of Prof. J. 
Morris Jones' edition). See also Lewis Morris, Celtic Remains, p. 96, " Coeten 
Arthur," i.e., King Arthur's Quoit. By this name a great many of those ancient 
monuments in Wales are called, which by the moderns are supposed to have 
been the altars of the Druids ; but in some places they are called " cromlech," 
pi. " cromlechau ". " Dolmen," it may be remarked, is not a term known in 
Wales. It has been employed since about 1800 by French archaeologists as the 
equivalent of the British cromlech, but there seems no reason why an English 
writer should use it, save that the French have confused matters by using 
" cromlech " in the sense of a stone circle. 

36 Mon. Ant. 47 (Lond. 47). Cf. Gibson, 676. 30 Gibson, 637. 


CHAP, acquired a marvellous hold upon the minds of the learned and 
has only been driven from the field within the last thirty or 
forty years. In Mono. Antiqua (1723) it appears fully grown 
the Druids, says the confident author, "had their Altars or 
Cromleche, on which they perform'd the Solemnities of Sacrifice 
and their sacred Rites of Aruspicy and Divination" ^ It was 
useless to urge that local tradition knew nothing of any such 
august associations, that the ordinary cromlech capstone was of 
such a form and so poised that no priest could have sacrificed 
upon it with dignity, and that many cromlechs had apparently 
been until recently covered ; little impression was made upon 
the fashionable creed until systematic research in the latter half 
of the nineteenth century showed that the cromlech only differed 
from the chambered sepulchral barrow in being the ruin of its 
former self the hard kernel which had defied the hand of time 
and had only to fear what man could do. 38 In the prehistoric 
cemetery near Auray in Brittany, cromlechs, according to Mr. 
Romilly Allen, " are to be seen in every stage of decay," 39 and 
Wales is not without examples of the process of destruction. 
Of the chambered mounds of the country, that at Bryn Celli 
Ddu has been stripped almost, but not quite bare ; the crom- 
lechs of Llugwy in Anglesey, of St. David's Head, of St. 
Nicholas in Glamorganshire, and of Llanamwlch in Breconshire 
show traces of the vanished mound or cairn ; 40 while the 
cromlechs themselves survive in every degree of dilapidation, 
from the complete chamber of " Cwt y Bugail " (Shepherd's 
Hut), near Y Ro Wen, Carnarvonshire, to the single supporter 
which is all that remains of " Yr Allor " (The Altar), near Rhos 
Colyn. 41 

Fascinating as they have always been to the inquirer, 
cromlechs cannot in the nature of things yield much evidence 
as to the men who built and used them. Human bones and 
pottery have in some cases been found within them, removing 

37 P. 6 9 (Lond. 69). 

38 A very full discussion of cromlech theories will be found in E. L. Barn- 
well's contributions on the subject to Arch. Camb. See III. vii. 49-55 ; xv. 118- 
131 ; IV. iii. 81-125 ; 89; V. i. 129-36. 

39 Arch. Camb. V. xvii. 221. 

40 Arch. Camb. III. xiii. (1867), 137 ; IV. iii. (1872), 143 ; IV. v. (1874), 72- 

41 Both these cromlechs are included in Mr. J. E. Griffith's book. I am in- 
formed that " Cwt y Bugail " is not the local name of the cromlech, but of a 
neighbouring cottage. 


all doubt as to their sepulchral character, but the only Welsh CHAP. 
cromlech excavated to any purpose in recent times has been 
that of Pant y Saer, near Llanfair Mathafarn Eithaf, Anglesey. 
Here in 1874 the Rev. W. Wynn Williams found, under a 
large flat stone not previously disturbed, the remains of several 
bodies buried in a contracted posture ; in the same cromlech 
were other human remains, bones of the ox and pig, sea shells, 
charcoal, and one bit of coarse pottery. 4a The arrangement 
of the bodies found underneath the slab makes it fairly certain 
that this was a neolithic interment, whatever doubt may attach 
to the remains found at a higher level, and, in general, it may 
be said that cromlechs are products of the Neolithic Age, 
though, when once erected, they may have been used for 
burial by the people of later epochs. 

Cromlechs are often found in groups, as at Plas Newydd, 
at Ystum Cegid and at Carreg y Gof (Pembrokeshire), and the 
idea is thus suggested that the neolithic raje dwelt together in 
little settlements, bound together for common defence. If it 
could be regarded as certain that the hill fortresses and the 
cliff castles, fenced with ditch and mound and dry stone wall, 
which are so frequently seen in Wales, had been planned and 
set in order for their wars by neolithic tribesmen, the question 
might be regarded as settled, and the New Stone Age might 
be treated as an era of village communities. But the frequent 
discovery of flint flakes and arrow heads in our primitive camps 
does not prove them neolithic ; these smaller trifles were made 
in stone, it appears certain, long after the introduction of metal 
for the fashioning of the more substantial weapons of a warrior's 
equipment, 43 and it is beyond doubt that fortresses of this 
kind were in general use as late as the time of the Roman 
Conquest. 44 That some of them may be neolithic is all, then, 
that can with safety be asserted. 

So far nothing has been said as to the physical type and 
race affinities of the neolithic dwellers in Wales. Yet if, as 
has been maintained, it is to this people the ancestry of the 
modern Welshman must in a large measure be traced, the 

"Arch. Camb. IV. vi. (1875), 341-8. 
4S Munro, Prehistoric Scotland (1899), p. 203. 

44 This has been conclusively proved in the case of the great walled fortress 
of Tre'r Ceiri, Carnarvonshire, by excavations carried on in the summer of 1903. 


CHAP, matter is clearly one of great interest. Investigations in the 
* " long barrows " of England have furnished a fairly definite 
picture of the neolithic occupant of the Wiltshire and Yorkshire 
uplands. He was short of stature, averaging about five feet 
five inches in height (the women were under five feet), small 
featured and small limbed. A low, narrow forehead, a thin, 
shapely nose, eyebrows and lips which projected but slightly 
beyond the plane of the face, gave an impression of mildness, 
upon which it might not have been safe to presume. Most 
important character of all, from the point of view of the 
anthropologist, was the oval form of the head, as looked at 
from above, i.e., the " long barrow " race were dolichocephalic 
or long-headed, as distinguished from the brachycephalic or 
round-headed races. Few communities, indeed, were more 
pronounced in their dolichocephaly, for the average proportion 
of the breadth of the skull to its length was as 71 or 72 to 
IOO, a proportion expressed in scientific language by the 
statement that the "cephalic index" was 71 or 72. 45 

If a comparison be now instituted between the few skeletons 
of Neolithic Age discovered in Wales and this general descrip- 
tion, it will appear that the Welsh remains are in respect of 
stature and the contour of the head of the same class as those 
of England, with the exception that the Welsh skulls are not 
so long, those from the Denbighshire and Flintshire caves 
having an average index, not of 71 or 72, but of 76 or 77.* 6 
This difference is not, however, important enough to warrant 
us in supposing that the neolithic population of Wales was 
not essentially the same as that of England. 

No help is, of course, to be expected from neolithic tombs 
in determining the question of the colour of the hair and the 
complexion of the people of the New Stone Age. But the 
probabilities are much in favour of black, curly hair and a 
brunet rather than blond aspect. Later immigrations seem to 
have been predominantly of fair-haired folk, and the facts point 
to the dark element in the British population as the oldest. It 
is in harmony with this view that dark eyes and hair are found 
in the greatest preponderance in the western parts of the British 

45 For the " long barrow race " see British Barrows, pp. 645-60 ; Prehistoric 
Scotland, pp. 448-52. 

46 Cave Hunting, p. 171. 


Isles and particularly in secluded spots, where there has been CHAP. 
very little disturbance of the population. Thus the tables 
compiled by Dr. Beddoe show Wales, in company with Corn- 
wall and the West of Ireland, as a land of dark folk, who no- 
where muster in greater strength than in the two old-world 
centres of Rhayader and Beddgelert 47 

When the neolithic inhabitants of Britain have been con- 
ceived of as a short, dark and long-headed people, it is not 
difficult to find relatives for them on the Continent. Tacitus long 
ago recorded the observation (doubtless) of Agricola, that the 
swarthy visages and twisted locks of the South- Welsh tribe of 
Silures pointed to their Iberian origin. 48 The Iberians, the 
oldest known inhabitants of the Spanish peninsula, are still the 
dominant element in this quarter of Europe ; " the present 
population of Spain," says Ripley, " is closely typical of that of 
the earliest prehistoric period ". He adds that " it is cranially 
not distinguishable either from the prehistoric Long Barrow 
type in the British Isles or from that which prevailed through- 
out France anterior to its present broad-headed population of 
Celtic derivation ", 49 The early neolithic race type was, in fact, 
uniform throughout Northern Africa, Spain, France, Italy and 
the British Isles, to go no further afield ; Western Europe was 
held by the race styled by some Iberian, and by others Medi- 
terranean. Its features and build are represented in modern 
Britain by the short, dark Welshman of South Wales, possibly 
its very qualities of soul and mind in the typical collier and 
" eisteddfodwr," impulsive and wayward, but susceptible to the 
influences of music and religion. Theories of this kind, which 
offer a simple explanation of a great variety of facts, are eyed 
with much suspicion by students whose work has made them 
familiar with the real complexity of race problems. But no 
more satisfactory explanation has yet been put forward of the 
race affinities of neolithic man as he is seen in Britain ; the 
Iberian or Mediterranean theory, indeed, has recently received 
from one quarter unexpected confirmation. It has long been 

47 Races of Britain (1885), pp. 185, 186. The results are conveniently mapped 
in Ripley's Races of Europe, p. 318. 

4814 Silurum colorati vultus torti plerumque crines et posita contra Hispania 
Hiberos veteres traiecisse casque sedes occupasse fidem faciunt " (Agricola, xi.). 

49 Races of Europe, p. 277. 


CHAP, known that Irish and Welsh, though as Celtic tongues they 
If draw the bulk of their vocabulary and inflexions from the 
great Aryan fountain-head, have a syntax which is entirely 
unlike that of the Aryan group of languages, and the conjecture 
was natural that this arose through the adoption of Aryan 
speech by a non- Aryan people, who perpetuated in the new 
tongue the idioms of the old, in the manner of modern Welsh- 
men whose mastery of English is imperfect. To Prof. 
Morris Jones is due the credit, however, of pointing out that 
these anomalous idioms, for instance, the habit of beginning 
the sentence with the verb (" dywedodd Arthur " for " Arthur 
said ") and the suffixing of pronominal endings to prepositions 
(ynddo = in him), are exactly paralleled in Egyptian and 
Berber, and that the pre-Aryan language involved must ac- 
cordingly have been of the so-called Hamitic family. 50 This 
much, then, is certain that before the advent of the first Celtic 
invaders, i.e., as will shortly be seen, the beginning of the 
Bronze Age, at least one of the languages (perhaps the only 
one) spoken in Britain was nearly akin to that of the Berber or 
Lybian race of North Africa. Should the attempt be made to 
establish an African connection in other than linguistic matters, 
there is no insuperable difficulty. The modern Berber is spare 
in figure and averages five feet seven inches in stature ; his skin 
is white and gives no suggestion of negro blood ; his hair is 
usually dark and the blond type is rarely seen ; though dolicho- 
cephalic, he is not more so than the men of the Denbighshire 
caves, with an average index of 76 or 75. "The Chawia," 
remark the authors of Libyan Notes (1901), "are, generally 
speaking, remarkably European in their appearance ; many 
might have passed for Irishmen or Scotchmen. The boys in 
particular, when about the age of fifteen or sixteen, would, 
if put into similar dress, be almost indistinguishable from 
English lads of the same age." The race is hardy, industrious, 
intelligent, and of independent spirit ; if due allowance be made 
for the effect of long centuries of growth and contact with 
civilisation, there seems no reason why it should not be taken 
as fairly representing the neolithic folk of Wales. Even the 
cromlechs are not wanting, and, though the art of building 
great stone chambers for the burial of the dead in course of 

50 See Appendix B. to W. People. 


time spread so far afield that as a race criterion it is somewhat CHAP. 
precarious, it is surely of some significance that the " dolmen " 
at Roknia in Algeria selected for illustration in Libyan Notes 
might very well have been taken, if left undescribed, for a 
typical cromlech from Anglesey or Pembrokeshire. 


Three events are associated with the Bronze Age in 
Britain, vis., the adoption of a new material for the making of 
weapons, the arrival of a new race, and the introduction of new 
burial customs. How far these three events are to be con- 
nected with each other, whether the new customs were those of 
the new race, whose victories were the natural outcome of 
their possession of the new weapons, is no doubt open to 
argument, but that bronze, the practice of cremation and a 
broad-headed people make their first appearance in these islands 
during the same epoch will hardly be denied. It will be con- 
venient first to deal with the new element in the population. 

While the investigation of the neolithic chambered tombs 
of England and Wales has never brought to light any but 
long skulls, the round barrows and cairns in which bodies, 
frequently, but not always, after a process of cremation, are 
placed without the protection of a chamber, have yielded great 
numbers of round or brachy cephalic skulls as well as those 
of the older form. The contrast between the two races goes 
beyond the shape of the skull, and is indeed as complete as it 
well could be. In height the new-comer averaged five feet 
eight inches, being several inches taller than his neolithic op- 
ponent Powerful of limb, with wide forehead, overhanging 
eyebrows, prominent cheek-bones and ample jaws, he was 
planned for immediate conquest, if we may not say for ultimate 
victory in the struggle of races. 61 In colouring he was most 
probably light; it is noteworthy in this connection that the 
hair and eyes of the eighty-two broad-headed men examined 
by Dr. Beddoe in Wales and the West of England were " very 
much lighter, generally speaking, than those of the rest of the 
local population ". 62 

There is every reason to suppose that this round-headed 

61 British Barrows, pp. 637-45. u Racts of Britain, p. 17. 

VOL. I. 2 


CHAP, invader, who perhaps first crossed the channel about twelve 
centuries before the Christian era, was the man who introduced 
Aryan speech, in its Celtic form, into Britain. Within recent 
years the Aryan question has undergone great modifications. 
It is no longer considered necessary, in order to explain the 
phenomenon of a great Aryan group of languages, that a single 
Aryan race, living in a remote past in some definite region of 
the East, should be supposed to have broken up in course of 
time into Indo- Iranian, Celtic, Teutonic, Hellenic and other 
branches and have scattered itself over the whole of what is 
now the area of Aryan speech. Prehistoric archaeology has 
shown that there is no evidence to support this theory of a 
great migration from east to west, that on the contrary the 
leading races of Europe have been settled in that continent 
from early neolithic times, and that Aryan languages are and 
have for ages been spoken by peoples of widely different 
physical type. What the philologist has now to explain, there- 
fore, is not the extension of a race, but the diffusion of a 
language, which was, no doubt, the language at first of a par- 
ticular race in some European or Asiatic district, but had the 
exceptional fortune to be adopted by other surrounding races, 
each of whom spoke it with its own accent and idiom, so that 
it became a group of dialects, instead of a uniform tongue. 
In view, then, of this altered aspect of the question, it is not 
necessary to contend that the Bronze Age invaders of Britain 
were of Aryan stock ; it suffices to show that they were nearly 
akin to continental peoples who spoke an Aryan language. 

At the close of the Neolithic Age a race of the same 
physical type as the men of the British round barrows is found 
in possession of the valleys of the Seine and Rhine, and even 
as far afield as Northern Italy and the upper valley of the 
Danube. Of this race it is clear that the round-barrow men 
formed an offshoot, attracted to Britain by causes which cannot 
now be ascertained. Now while the portion of this stock 
which was settled in the valley of the Po became the progenitor 
of the Umbrian people and the creator of Latin, the bulk of it 
undoubtedly went to make the great Celtic nation of history, 
speaking an Aryan language closely related to Latin, out of 
which Welsh and Irish have in course of time developed. The 
term Celtic is perhaps unfortunate, though there can be now no 


question, in the domain of philology at least, of substituting CHAP. 
another for it. It has been often pointed out that the name l ' 
Celt was never recognised as a national designation by the 
whole of the family of tribes to whom we nowadays apply it ; 
in Britain, for instance, it was quite unknown, and, until the 
rise of comparative philology, no one would have dreamed of 
looking for a Celt in these islands. Historically, the name 
is that which the south-western wing of the Celtic-speaking 
community bore among themselves when they first touched 
the Mediterranean ; 53 owing to the establishment in this region 
at an early date of a Greek colony at Marseilles, it became 
widely diffused as the Greek term for the whole race, nor was 
it until a later age that the name which the more northerly 
tribes gave themselves, viz., Galatae, became generally known, 
and, owing to its adoption by the Romans in the form Galli, 
finally ousted the other as a general designation. In choosing 
to speak, therefore, of the primitive Celtic race, philologists use 
a phrase which is far from exact, but, so long as the true 
position of the " Celts of history," the dwellers between the 
Seine and the Garonne, is kept in mind, no confusion can arise 
from the use of a term which has won for itself, as a symbol 
of what is common to Gael and Brython, a very secure position 
in the English language. 

The primitive Celtic race was, however, not homogeneous. 
No fact in connection with this group of languages is better 
known than that it falls into two main divisions, the Goidelic 
(to use the convenient terms coined by Sir J. Rhys) 64 and the 
Brythonic. To the former belong the Celtic tongues spoken 
in Ireland, Scotland and the Isle of Man, to the latter those of 
Wales and of Brittany, together with the extinct Cornish. 
Within the two divisions the linguistic barriers are not formid- 
able ; the gap is widest between Welsh and Breton, though 
even in this case there is no reason why we should look back 
more than fifteen centuries for the point of separation. But 
the cleavage between the two divisions themselves goes deep 
and shows itself in matters of the first importance. One of 
the best-known points of divergence is the treatment of primi- 
tive QU ; while Goidelic long retained the original sound (it 

" See page 31 below. M In the first edition of Ctltic Britain (1882). 

2 * 


CHAP, is still heard in Manx), only reducing it to C within historic 
times, Brythonic had in all words in which it occurred converted 
it into P, by the double process of dropping the guttural and 
modifying the labial, before the Brythonic peoples appeared 
on the historic stage. Thus early Celtic " ^aetveres " ( = four) 
becomes in Irish " reathair," but in Welsh "/edwar " and in 
Breton "/evar"; old Welsh "ma/>" answers to Irish " ma<r," 
which in the Ogam inscriptions is found as " magui " (in the 
genitive), with the primitive sound unaltered. The differences 
between Goidelic and Brythonic, of which one example has 
been given, are so numerous and fundamental that it must be 
supposed that the two peoples who spoke these two forms of 
the language were either long separated from each other or 
were not originally, despite the use of a common tongue, of 
the same stock. Which, then, it will be asked, was the branch 
which invaded Britain at the beginning of the Bronze Age? 
The answer can hardly be a matter of doubt if the relative 
position of Goidels and Brythons in our island group since the 
beginning of history be present to the mind. The Goidelic 
has always lain to the north-west of the Brythonic country, 
which has interposed between it and the Continent ; before it 
was proper (if it is now) to speak of a " Celtic fringe," there 
was a Goidelic fringe, the remnant of an early Celtic migration 
driven over mountain and sea by the pressure of a newer. 

It was at one time believed that the Goidelic migration 
to Britain was so complete as to leave no trace behind in the 
Celtic area of the Continent. But of late the view has been 
advanced that the historical Celts, the inhabitants of Central 
Gaul, were to a great extent Goidelic in speech and that herein 
lay the difference between them and their Galatic neighbours, 
whose language beyond question belonged to the Brythonic 
branch. 55 

The characteristic weapon of the Age of Bronze was the 
bronze axe. At first, it was of the simplest form, reproducing 
in metal the outlines of the neolithic celt ; perfectly flat, it had 
no means of attachment to its wooden handle, which must 
accordingly have been cleft to receive it. As the art of the 
smith developed, more serviceable forms were devised ; in the 

88 For the whole subject see W. People, chap, i., and Rhys, Celtce and Galli 


palstave type there are stop-ridges which prevent the head of CHAP, 
the axe from slipping too far back into its handle and a loop 
which enables head and handle to be firmly fixed by means of a 
thong to each other. Last of all came the socketed axe, a shell 
of metal into which the wooden haft was inserted. Side by side 
with the development of the axe may be traced that of the 
sword. The neolithic knife was first reproduced in bronze in 
the form of a short squat dagger, secured by rivets to a wooden 
or horn handle ; this grew in course of time into a long and 
graceful , sword, of which the hilt as well as the blade might 
occasionally be of bronze. At the same time the dagger 
underwent another development and grew into a bronze spear 
head, having a hollow stock for the insertion of the wooden 
shaft. The flint arrow head of the Neolithic Age held its own 
undisturbed throughout this period, for no one seems to have 
thought it wise to waste good bronze in the making of a missile 
which the owner was quite likely to lose on the first occasion 
of using it. 

It is not probable that many, if indeed any, of these 
developments were due to the ingenuity of British artificers. 
Like the discovery of bronze itself, the revelation to the neo- 
lithic craftsman of the happy effect upon his unmanageable 
copper of a moderate admixture of tin, they must be referred 
to the Continent, with which the insular Celts, through their 
kinsmen on the Rhine and Seine, no doubt maintained a con- 
nection. But if the ideas came from abroad, there is no lack 
of evidence that the workmanship was to a great extent 
domestic. The identity of the word for " smith " in all the 
Celtic languages suggests that the craft was known to the 
Bronze Age men at the time of their arrival here ; 66 for later 
periods there is the still stronger proof afforded by the dis- 
covery of the tools employed, including the moulds in which 
the various weapons were cast. Many articles were no doubt 
imported, and this would especially be the case in the late 
Bronze Age, when work was being produced on the Continent 
in which beauty of design was considered no less than mere 

The custom of cremation made its appearance in Britain in 

"See Urk. Spr. s.v. "goban," 114 (Irish "goba" = Welsh and Cornish 


CHAP, the early part of the Bronze Age. Whatever its precise 
origin, it had no doubt a religious significance, and, far from 
implying contradiction of the neolithic idea, that the existence 
of the dead was a continuance in the spirit world of the life 
once lived in this, it may well have been designed for the pur- 
pose of giving that idea a more reasonable expression. In 
order that the human spirit might range freely in the new 
sphere of its being, it was right that it should be delivered 
from the incubus of the body. The corpse was first burnt to 
ashes on a funeral pyre. This rite was still remembered by the 
Welsh at the date of the composition of the so-called " his- 
torical " Triads, for in them it is told how a party rode on the 
back of one horse to see the smoking bier of the men of 
Gwenddoleu who had fallen in the fight of Arderydd. 57 The 
remains were then in most cases placed in a funeral urn, which 
was a roughly shaped crock about a foot or a foot and a half 
high, decorated with the characteristic patterns of the period. 58 
As in neolithic times, implements and trinkets of various kinds 
were buried with the dead, though it is remarkable that hardly 
any bronze axes or swords or spear heads have been found so 
interred in these islands they were perhaps too valuable to 
be spared. In the disposition of the urn, no set custom seems 
to have been observed ; sometimes it was enclosed in a " cist " 
or small receptacle of flat stones, sometimes it rested mouth 
downward on a single slab, and sometimes it had no protec- 
tion whatever. Over the whole a great barrow or cairn was 
raised, which, as no passage had to be provided to a perma- 
nent grave-chamber, was naturally round in figure. The true 
characteristic of the Bronze Age burial, as seen in the greater 
part of Britain, is not, however, the shape of the covering 
" crug " or " tomen," but the absence of an accessible chamber. 
Though the fashion of burning the dead spread during this 
age to every corner of the British Isles, and even established 
itself in remote districts among people who long after the close 
of the neolithic period still built sepulchral chambers, 59 it did 

87 " Ar eil marchlwyth aduc coruann march meibon differ gosgorduawr . . . 
y edrych ar vygedorth llu gwendoleu yn arderyd " (Triad in Mob. 301). 

88 The art of the Bronze Age is the subject of chap. ii. in Mr. Romilly Allen's 
Handbook to Celtic Art (London, 1904). 

59 Prehistoric Scotland, p. 478. 


not at once displace the older custom of simple burial. Bronze CHAP. 
Age barrows contain unburnt skeletons as well as bones which 
have undergone cremation, and evidence which would other- 
wise be wanting is thus obtained as to the physical character- 
istics of the population during this period. The old race and 
the new are found lying side by side, so that, while there can 
be little doubt that the vigorous, athletic Celt was everywhere 
master and lord, he must in many parts of the country have 
spared for servile labour the older inhabitants of the soil. He 
would especially need their help, if, as there is good reason to 
think, he practised a primitive agriculture. The neolithic life 
was, probably, purely pastoral, but comparative philology 
shows that the peoples among whom the European form of 
Aryan speech was evolved were familiar with the tillage of 
the soil, 60 though their implements were of the rudest, and this 
conclusion is supported by the discovery of grain-crushers in 
the round barrows of the Yorkshire wolds. 61 In other re- 
spects, also, the civilisation which was the fruit of the union of 
the two races was an advance upon that of the Age of Stone ; 
the pottery made was of a better kind and was stamped with 
geometrical patterns ; ornaments were more elaborate, among 
them being handsome necklaces of jet, such as that of which 
the remains were found in 1828 in a Bronze Age grave at Pen 
y Bone, near Holyhead ; 62 while the occupation of hill fort- 
resses with their round stone huts shows the existence of a 
tribal life, presided over by the tribal chieftain, the Celtic rix or 
"rhi". 63 

The sepulchral monument did not always in this age take 
the form of a cairn or barrow. Investigation of the stone 
circles of Great Britain and Ireland has led to the conclusion 
that very many of them are Bronze Age burying-places. 
Though the greater triumphs of the megalithic builder, such as 
Stonehenge and Avebury, can hardly be disposed of so lightly, 
and may very well have had some religious purpose, the ordin- 
ary circle as seen in Wales, for instance, on the hills above Pen- 
maen Mawr or in the parish of Llanidan, Anglesey, was no 
doubt a place of burial. Barrows, both chambered and un- 

80 See part iv. chap. v. of Schrader's work. 

61 British Barrows, pp. 114-5. 

82 Ancient Stone Implements, p. 459. w Urk. Sfir. 230. 


CHAP, chambered, are often found to have a ring of standing stones 
around them, so that in the stone circle a part of the old form 
of monument survives, after the barrow itself has been aban- 
doned as unnecessary. The simplest memorial of all is the 
single standing stone or " maen hir " (long stone), which may 
have had many purposes from time to time, but was certainly 
often used to mark the resting-place of the dead. As in the 
case of cromlechs, the traditional names of " meini hirion " are 
no safe guide to their early history : " Llech Idris" (The Stone 
of Idris), 84 " Carreg y Bwgi " (The Bogey Stone) 65 and " Maen 
Beuno " (The Stone of Beuno), 66 for instance, but preserve the 
speculations of a peasantry who knew nothing of the original 
design of the monument. 

Abundant remains of the civilisation of the Bronze Age 
have been discovered in Wales, and this in all parts of the 
country, so that it is clear that the conquering race of Goidels 
in time made their influence felt in every corner of it. But 
each successive tide of culture which has washed the shores of 
Britain has made a belated appearance in Wales, and hence it 
need not be supposed that the new era began in these western 
highlands until long after it had obtained a firm foothold in 
the south-eastern plain. That the Bronze Age of Britain was 
far advanced ere its customs rooted themselves in Wales may, 
indeed, be inferred from the rarity in Welsh burials of the 
period of any other method of disposing of the dead than 
cremation, 57 and from the prevalence of highly developed forms 
among the bronze weapons which have been collected on 
Welsh soil. In the life of the little community whose home 
and burying-place under the brow of a limestone cliff were 
recently excavated by Prof. Boyd Dawkins near Newmarket 
in Flintshire, the first contact with the new culture seems to 
be brought to light. 68 Charcoal, the remains of domestic 
animals and a polished flint flake ; great numbers of human 
bones huddled together in a square sepulchral chamber here 
was nothing which might not be neolithic. But the skulls, 

64 Near Trawsfynydd. 

65 Between Cellan and Llan y Crwys (Arch. Camb. IV. ix. 325). 

66 Arch. Camb. III. iii. (1857), 299-300. 

67 British Barrows, p. 21 (note i). 

88 Arch. Camb. V. viii. (1891), 71-2 ; Archczological Journal, 1901, pp. 322-41. 


though for the most part long, included some of which the CHAP. 
breadth indicated the presence of the newer race, and, most 
convincing proof of all, the pottery was of unmistakable 
Bronze Age pattern. Like many conquering stocks, the 
Goidels who settled in Wales may have been a minority, a 
military aristocracy ; it could hardly have fallen out otherwise 
that in a later age the Silures should have still vividly recalled 
in aspect and build their Iberian kinsmen. Nor is it likely, 
from what is known of the Welsh tribes in the time of Caesar, 
that Aryan agriculture made much headway in the country. 
But the pottery, the weapons, and the burial usages of the 
Bronze Age became ultimately as common in Wales as in the 
rest of Britain. In Pembrokeshire, for instance, where the 
number of cromlechs shows that neolithic ideas survived in 
strength to a late period, as compared with the time of their 
disappearance from the rest of South Wales, over thirty round 
barrows, locally known as " tumps," have been excavated and 
in almost all the remains found have been those of cremated 
bodies. 69 In the mountain-locked glen, sloping seaward, 
where now stands the village of Penmaen Mawr, ensconced 
in what was, until the making of the Chester and Holyhead 
railway, one of the most secluded spots in North Wales, a 
Bronze Age burial was in 1889 accidentally brought to light 70 
The urns, shaped and decorated in the usual manner of the 
period, were found resting mouth downward on flat stones ; 
they were full of calcined bones. Further, to put beyond all 
doubt the epoch to which they belonged, in each of two of 
them a small bronze pin was discovered. 71 

NOTE TO CHAPTER I. ii. Cantref y Gwaelod. 

The story of the Submerged Hundred first makes its appearance in a poem 
in the Black Book of Carmarthen (535, 543), written in the last part of the book, 
and, therefore, about 1200. This poem has been often translated ; there is, for 
instance, an English version in Meyrick's History of Cardiganshire (2), 153 and a 
modern Welsh one in Cymru Fu (p. 6). The best is, however, the most recent, 

89 Lit. Bug. p. 29. 

Arch. Camb. V. viii. (1891), 33. 

71 It should be observed that Dr. T. Rice Holmes, in his recently published 
Ancient Britain and the Invasions of Casar (Oxford, 1907), challenges many of 
the accepted views as to the prehistoric civilisation of these islands, and, in par- 
ticular, the view that the Early Bronze Age invaders were Celts. The whole 
subject must be regarded as still under discussion. 


CHAP, that of Sir J. Rhys in the Cymmrodorion Transactions for 1892-3 (pp. 14-16), from 
I. which it appears that the Plain of Gwyddneu was overwhelmed by the sea by 
reason of the wickedness of its inhabitants, who had given themselves up to eat- 
ing and drinking and insolent pride of heart. The person who let loose this 
judgment upon the land was a maiden, perhaps called Margaret ("Mererid"), 
who at a time of feasting suffered the waters of a magical well which was under 
her charge to escape and overflow the country round. What the share of " wit- 
less Seithennin " was in the catastrophe is not apparent; he survives it, however, 
and is called upon to survey the melancholy scene. Such was the primitive 
story ; it is supplemented in one point by the compiler of the earliest form of the 
Pedigrees of the Saints (also dating from about 1200), who speaks of five " Saints " 
as sons of King Seithennin of the Plain of Gwyddno, whose realm was swallowed 
up by. the sea (Myv. Arch. II. 24 [416]). For the germ of the modern legend, 
which is in many ways a very different one, we have to look to the third series 
of Triads, belonging to the sixteenth century ; the third of the Three Arrant 
Drunkards of Britain (a festive group unknown to the older triadic literature) is 
there said to be Seithinyn the Drunken, King of Dyfed, who in his cups let the 
sea loose over the Lowland Hundred, a region of fair cities and the patrimony of 
Gwyddno Garanhir, King of Ceredigion (Myv. II. 64 [404]). The well maiden 
has now disappeared, Seithennin has become the author of the mischief, and the 
drowned kingdom is no longer his, but that of his neighbour Gwyddno. But 
the famous embankment has still to be introduced into the story. It was to the 
antiquary Robert Vaughan of Hengwrt (1592-1667) that the idea first occurred 
of connecting the story of the Lowland Hundred with the natural causeway near 
Harlech called by the peasantry, in that age as in this, " Sarn Badrig" or St. 
Patrick's Causeway. The popular explanation no doubt was that this was the 
saint's private road home to his beloved Ireland, but for Vaughan it is " a great 
stone wall made as a fence against the sea," which he has no difficulty in sup- 
posing to have once been a rampart of the buried realm. Lewis Morris, in the 
next century, took the same view, and, remembering the poem in the Black Book, 
added a suggestion of his own, " that by drunkenness the flood-gates were left 
open " (Celtic Remains, p. 73 ; cf. p. 390). But one more touch was needed to 
give the narrative its modern form ; the business of the flood-gates must be speci- 
ally laid at the door of Seithennin, who must play the part of the drunken lock- 
man. This is done in Owen's Cambrian Biography (1803) ; under the patronage 
of so influential a student of Welsh antiquities, the story as thus rounded off 
won great popularity and furnished an attractive theme for literary treatment. 
Englishmen were made familiar with it by the fascinating pages of The Misfor- 
tunes of Elphin ; for Welshmen it was vigorously told in the verse of Hiraethog 
and leuan Glan Geirionydd. 



(Use has been made of Elton, Origins of English History ; Evans, Coins 
of the Ancitnt Britons ; Rhys, Celtic Britain ; Rhys and Brynmor-Jones, 
Welsh People ; A. J. Evans, Rhind Lectures, as well as of the original author- 
ities, Caesar, Diodorus Siculus, Strabo and Ptolemy.) 


As the Bronze Age draws to its close, there is more and more CHAP, 
evidence of trade intercourse between Britain and the outer n - 
world. The first civilised visitor to these shores would seem 
to have been the Phoenician trader, who sailed from Southern 
Spain to the south-western corner of Britain : 

There, where down cloudy cliffs, through sheets of foam, 
Shy traffickers, the dark Iberians come, 
And on the beach undid his corded bales. 

It is true that the Cassiterides, the " Isles of Tin," whence the 
Phoenicians obtained great quantities of this metal, so much 
in request for the making of bronze, are no longer identified 
with the Scilly Isles, but it is suggested with much probability 
that they were the British Isles themselves. 1 The Phoenicians 
were careful to conceal their situation from the knowledge of 
the rest of the world, so much so that Herodotus in the fifth 
century B.C. doubted their very existence, 2 and thus it came 
about that when they finally came to light, as the result of the 
voyage of Pytheas, under another name, geographers still con- 
tinued to locate the Isles of Tin off the north coast of Spain and 
distinguished them from the isles of Albion and lerne. Side by 

1 The question of the situation of the Cassiterides is fully discussed in 
Elton's Origins of English History, chap. i. But the view there put forward is 
now abandoned in favour of that of Miillenhoff, for which see W. People, p. 61 
(note) ; Arbois de Jubainville, Cours de Litterature Celtique, xii. (1902), pp. 
4-11 ; Eng. Hist. Rev. xix. (1904), p. 140 (note). 

2 Otire v^ffovs olSa KaffffirtplSas tovffas (iii. 115). 



CHAP, side with this intercourse with Phoenician civilisation, which 
seems to have had little or no effect upon the inner life of 
Britain, there was communication, too, with the Celtic peoples 
of the mainland. About six or seven centuries before Christ 
the Celts of the Continent had attained a high degree of skill 
in the designing and fashioning of weapons and ornaments ; 
it was the era of what is known, from one of its important 
seats in the Noric Alps, as the Hallstatt culture, 3 and the im- 
portation into Britain in the late Bronze Age of bronze articles 
of the Hallstatt type shows that there was at this time a trade 
route which carried such articles across the Straits of Dover. 
Britain was gradually being drawn into contact with the organ- 
ised communities which, clustering around the Middle Sea, 
formed the civilised world of the day. 

In the middle of the fourth century B.C. the Greeks, wide 
as was their geographical knowledge, had no certain knowledge 
of the islands in the Ocean beyond the region of the Celts. 
But commercial rivalry came to the aid of geography. The 
adventurous traders of the Greek colony of Massalia (Marseilles) 
in Southern Gaul, having vainly sought to obtain access to the 
regions beyond the Pillars of Hercules our Straits of Gibraltar 
whence their Carthaginian competitors drew such abundant 
stores of tin, resolved to send out into the unknown Ocean an 
expedition which should either open up for them this market 
or find a new one in the mysterious North-west. A learned 
citizen of the place, one Pytheas, 4 an astronomer, was placed 
in charge, and under his guidance a voyage was made which 
not only achieved the immediate commercial purpose, but also 
added vast regions to the known world of Greek geographical 
science. Having sailed right round the Spanish coast, the ex- 
pedition directed its course to Brittany and thence to the 
Straits of Dover; Pytheas not only landed in Britain, but 
spent a considerable time in the island, making himself fully 
acquainted with at least the south-eastern portion and ascertain- 

3 There is a concise account of the discoveries at Hallstatt and of their 
significance in the second edition (1900) of Munro's Rambles and Studies in 
Bosnia and Herzegovina, chap. xi. 

4 The works of Pytheas have long been lost, and our knowledge of his 
doings is entirely derived from the incidental references of later (and often hos- 
tile) writers. Full justice is meted out to him in Elton's Origins of English 
History, chaps, i. and ii. 


ing that there would be no difficulty in obtaining from the CHAP, 
mines of the Cornish peninsula an ample supply of tin for the 
purposes of his employers in Marseilles. With good reason is 
he regarded as the discoverer of Britain, its Columbus, who 
first made it known to learning and civilisation. 

The Brythonic Celts had not yet, so far as can be gathered, 
made any settlements on the northern side of the channel. 
They held the opposite coast, from the estuary of the Rhine to 
that of the Seine, but were still a purely continental people. 
Yet it is probable that the two great islands which Pytheas 
brought within the ken of Greek merchants and scholars were 
even then known as the Pretanic Isles (ai TIperravLical vrja-oi), 
a form preserved for many centuries by writers on geography. 6 
The truth is, this term has in origin no connection with 
Britain or the Britons ; it is, on the other hand, closely related 
to the Ynys Prydain of the Welsh, who never, save under the 
influence of a misguided purism, speak of Ynys Prydain. 
Prydyn, the mediaeval Welsh name of Scotland, is clearly the 
same word, and so too, Celtic philologists assure us, the Irish 
name for the Picts, viz., Cruithni. 7 Thus the Pretanic Isles 
may mean the Pictish Islands and, as the form (with its/ for 
the Goidelic gu) is Brythonic, it receives the most natural ex- 
planation if we suppose it to be a name given to the group by 
the Brythonic or Galatic dwellers on the northern coast of 
Gaul (who perhaps learnt it from Goidelic predecessors), on 
account of the large " Iberian " or Pictish element in its popula- 

When it was desired to distinguish the two islands, the one 
was called Albion, the other lerne. According to the elder 
Pliny, Albion was known to be a more ancient name than 
Britannia, 8 and this is confirmed by the fact that, in the forms 
Alba, Alban (whence the Welsh " yr Alban " for Scotland), it 
has long been used by the Goidelic peoples to denote either 

"This, the former opinion of Sir John Rhys (W. People, p. 75), is now 
abandoned by him (Celt. Br. (3), p. 4). 

"Mtiller believes UptrraviK^ to be the true reading throughout Strabo's 
Geography (Firmin-Didot edition, 965). It is also the form regularly used by 
Marcian of Heraclea (see Muller's edition of the Minor Greek Geographers, 
Paris, 1882, vol. i. pp. 516-7). 

7 Urk. Spr. 63 ; W. People, pp. 76, 79. 

8 " Albion ipsi nomen fuit, cum Britanniae vocarentur omnes " (Nat. Hist. iv. 


CHAP, the whole or a substantial part of Britain. Albion was the 


" white land," 9 whose gleaming cliffs were the constant wonder 
of the continental Celts as they gazed across the blue waters 
of the Channel, and thus the mediaeval Welsh scribe, though he 
knew nothing of Celtic philology and was thinking only of the 
Latin " albus," was happily inspired when he rendered it " Y 
Wen Ynys " (The White Island). 10 


It was, perhaps, about the middle of the third century B.C. 
that the Brythons began to settle in Britain. 11 Their arrival 
marks the appearance in the island of the third and last of the 
three considerable race elements which have gone to the making 
of the Welsh people. If it be contended that the first of these, 
the Neolithic, was the most important in respect of its contribu- 
tion to the national physique and character, and that the second, 
the Goidelic, was the source of the early political and social 
institutions of the Welsh, it cannot be denied that it is to the 
third we owe the Welsh language. Notwithstanding the alien 
idioms to which reference was made in the last chapter as 
traceable to an African origin, Welsh is historically an Aryan 
and a Celtic tongue, the lineal descendant of the speech of the 
men who, as Caesar says, 12 first came hither, on hostile and 
plunderous errands, in the age immediately preceding his own. 

In the well-known exordium to the " Gallic War," Caesar 
gives this people the general name of Belgae. But it may be 
doubted whether this was really an ancient race-name which 
they recognised among themselves. Caesar elsewhere 13 states 
that most of the Belgic settlers in Britain retained as colonists 
the tribal names which they had borne at home, and this is 
known in some instances to have been the case ; there were, 

9 Urk. Spr. 21 ; IV. People, p. 77. 

10 " Ar amser hwnnw y gelwit hi y wen ynys " (Brut y Brenhinoedd in Bruts, 
58 ; cf. also 40). 

11 So Dr. A. J. Evans in his Rhind Lectures. Cf. W. People, p. 5. As a 
result of the Aylesford finds (Archczologia, Hi. 317-88), archaeologists are now 
disposed to regard the Belgic settlement as a late Brythonic invasion of about 
150 B.C. 

l *Bell. Gall. v. 12: " lis qui praedae ac belli inferendi causa ex Belgio 
transierant . . . et bello illato ibi permanserunt atque agros colere coeperunt ". 

13 Ibid. : " Qui omnes fere iis nominibus civitatum appellantur quibus orti 
ex civitatibus eo pervenerunt ". 


for instance, Paris! to the north of the Humber as well as CHAP. 
Parisii on the banks of the Seine, Atrebatii south of the Thames 
as well as Atrebates where Arras now stands. 14 But, strangely 
enough, while there was in Britain a tribe of Belgae occupying 
what may be called the Bath and Winchester region, 16 no 
corresponding tribe appears in Gaul, the name being applied 
instead to a great tribal group, by no means closely knit 
together. It would almost seem, therefore, as if the original 
Belgae had, as Sir J. Rhys suggests, 16 transferred themselves 
lock, stock and barrel to Britain, but had left behind them in 
Northern Gaul, in the wide extension of their tribal name, a 
lasting memorial of the military supremacy which they had 
once exercised in the region. 

If any general name was current among the Brythonic or 
Belgic tribes of Northern Gaul, it was probably Galat, or, if the 
Latin be preferred to the Greek form, Galli. Diodorus of Sicily, 
writing in the first century B.C., essays to correct the popular 
misuse of the terms Celt and Gaul. 17 The Celts (Ke\Tov9), he 
says, are those who inhabit the " hinterland " of Marseilles and 
the regions adjacent to the Alps and the Pyrenees ; the country 
bordering on the Ocean and the Hercynian Forest and away 
east to Scythia belongs to the Gauls (raXara?). It is the 
Romans who have caused confusion by massing all together as 
Gauls (JaXaTa? for Gallos). So wide, in fact, had the applica- 
tion of the name now become in Latin usage that Caesar could 
find no better distinctive name for these northerners than 

The vogue of the Galatic name was the result of the career 
of world-wide conquest upon which the Galatae embarked about 
the beginning of the fourth century B.C. Celtic scholars derive 
the word from a root which in Irish has yielded "gal," valour, 
and in Welsh "gallu," to be strong, and thus explain it as 
signifying " strong men," " braves ", 18 Few races of fighters, 
indeed, have left a more definite impression of themselves in 
the pages of antiquity. For the invasion and settlement of 
South-eastern Britain was but one of their achievements ; the 

14 All four tribes are mentioned by Ptolemy (II. iii. 10, 12 ; viii. 10 ; ix. 4). 
18 B&/xeu KO.I w6\fis' *l<TKa\is, "fSara Sfpftd, OuWa (Ptol. II. iii. 13). 
18 W. People, p. 6. " Diod. Sic. v. 32. 

18 Urk. Spr. 107 ; Celt. Br. (2), 267, 298. Whitley Stokes appears to regard 
Gallus as a word of different origin (Urk. Spr. 108). 


CHAP, sack of Rome in 390 B.C. was their work and the attack 
upon Delphi in 278 B.C., while the name Galatia preserves to 
all time a record of their great settlement in the heart of Asia 
Minor. In some of these movements the Celts strictly so 
called may have borne a part, but the originating impulse was 
Galatic, spreading in every direction from the valley of the 
Rhine. Thus the figure of the Galatic warrior looms large 
in classical literature, and there is little difficulty in forming a 
mental picture of the kind of man who came over in the 
Brythonic host of invasion. 19 He was of commanding height ; 
Strabo testifies to seeing at Rome British lads who were six 
inches taller than the tallest men in the city. The women, 
inured to the same perils and even sharing the rough fortune 
of battle, were built on the same scale, and possessed all the 
spirit and daring of their spouses. The limbs of the Galat 
were long and vigorous, but he was not well put together ; 
his great, unwieldy frame, formidable by reason of its weight, 
had not the compact agility of the more resourceful southerner. 
He had the clear red and white complexion of the true blond ; 
his hair was yellow or yellowish-brown, and so proud was he 
of its golden tint that, in Gaul, at least, he was in the habit of 
heightening the colour by means of a dye. His ample locks 
were bound together in a mass above and behind the head ; 
his chin was bare, but long moustaches drooped over the 
mouth. His under-garment was a sleeved tunic, while on his 
legs he wore the characteristic Gallic breeches, fitting loosely 
round the ankle ; on his shoulders was a thick woollen cloak, 
fastened in front by a brooch of delicate workmanship. Such 
was the outer man ; in disposition the Galat was impetuous, 
boastful, self-indulgent, a shedder of much blood, but brave, 
hospitable, true to his honourable ideals and at bottom of 
childlike simplicity. He was lavish in expense upon his own 
adornment and that of his arms and warlike accoutrements, 
but no money could buy from him the embalmed heads which 
had been handed down to him as the grim trophies of his 
ancestors' valour. 

This portrait reveals to us, as was intended by the polished, 

19 The account which follows is based upon Diod. Sic. v. 26-31, and Strabo, 
IV. iv., the Firmin-Didot edition having been used in each case. 


self-restrained observers to whom we owe it, the lineaments of CHAP, 
a barbarian, but the race was nevertheless one, as these writers 
themselves allow, of great capacity. In several important 
respects, it had already made great advances in culture. Of 
these the chief was perhaps the adoption of iron instead of 
bronze for the manufacture of weapons and cutting imple- 
ments. The age of the Hallstatt culture, though it coincided 
with the late Bronze Age of Britain and is only represented in 
this country by bronze objects, was one in which considerable 
progress was made in Central Europe in the adaptation of 
iron to warlike and domestic uses, and thus, when the time 
came for the Galatic tribes to play their part in the history of 
the West, they were fully equipped in this respect Sword, 
lance-head and javelin were of iron ; so too, in many cases, the 
metal fittings of the long shield which protected the warrior's 
whole body and the metal brooch which pinned his cloak 
around him. It is most probable that the use of iron was 
introduced into Britain by the Brythonic invader, and that the 
beginning of the British Iron Age thus coincides with the era 
of the Brythonic settlement. 

This is certainly the conclusion suggested by the small 
number of barrows which can be assigned to the period between 
the introduction of iron and the arrival of the Romans and by 
the entire absence of barrows of the kind in Scotland and 
Ireland, districts only slightly affected by the incoming of the 
Brythons. But it was not the use of a new material alone 
which distinguished the Galatic peoples as a progressive 
community ; their skill in the working of metals, of bronze, 
of silver, and of gold, no less than of iron, furnished them 
with numberless articles of ornament, of which the artistic 
beauty of design was every whit as remarkable as the ex- 
cellence of the workmanship. Nothing more free and graceful 
can be imagined than the best decorative work of this age, 
known sometimes as the Late Celtic, but on the Continent 
associated with the name of La Tene, a Galatic fortress on 
Lake Neuchatel in Switzerland which has yielded abundant 
remains of the period. Though they lived in roughly timbered 
huts of wattle-work and sat down to their meals on littered 
straw covered with rugs of wolf- or dog-skin, the Galatic 
chiefs had torques of gold about their necks, golden bracelets 
VOL. I. 3 


CHAP, on their arms, and arms of offence and defence which, whether 
of bronze or of iron, were richly moulded and incised by 
artists of genuine ability. 20 

Two other evidences there are of the progress in civilisation 
which the race had achieved at the time of its settlement in 
our island its mastery of horsemanship and charioteering and 
its possession of a coinage. Strabo bears emphatic witness to 
the former point ; the Gauls, he says, fight best on horseback 
and supply Rome with its most efficient cavalry. He speaks, 
too, as does Diodorus, of their use of chariots in war, and it will 
be remembered how Caesar, when he came into contact with 
this kind of fighting in Kent, was struck by the ease with 
which his enemies managed their horses and by the reckless 
daring of their movements. 21 Nor is archaeological evidence 
wanting of the important place occupied in Galatic life by the 
horse and the wheeled vehicle. Horse trappings, and in par- 
ticular bridle bits, are among the commonest objects found 
among Late Celtic remains, and at La Tene an entire chariot 
wheel was discovered. The series of Gaulish coins begins 
about 300 B.C., and is altogether derived by imitation from a gold 
coin issued about the middle of the previous century by Philip 
II. of Macedon ; 22 yet, though in this direction little artistic 
originality was displayed, the mere existence of a coinage 
shows that the multiplying needs of a community advancing 
towards civilisation were making the old methods of barter 

There is little difficulty in understanding how a race thus 
equipped overbore the opposition of the earlier inhabitants of 
Britain and established flourishing colonies in the most attrac- 
tive parts of the island. Nor need there be much hesitation 
as to the region where they first found a foothold. They 
would naturally direct their skiffs in the first instance to Kent, 
and, in conformity with this, Caesar states that Kent was 

20 The questions connected with the introduction of iron into Britain and the 
development of the Late Celtic style of decoration are discussed by Dr. Munro 
in Prehistoric Scotland, chap, vii., and Rambles and Studies in Bosnia and Herze- 
govina, second edition, chap. xi. See also A. J. Evans, Rhind Lectures, and 
J. Romilly Allen, Celtic Art in Pagan and Christian Times. 

21 Bell. Gall. iv. 33. 

22 Evans, Coins of the Ancient Britons, 1864 (Supplement published in 
1890), pp. 23-5. 


much the most civilised part of Britain and in its customs CHAP. 


differed little from Gaul. 23 This being the case, it is some- 
what remarkable that the inhabitants of this district were not 
known by any tribal name connecting them with a Belgic 
tribe on the other side of the water. They were simply 
termed Cantii, from the place name Cantion, which is said, like 
Albion, to signify " white land ", 24 This anomaly has led to the 
ingenious conjecture that the true name for them was Brittones, 
that as the van of the Belgic invasion they migrated bodily 
from their first homes at the mouth of the Somme (where 
there is still a village called Bretagne) and that this was how 
their continental kinsmen came to apply the British name to all 
inhabitants of our island. It must be remembered that the 
Brythonic-speaking members of the Celtic family have always 
dubbed themselves Brython, representing an old Brittones, 
while the Goidelic-speaking members have in like manner 
agreed to call them Bretain, for an older Brittani, so that, 
whether the Kentishmen be the original Britons or not, the 
term is properly confined to the Belgic or Galatic immigrants 
of the Iron Age, who may henceforth be spoken of as Bry- 
thons. Why both the Greeks and the Romans should have 
preferred the Goidelic to the Brythonic form of the name 
(Greek BperravoL, Latin Brittani and, less correctly, Britanni) 
has not yet been satisfactorily explained. 25 

The descriptions which Diodorus and Strabo give of Galatic 
life and manners both on the Continent and in Britain are be- 
lieved to owe much to the travels of Posidonius, the Stoic 
philosopher who traversed Spain, Gaul, and probably Britain 
at the beginning of the first century B.C. These travels, how- 
ever, are not extant, and thus the first authentic account 
of the Brythons in their island home is furnished by Julius 
Caesar. When he made his famous expeditions in 5 5 and 54 
B.C., they were firmly established in the south-eastern region. 
In this, termed by him the " maritime part" of Britain, popula- 
tion was dense, agriculture flourished, sheep and cattle were 

""Ex his omnibus longe sunt humanissimi qui Cantium incolunt . . . 
neque multum a Gallica differunt consuetudine " (Bell. Gall, v. 14). 

84 W. People, pp. 77, 78. 

"See for the whole question Celt. Br. (2), pp. 205-13, with the modifications 
in W. People, pp. 6, 75-7. 



CHAP, to be seen in great abundance. 26 Arriving at the season of 
harvest, his troops were able on the first expedition to supply 
themselves with corn at the expense of the British farmer. 27 
Nor was the art of war neglected ; the progress of the Roman 
general was seriously opposed by men who were skilful in the 
use of chariots and of horses and in the fortifying of positions 
of natural strength ; some amount of political combination 
there was also, for a number of the tribes had united to resist 
the Romans under the leadership of Cassivellaunus, the first 
dweller in these islands whose name has been handed down by 
history. Much, no doubt, remained to show that the Bry- 
thons were but a partially civilised race, for instance, the habit 
of staining the body blue, indulged in by the Brythonic 
" braves " when they went forth to battle, 28 and certain singu- 
lar marriage customs. Yet the general view presented by 
Caesar's narrative is undoubtedly that of a thriving and busy 
people, whom there is no difficulty in imagining as the makers 
and buyers of the beautiful creations of Late Celtic art. 

Behind this prosperous Brythonic foreground, Caesar reveals 
to us, in dim and shadowy tints, a background of savage life 
of which, in all probability, he knew nothing from actual ob- 
servation. " The men of the interior," he says (distinguishing 
them from those of the coastwise lands), " for the most part 
sow no corn, but live on milk and flesh and clothe themselves 
in skins." 29 They are, in short, a pastoral and not an agri- 
cultural people, alike in their food, their clothing and their 
habits. Thus is briefly described the condition at this epoch 
of the older inhabitants of Britain, those who, according to 
their own traditions, were the autochthonous children of the 
island, sprung from its very soil, 30 but whom history recognises 
as Iberians and Goidels, descended from prehistoric immigrants. 
It is true that the Goidelic race, whatever may be said as to 

26 " Hominum est infinita multitude creberrimaque aedificia . . . pecorum 
magnus numerus " (Bell. Gall. v. 12). 

27 See especially Bell. Gall. iv. 32 (" frumentatum missa") and v. 17 
(" pabulandi causa "). 

28 " Omnes veto se Britanni vitro inficiunt " (v. 14) makes it clear that the 
Brythons are included. 

*Bell. Gall.v. 14. 

30 " Pars interior ab iis incolitur quos natos in insula ipsi memoria proditum 
dicunt" (v. 12). 


the Iberian, practised agriculture, but, apart from the very CHAP. 
summary character of Caesar's account and his qualification 
" plerumque," one must remember that the seizure by the 
Brythons of the best of the corn-growing regions would tend 
to discourage the tillage of the soil among a race thrown back 
upon the great pasture lands of the North and West. The 
history of Wales shows clearly that a people familiar with 
agriculture may be led by the physical conditions to relegate 
it to a secondary place in the tribal economy and to give its 
main strength to the rearing of sheep and cattle. 


During the century which followed Caesar's attacks upon 
Britain, the Brythons continued to advance in civilisation and 
to encroach more and more upon their Goidelic and Iberian 
neighbours. It was during this period that they developed an 
inscribed coinage, 31 and the range of Brythonic influence is well 
shown by the wide extent of country over which specimens of 
this coinage have been found. South-east of a line drawn from 
the Wash to the Bristol Channel it was apparently (save in the 
western peninsula) in general use, while occasional coins strayed 
as far as the Tyne and the west of Cornwall. 32 But no British 
inscribed coin has yet been found within the limits of modern 
Wales, and the impression is thus created, which is in the main 
confirmed by other considerations, that this district remained 
little affected by the Brythonic immigration and the culture 
associated with it down to the very eve of the Roman conquest. 
It was part of Caesar's " interior," a stronghold of primitive 
ideas, of barbarous customs and of simple modes of life. 

Using the information to be derived from the geographer 
Ptolemy and other writers who dealt with Britain after its 
occupation by the Romans, one may without much hazard 
attempt an outline of the tribal divisions of pre-Roman Wales. 
Along the northern shores of the Bristol Channel dwelt the 
Silures, already referred to as an Iberian people. The only 
place which is certainly known to lie within the ancient limits 

31 " The introduction of a legend on British coins does not appear to have 
taken place until about the period of the accession of his (i.e. Commios's) sons ' : 
(Evans, Coins, p. 156). 

39 See map in Evans, Coins, suppl. 


CHAP, of this tribe is Caerwent in Monmouthshire, 33 but, in view of 
the important part which they played in the resistance to the 
Romans, no narrower theatre can be assigned to them than the 
modern counties of Monmouth, Glamorgan and Brecknock. 
Celtic scholars have so far been baffled in the attempt to ex- 
plain the name from Brythonic or Goidelic sources, 34 and to 
this extent the theory that the race was predominantly non- 
Aryan finds confirmation. Yet traces of Celtic influence are 
not wanting also ; one of the principal rivers of the district 
bore the Celtic name of Isca, represented by the modern Usk, 
and, whatever the meaning of the name Venta by which the 
tribal centre was distinguished, its occurrence elsewhere leaves 
little doubt as to its Celtic origin. It may be inferred, too, 
from their willingness to accept, in the great struggle with 
Ostorius Scapula, the leadership of the Brython Caratacus, of 
the Catuvellaunian royal house, that the Silures set no impass- 
able barrier between themselves and the Celtic tribes of the 
South-east. Yet, when all is said, they appear as a race apart, 
confronting their foes with that stubborn spirit, incapable of 
being melted by kindness or quelled by seventy, which there 
are good grounds for regarding as a specially Iberian character- 
istic. 35 Hemmed in by the Forest of Dean on the east and 
thus parted from the Dobuni, separated from the Belgae and 
the Dumnonii by the Severn Sea, they held a position of 
great natural strength and made full use of its military ad- 

To the west of the Silures came the Demetae. Ptolemy 
alone of ancient writers mentions them, but their existence is 
placed beyond doubt and their situation indicated by the per- 
sistence of the name Dyfed, which is derived from the old form 
through the early Welsh Demet. Dyfed has had various 

33 < Venta Silurum " is found both in the Itinerary of Antoine and the Geo- 
grapher of Ravenna. I know of no ancient authority for the form " Isca Silurum," 
and indeed the tribal name appears to have been added in such cases, not for the 
purpose of distinguishing places of the same name, but in order to indicate the 
tribal centre. Ptolemy only assigns to the Silures the town of Bofa\aiov , which 
is reasonably, but not certainly, identified with the Burrium ( = Usk) of the 

34 Celt. Br. (2), p. 306. 

35 Strabo points out (IV. iv. 2) that the Romans found the Iberians much 
more difficult to conquer than the Gauls. They husbanded their forces and 
fought in small detachments, \rj<rrpiKu>s voKefiovvrfs. 


meanings at various epochs in Welsh history, but it has never CHAP, 
ceased to stand for a considerable region in South-western 
Wales. In conformity with this, Ptolemy 3 " assigns to this 
tribe the station of Maridunum, which can be none other than 
our Carmarthen. As his Luentinum is further supposed to be 
Llanio in Cardiganshire, it would appear safe to locate the 
Demetae in the three modern counties of Carmarthen, Cardigan 
and Pembroke. The tribe seems to have been reckoned for 
all purposes, save those of precise geography, as a section of 
the Silures, who probably held them in a kind of subjection ; 
thus, according to the elder Pliny, 87 the shortest passage from 
Britain to Ireland, for which he gives the very low estimate of 
thirty Roman miles, was " a Silurum gente ". If this be so, it 
is right to extend to the Demetae the description given by 
Tacitus of the dark -haired Silures and to regard them also as 
a race of Iberian origin, as indeed is suggested by the figures 
of Dr. Beddoe, who found as dark a population at Carmarthen 
as at almost any Welsh centre at which he made his observa- 
tions. 38 Celtic influences had no doubt been brought hither, 
as further to the east, by the victorious Goidels ; a Celtic name, 
Maridunum, was given to the fortress which stood at the head 
of the tidal portion of the Towy ; but there is nothing to show 
that the Brythons had made their way into this region before 
the Romans appeared upon the scene. 

It is not easy to say what people inhabited the north- 
western corner of Wales, the district which was in later times 
the stronghold of Welsh independence. Ptolemy here fails us, 
for, though it has until recently been taken for granted that he 
assigned to the Ordovices the whole of North Wales, the places 
mentioned by him as belonging to this tribe lie to the east, 
near what is now the English border, and in reality this part 
of his map, save for the name of a cape and that of a river, is 
blank. 39 Of Segontium and Conovium, stations in the district 

38 II. iii. 12. 37 Nat. Hist. iv. 102. s8 Races of Britain, p. 185. 

39 This does not appear from the map of Ptolemy's Britain contributed by 
Dr. Henry Bradley to Archaologia (xlviii. 379). But in this the Toisobios River 
is made to precede the cape of the Gangani and to belong to the north coast of 
Wales, whereas the better reading adopted by Miiller reverses this order and 
so puts the mouth of this river somewhere near Portmadoc, an arrangement 
which has the further advantage of bringing Ptolemy's M<W xf}<ros into closer 
connection with his North Wales. 

40 tit STORY OF* 

CHAP, of which mention is made elsewhere, he had apparently not 
heard. The same impression, that Anglesey and Snowdonia 
were not in the occupation of the Ordovices, is conveyed by the 
passage in the Agricola of Tacitus 40 which tells of the 
subjugation of the latter ; they were all but annihilated in the 
first campaign of the new general, who then, the narrative 
continues, mindful of the necessity of pressing home his advan- 
tage and of not allowing the first impression of terror to wear 
off, set himself seriously to the task of reducing Anglesey. 
Only one clue is obtainable as to the condition of affairs at 
this time in the Snowdon region, and this is furnished by the 
Welsh name of the district, viz., Gwynedd. This is un- 
doubtedly ancient, appearing, as it does, in the form " Vene- 
dotis" (a genitive, presupposing some such nominative as 
" Venedas"), in an inscription at Penmachno, Carnarvonshire, 
of the sixth or seventh century. 41 The meaning is indicated 
by the cognate Irish word, " fine," a tribe or sept, and Gwynedd 
would thus denote a group or confederacy of tribes. If this 
interpretation be correct, it will not be matter for wonder that 
no general tribal name can be found which covers the whole 
district ; the task of the investigator will rather be to distinguish 
the individual tribes included in the group. Nor is it out of 
the question to achieve something in this direction. Ptolemy 
himself, when he gives the title of " Cape of the Gangani " 42 to 
the extreme point of the Lleyn peninsula, supplies the name 
of one tribe and suggests its Goidelic (or Iberian) origin, for 
there were Gangani in the West of Ireland. The chronicler who 
speaks of Degannwy as " arx Decantorum " 43 furnishes another, 
and again a parallel form, " Decantae," is forthcoming, this time 
from the Pictish North. The " Decangi " of Tacitus, dwelling 
not far from " the sea which looks towards Ireland," 44 may be 
regarded as a third Venedotian tribe, and, if the view of Sir J. 
Rhys should be accepted, that the true form is " Deceangli," 45 

40 Cap. 18. 

41 Arch. Comb. IV. ii. (1871), 257-8. See Celt. Br. (2), p. 311 ; Urk. Spr. 270. 
The inscription is No. 135 in Hiibner's Inscriptiones Britannia Christiana:. 

42 II. iii. 2. Pcry- avcav is the reading of most of the MSS. and is much to be 
preferred to Miiller's fanciful Katayyavuv. 

43 Harl. MS. 3859, printed in Y Cymmrodor, ix. (1888), p. 163. " Arx " is 
supplied from other sources. 

44 Annals, xii. 32. 45 W. People, p. 94. 


would naturally be assigned to Tegeingl or Northern Flintshire. CHAP. 
Llanfihangel Din Silwy and Llanfair yng Nghornwy may 
suggest that there were Selgovae and Cornavii in Anglesey, 
while Forth aethwy and Din daethwy appear to be the " creek " 
and the "stronghold" of another of the tribes of the island. 
Of the whole group, no less than of the Silures and the Demetae, 
it may be said that the primitive Iberian element was pro- 
minent in it, tempered, no doubt, by a Goidelic admixture; 48 
here was the peculiar home of Druidism, hither fled those to 
whom the civilisation of the Roman province was irksome ; 
under the shadow of the venerable oaks of Mona the very air 
was redolent of tradition and ancestral calm. 

The fourth region of Wales, corresponding roughly to the 
modern counties of Montgomery and Radnor with the adjacent 
portions of Merioneth and Denbigh, was occupied by the 
Ordovices, who, strong in the difficult nature of their country, 
offered as fierce a resistance to the invader as did the Silures, 
and were among the last tribes in Southern Britain to lay 
down their arms. But, unlike the Silures, they represented 
not the older but the newer elements in the population of the 
island. Celts they undoubtedly were, for the tribal name is 
derived from the Celtic root ord-, meaning "mallet" or 
hammer ; 47 they were the " hammer-men," wielding in battle 
the axe-hammers of stone which were in common use during 
the Bronze Age and probably much later. But more than 
this ; they are also believed to have been Brythons, leaders of 
that Brythonic forward movement which in the first century of 
our era was making rapid progress, the people, in short, from 
whom the rest of Wales learnt the Welsh language. 48 That 
the tribe which held the land to the east, the Cornavii of our 
Cheshire, Shropshire and Staffordshire, were of the Brythonic 
race is known from the name of one of the places within their 
bounds, viz., Pennocrucium (now Penkridge), which is pure 
Brythonic for " the head of the mound " or " the place of the 
chief of the mound ", 49 Testimony of the same direct kind 

48 The discoveries made in 1903 in the Venedotian stronghold of " Tre'r 
Ceiri," crowning the eastern spur of the Rivals in Carnarvonshire, show that the 
productions of Late Celtic, and even of Roman art, were making their way into 
this region during the first century of our era (Arch. Camb. VI. iv. (1904), 1-16). 

47 Celt. Br. (2), p. 303. "Ibid., pp. 87, 217-8. 

* Rhys, Hibbert Lectures, publ. 1888, pp. 202-4 ; W. People, p. 7. 


CHAP, is not forthcoming in the case of the Ordovices, but two facts 
make it highly probable that this was the centre from which 
Brythonic influences radiated into North and South Wales ; 
the district is entirely free from those evidences of a late 
Goidelic occupation which occur throughout the rest of Wales, 
and tradition knows nothing of any conquest of it in post- 
Roman times such as elsewhere is attributed to the prowess 
of the Brython. The men of Powys, to give the district its 
mediaeval name, were of ancient standing in the land and 
must have been already in possession when the Romans 
appeared on the banks of the Severn and the Dee. 

Of the degree of civilisation attained by the Ordovices 
before their subjection to the Roman yoke a fair notion may 
be formed from the remains discovered on the site of the 
Glastonbury lake village in 1892 and subsequent years. 50 The 
date at which the history of this settlement comes to an end 
is shown by the fact that no trace of Roman influence has yet 
been observed by the explorers who have examined the site, 
while on the other hand the culture to which the remains bear 
witness is unmistakably Late Celtic in character. No coins 
have yet come to light, so that the lake dwellers appear to 
have resembled the Ordovices in making little or no use of 
money ; in other respects, their civilisation has a thoroughly 
Brythonic aspect and may well represent the normal state of 
things all along the Brythonic frontier. 

The site of the village, lying a little north of Glastonbury, 
is now dry, but in former days the Mear Pool, of which a 
remnant survived in the sixteenth century, no doubt spread 
over the whole of the low-lying region between the Hartlake 
and the Brue. Sixty to seventy low mounds mark the position 
of as many round or oval houses, which once rose out of the 
water, resting on foundations of timber and brushwood which 
were secured in their places by rows of wooden piles driven 
into the peat of the lake bottom. The floors of the huts were 
of clay and in the centre of each was the hearth. In diameter 
they varied from 20 to 35 feet. The walls were of timber, 

50 For the account I give of the Glastonbury lake village I am indebted to 
Mr. Arthur Bulleid's paper in the Proceedings of the Somerset Natural History 
and Archczological Society for 1894, an< i to tne reports made to the British As- 
sociation from 1893 to 1899 by the special committee appointed to superintend the 


filled in with wattle and daub ; in one case a doorstep and CHAP, 
threshold were found intact The dwellers in these huts are 
shown to have been tillers of the soil by the discovery of an 
iron reaping-hook and of a sickle of the same material, each 
having its wooden handle attached ; the bones of their domestic 
animals ox, goat, sheep, pig, horse, dog, fowl bear witness 
to a fully developed system of rural economy. ^An iron horse- 
bit and parts of a wheel show that they understate? horseman- 
ship and the use of wheeled vehicles ; appliances for weaving 
have also been brought to light. Among the objects of 
domestic use mention should be made of the querns or hand 
mills for grinding corn, the staved buckets and the Late Celtic 
pottery. Bronze ornaments for the person, together with a 
fragment of a bronze mirror, illustrate the lighter side of this 
Brythonic culture, and are in keeping with that love of adorn- 
ment which the grave Strabo notes 51 as a Galatic weakness. 
Finally, the very boats and ladders by means of which this 
islanded community kept up communication with the outer 
world have been preserved, so that the picture lacks nothing 
to make it complete. 


In dealing with the subject of this section, one is confronted 
on the threshold, with the difficulty of reducing to sober pro- 
portions a theme upon which more misguided ingenuity has 
been expended than, perhaps, any other within the range of 
Welsh history. Edward Davies's treatise of six hundred pages 
upon The Mythology and Rites of the British Druids is 
but one of many works in which the creed, rites, institutions 
and monuments of the Druids are described with a particularity 
for which nothing in the ancient authorities gives any warrant, 
and which is, in truth, only attained by the unrestrained 
exercise of the imagination. History has little to tell of the 
Druids of Gaul, still less of those of Britain ; except for some 
three chapters in the Gallic War of Caesar, what is known 
of the order is derived from the scanty incidental notices of 
writers who do not seem to have had any first-hand knowledge 
of the facts. 

Undoubtedly, the Druidic system was one which deeply 

81 IV. iv. 5. 82 London, 1809. 


CHAP, impressed itself upon the imagination of the Roman world ; 
otherwise (so slight was the interest taken in barbarian re- 
ligions) no account whatever of its features would probably have 
been preserved. Its ruthless immolation of human victims 
startled a society which was learning to be humane ; the curious 
in matters of philosophy were attracted by the Druidic doctrine, 
recalling that of the Greek Pythagoras, as to the transmigration 
of souls ; while statesmen took note of the completeness of the 
Druidic organisation, its hold upon the mass of the people, the 
thoroughness of its educational discipline. As a whole, the 
system was to the Roman mind strange and uncanny ; a fact 
which lends support to the view 53 that it was not of Celtic 
origin, in spite of its prevalence in the Celtic lands, but had 
been taken over from some more primitive race (such as the 
Iberian) which practised rites and observed customs long since 
discarded by the Aryan peoples. According to Caesar, its 
real home was Britain, whither accordingly inquirers went who 
desired to drink at the undefiled fountain-head. In Britain, its 
stronghold indeed, the only district with which it is expressly 
associated is found to be the island of Mona, where the 
Iberian element is on other grounds believed to have been im- 
portant, and thus another reason is furnished for regarding 
Druidism as an importation into the Celtic world. The diffi- 
culty of explaining the word "druid" from Celtic sources 
points in the same direction. 54 

If this view be accepted, it follows, as indeed is suggested 
by Caesar, that the comparatively enlightened and philosophical 
form of Druidism described at length in the sixth book of the 
Gallic War is a modification of the true and unadulterated 
Druidism, as depicted for us, with brief but vivid touches, 
by Tacitus in his account of the attack of Suetonius upon 
Anglesey. 55 It may well be suspected that the highly de- 
veloped organisation which enabled the whole Gallic order to 
meet annually, in the country of the Carnutes (the district of 

83 Celt. Br. (2), p. 69; W. People, p. 83. 

64 According to Dr. Whitley Stokes (Urk. Spr. 157), the etymology of 
" druid " is still quite uncertain. No connection can be established with " derva," 
the Celtic original of "derw" (oak), or its congeners " dari," " darik," "daru," 
of the same meaning. 

6 Annals, xiv. 30. 


Chartres), under the presidency of an elective Arch-druid, for CHAP, 
the trial and determination of causes from all parts of Gaul, 
had been devised not long before Caesar's day in the interests 
of some political movement. The account given of Druidism 
in the Gallic War is believed to have been based to a great 
extent on information supplied to Caesar by his friend Divi- 
ciacus the yEduan, himself a prominent member of the order 
and yet an admirer of Roman civilisation. 66 With such a 
man it would be a point of honour to put the very best face 
upon the Druidic system, and, while he would recognise that it 
was useless to strive to conceal the darker aspects of its ritual, 
the merits of the Druids as teachers, judges, philosophers and 
peacemakers would be emphasised by him in the fullest 

The Druids of Britain were men of a grim and forbidding 
aspect, skilful in the use of the terrors of religion, and needing 
no association with inhuman rites to appear awful in the eyes 
of Roman soldiers. The gods to whom they appealed with 
uplifted hands were fierce powers whose wrath could oftentimes 
only be appeased by the shedding of human blood and whose 
will was declared by the movements of tortured human 
victims. No temples were raised for the worship of these 
deities, but they were served in forest sanctuaries, where the 
thick leafage and undergrowth shut out the gaze of the vulgar 
and cast a gloom over the savage altars. Such is the picture 
handed down by Tacitus, and little can be added to it save by 
way of conjecture. It may be surmised that in Britain, as in 
Gaul, the Druidic order included a guild of " bardi " or pro- 
fessional minstrels, for this is the readiest explanation of the 
fact that when " beirdd " first make their appearance in Welsh 
history, in the Laws of Hywel the Good, they are found to be 
united in a regular organisation. Another character assumed 
by the British Druid was certainly that of diviner and 

M Diviciacus (a form which is to be preferred to the Divifiacus of the 
ordinary editions of Caesar ; see Celt. Br. (2) p. 292) was an ^Bduan noble in con- 
stant attendance upon Caesar and high in his confidence during his wars in 
Gaul. He had shown himself a friend of Rome for many years and had visited 
the city to seek the help of the Senate against a neighbouring tribe (Bell. Gall. 
vi. 12). That he was a Druid and professed knowledge oi the future we know 
from a reference in Cicero's De Divinatione (i. 41). 


CHAP, magician; this is the ordinary sense of the word "derwydd" 
in early Welsh literature, as in the lines of Cynddelw 57 

Nis gwyr namyn Duw a dewinion byd 

A diwyd Derwydon 

O eurdorf eurdorchogion 

Ein rif yn Riweirth afon 

(" None knoweth, save God and the world's diviners and Druids assiduous, 
how many were of us, that golden torqued host, at the river of Rhiweirth.") 

But it will not do to assume, without trustworthy evidence and 
against historic probability, that the Druids worshipped one 
God, that they sacrificed on cromlechs, that their lore was cast 
into the triadic form, and that it is represented in spirit and scope 
by the documents put forward in their name by the bards of 
Glamorgan in comparatively recent times. 68 The idea that 
Druidism was a British variety of the religion of the patriarchs, 
presided over by dignified and benevolent sages, was begotten 
by the political philosophy of the eighteenth century, which in its 
superficial knowledge of savage life idealised the "state of nature" 
and endowed with all the virtues the rude denizen of steppe 
and forest. "What may be considered as the foundation of the 
Order was the doctrine of Universal Peace and. Good Will". 
When William Owen Pughe writes thus in his Sketch of 
British Bardism^ he is under the influence of the idea to 
which Locke gives expression in the words "the woods and 
forests, where the irrational untaught inhabitants keep right by 
following Nature, are fitter to give us rules than cities and 
palaces ". 60 All ancient testimony is against him ; Druidism, 
in short, represents, not the high-water mark of early British 
civilisation, but a survival from the less civilised past. 

My v. Arch. I. 212 (155). The modern form " Rhiwarth " obscures the 
derivation of the name from that of the Eirth, which here falls into the Tanat. 

68 These documents have been thoroughly examined and their true history 
elucidated by Prof. J. Morris Jones in Cymru, vol. x. (1896). 

5a Heroic Elegies of Llywarch Hen (London, 1792), xxv. 

80 Civil Government, bk. i. chap. vi. 




(The chief authority for the events narrated in section i. of this chapter is 
Tacitus; to the editions of the Annals and the Agricola by Mr. Furneaux I am 
accordingly much indebted. Among other works used mention may be made of 
Mommsen's Provinces, Rhys's Celtic Britain, Evans's Coins of the Ancient 
Britons, Henderson's Life and Principate of Nero (1903).) 

CAESAR attacked Britain, he tells us, because he wished to cut CHAP, 
off from the tribes of Northern Gaul the assistance which in 
their struggles against the power of Rome they constantly 
received from this island. 1 This reason may well have been the 
real one, for though he does not record many instances of 
British support afforded to the Gauls, 2 there was between the 
Brythonic settlers and the Belgae of the Continent so close a 
connection in point of race and civilisation as to make joint 
action in war and politics easy and, in fact, inevitable. The 
Straits of Dover formed at this time no race frontier, but 
flowed between two sections of the Belgic or Galatic world. 
No doubt, in the course of the preparations for this somewhat 
hazardous enterprise, other considerations were put forward, 
fitted to recommend it to the rank and file of the army ; the 
wealth of Britain, its mines of silver, the pearls of its coasts, 
were not allowed to be forgotten ; 3 but, so far as the general 

1 " Quod omnibus fere Gallicis bellis hostibus nostris inde subministrata 
auxilia intellegebat " (Bell. Gall. iv. 20). 

8 Certain chiefs of the Bellovaci, having failed in a movement against Rome, 
fled to Britain (ibid. ii. 14). The Veneti sought aid from Britain in 56 B.C. (ibid. 
iii. 9). 

3 Evidence of the disappointment which naturally followed is furnished by 
the letters of Cicero, whose brother Quintus served in the second expedition. 
Writing to Atticus, he says (Ad. Att. iv. 16) : "etiam illud iam cognitum est, 
neque argenti scripulum esse ullum in ilia insula neque ullam spem praedae nisi 
ex mancipiis " ; cf. " nulla praeda " in iv. 17. In a letter written to Quintus 



CHAP, was concerned, the main object of the two expeditions was 
probably that which he achieved, namely, to convince the 
British tribes that Rome was a dangerous power with which to 
meddle, and that their safety lay in abstention from interference 
with the affairs of Gaul. 

It is, of course, not unlikely that, if Caesar had found the 
British resistance less formidable than it proved to be, he might 
have brought at least a portion of the island under the sway of 
Rome, but there is nothing to show that the idea of a perma- 
nent conquest had in any way taken possession of him. When 
he withdrew from Britain in 54 B.C., he did indeed stipulate 
for the payment of a fixed annual tribute to the Roman people, 
but the unreality of this arrangement must have been patent to 
him, and it was made, not in order that a new province might 
be added to the Roman Empire, but with the view of furnishing 
an excuse, should necessity arise, for a fresh attack upon the 

Whatever schemes may have floated in the minds of Caesar 
and those who followed him as heads of the Roman state, it is 
certain that after 54 B.C. no serious attempt was made to 
conquer Britain for almost a hundred years. On two occasions 
Augustus planned an invasion of the island, 4 raising high 
the hopes of his admirers ; Virgil saw distant Thule enter 
his service 5 and Horace exclaimed that this new triumph 
would raise him forthwith to the rank of a god. 6 But more 
urgent business drew him aside from his purpose and he was 
content, when his power became firmly established, to extend 
a benevolent patronage to fugitive British princes, grounding 
his inaction on the plea that the Empire was growing too big 
to be manageable. 7 Tiberius very readily fell in with this idea 

(ii. 16) when news had just arrived of the landing, he says that the enterprise 
" plus habet tamen spei quam timoris, magisque sum sollicitus exspectatione 
quam metu," but after the receipt of another letter from Britain "de Britannicis 
rebus cognovi," he writes (iii. i), " ex tuis litteris nihil esse nee quod metuamus 
nee quod gaudeamus ". 

4 Dio Cassius, xlix. 38 ; liii. 22, 25. 

6 " Tibi serviat ultima Thule " (Georgics, i. 30). Cf. Horace, Epodes, vii. 7, 
and Georgics, iii. 25. All three references belong to the period 35-30 B.C. 

6 " Prsesens divus habebitur 
Augustus adiectis Britannis 
(Horace, Odes, III. v. 3. Cf. I. xxxv. 29. The year is 27 B.C.) 

7 Agricola, 13. 


and left Britain alone ; Gaius, familiarly known as Caligula, CHAP, 
revived the notion of a conquest of the island, only to make it 
ridiculous by his madcap tricks ; hence it was not until 43 A.D., 
in the early years of Claudius, that the Britons found arrayed 
against them, with formidable intent and real purpose, those 
legions which had tamed the rest of the Western world. 

The opportunity for the expedition was probably afforded 
by the death of Cunobelinus. This prince is shown by the 
evidence of his coinage to have borne sway in the South-east of 
Britain for a long period at the beginning of the first century 
of our era. His capital was at Camulodunum, the modern 
Colchester, as appears not only from the story of the Roman 
conquest, but also from the inscriptions upon his coins. 8 That 
he belonged, however, to the Catuvellauni of mid-England and 
not to the Trinovantes of Essex, is clear from the fact that his 
father Tasciovanus coined at Verlamium (St. Albans) 9 and 
that his sons are said by Dio Cassius to be Catuvellauni. 10 It 
is not too much to suppose, therefore, that he was of the family 
of Cassivellaunus, the doughty antagonist of Caesar, and that 
he carried on his dynastic traditions, having perhaps shifted 
the seat of government for convenience of access to the sea. 
Everything goes to show that he was a powerful, wealthy 
ruler, with abundant resources and means to enforce his will. 
He is last mentioned in A.D. 40, as having expelled from 
Britain his son Amminius, whom Gaius ostentatiously received 
but in no way helped to recover his own ; u when the Roman 
army in another three years attacks the island, it is confronted, 
not by the old king, but by two of his sons. It is obvious 
that the removal of so conspicuous a figure in the political life 
of Britain must have led to great confusion and change ; de- 
pressed parties would regain their courage, dynastic and tribal 

'Evans, Coins of the Ancient Britons, pp. 284-348, 555-73. 

9 Ibid., p. 225. 

10 This point was overlooked by Mommsen (Provinces, i. 171). The view 
taken in the text is that of Rhys (Celt. Br. (2), pp. 26-8). 

11 Suetonius, Gaius, 44. The " Adminio Cynobellini Britannorum regis filio " 
of this author was copied by Paulus Orosius (about 416) in the form " Minocy- 
nobelinum Britannorum regis filium " (Hist. vii. 5, ed. Zangemeister). This pro- 
duced, in the pages of Nennius, a " Bellinus " son of " Minocanni " (cap. 19), who 
figures conspicuously in later Welsh legend as Beli Fawr ap Manogan (Mab. 
26, 88, 93). To Prof. Zimmer is due the credit of having worked out the true 
pedigree of this mythological impostor (Nennius Vindicates, 271-3). 

VOL. I. 4 


CHAP, ambitions would revive ; this was, it can hardly be doubted, 
the moment chosen by the Romans as the most suitable for 
their enterprise. 

The expedition was planned on a large scale. Four 
legions, the Second (Augusta), the Ninth, the Fourteenth and 
the Twentieth, 12 were detached from the Rhine and the Danube 
frontiers (where two new legions took their place), and formed, 
with a very large body of auxiliary troops, an army of invasion 
of not less than 50,000 men. At its head was set a capable 
general, Aulus Plautius Silvanus, who had been for many years 
in the service of the state and who proved himself on this 
occasion fully worthy of the charge laid upon his shoulders. 
He had under him able lieutenants, notably Flavius Vespas- 
ianus, commander of the Second Legion, who was in later life 
to wear the imperial purple. Claudius himself crossed the 
Channel when the campaign was well advanced and took part 
in the storming of Camulodunum, but, though he thus earned 
the title of Britannicus and had other distinctions bestowed 
upon him by the senate, the sixteen days he spent in the 
island 13 can have had little effect upon the course of the cam- 
paign, and too much honour is done him when reference is 
made to the " Claudian " conquest. Important as the enter- 
prise was, there is but one extant account of it, that of Dio 
Cassius, the no doubt full narrative of Tacitus being in one of 
the lost books of the Annals, and from Dio's pages it is difficult 
to construct a story of the struggle which will suit the geography 
of South-eastern Britain. It remains beyond doubt, however, 
that the invaders were opposed by two of the sons of Cunobe- 
linus, Caratacus and Togodumnus, the latter of whom died in 
the course of the war, and the inference is a safe one, that the 
tribe specially singled out for attack was the Catuvellauni, 
whose power was shattered and whose rule over other tribes 
came to an end. The seat of their authority, Camulodunum, 
became the centre of the Roman province, where in a few 
years a temple was set up in honour of the deified Claudius 
and a colony of veterans was established. Among the tribes 

12 A full discussion of the difficult problems connected with the movements 
of these four legions during the era of conquest will be found in Engl. Hist. Rev. 
xviii. (1903), pp. 1-23. 

13 Dio Cassius, Ix. 23. 


who were led by jealousy of the Catuvellauni to espouse the CHAP. 
Roman cause and who were thus included in the new province 
without having been first humbled by defeat were the Eceni 
of our Norfolk and Suffolk, 14 the " Boduni " of uncertain 
situation, 15 and the subjects of a certain Cogidumnus, 18 prob- 
ably the Regni of Sussex. The subjugation of the south- 
west of the island appears to have been the work of Vespasian, 
who is expressly said to have conquered the Isle of Wight n 
and whose legion (the Second) is ultimately found at Isca in 
the country of the Silures. 

Plautius returned to Rome in A.D. 47 and was voted the 
exceptional honour of an ovation, being the last private 
citizen who received it under the Empire. With the accession 
to office of the new legate of the province, Fublius Ostorius 
Scapula, the war enters upon a new phase and the Roman 
forces are marshalled against the less civilised inhabitants of 
Britain, the tribes of the North and the West. They had so 
far had to deal only with the Brythonic or Galatic settlers of 
the South-east, who had offered a stout resistance, but, when 
once thoroughly defeated, were not unwilling to become sub- 
jects of the great Roman Empire. An agricultural people, 
they could not carry on the struggle indefinitely without risk 
of starvation ; a partly civilised people, they were not insen- 
sible to the attractions of Roman culture. With the wild 
pastoral tribes of the hills, half Goidelic and half Iberian, it 
was otherwise. As in later ages, the fact that they had no 
stake in the soil, no rich crops to rot in the ground, no well- 
built houses and barns which an enemy might burn, gave them 
infinite mobility ; nothing was gained by an inroad on their 
homes ; they must themselves be hunted out and brought to 
bay. Moreover, Rome could not tempt them ; in the barbaric 
simplicity of their life their one fierce desire was for indepen- 
dence. Hence, it is not at all surprising that this second stage 
of conquest proved much more tedious and harassing than the 
first, and that the slow progress made by the Romans in the 

14 " Valida gens nee proeliis contusi, quia societatem nostram volentes acces- 
serant " (Annals, xii. 31). 

15 Dio, Ix. 20. 

18 " Quaedam civitates Cogidumno regi donatae : is ad nostram usque 
memoriam fidissimus mansit " (Agricola, 14). 
17 Suetonius, Vespasianus, 4. 



CHAP, subjugation of the more backward tribes even encouraged the 
dwellers in the settled South-east to rise in revolt against 
their new rulers. 

When Ostorius assumed command, he found Britain south 
of the Humber and east of the Severn nominally subject to 
Roman rule, the Brigantes who dwelt north of the former 
river in alliance with Rome, but the tribes of what is now Wales, 
from the mouth of the Dee to that of the Wye, in a state of 
active hostility, which showed itself in raids upon the recently 
conquered districts. His first step was to make sure of the un- 
easy allegiance of the province itself, the lands bounded by the 
Severn and the Trent, 18 and this he did not accomplish with- 
out ruffling the susceptibilities of the friendly Eceni, who then 
first learnt that, though never subjugated, they were a subject 
people. Next, he attacked the Decangi, who had probably 
been among the principal raiders of the border, and penetrated 
so far into their country as almost to reach the shores of the 
Irish Sea ; he repaid with interest their ravages, but they, true to 
their highland methods, would not be enticed into the open 
field, so that their strength was not really laid low. At this 
stage, troubles among the Brigantes showed Ostorius the in- 
security of his position in this part of Britain ; having composed 
them, he retired southward, and, if the narrative of Tacitus 
gives a correct impression, henceforth devoted his energies to 
the conquest of the Silures. 

The strength of this tribe lay in their untamable, resolute 
spirit and the rugged character of the country they occupied. 
But they had at this moment the additional advantage of the 
leadership of Caratacus, who, when all had been lost in the 
East, had thrown in his lot with the rough hill-folk of the West, 
strangers to him in speech and in manners, but his comrades 
in love of liberty. He seems to have spent some three years 
among the Silures and their neighbours, the Ordovices, organ- 
ising the resistance to the Romans and rendering infinitely 
difficult the task which Ostorius had set himself. Of the details 
of the struggle nothing is known, though it may be that some 
faint tradition of it is preserved in the name " Caer Caradog" 

18 I follow the suggestion of Bradley : " cunctaque cis Trisantonam et Sab- 
rinam fluvios cohibere parat ". See Furneaux, Annals of Tacitus, vol. ii. p. 252, 
and Celt. Br. (2), p. 80. 


borne by more than one hill-fortress in Southern Shropshire. 19 CHAP. 
It ended in A.D. 5 1 in the signal defeat of Caratacus, at a spot 
which it is now impossible to identify, since the description of 
Tacitus carries the reader no further than this, that it was a 
hill-fortress in the country of the Ordovices, protected on one 
side by a river not easy to ford. 20 Though the capacity of the 
Silures for resistance was not seriously diminished, they now 
lost their leader, who fled for safety to the Brigantes and by 
them was handed over to his foes. 

In the opinion of the authorities, the capture of this famous 
chieftain, who had for so many years bidden defiance to the 
Roman arms, afforded the proper opportunity for celebrating 
in the capital with due ceremony the victories won in Britain. 
Caratacus was led in triumph, with his kinsmen and dependants, 
through the streets of Rome ; Claudius set up a triumphal 
arch to commemorate the submission of eleven kings and the 
first extension of Roman rule beyond the Ocean. But the 
course of events soon showed that, so far as the tribes of the 
West were concerned, the struggle was but just begun. Ostorius 
thought it requisite to establish in the Silurian country a per- 
manent camp, which was no doubt fixed at Isca, the home in 
later times of the Second Legion, but he found the utmost 
difficulty in carrying out his design. The enemy set upon the 
soldiers to whom the work had been entrusted, cut off foraging 
parties, put the auxiliary forces to flight, and, by the system 
of guerilla warfare afterwards so well known in Wales, even 
shook the stability of the legions. In the midst of this ob- 
stinate warfare Ostorius died, overpowered, as was generally 
supposed, by the weight of the difficulties with which he had to 
contend. Immediate advantage was taken by the Silures of 

19 There is a Caer Caradoc near Church Stretton (Penn. iii. 271-3), as well 
as one near Clun (Gibson, 541, 551), which is also known as Gaer Ditches. The 
" cair caratauc " of the Civitates of Nennius (p. 211 of Mommsen's edition) must, 
on the other hand, have been, like other " caerau " in the list, a Roman city or 
station : it is mentioned in the Book of Taliesin (" O gaer glut hyt gaer garadawc," 
IV. Anc. Bks. ii. p. 194), and by Geoffrey of Monmouth (Hist. Reg. vi. 15 ; viii. 9) 
and " Brut y Brenhinoedd " (Bruts, 140, 165) is arbitrarily fixed at Salisbury. 

* Notwithstanding the paucity of data, the question of the site of this battle 
has been hotly debated among antiquaries. The favourite suggestions have been 
the Gaer Ditches (Humphrey Llwyd, Camden), Coxwall Knoll in the same dis- 
trict (Hoare, Roy), Cefn Carnedd, near Llanidloes (Hartshorne), and the Breiddin 
near Welshpool (Arch. C. II. ii. (1851), 122-43 '> IV. x. (1879), 272-83). 


CHAP, the withdrawal of the strong hand of the governor ; before his 
successor could arrive, the Second Legion had suffered a grave 
reverse, and the first business of Aulus Didius Gallus was to 
determine the lines of his frontier policy. He resolved to 
abandon the attempt to subjugate the Silures ; during the five 
years of his governorship (A.D. 52-57) the limits of the province 
remained practically unaltered. 21 Old age and satisfied am- 
bition are said to have had not a little to do with the adop- 
tion of this conservative attitude, but the total failure of the 
persistent activity of Ostorius shows that it was not an un- 
reasonable one. 

With the arrival of Quintus Veranius, the fourth of the 
governors of Britain (A.D. 58), the policy of extension is re- 
sumed and the Silures are again attacked. But death cut short 
the schemes of this governor before the end of his first year of 
office, and the idle boast of his will, that in two years more he 
would have reduced the whole country to subjection, suggests 
that he had taken no true measure of the task before him. 
He was followed by a soldier of the first rank, Caius Suetonius 
Paulinus, who had won a great reputation by his achievements 
in Mauretania and whose appointment to Britain signified that 
the war was to be vigorously carried on. For a while he was 
remarkably successful ; leaving the Silures, it would appear, to 
themselves for the time being, he directed his attention specially 
to the north-western side of the province, which was exposed 
to the attacks of the Ordovices, of the Decangi and, as the 
result of a change in tribal policy, of the Brigantes. It was 
during this period, if not earlier, that strong camps were estab- 
lished at Lindum (Lincoln), Deva (Chester) and Viroconium 
(Wroxeter), which became the head-quarters of the Ninth, the 
Twentieth and the Fourteenth Legions respectively ; the second 
and the third of these stations were connected with each other 
and with the Kentish coast by the great military road known 
to later ages as Watling Street Having thus firmly established 
his position at the mouth of the Dee, Suetonius resolved to 

21 " Didius Gallus par ta a prioribus continuit, paucis admodum castellis (i.e., 
forts, not legionary camps) in ulteriora promotis" (Agr. c. 14). Cf. Ann. xii. 
40 : " arcere hostem satis habebat". With this testimony before one, it is diffi- 
cult to believe that " Didius seized the occasion to advance the head-quarters of 
the Fourteenth and Twentieth Legions from Viroconium to Deva " (Henderson's 
Nero, p. 205). 


undertake the conquest of Mona, where the fertility of the soil CHAP. 
and the protection given by the great mountain barrier of 
Snowdonia enabled a large population to defy the arms of the 
invader and where the rites of Druidism were still practised 
undisturbed. He rightly judged that a blow struck here would 
have a moral effect quite out of proportion to the actual injury 
inflicted on the foe. In A.D. 60 or 6 1 22 the island was accord- 
ingly attacked ; flat-bottomed ships were specially constructed 
in order to carry the legionaries across the shallows of the 
Menai Straits, while the auxiliary cavalry, accustomed probably 
to crossing tidal waters in their own homes, contrived to make 
the passage on the backs of their horses. Tacitus, as his 
manner is, gives hardly any clue as to the spot at which the 
crossing took place ; the sands between Aber and Beaumaris, 
where a good deal of the traffic for Ireland formerly entered 
Anglesey, may be as confidently suggested as the narrower 
reaches near Port Dinorwic favoured by many antiquaries. 28 
For the moment, the expedition was entirely successful ; 
startled and awe-struck as the Roman soldiers were for a few 
brief moments at the grim and forbidding spectacle which 
awaited them on their landing, their practised valour soon 
won the day ; the Druids and their attendant host were over- 
whelmed, the sacred groves were cut down, and measures were 
at once taken for holding the island by means of a permanent 
camp and garrison. 

But once again the forward policy received a severe check. 
At the moment of his triumph Suetonius received the news of 
the great rebellion of the south-eastern tribes, led by the Eceni 
and their heroic mistress, Boudica. 24 The struggle which 

22 A.D. 6 1 is the date implied in the narrative of Tacitus, but good grounds 
have been advanced for believing that the rebellion really belongs to the previous 
year. (See Henderson's Nero, pp. 477-8.) 

23 The Rev. John Davies, rector of Newborough (d. 1695), first suggested 
the neighbourhood of Llanidan (Gibson, 675) and was followed by Rowlands in 
Mona Antiqua, 98 (99), who cites " Pant yr Yscraphie " as a place name in 

24 " Boadicea " is not found in any MS. and must, therefore, in any case be 
rejected. Of the various forms actually recorded, Boudica is most favoured by 
Celtic scholars, as a likely derivative of the root " boudi," victory, seen in Welsh 
" budd " and " buddugol " (Urk. Spr. 176 ; Celt. Br. (2), p. 282). Thus it was a 
fortunate hit of Theophilus Evans when in his " Drych y Prif Oesoedd " (1716) 
he styled the queen " Buddug ". His other renderings of the old names, e.g., 
" Ploccyn " for Plautius, are less felicitous. 


CHAP, followed was a decisive one in the history of these islands , 
[II> had the insurgents defeated Suetonius, Roman rule must have 
come to an end in Britain. But even the signal victory which 
the general won, over a force vastly outnumbering his own 
and flushed by its early successes, left the situation greatly 
modified. He fought with the Fourteenth Legion, together 
with a detachment of the Twentieth and some auxiliaries ; 
hence, the main body of the Twentieth was no doubt left to 
hold Deva, while the Second (from which Suetonius, through 
the inaction of the prefect of the camp, then in temporary 
command, derived no assistance) remained stationed at Isca. 
But, though the borders of the Western tribes were thus firmly 
held, there could be now no question of any aggressive move- 
ment against them, when the whole province required from 
the government the most anxious supervision and the most 
skilful treatment, lest the wounds left by the great encounter 
should become open sores. Suetonius, fit instrument as he 
was of the policy of aggression and conquest, was now out of 
place in a post which called for the exercise of powers of 
diplomacy and pacification. He was recalled, and Publius 
Petronius Turpilianus was sent in his stead to govern Britain, 
with instructions to show clemency to the subject peoples and 
to embark on no new enterprise against the still unconquered 
tribes. When after a year or two he returned to Rome, his 
place was filled by the appointment of another man of peace, 
Lucius TrebelHus Maximus, who was so little of a soldier as 
not to be able even to earn the respect of his troops ; they 
scorned the miser whose chief purpose seemed to be to amass 
wealth at the expense of the province, and under him and his 
no less weak, though more amiable, successor, Marcus Vettius 
Bolanus, the legions of Britain were idle and disorderly ; the 
purpose of completing the conquest of the island might seem 
to have been finally abandoned. 

It was reserved for a new imperial dynasty, the Flavian, to 
extend the province to its natural limits by the subjugation of 
all the tribes of the island save those of the extreme North. 
Vespasian was first and foremost a soldier ; though his latest 
honours had been won in the East, he had fought with dis- 
tinction in Britain under Aulus Plautius, and thus it was natural 
that his accession to the supreme power should mark a new 


epoch in the attitude of the government towards the British CHAP. 
question. Energetic measures are once more in favour, as is 
seen in the appointment as governor in A.D. 71 of the emperor's 
friend, Quintus Petilius Cerialis, who had crushed a formidable 
rebellion in Gaul and held a high position as a soldier. Petilius 
justified his appointment by breaking the power of the Brigantes, 
the leading tribe of the island, whose independence had been 
a menace to the peace of the province for twenty years. On 
his return (A.D. 74), another man of mark was sent here by 
Vespasian, namely, Sextus Julius Frontinus, the author of 
extant works on the art of war and the construction of aque- 
ducts, a man who happily combined literary with practical 
ability. He was, too, of a noble simplicity of character ; while 
careful in the evil days of the tyrant Domitian to give no 
reasonable ground of offence, he refrained from flattery, and 
the words deserve to be recorded in which he declined the 
honour of a monument to his memory : " The expense is un- 
necessary ; our memory will endure if we have deserved it by 
our life ", 25 Such was the general who now grappled success- 
fully with the task which had proved too thorny for his pre- 
decessors, the conquest of the Silures ; two or three lines in 
the Agricola, as ill luck will have it, are the sole record handed 
down of what must have been a very important series of cam- 
paigns, ending about A.D. 78 in the complete occupation of 
our South Wales. Of the work of Frontinus all that can be 
said is that it was never necessary to repeat it. 

The third of the distinguished soldiers appointed by Ves- 
pasian to the governorship of Britain was the most famous of 
all. Julius Agricola had already seen a good deal of service 
in the island. At the beginning of his official career he had 
been here as a military tribune, when he enjoyed the special 
confidence of the governor, Suetonius Paulinus ; he had gone 
through the great rebellion, and it may be surmised that the 
moderation and justice of his rule as a governor were due in no 
small degree to his experience in that year of the fatal effects 
of rapine and oppression. After filling the usual public offices 
in Rome, he had returned to take the command of the 
Twentieth Legion, stationed at Dva, and had proved a loyal 
subordinate to two commanders of such diverse types as Bolanus 

98 Pliny, Letters, ix. 19. 


CHAP, and Cerialis. In A.D. 78, after some experience as governor 
of Aquitaine and a subsequent nomination to the consulship, 
he was chosen to complete the conquest of Southern Britain. 
His admiring son-in-law dwells with filial pride upon the vigour 
with which he entered upon the duties of his office. The 
Ordovices were on the warpath and had just annihilated almost 
the whole of a detachment of auxiliary cavalry which had been 
stationed within their limits ; the summer was drawing to its 
close, and it was, in the opinion of all, too late for measures of 
reprisal. Nevertheless, Agricola resolved, immediately upon 
his arrival, to attack the offending tribe and moreover to inflict 
upon them so severe a chastisement as finally to deprive them 
of their power to injure the province. With a force composed 
of picked men from the legions under his charge, he breasted 
the hills which cluster round the Upper Severn and the Dee 
and utterly defeated in their mountain fastnesses this race who 
were not to be tempted into the open country. Determined 
to press to the uttermost the advantage he had won, he pushed 
on for Mona, 26 which he had seen slip from the hands of 
Suetonius in the moment of victory. This time there were no 
boats in readiness for such an enterprise, undreamed of alike 
by friend and foe, but the auxiliary cavalry made short work 
of the passage, and the suddenness of the attack placed 
the island at Agricola's mercy. Thus at the beginning of 
the year 79 the whole of our Wales had been reduced and the 
general was free to devote himself to those more ambitious 
schemes of conquest in the North which engrossed the remain- 
ing six years of his governorship. He probably spent the 
winter of 78-79 at Deva, and it may be that we have a concrete 
memorial of his activity there in the lead piping now in the 
Grosvenor Museum at Chester which bears his name as " Legatus 
Augusti pro praetore," z.., governor of the province, and a 
date equivalent to A.D. 79. 27 

26 Pennant is responsible (ii. 27) for the view that the general passed on his 
westward march through a gap in the Clwydian range which he calls " Bwlch 
Agricla," but which locally seems to be known as " Bwlch criglas " and " Bwlch 
saeth cricaeth " (Arch. Camb. II. i. 89). The name Agricola could not have 
come down unaltered in such a connection ; it yields Aircol in old Welsh (Celt. 
Br. (2), p. 256) and in modern Welsh would have been still further modified. 

27 The piping was found in 1899 on the north side of Eastgate Street, 
Chester. The inscription, in its perfect form, runs " IMP VESP vim T IMP vn cos 


After an obstinate struggle of rather more than thirty CHAP, 
years, the Western tribes thus came under the direct rule of 
Rome. No more is to be learnt from written records as to 
their fortunes until they emerge to view once again at the be- 
ginning of the fifth century. But, if their history is in this 
sense a blank during the three centuries of Roman occupation, 
much light has nevertheless been thrown upon this period by 
the labours of archaeologists who have studied its remains, and 
with the aid of their researches it is possible to form some idea 
of the leading features of the Roman occupation of Wales. 


(In all matters relating to the Roman occupation of Britain, Prof. Haver- 
field's authority is paramount, and I have accordingly based much of what I have 
to say upon his writings, only regretting for my own sake and that of other 
students that he has never been prevailed upon to embody his unrivalled know- 
ledge of the subject in the form of a textbook. The sections contributed by him 
to the illustrated edition of Social England (vol. i. 1901) have been of great 
service, as also the chapter on Roman Britain in Authority and Archeology (ed. 
Hogarth, 1899). I have, of course, made use of the seventh volume of the 
Berlin Corpus Inscriptionum (ed. Hiibner, 1873), of Westwood's Lapidarium 
Wallia (Oxford, 1876-9), and of the papers in Archezologia Cambrensis and the 
Antiquary dealing with Roman sites in Wales and Monmouthshire.) 

While the Roman Empire may be regarded from some 
points of view as a military despotism, its headship the prize 
of the successful soldier, its basis military force, it was by no 
means the case that throughout its whole extent it presented a 
military aspect. Broad regions lying at its heart Gaul, Italy, 
Greece, Asia were without any garrison, the loyalty of the 
inhabitants being sufficiently assured, and it was only in the 
frontier provinces, such as Syria, Pannonia and the two 
Germanics, which were newly conquered and exposed to bar- 
barian attack, that Rome showed herself as the true mother of 
armies. The legions which maintained her power were dis- 
posed in a circle around the peaceful, well-governed provinces 
of the centre ; they were the watch-dogs which kept the 

Of the provinces in which, as the result of their position 
on the border, the military element predominated, Britain was 

CN IVLIO AORICOLA LEO Avo PR PR ". See Haverfield's Catalogue of Roman 
Stones in the Grosvenor Museum (1900), pp. 86, 87, 127. 


CHAP, one of the most important. Not only the fierce and stubborn 
valour of the conquered peoples, but also the presence in 
Caledonia and Hibernia of barbarians yet unconquered, made 
it necessary to maintain a large force in the island ; during the 
period of conquest, four legions, after the governorship of 
Agricola, three, were stationed here, together with a large 
number of auxiliary troops, so that the army of occupation 
rarely fell below a tenth of the whole Roman force. But 
the distinction between garrisoned and ungarrisoned districts 
which is so marked in the Empire as a whole reappears within 
the bounds of Britain itself. The troops were concentrated in 
those parts of the province which lay open to barbarian in- 
vasion ; one legion at York, with a great force of auxiliaries 
on or about the Wall, opposed a firm front to the Caledonians, 
one at Deva and one at Isca, with other auxiliary troops, 
guarded the Western coast against Irish irruptions. Meanwhile, 
to quote Mr. Haverfield, the highest authority on the subject, 
" the Midlands and the South-east of Britain were almost as 
empty of soldiery as Italy itself. They contained a peaceful 
population which was not unacquainted with Roman speech 
and culture." 28 Though Britain was a poor province, which 
could show nothing to compare with the rich and busy life of 
Antioch, Lyons or Alexandria, it was not, in the settled South- 
east, without civil life of a kind, and it is only of the North 
and West that one can say without reserve that the occupation 
was that of a garrison in an unfriendly country, holding itself 
in constant readiness for defence. 

One region in the West had, indeed, an "exceptional history. 
So far as the available evidence goes, it would appear that the 
Romans, during the second and third centuries, left the 
peninsula of Devon and Cornwall to itself. There is no trace 
of any military occupation of the district, no proof, on the 
other hand, that its dwellers quietly accepted Roman civilisa- 
tion. While it is probable enough that Vespasian fought in 
this region, his successors apparently left the Dumnonii to their 
own devices, finding, it may be supposed, that they were not 
likely to disturb the peace of the province. The tin mines 
had for some reason ceased to yield their ancient tribute of 

28 Social England, illustrated ed., i. p. 83. 


ore, and thus what might otherwise have been a strong motive CHAP, 
for conquest was taken away. 28 

North of the Severn Sea it was very different. With the 
exception of the neighbourhood of the Wall of Hadrian, there 
is no part of Britain which affords such ample evidence of the 
presence of the Romans as an army in possession as Wales and 
the Welsh border. The spade brings constantly to light traces 
of the activity of the conquerors in this district, and they are 
almost invariably traces of military occupation. A survey of 
the country, dealing with the remains of the Roman period, 
will make it clear how thorough that occupation was, and will 
suggest the inquiry what the special dangers were against 
which the province needed to be in this quarter so carefully 

At the point where the river Dee, emerging from the sand- 
stone bluffs through which it has urged its rapid course, swings 
round into what is now the great Roodee meadow, but was 
once a broad stretch of sand and the head of the river's ex- 
tensive estuary, the lines were early laid down of the military 
station of Deva. 30 Paulinus, it has already been suggested, 31 
had fortified the spot before he advanced on Mona, and one of 
the tombstones preserved in the Grosvenor Museum, that of a 
centurion who ended a long military career in the Twentieth 
Legion, is believed to belong to this period. 32 It is not likely 
that the Cornavii, in whose territory the new station stood, had 
previously any important settlement on the spot, for it seems 
to have had no British name, and the Roman name, as in so 
many other cases, was adopted from that of the river on which 
the camp stood, the Deva or the " divinity," if we may accept 
the explanation of Sir John Rhys. 33 From the first it was the 
permanent seat of the Twentieth Legion, Victorious, Valerian, 
of which the name appears on countless titles which have been 

99 For the history of Devon and Cornwall during the Roman occupation see 
the Archaological Journal, xlix. (1892), pp. 176-81, and the Edinburgh Re- 
view, April, 1899, p. 389. 

30 The best handbook to Roman Chester is Haverfield's Catalogue of the 
Roman Inscribed and Sculptured Stones in the Grosvenor Museum (Chester, 

81 See p. 54. w Catalogue, No. 54. 

33 Celt. Br. (2), pp. 291-2. Whitley Stokes appears to question this etymo- 
logy (Urk. Spr. 145). 


CHAP, discovered on the spot. Another legion, the second " Adiutrix," 
l ' has indeed left many signs of its presence in Deva, but its stay 
in the island is known to have been short, hardly more than 
fifteen years, and during this period the camp may have had a 
double garrison. 34 It was the Twentieth which, remaining in 
occupation until the end of the Roman period, gave the place 
its British name of " Caer Lleon," the " fort of the legion " : 35 
twenty-five tombstones of its officers and men have at various 
times come to light, and fragments of roof ornament show that 
many a building in the settlement displayed the boar which was 
the badge of this celebrated corps. 36 Its members were drawn, 
citizens of the Roman state though they all were, from the 
most diverse parts of the Empire ; Syria, Spain, Gaul, Germany, 
and Noricum were represented in its ranks, though the Western 
element, as was but natural, largely predominated. Twenty 
years' service on the banks of the Dee would tend to make 
them all men of Deva, the place where their life-work was 
done and their comrades and kinsfolk were buried. 

It is not possible to say with precision what part of the 
modern city of Chester was occupied by the original fortress. 
Its centre probably lay not far from St. Peter's Church, where 
the four ancient streets of the city met at right angles, for it is 
in this neighbourhood that the foundations have been dis- 
covered of the more important buildings of the settlement. 
The four streets may themselves be the direct successors of the 
principal ways of the camp, as originally laid down in con- 
formity with Roman practice. What is now, however, certain, 
as the result of discoveries made in 1887 and subsequent years, 
is that the existing walls, from Morgan's Mount on the north 
side to Newgate on the east, represent the defences of Deva in 
that direction, as erected (either by way of extension or on old 
foundations) about the end of the second century of our era, the 
lower courses being in point of fact the original Roman work. 
The special value of these discoveries was that they brought to 

34 Catalogue, Introd. pp. g-io. 

85 The earliest occurrence is in Bede (H.E. ii. 2 : " Carlegion "), who, how- 
ever, wrongly explains it as " ciuitatem legiom^w ". See also the list of cities 
in Hist. Britt. 211 (c. 66a " Cair Legion"). Leon in Spain was similarly so 
styled as the head-quarters of the legion VII. Gemina (Eng. Hist. Rev. ii. (1887), 
p. 645). 

a6 Catalogue, No. 200. 


light a great number of Roman tombstones which, on the CHAP, 
occasion of this rebuilding of the wall, had been taken from 
an adjacent cemetery and used as rough material to form a 
backing for the well-worked masonry of the surface. Much 
new material thus became available for the study of Roman 
Chester, while at the same time, as the work of excavation 
proceeded, no room was left for doubt as to the genuinely 
Roman character of the walling. The northern and eastern 
limits of Deva during the third and the fourth century are 
therefore well ascertained ; it is still a matter of conjecture how 
its western and southern walls ran, though most antiquaries 
agree that the Castle heights were not included. 

As a military station of the first class, Deva had an im- 
portant position in the system of roads which, in this country 
as elsewhere in the Empire, was a conspicuous feature of 
Roman civilisation. 37 The earliest road laid down was no 
doubt that by which the fortress was approached from the 
south, known in later times as Watling Street and connecting 
the station with Viroconium, Londinium, and the ports of 
embarkation for Gaul. This road crossed the Dee, probably 
by a ford, not far from the old bridge, which is mediaeval, 
and thence ran southward to Aldford, where there was another 
ford, across which the traveller made his way to the unidentified 
station of Bovium and on to Wroxeter. The course of the 
road from Chester to Aldford is well ascertained, and portions 
of the original paving have been discovered, made of the red 
sandstone ot the district. Another road, leaving the fort by 
the East Gate, ran by way of Condate, which is perhaps 
Kinderton, near Middlewich, to Manchester, and thus supplied 
the means of communicating with another great military centre, 
Eboracum, the seat of the Sixth Legion. The North Gate 
was the starting-point of a road which led to the fort at 
Wilderspool on the Mersey, and thence to the stations which 
guarded the western shores of the country of the Brigantes. 
From the road-book known as the Itinerary of Antonine it 
appears that a western road also left Deva, passing through 
the station of Varae, not yet identified, to Caerhun and 
Carnarvon. This in all probability branched off from the 

87 For an account of the Roman roads running out of Chester see Watkin, 
Roman Cheshire (1886), chap. iii. 


CHAP. Wroxeter road some little distance to the south of the fortress, 

for direct access to the west was blocked by the wide spreading 

sands and marshes of the estuary of the Dee. 

Strongly posted, with a great river defending its southern 
and western sides, protected by massive stone walls, garrisoned 
by a strictly Roman force of over 5,000 heavy armed infantry, 
Deva was above all else a place of arms, where everything 
was subordinated to military considerations. It was a standing 
camp, within the walls of which none but soldiers might dwell, 
and though it had, no doubt, like other fortresses of the kind, 
its outlying civilian settlements in the suburbs, there is nothing 
to show that these were of special importance. Such trade as 
was carried on was probably conducted in the interests of the 
garrison ; the place was not, so far as we know, a colony, with 
citizen rights, like the legionary fortress at Eboracum it was 
a great stronghold and nothing more. 

On the western side of the Dee estuary lay the country of 
the Decangi. Little has been found in this district in the way 
of direct testimony to the Roman military occupation, and 
even the course of the road to Caerhun has not yet been made 
out, for the identification of Varae with Botffaru, the " bod " 
or dwelling of Tyfarru, has no philological and scarcely any 
other basis. 38 But evidence is not wanting to show that, as 
might be expected in a region so close to Deva, the Romans 
were in full possession of the country and that their enterprise 
turned its natural resources to practical use. Great quantities 
of slag, of half-melted lead and of lead ore have been found 
along the coast in the neighbourhood of Flint, and with them 
Roman coins and trinkets, showing that the smelting of lead 
drawn from the neighbouring hills was in Roman times a 
regular industry in this district. 39 It was here that the pigs of 
lead bearing the legend " Deceangi " (often in a curtailed form) 

38 Camden (Britannia, 602) was the first to argue that " Bod Vari " must 
be the "Var(ae)" of the Itinerary (p. 231), which is said to be 32 miles from 
Deva and 18 from Conovium. The old forms are, however, " Boteuuarul " 
in Domesday (Cheshire, i. 26gai), " (B)ottyfarru " (Myv. Arch. II. 25 (416)), 
" Buttanari " (Monasticon, ii. 387), " Bottewarrn " (Mont. Coll. xxi. 332), and 
" Bottervarrn" (Tax. Nick. 287), all implying some such form as Bod Tyfarru. 
Traces of Roman occupation are said to have been found at Pont Ruffydd, some- 
what to the west of the church (Arch. Cantb. III. v. (1859), 128). 

39 Penn. i. 71-2, 93-9. 


which have been found in various parts of Britain were made CHAP. 

up and stamped so as to show the date and place of origin. 

None have so far come to light in Flintshire itself, but two, 
now preserved in the Grosvenor Museum, have been found in 
the vicinity of Chester. 40 All the known specimens are of 
early date and show that lead mining was being briskly carried 
on in the country of the Decangi within a few years after the 
supposed date of the foundation of Deva. 41 

The wild North-west of Wales, known to later ages as 
Gwynedd and inhabited in the Roman period by a mixed 
Iberian and Goidelic population of primitive habits, was held 
for the Empire by a group efforts, similar in plan and purpose 
to those which clustered round the great Wall in the North of 
Britain. Forts of this type reproduced, on a much smaller 
scale, the features of the legionary fortresses ; like them they 
were of rectangular form, with rounded corners, and were en- 
closed by a substantial stone wall. They were permanently 
garrisoned, either by a detachment of the nearest legion, or, 
more often, by a body of those auxiliary troops raised by the 
Romans from among the subjects of their Empire who were not 
citizens. At least five such forts are known to have been 
built in the old Principality of North Wales represented by the 
modern counties of Anglesey, Carnarvon and Merioneth, viz., 
at Caerhun, Carnarvon, Tomen y Mur, Caer Gai and Pennal, 
and the roads which secured their communications are still in 
a large measure traceable. 42 

The first named of these forts protected the passage of the 
Conway and accordingly took its name of Conovium from that 
of the river. Though Caerhun must be compounded from the 
personal name Rhun, and cannot be, as Camden 43 and even 
Edward Llwyd 44 supposed, a corruption of " Caer hen," the 
ancient city, its identity with the " Conovio " of the Itinerary 
has never been questioned, and the discovery of the name on 
a milestone in the district has lately set the matter beyond 
possible doubt. The fort itself occupies rising ground a little 

40 Catalogue, Nos. 196, 197. 41 C./.L. vii. Nos. 1204-6. 

42 Few of the Roman forts of Wales have as yet been scientifically excavated, 
and accordingly it may seem rash to assume that they were all occupied in the 
same age and formed part of one system. The evidence at present available 
seems, however, on the whole to favour this conclusion. 

43 Britannia, 597. 44 Gibson, 670. 
VOL. I. 5 


CHAP, distance from the western bank of the river, which is here tidal 
L and easily navigable by small craft ; it forms a square of about 
one hundred yards, in the centre of which now stands the 
parish church of St. Mary. The principal discoveries appear 
to have been made in " Erw Gaer," or the Castle acre, which 
lies between the fort and the river. Here the foundations of a 
building were laid bare ; it contained the usual arrangements 
for heating by means of a hypocaust or cellar furnace, and tiles 
bearing the stamp of the Twentieth Legion afforded evidence 
that it was meant for the use of that legion. Pottery of the 
Roman period was found in abundance, together with a great 
cake of copper, weighing over forty pounds, upon which was 
a Latin inscription, showing that it was private property and 
probably on its way to Rome. 45 

Caerhun stands in an amphitheatre of mountains, among 
the loftiest in Eryri, which seem to forbid all progress to the 
west. Through these the road from Deva to Segontium 
threaded its way by the pass of Bwlch y Ddeufaen, rising 
as it did so to a height of nearly 1,400 feet above the sea- 
level. Passing over desolate moors which have never known 
the plough, it is in this portion of its course easily re- 
cognised ; the raised " sarn " or causeway marks it out at once 
from the ordinary British or mediaeval trackway. It descended 
to the coast between Aber and Llanfair Fechan, and here was 
found in 1883, on the farm of Rhiwiau Uchaf, the milestone 
already referred to which marked the eighth (Roman) mile 
" a Kanovio ", 46 The stone was erected in the time of the 
Emperor Hadrian (between A.D. 119 and A.D. 138), but there 
is no need to suppose that this was the age in which the road 
itself was laid down ; such monuments were of necessity re- 
newed from time to time, and a second milestone was, in fact, 
discovered in the immediate neighbourhood of the first which 
had been set up about A.D. 200, under Septimius Severus. 47 
From Aber to Carnarvon the line taken by the road has not been 
ascertained, but its general direction may be inferred from the 
discovery of a milestone of Caracalla (A.D. 211-217), near Ty 

45 Penn. i. 87-8, iii. 136-7; Archaologia, xvi. 127-34; C.I.L. vii. No. 1200; 
Watkin, Roman Cheshire, p. 122. 

46 Arch. Camb. IV. xiv. (1883), 170-1; Mont. Coll. xvii. 282-91. 

47 Ibid, xvii. 291-6. Both stones are now in the British Museum. 


Coch, Bangor, 48 and one of Decius (A.D. 249-251), near Llan- CHAP. 
ddeiniolen. 49 IIL 

The fort of Segontium bore a Celtic name and was perhaps 
so called from the river Saint, 50 which here makes a great 
curve to the south ere it falls into the Menai Straits. It did 
not occupy the same position as the modern town of Carnar- 
von, but was placed on the higher ground within this curve, 
where the parish church of Llanbeblig now stands. It was of 
considerable size, covering about 6 acres, and had, no doubt, 
some importance as the terminus of the North Welsh road and 
one of the ports giving access to Ireland. The coins found on 
the spot range from the time of Vespasian to that of the Con- 
stantines, so that it may be assumed it was held from the end of 
the first century, when Agricola, perhaps, built the fort, until 
at least the middle of the fourth. One inscription has come to 
light, recording that about A.D. 200 the first cohort of Sunici 
repaired the conduit which supplied the place with water. 
As this auxiliary force, drawn in the first instance from a 
people who dwelt around the lower reaches of the Meuse, is 
known to have been in Britain in the age of Hadrian, at the 
beginning of the first century, the suggestion may be hazarded 
that it formed the permanent garrison of Segontium. 51 

A passing reference should be made to the Roman occu- 
pation of Anglesey. No inscription or other clear witness to 
the presence of Roman soldiery has been found in the island, 
but it does not admit of doubt that Agricola's conquest of 
Mona was permanent, and at least three sites are with some 
probability pointed out as those of Roman military works 
The very name of Caergybi (Holyhead), recalling as it does 
Caerhun, Caergai, Caersws, Caerfyrddin and many others, 
carries with it this suggestion, and serious consideration has 
been given to the view of Pennant that the churchyard walls 

48 Cambrian Quarterly Magazine, iv. (1832), p. 515; Arch. Camb. I. ii. (1847), 
502; Lap. W. 183 ; C.7.L. vii. No. 1164. 

* 9 Camb. Quart. Mag. iv. p. 515 ; Lap. W. 173-4; C.I.L. vii. No. 1163. 

80 Saint represents the old Welsh Segeint, seen in the " Cair Segeint " (which 
may or may not be Carnarvon) of Hist. Britt. 166 (c. 25) and 211 (c. 66a), and 
this again goes back to Segontion (Celt. Heath, pp. 272-3). 

81 Penn. ii. 412-4; Lap. W. 172-3; C.I.L. vii. No. 142; Arch. Camb. I. i. 
(1846), 75-9, 177-82, 284-9 ; iii. (1848), 362-3 ; iv. (1849), 150 ; II. iv. (1853), 
71-2 ; VI. v. (1905), 73-6. For the Sunici see Tac. Hist. iv. 66, and Pliny, Nat. 
Hist. iv. 31. 



CHAP, were originally those of a small Roman fort which protected 
the harbour. 52 At Rhuddgaer, too, the "red fort " opposite 
Segontium, Roman coins and pottery have come to light, 53 
while similar finds at Caerleb, in the parish of Llanidan, lend 
support to the view that this was also a Roman station. 54 

From Segontium a road no doubt ran to the next fort to 
be mentioned, which was situated not far from Trawsfynydd 
in Merionethshire and was known in the Middle Ages as Mur y 
Castell, 55 but nowadays bears the name of Castell Tomen y 
Mur. This line of communication is, however, no longer 
traceable ; 56 only the northern and southern approaches to 
this fort, each known by the traditional name of Sarn Helen, 
i.e., " Helen's Causeway," 57 have been satisfactorily made out. 
The former, starting, no doubt, from Conovium, is a well-marked 
track over the hills between Dolwyddelan and Festiniog, passing 
through Bwlch Carreg y Fran and giving its name to the ford 
of Rhyd yr Halen. The latter, running due south, is visible 
at Pen y Stryd (Street End) and for many miles along the 
ridge which separates the valley of the Eden from that of the 
Cain. What name was borne by the station which was thus 
linked with the rest of the Roman world it is useless to con- 
jecture ; those who imagine it is to be found in the Heriri 
Mons of " Richard of Cirencester " overlook the certainty 
that this form was taken by Bertram, the forger of that work, 
straight out of Nennius, who uses it for the Snowdonian region. 58 
Tomen y Mur 89 is placed on the south-western slope of a hill 

52 Penn. iii. 75 ; Arch. Camb. IV. i. (1870), 359. 

53 Arch. Camb. III. ii. (1856), 326-8; iii. (1857), 218-9; vii. (1861), 37. 
M Ibid. xii. (1866), 209-14; IV. ii. (1871), 63-4. It should be added that 

there is evidence of the working of the Parys Mountain copper mines in Roman 
times (C.7.L. vii. No. 1199). 

05 Mab. 74; Bruts, 293 ; Myv. Arch. II. 547 (672), 603 (733). 

56 There seems to be a bit of genuine Roman paving on the moors above 
Bwlch Gwernog. 

67 The northern road was known by this name in the time of Camden 
(Britannia, 593), while Edward Llwyd called attention to its occurrence also 
in the parish of Llanbadarn Odwyn, Cardiganshire, and between Brecon and 
Neath (Gibson, 661). The association of this form with ancient lines of com- 
munication is certainly very remarkable ; Llwybr Hilyn and Moel Hilyn (not 
" Heilyn," as in the maps), near Llanfyllin, Pen Ffordd Elen, near Penygroes, 
and Sarn Hwlcyn, near Newmarket, may be cited as cases in point. 

68 Hist. Britt. c. 40, where " in montibus heriri " is the reading of many MSS. 

69 Penn. ii. 258; Arch. Camb. IV. ii. (1871), 190-202 ; C.I.L. vii. Nos. 143, 
144, 145 ; Lap. W. 155. 


which commands a wide view of the surrounding country ; its CHAP, 
situation proclaims its character as a post of observation. It 
was a small fort, of not more than half a acre in extent, but 
very carefully built of non-local stone of the requisite hardness, 
laid in unmortared but closely fitting courses. There is nothing 
to show what body of troops was stationed here, but a number 
of centurial stones have been discovered, recording the names 
of Julius Mansuetus and of Perpetuus, two centurions whose 
men had charge of certain sections of the building work. A 
portion of the eastern gateway is still intact, testifying to the 
care with which this work was done. The purpose of the 
" tomen " or " mound " which occupies an angle of the walls 
is obscure ; probably it is no part of the original fortress, but 
was added in the Middle Ages, when the place was a residence 
of the chieftains of Ardudwy. 

Caergai, 60 the next of the forts of this district to claim 
attention, crowns a slight rise at the head of Bala Lake and 
probably marks the point, though the Roman roads of the 
district have not been satisfactorily traced, where the road 
from Chester to Tomen y Mur branched off from that which 
led to Pennal and the country of the Demetae. Camden 81 
regarded the place as so called because built by a Roman 
Caius, but the Cai with which tradition connects it is Cai Hir 
ap Cynyr, the " Sir Kay the Steward " of the Arthurian 
romances, of whom Spenser says, referring to Caergai, 

His dwelling is low in a valley greene, 
Under the foot of Rauran mossy hore. 62 

The name is thus post-Roman in origin, like Caerhun and 
Caergybi, and the history of the fort can only be deduced 
from the remains found on the spot. Until 1885 these were 
limited to coins, tiles and pottery, but in that year the lower 
part of a monument of some description, which was possibly 
sepulchral and contained a complete inscription, came to light. 
It gives the name of Julius, son of Gavero, a soldier of the first 
cohort of Nervii. 63 This detachment of auxiliary troops, drawn 

80 Penn. ii. 220; Camb. Reg. i. 191 ; Arch. Camb. V. ii. (1885), 196-204. 

81 Britannia, 593. 

62 Faery Queene, bk. i. canto 9. " Rauran " comes from Saxton's map of 
Merionethshire (1578), which places " Rarau uaure Hill " (Yr Aran Fawr) where 
Arenig should be. 

83 Chester Catalogue, No. 210. 


CHAP, from the North of Gaul and recorded as serving in Britain in 
the time of Trajan, 64 may well have formed the garrison of 

The fort at Cefn Caer, 65 a little to the south-east of the 
village of Pennal, has so far yielded no inscriptions, but the 
quadrangular form of the defences, the coins and other small 
objects discovered, and, last of all, the remains of a hypocaust 
uncovered in 1866, are unmistakable evidence of its character. 
Like Tomen y Mur, it was built of stone conveyed from a 
distance. Its purpose was clearly to protect the passage of 
the Dovey ; a pitched way or " sarn " leads to the bank of 
the river, which affords ready access for light vessels to the 
sea and at the same time is fordable at low water. Of this 
fort also the Roman name is unknown, for Camden's suggestion 
that the name of the neighbouring town of Machynlleth points 
to its being the " Maglona " (really Magloz/a) where a garrison 
was kept in the last years of the Roman occupation, will not 
hold water for a moment. The two names can have no con- 
nection, since Machynlleth is easily explained as the Plain 
of Cynllaith, and wherever Maglova was, the context in which 
the name occurs makes it necessary to look for it in the North 
of England. 66 

There is, therefore, good evidence that the Romans were 
in permanent military possession of Gwynedd during the 
greater part of their stay in this island. The case is not 
otherwise if we turn to the country of the Ordovices, which 
became the Powys of a later age. Four forts in this district, 
viz,, at Caergwrle, Y Gaer, near Montgomery, Caersws and 
Castell Collen bear witness to the thoroughness of the Roman 
conquest. It should at once be said that there seems no pro- 
spect at present of the two towns assigned by Ptolemy to the 
Ordovices being identified ; no one has yet ventured to locate 
Brannogenium, and, though in the case of Mediolanium there 
is the additional evidence of the Itinerary and the help of the 
name itself, which denotes "the place in the midst of the 
plain," 67 antiquaries have, from the days of Camden to our 
own, propounded so many solutions of the problem that it is 

b4 C./.L. vii. No. 1194. 

65 Camden, Britannia, 589; Gibson, 651 ; Arch. Camb. III. xii. (1866), 542. 

66 Not. Dig., 209. 67 Urk. Spr. 207, 236. 


not unreasonable to regard it as in the light of present know- CHAP. 
ledge insoluble. Only modern names can therefore be given 
to the four forts now to be described ; yet their Roman origin 
is none the less certain. Caergwrle, 88 on the banks of the 
Alun, was an outpost of Dva ; the tiles found on the spot 
show that it was built by a detachment of the Twentieth 
Legion. Here, as in the Flint district, the smelting of ore was 
apparently carried on, and the remains dug up at Bryn lorcyn, 
a little to the south of Caergwrle itself, seem to suggest a 
fairly flourishing settlement. A wide gap separates this from 
the two Montgomeryshire forts, and none of the efforts made 
to bridge it by conjectures as to the road system can be pro- 
nounced quite satisfactory. It would no doubt go a long way 
to clear up the topography of this district if Mediolanium 
could with any confidence be identified. Of the two forts 
which secured the upper valley of the Severn the eastern 
stands on the English bank of the river, a little below the 
historic ford of Montgomery. Locally, it is known as " Y 
Gaer," M which probably stands for some longer form like 
Caerhun or Caergwrle, but for " Caer Flos," a name by which 
it is sometimes distinguished, there appears to be no older 
authority than the Ordnance Survey map published in 1 836. 
It is of considerable extent, occupying about 5 acres, but no 
remains of any kind have so far been unearthed, so that it is 
possible that it was not a standing camp, but merely served 
for the purposes of a single campaign. Of greater interest is 
the western fort, stationed at the meeting-place of three 
valleys which give access to the heart of the Plynlimmon 
mountain region. Caersws, 70 which there is reason to think 
is a shortened form of some such name as Caerswyswen, was 
built of red sandstone on a low-lying site washed on the 
southern side by the Carno and the Severn. It was almost 
a square, with rounded angles and an area of about 7 acres. 
The Cambrian railway station now occupies its south-west 
corner and the farm-house of Pendre its centre. The inscrip- 

68 Camden, Britannia, 605 ; Penn. ii. 46 ; C.I.L. vii. No. 1225 ; Lewis, 
Top. Diet. s.v. Hope; Arch. Camb. IV. ii. (1871), 97-8. 

89 Penn. iii. 199; Mont. Coll. xvii. (1884), 105-8. 

70 Gibson, 653 ; Penn. iii. 192 ; Arch. Camb. I. iii. (1848), 91-6 ; Mont. Coll. 
ii. (1869), 46-65 ; C.I.L. vii. Nos. 1243, 1336 (533). For the name see L. G. 
Cothi, i. 12 ; Camb. Quart. Mag. i. p. 34. 


CHAP, tions which have come to light from time to time are so elliptic 
as to afford no information as to the character of the garrison, 
but, as in the case of Caerhun, the remains have been dis- 
covered between the fort and the river of a building which was 
probably a bath-house, and which is shown by the coins found 
in its ruins to have been occupied as late as the second half of 
the third century. The roads leading out of Caersws have 
been partially traced, and one in particular is well preserved, 
namely, that which strikes westward, along the valley of 
Tarannon. This can hardly have been a line of communica- 
tion with any other fort, and is most naturally explained as 
the road to the lead mines which it is believed the Romans 
worked in the district of Dylife. 

The southward road from Caersws led to Castell Collen, 71 
a fort on the banks of the leithon, not far from Llandrindod 
Wells, which held in subjection the inhabitants of our Radnor- 
shire. It is of the usual rectangular form, with rounded angles, 
and covers about 4 acres. At present, the defences have the 
appearance of having been roughly thrown together and no 
stone walling is to be seen ; but this is not surprising when it 
is borne in mind that the ruins have for years been used as a 
quarry, from which the farmers of the district have supplied 
themselves with building stone. No inscriptions have been 
found at Castell Collen itself, but it is most likely that the 
centurial stone, recording the name of a centurion Valerius 
Flavinus, which was found in pulling down the walls of the 
neighbouring church of Llanbadarn Fawr, was originally from 
this place. Coins, pottery and tiles have been found, and the 
course of the road which runs south to the fort near Brecon 
is well marked across the common at Howey and Llandrindod. 

It may be of advantage, in order to give more definite 
outlines to this survey of Roman Wales, to cast a brief glance 
at the condition during this period of what is now the Welsh 
border. Its northern end was occupied, it has been shown, by 
a strong military post, and it will shortly appear that the same 
may be said of the southern extremity. But the intervening 
region is remarkably free from traces of military occupation. 
In Herefordshire there is no Roman fortress, unless there be 

71 Hoare's Giraldus Cambrensis (1806), I. clvi. ; Radnorsh. (2), pp. 119, 359 ; 
Lap. W. 240. 


one at Leintwardine, which is pretty certainly the " Bravonium " CHAP, 
of the Itinerary. The same is true of Shropshire and South 
Cheshire. The type of settlement found in these regions is 
one which is common in South-eastern Britain, but of which 
no example occurs within the limits of the twelve counties of 
modern Wales, namely, the Romano-British town, inhabited 
by a civil population, in which the British element was large, 
while the civilisation was Roman. Shropshire furnishes an 
excellent instance of this type in Wroxeter, 73 the Viroconium 
of the Romans, once the most important centre of the district, 
though long superseded by Shrewsbury. Among the inscrip- 
tions yielded by Wroxeter are two which commemorate soldiers 
of the Fourteenth Twin Legion, 74 and it has therefore been 
supposed that in the early days of the occupation it was the 
permanent station of that corps. While this may well have 
been the case, the withdrawal of the legion from Britain about 
A.D. 70 must have put an early end to the military importance 
of Viroconium, which henceforward is a purely civil centre. 
A British settlement had existed here before the conquest, in 
connection, no doubt, with the ford across the Severn, and had 
apparently been the tribal centre of the Cornavii. 75 In Roman 
times the place was a straggling town, spreading over 170 acres 
of undulating land, which in the later days of the occupation 
were enclosed within roughly built walls for protection against 
barbarian attack. In its centre were public buildings of some 
pretensions, including a large town hall and a complete set of 
bath-houses. The multitude of small objects discovered during 
the process of excavation, lamps, earthen and glass vessels, 
finger rings, brooches, statuettes, door keys and workmen's 
tools, bears ample witness to the busy and many-sided life 
of this half Roman, half British community. Moreover, it was 
a life which underwent no eclipse until the end of the period 
of Roman occupation, as is proved by the discovery, in as- 
sociation with its remains, of coins of the second half of the 

78 Archaological Survey of Herefordshire (Westminster, 1^96), p. 4. 

73 The remains at Wroxeter have been described by J. Corbet Anderson 
(London, 1867), Thomas Wright (London, 1872), and G. E. Fox (Archaological 
Journal, liv. (1897), pp. 123-73). 

74 C.I.L. vii. Nos. 154, 155. 

78 I infer this from the form " U(t)riconion Cornoviorum " handed down by 
the Geographer of Ravenna. 


CHAP, fourth century. Wroxeter, in short, supplies abundant material 
for those who would argue that the Britons were, within the 
limits of the province, thoroughly Romanised. 

Returning, however, to our survey of Wales, we are once 
more on military ground. In the country of the Demetae no 
certain traces of Roman settlement are to be found save the 
vestiges of two, perhaps three, forts which, with their connect- 
ing roads, served -to maintain order in the district. A road 
ran south from Cefn Caer into our Cardiganshire 76 which, 
from the Wyre onwards, is clearly traceable and bears the 
significant name of Sarn Helen. 77 At the passage of the Teifi 
it was guarded by the fort of Llanio, 78 which may be the 
Luentinum (or Luentium) of the geographer Ptolemy, though 
the philological connection is not at all obvious. The ground 
is now partly occupied by the farm-house and buildings of 
Llanio Isaf and the lines of the original fortress are difficult to 
trace. But four inscriptions which have been dug out of the 
ruins point to the site as a military one, and one mentions (as 
also does the inscription in the wall of Llanddewi Brefi church, 
which is not far off) the second cohort of Astures, a body of 
Spanish auxiliaries which is more closely connected with the 
fort of Aesica (Great Chesters) on the wall of Hadrian, 79 but 
may well have occupied Llanio at one period of its history. 
The other undoubted fort of this district was situated at Car- 
marthen, 80 at the head of the tidal portion of the Towy, in a 
strategic position which has many times proved of crucial im- 
portance in the history of Wales. It was a hill-fort of the 
Demetae, known by the Celtic name of Maridunum ; under 
the Romans it became a centre of roads which branched out to 
the north, the east and the south-east In the days of Giraldus 

76 The fort placed by Haverfield (map in Social England, vol. i.) near " Llan- 
badarn Vawr," Aberystwyth, owes its existence to confusion with Castell Collen, 
which is near Llanbadarn Fawr, Radnorshire. 

77 Gibson, 661. 

Ibid. 645; C.I.L. vii. Nos. 148-50; Lap. W. 141, 142, 143; Ephemeris 
Epigraphica, vii. p. 285 ; Arch. Camb. IV. ix. (1878), 353 ; V. v. (1888), 297-317 ; 
vi. (1889), 180-1. 

79 C.I.L. vii. Nos. 732, 1228. The cohort was in Germany in the first cen- 
tury A.D., but appears in Britain as early as A.D. 105 (C.I.L. vii. No. 1194). 

80 Gir. Camb. vi. 80 (Itin. i. 10) ; Charter of Henry II. in Carm. Cart. No. 
78 ; Camden, Britannia, 578 ; Lap. W. 98-9 ; Ephemeris Epigraphica, iii. 139 
Antiquary, 1897, p. 231. 


Cambrensis the walls were still standing and their circuit was 
known as " the old city " ; but the bustling civic life of seven 
centuries has obliterated all traces of them, and it is now only 
possible to conjecture that Maridunum lay to the east of 
mediaeval Carmarthen, St. Peter's Church and the Priory, but 
not the Castle, being within its limits and the Parade re- 
presenting its river front. Remains of the usual description 
have been found from time to time, but little of the history 
of the place can be inferred from them, save that it was still 
occupied at the beginning of the fourth century. 

West of Maridunum the Romans do not seem to have 
ordinarily travelled. The coast road to Isca is made to start 
here in the Itinerary, as though no one who used a road-book 
would be likely to have any business further west. Nor have 
judicious antiquaries been able to find either Roman settlements 
or Roman roads in Pembrokeshire. 81 The Dimetian peninsula 
seems, like the Cornish, to have been left to itself. But, in 
addition to the north road to Llanio, two lines of communica- 
tion connected Carmarthen with the east. One of these led to 
the forts on the Loughor and the Neath, in the Silurian 
country; the other took the line of the Towy and, passing 
not far from Llandeilo, as is shown by the milestone of the 
Emperor Tacitus (275-276) found in the neighbourhood, 82 
kept the course of the river until it reached the hillock near 
Llandovery, now crowned by the church of Llanfair ar y Bryn. 
Here there is some evidence of the existence of a third 
Dimetian fort. 83 From this point the road, joined no doubt by 
a branch from Llanio which can be clearly traced thence as far 
as the ancient mines of Gogofau on the Cothi, 84 struck eastward 
across the watershed. Not far from the highest point, near the 
old Black Cock Inn, a milestone has been discovered bearing 
the name of Postumus (258-26S). 85 Thence the valley of the 
Usk was followed to the fort at Y Gaer, near Brecon. 

The character and extent of the Roman occupation of the 
country of the Silures can best be understood by beginning 

81 Lit. Eng. pp. 37-48. A Roman site has lately been discovered at Cwm 
Brwyn, near Laugharne (Arch. Camb. VI. vii. 175). 

8? Arch. Camb. III. iv. (1858), 346 ; Lap. W. 98 ; Eph. Epigraphica, iii. 139. 

88 Hoare's Giraldus, I. cl. 84 Lewis, Top. Diet. s.v. Cayo. 

88 Hoare's Giraldus, I. cl. ; Breconsh. (2), p. 493 ; C.I.L. vii. No. 1161. 


CHAP, with the great military centre of the district, the legionary 
fortress of Isca. 86 Here, as in the case of Deva, it is unlikely 
that a British settlement preceded that of the Romans, since the 
latter had recourse, for the purpose of naming the site, to the 
name of the river, the Usk, in Welsh the Wysg, which winds 
majestically past it to the sea. It has already been suggested 
that the station was first established by Ostorius 87 and that it 
remained in Roman hands during the long struggle with the 
Silures. To this period, perhaps, belongs the building dis- 
covered in 1877, in which were found four coins of Vespasian. 88 
But the substantial evidence of the importance of Isca is mainly 
of the second and the third century, when it was beyond any 
doubt the standing camp of the second Augustan Legion and a 
sister fortress to Deva, and thus acquired its Welsh name of 
Caerllion, 89 which English tongues have turned into Caerleon. 
A large number of inscriptions have come to light, bringing out 
clearly, on the one hand, the purely military character of the 
settlement, and, on the other, the manifold interests represented 
in the head-quarters of a legion. Flavius Postumius, legate of 
the legion or of the province, restores the temple of Diana. 
The barracks of the seventh cohort of the legion are rebuilt 
from the foundations about A.D. 2 5 5 by order of the governor 
of Britain and the local commander. One prefect of the camp 
raises a votive altar to Fortune ; another joins his two sons 
in a dedication which has in view the weal of the emperor 
Septimius Severus and his sons, Caracalla and Geta. A 
standard bearer of the Second Legion, whose native place is 
Lyons, is buried by his heir. Tadia Exuperata raises, beside 
her father's grave, a monument to her mother and also to her 
brother, who died on an expedition to Germany. These in- 
scriptions have been found, not only at Caerleon itself, but 
also at various places in the immediate neighbourhood, such as 
Bulmore, Kemeys Inferior and St. Julian's, showing that the life 
of the fortress overflowed into the surrounding country and 
made the settlement one of the busiest in Roman Britain. 

86 Hen. Hunt. p. 7 ; Gir. Camb. vi. 55-6 (I tin. i. 5) ; J. E. Lee, Isca Silurum, 
London, 1862 (Supplement, Newport, 1868); C.I.L. vii. 36-42; Lap. W. 

87 P. 53. 8S Arch. Camb. IV. viii. (1877), 161. 

89 Gildas (c. 10) is responsible for the incorrect form, " Legionww urbs ". 


The outline of the walls can still be traced, though there CHAP, 
are few traces of the magnificence which Giraldus Cambrensis 
saw, with the eye, possibly, of the historic imagination, and 
not that of sight, for the duller Henry of Huntingdon asserts, 
half a century earlier, that hardly anything of the walls is to 
be seen. They enclosed a square of about fifty acres, in the 
midst of which now stand the church and village of Caerleon ; 
each side was pierced by a gateway, the position of which can 
still be determined. The Castle Mound, situated without the 
walls at their eastern angle, has been proved to be a mediaeval 
work ; it partially covers the ruins of Roman buildings. The 
circular earthwork which lies between the south-western wall 
and the river has, on the other hand, been very generally 
regarded as a place assigned to the sports and exercises of the 
soldiery ; it occupies a plot of ground which, hemmed in as it 
is by the river, must have been secluded and well suited for 
such a purpose. Roads led from Isca in at least three direc- 
tions, eastward to Venta, northward to Burrium, and westward 
to the forts of our Glamorganshire. A brief account of the 
stations thus connected with the legionary centre will bring 
this survey to a close. 

Two of the forts which may be regarded as western out- 
posts of Isca have recently been excavated with interesting 
results. It had long been suspected from the situation, the 
name and the visible remains of the ancient stronghold of 
Cardiff that it had commenced its history as a Roman station, 
and this has now been proved by the spade-work set on foot 
by the late Marquis of Bute, the owner of the Castle. 90 The 
Roman wall, over 10 feet thick, was found embedded in the 
great earthen rampart which had been thrown over it by later 
defenders of the spot to the extent of nearly two-thirds of its 
length. At regular intervals along it appeared polygonal 
bastions, and the north gateway, with its guard chambers, was 
found to be in exceptionally good preservation. It had been, 
apparently, long occupied and had been altered by the ad- 
dition of the bastions at the close of the period of Roman 
occupation. The Cardiff fort was of some size, covering about 
10 acres ; in the hills to the north was a smaller station, of 

90 Archaologia, Ivii. (1901), 335-52 (J. Ward). 


CHAP, some 3-^ acres, near the church of Gelligaer, 91 of which the 
walls and gateways have also been remarkably well preserved. 
" The raised sill of one of the portals is to all appearances," 
to quote an account of the excavations of 1900, "as fresh to- 
day as when the camp was abandoned sixteen or seventeen 
centuries ago. It shows the hollows worn by the chariot 
wheels, the sockets in which the pivots of the great doors 
turned, and the square hole into which the great bolt shot to 
make all fast for the night, or when danger approached." ! 
Close by a road, known as Heol Adda (Adam's Street), runs 
along the ridge in the direction of the fort near Brecon. 93 

Forts, no doubt, guarded at intervals the road which ran 
west from Cardiff to Carmarthen, but their sites have not yet 
been ascertained, and all that can safely be said of them is 
that the Nidum and Leucarum of the Itinerary must have 
stood on the banks of the rivers so called, the Nedd and the 
Llwchwr of to-day. The portion of the road between the 
Ogwr and the Nedd has been specially productive of mile- 
stones ; one found near Pyle was inscribed with the name of 
the usurper Victorinus (about 267), 94 the two from the neigh- 
bourhood of Aberavon belonged to the age of Diocletian (284- 
3I3), 95 while the fourth, recently discovered at Melin Crythan, 
also bore the name of this emperor. 96 

From the mouth of the Nedd a road struck across the hills 
to the north-east, which is still clearly traceable for a large 
part of its course under the name " Sarn Helen ". 9T Its des- 
tination was the fort in the valley of the Usk, two or three 
miles above Brecon, which has more than once been mentioned 

91 Gelligaer was excavated with scientific precision by the Cardiff Natur- 
alists' Society in 1899-1901. A full account of the discoveries, compiled by Mr. 
John Ward, was issued in 1903 (The Roman Fort of Gellygaer, London). The 
fort, though very complete in its details, is not believed to have been held for 
any long period. 

Arch. Camb. VI. i. (1901), 59. 

93 At Penydarren, Merthyr Tydfil, which lies upon the line of this road, 
traces of an important Roman settlement have lately been discovered, though its 
character has not yet been determined. See Arch. Camb. VI. vi. (1906), 

94 C.7.L. vii. No. 1160; Lap. W. 40. 

98 C./.L. vii. Nos. 1158, 1159; Lap. W. 38,41; Academy, ist Aug., 1896 
(p. 86). 

96 Antiquary , 1894, i. p. 245. 97 Gibson, 661. 


as the goal of Roman roads in Mid Wales. " Y Gaer," M as CHAP. 
it is locally known, had once a longer Welsh name, which it 
does not seem possible to restore ; its Roman name, too, is 
unknown. It was a large fort of 6 acres, with walls 7^ feet 
thick, and occupied a pleasant site near the confluence of the 
Ysgir and the Usk. Tiles inscribed with the name of the 
second Augustan legion furnish evidence that it was an outpost 
of Caerleon, while the military character of the settlement 
further appears from two tombstones from the immediate 
vicinity. The one," known for centuries as " Maen y Moryn- 
ion " (" The Maidens' Stone "), stands on the line of the Roman 
road which runs east and west a little to the north of Y Gaer, 
and, though sadly weather-beaten, still shows the sculptured 
outlines of a Roman soldier and his wife. The other 10 was 
discovered in 1877 in a field about a mile away, and is in- 
teresting as the gravestone of an officer or soldier of the 
cavalry troop of Vettones, Spanish auxiliaries known to have 
been in Britain from about A.D. ioo, 101 the approximate date 
of the inscription. It is possible that the troop may in the 
first century A.D. have formed the garrison of the station, 
though later on it is found in the neighbourhood of the Wall 
of Hadrian. 102 

The eastward road from this fort followed the course of 
the Usk, in a line still easy to trace, to another, also known as 
" Y Gaer," in the vale of the Rhiangoll, not far from Tretower. 103 
Theophilus Jones was the first to point this out as a Roman 
site and to show how it broke the long stretch of road from 
the Ysgir to the Gefenni. The ramparts, of the usual rect- 
angular form, are well marked, and tiles, coins and fragments 
of building material supply the evidence of Roman occupation. 
Nor can there be much doubt that the two stones found at 
Tretower and bearing the names of the centurions Valens and 
Peregrinus were taken from the ruins of Y Gaer. At Aber- 
gavenny few Roman remains have come to light, and it is 

9 *Breconsh. (2), p. 218; C./.L. vii. No. 1222; Arch. Comb. II. ii. (1851), 
167 ; IV. ix. (1878), 235 ; VI. iii. (1903), 12. 

"Gibson, 593 ; C./.L. vii. No. 152; Lap. W. 57. 

100 Lap. W. 75-6; Ephemeris Epigraphica, iv. (1881), 198 

101 C./.L. vii. No. 1193. 

102 Ibid. No. 273 ; Ephemeris Epigraphica, vii. Nos. 979, 980. 
103 Brecon}h. (2), pp. 416-7 ; C./.L. vii. Nos. 146, 147; Lap. W. 54. 


CHAP, therefore not certain that a fort here guarded the pass which 
leads to the Herefordshire plains. At the same time the exist- 
ence at this spot of a station of some kind or other, styled 
Gobannium from its position on the Gefenni, is clear, not only 
from the Itinerary ', but also from the direction of Stone Street, 
the Roman road which runs south from Kenchester (Magna) 
and has been traced as far as Abbey Dore. 104 At Usk, 105 the 
Burrium of the Itinerary, there was certainly a fort, in the 
neighbourhood of the present county gaol, where Roman tiles 
and pottery have been found, together with fragmentary in- 
scriptions which tell of the activity at this spot of the ever- 
busy Second Legion. 

West and north of Caerleon, therefore, the military neces- 
sities to which it owed its existence had everywhere to be 
recognised, to the all but absolute exclusion of Roman civil 
life. It was only necessary to travel a few miles eastward to 
enter a different atmosphere. Caerwent, it may be inferred 
from its name of Venta Silurum, was a tribal centre before it 
fell under Roman sway ; 106 its ruins have hitherto furnished no 
evidence of the presence of a garrison in the Roman period, but, 
so far as they have been excavated, suggest that it retained its 
importance as a place of peaceful resort and habitation, where 
luxury was not unknown. True, it was girt with massive 
walls, but it can hardly have been necessary to build a fortress 
of 40 acres in extent within 10 miles of Caerleon, and the 
explanation of these defences no doubt is that, as in the case of 
Wroxeter, they were raised at a late period of the Roman 
occupation, when the old military system could no longer give 
proper protection from barbarian attack, and each city had to 
transform itself into a camp of refuge. From Caerwent it was 
no long journey to Bath, the Aquae Sulis of the Romans, then 
as now famous for its medicinal waters and frequented by 
visitors from all parts of the province; Gloucester, too, 
was not far off, one of the few British towns which had 
Roman civic rights the whole region had been not merely 

104 Arch&ological Survey of Herefordshire, p. 14. 

10B Camden, Britannia, 568; Arch. Camb. I. i. (1846), 188 ; Ephemeris 
Epigraphia, iv. (1881), 198, 206. 

106 Proof of this has of late been afforded by the discovery of an inscription 
(Athenaum, 26th Sept., 1903). 


subdued by Rome but had become an integral part of her CHAP. 

TT* 111* 



A question has now to be considered which thus far has 
been only indirectly touched upon, but which must be answered 
ere any wide view can be taken of the course of development 
of Welsh history. What was the relation of the subject popula- 
tion of Wales, the conquered Silures, Demetae, Ordovices and 
men of Gwynedd to this Roman civilisation which established 
itself at so many points in the country ? Were they Romanised 
as well as subdued, accepting, as other provincials did in 
Western Europe, Roman institutions, learning the Latin tongue 
and merging their distinctive tribal features in the general life 
of the Empire ? Or did they live in tribal isolation, retaining 
their Celtic speech and institutions and learning little from the 
soldiers who moved in and out among them? These are 
questions which will not be answered for us by any contempor- 
ary record ; for, save for some geographical particulars, nothing 
is to be gleaned from the literature of the day as to the internal 
history of Wales from the age of Tacitus to that of Gildas. 
Only the evidence supplied by archaeology, by philology, and by 
the later history of the country is at our service in this inquiry, 
and this is in many ways so precarious that it is not surprising 
that upon it widely different views should be founded. 

It is desirable, however, at the outset to separate entirely 
the question here under discussion, namely, the extent of 
Roman influence upon Wales, from another and much more 
complicated issue, namely, the extent of that influence upon 
the province of Britain. Enough has been said in the previous 
section to show that the Roman occupation of Wales differed 
essentially in character from that of the rest of Southern 
Britain and only found a parallel in what is now the North of 
England. The presence of troops throughout the country 
inevitably suggests a state of things very unlike that which 
must have prevailed in the districts where it was not necessary 
to brandish the sword. Moreover, the fact that Wales to this 
day speaks a Celtic language and long retained an archaic 
system of laws and institutions raises a presumption as to what 
must have happened here which cannot be extended to the 
VOL. i. 6 


CHAP, south-east, for it is only matter of conjecture what language 
and what institutions were there overwhelmed by the tide of 
Teutonic invasion in the fifth and sixth centuries. It would, 
indeed, be not at all wonderful, but in conformity with the 
earlier and the later history of Britain, if it should ultimately 
appear that during the Roman period the corn-lands of the East 
went a different way from the rough hill-pastures of the West. 

The first point, then, which has a bearing on the question 
of the relations of conquerors and conquered in Wales is the 
military purpose of the two standing camps of Isca and Deva, 
with their network of dependent forts. No one who knows 
anything of the Roman army under the Empire will need to 
be told that two legions were not quartered for at least a 
couple of centuries on what is now the Welsh border without 
very good military reasons. There were but three in Britain and 
not many more than thirty in the whole Empire. It is a natural 
suggestion that, as the Sixth Legion at York, with its auxiliary 
forces, protected the north of the province from the barbarians 
of Caledonia, so the Second and the Twentieth were responsible 
for the defence of the west against possible attacks from Ireland. 
But, though they no doubt performed this function, it is diffi- 
cult to think of it as their sole duty. While the northern 
peril was real and instant, calling for unceasing vigilance and 
for fresh measures of defence under each successive emperor, 
the western is not once mentioned in any record of the time, 
and its existence is purely a matter of inference. It was 
surely for more potent reasons that these fortresses were main- 
tained on so large a scale and at so great a cost to the Empire. 
Just as the northern force had to reckon, not only with attacks 
from without, but also with the disaffection of the Brigantes, 
who in the time of Hadrian annihilated the legion quartered 
at York and for many years afterwards were restless and 
defiant, 107 so the legions of Chester and Caerleon were stationed 
there quite as much to keep in subjection the Ordovices and 
the Silures, whom it had taken so many years to subdue, as 
to guard the coasts from piratical incursions. 

But a people whom it was necessary thus to overawe can 
hardly have been ordinary peaceable subjects of the Empire. 
That they enjoyed political independence it is, of course, absurd 

107 C./.L. vii. 64; Pausanias, Grtzc. Descr. viii. 43. 


to imagine ; they no doubt bore the burdens of Empire, paid CHAP, 
tribute, worked mines, and furnished recruits for the auxiliary 
part of the army. But they so far held aloof from Roman 
civilisation as to be a possible source of danger to the province, 
and the very thorough way in which their country was covered 
with military stations, while it allowed no room for tribal 
leadership and tribal warfare, shows that they could not be 
left to themselves, to work out their own destiny, so strong 
was the old tribal feeling and so easy the return to the condi- 
tions which obtained before the conquest. 

The positive evidence supplied by the character of the 
Roman remains in Wales may be supplemented by the nega- 
tive evidence afforded by what is not to be found there. It 
has already been stated that no Romano-British town of the 
type so common in England, no Silchester or Wroxeter or 
Bath, has yet been discovered in Wales strictly so called, though 
Monmouthshire furnishes an instance in Caerwent. It has 
now to be added that another feature of the orderly, peaceable 
south-east, namely, the "villa" or Roman country house, is 
also notable by its absence, save in the district nearest Caerleon. 
At the end of the eighteenth century the foundations of a 
building were unearthed at Llanfrynach, 108 Brecknockshire, 
about half-way between the two " Gaers " of the county, 
which have since been destroyed, but appear to have been the 
remains of a " villa " . The discovery of a villa at Llantwit 
Major was also recorded in i888. 109 Elsewhere nothing has 
come to light to show that it was usual for wealthy civilians, 
unconnected with the army, to live at their ease in Wales as 
they did in the districts nearer Gaul. Here, again, the analogy 
of the North of England suggests itself, for in the country of 
the Brigantes no " villa " has been found much to the north 
of York. 110 So far as the evidence goes, the Romans had only 
one interest in Wales beyond the military, namely, the mines 
which native labour made profitable to the imperial exchequer. 
But for these they might have been content to make the line 

108 Breconsh. (2), pp. 19, 462. 

109 Transactions of the Cardiff Naturalists' Society, vol. xx. (1888), pp. 49-61. 
The remains discovered at Ely, near Cardiff, in 1893 are now held to mark the 
site of a small fort (Archaologia, Ivii. (1901), 335). 

110 Edinburgh Review, April, 1899, p. 386. The article is now understood to 
be fiom the pen of Mr. Haverfield (Eng. Hist. Rev. xix. (1904), p. 629, note 10). 



CHAP, from Isca to Deva the western frontier of the province and 
to justify the abandonment of Wales in the same terms as 
were used by Appian m to defend the withdrawal from Cale- 
donia " we have the really valuable part of the island, and 
stand in no need of the rest ". 

The existence, again, of the Welsh language is a fact of 
the first importance in its bearing on the question of Roman 
influence in Wales. If, when the light of history once more 
discloses, at the beginning of the Middle Ages, the peoples of 
Cornwall, Wales and Strathclyde, they are found speaking not 
a Romance but a Celtic language, the conclusion seems inevit- 
able that they had never ceased to do so, having escaped the 
influences which elsewhere in Western Europe made Latin 
universal. Efforts have been made to explain Welsh as a 
language introduced at the close of the Roman period by 
Brythonic invaders from the region of the two Walls, 112 but 
though this invasion undoubtedly took place and materially 
aided the Brythonic form of the Celtic speech in its struggle 
with the Goidelic, it is made to assume undue proportions 
when it is deemed capable of wiping out Latin throughout the 
whole of Wales and introducing Welsh in its place. It must, 
therefore, be concluded that the inhabitants of Wales were so 
far divorced from the main current of Roman life as to speak 
a separate language, which linked them with the customs and 
traditions of the past. 

But, though they retained their Celtic speech, the Welsh 
tribes were far from being unaffected by the influence of the 
Latin which must have overflowed their country with the com- 
ing of the soldiers. Welsh contains a very large number of 
Latin loan-words, of which some, no doubt, came in with the 
acceptance of Christianity and a ! few as learned book-words, 
but most as the result of contact with Roman civilisation 
during the three centuries of military occupation. 113 They 
entered into the language in the same way as the English loan- 
words which abound in modern spoken Welsh, and which, 

111 Hist. Rom., prooem. 5. 112 W. People, p. 503. 

113 For the material employed in the following paragraphs I am chiefly 
indebted to Les Mots Latins dans les Langues Brittoniques, by M. Loth (Paris, 
1892). The lists on pages 42-45 I found especially useful. In the matter of the 
Celtic roots I followed, of course, the Urkeltischer Sprachschatz of Dr. Whitley 
Stokes (Gottingen, 1894). 


despite the wrath of the purist in such matters, are steadily CHAP, 
working themselves into an assured position in the language. 
Now it is obvious that these loan-words do more than bear a 
general testimony to intercourse between the Welsh and their 
conquerors ; they indicate the points of contact. From them 
it may be inferred in what regions of life and thought there 
was borrowing from the richer civilisation of the conquering 
people, while their absence in certain other spheres may be 
taken as evidence that in these foreign influence did not tell. 
It has, of course, to be premised that, on the assumption made 
in these pages, namely, that the native speech of most of the 
Welsh tribes was Goidelic, evidence drawn from the Latin 
element in modern Welsh can only hold good for the tribe 
who spoke Brythonic, to wit, the Ordovices of Mid-Wales. 
But this need not in any way baulk us ; since there is no 
reason to think that there was anything exceptional in the 
relations between this tribe and the Roman power. 

The first domain in which there is evidence of extensive 
borrowing is the military. In the names of arms of offence 
and defence there was no great change ; the sword (cleddyf), 
the shield (ysgwyd), the javelin (gwaew) retained their Celtic 
names, 114 and only in the name of the arrow (saeth from 
sagitta) is there any indication of the adoption of a new weapon. 
What probably happened in this particular case was the sub- 
stitution of an iron tip for the old arrow-head of stone or 
horn. 115 But in military engineering and the like the old race 
had everything to learn from the new. " Mur " (wall), " magwyr " 
(the same), " ffos " (trench), " castell " (castle), " pont " (bridge), 
" pabell " (tent) are loan-words which show how considerable 
was the borrowing in this direction. For two of the most 
characteristic products of Roman military activity the Brython, 
indeed, used native names ; the fort which held him to 
obedience he called " caer," 11C the paved road which made 
his mountains passable he termed " sarn," m or causeway. 

114 Urk. Spr. 82, 309, 281. 

118 Schrader, Prehistoric Antiquities of the Aryan Peoples, tr. Jevons 
(London, 1890), p. 233. 

116 Even this may be from " castrum," though Stokes (Urk. Spr. 74) gives 
a native " kastro- ". Prof. Morris Jones considers that the problem has not as 
yet been solved. 

"Ubid. 313. 


CHAP. But in general it may be said that the evidence of language 
supports the conclusion, to which on other grounds it is natural 
to come, that in all that concerned warfare the tribesmen of 
Wales were highly susceptible to Roman influence. One 
remembers that there is no gift at the disposal of European 
civilisation which is better appreciated by a primitive race than 
the modern rifle. 

Another sphere in which the influence of the conquerors 
was paramount was that of letters. The Celts had no native 
alphabet or system of writing ; in Gaul, Caesar states, their 
men of learning, the Druids, used Greek characters (no doubt 
borrowed from Marseilles) for business purposes, and thought 
it wrong to commit their professional lore to writing in any 
form. 118 When the Brythons began to make use of an inscribed 
coinage they adopted the Roman alphabet, and it will be seen 
in a later chapter that, after the withdrawal of Roman troops 
from the country, the inhabitants of Wales used in their 
sepulchral inscriptions not only the Roman letters, but also 
the Latin tongue. It is not surprising, therefore, to find in 
Welsh a large number of loan-words which have to do with 
reading and writing ; such are " llyfr " (book), " llythyr " 
(letter), "ysgrif" (script), " ysgrifennu " (to write), "lien" 
(what is read), and " awdwr " (author). Be it observed that it 
is the absence of literature in the strict sense, of written com- 
positions, to which these words bear witness ; the Welsh tribes- 
men had beyond a doubt their skilled poets and musicians, 
whose flights of song were independent of literary aid, for the 
leading terms connected with these two arts, such as " bardd " 
(poet), " prydydd " (maker), " cerddor " (musician), " crwth " 
(crowd), "telyn" (harp), " cathl" (song), " tant " (string), 119 
are of purely Celtic origin, being, for the most part, common 
to Brythons and Goidels. But the songs and tales and sayings 
of these minstrels, with whom were deposited the remnants of 
Druidic tradition, were handed down orally, and it was only 
from the foreigner that any knowledge of letters was to be 

A third class of important loan-words is connected with the 
home and its material civilisation, and witnesses unmistakably 

118 Bell. Gall. vi. 14. 

119 Urk. Spr. 162, 60, 80, 99, 69, 128. For " telyn " see W. Ph. (2), 184. 


to the higher standard of comfort which was brought in by the CHAP. 
Romans. The primitive Celt had his " ty " (house), with its 
" to " (roof), its " aelwyd " (hearth), and its " drws " (door) 
secured by a "clo" (bolt). 120 But "ffenestr" (window), 
" pared " (partition), " ystafell " (chamber), " post " and " col- 
ofn " (pillar), " trawst " (beam), and " ceibr " (rafter) are all of 
Latin origin, so that it must be concluded that familiarity with 
Roman methods of building led to greater solidity of construc- 
tion among the Welsh in the domestic as in the military sphere. 
It was much the same with the household furniture and 
utensils. The bronze or iron cauldron (pair), in which the 
meat of the household was boiled, was, with its attendant 
spit (ber), 121 an ancient feature of Celtic civilisation, which 
enters into many of the early legends of the race ; of primitive 
origin, too, were the earthen pot (crochan) and the spoon or 
ladle (llwy). 122 But the knife (cyllell), the dish (dysgl), the 
metal pan (padell), the gridiron (cradell), the bowl (ffiol), the 
chest (cist, arch), the candle (cannwyll), the three-legged stool 
(trybedd), the couch (lleithig) and the chair of state (cadair) 
were all gifts of Roman culture to the Welsh tribes. How the 
old influences blended with the new is well seen in the bronze 
strainer found at Cyngadel, near Laugharne, in the early part 
of the last century. 123 The vessel is of a familiar Roman type, 
and, in view of the fact that it contained coins of Carausius, 
must have been made for an owner accustomed to Roman 
ways at the end of the third century of our era. But in its 
design it reveals the hand of a native artificer, whose models 
were not those in vogue among the metal workers of the Em- 
pire, but the graceful creations of Late Celtic art an art by 
this time extinct in the Romanised portion of Britain but still 
vigorous among the independent tribes of Caledonia and 

Significant as these groups of loan-words are, they are not 
more so than the absence of important borrowing from the 
Latin in other spheres in which the ruling people might have 
been expected to dominate their half-civilised subjects. Hardly 
any of the technical terms of Welsh law and politics, as they 
emerge in the Middle Ages, are derived from Latin sources ; 

180 Urk. Spr. 126, 127, 7, 158, 103. 1J1 Ibid. 61, 170. 

"* Ibid. 99, 241. Arch. Com*. VI. i. (1901), 24-38. 


CHAP, popular speech may employ such words as " ciwdod " (folk), 
" pobl " (people), " estron " (foreigner), and " gwig " (village), 
but the language of the courts and of the text-books knows 
(with the exception of " tyst," a witness, and a few others) no 
such alien terms ; its " gwlad " (country), " tref " (settlement), 
" alltud " (stranger), " cenedl " (kindred), " aillt " (serf), " bren- 
hin " (chieftain), and "brawdwr" (judge), 124 in short, the names 
of the cardinal features of the Welsh system of law and society, 
are of undoubtedly native origin. The impression thus created, 
that Roman law, so potent a force in every 'country in which 
Roman influence really made itself felt, found no permanent 
foothold in Wales, will be deepened upon examination of the 
substance of the Welsh codes. Though it is possible to discern, 
especially if comparison be made with Irish legal ideas and 
methods, the effect of intercourse with a people among whom 
the central authority was strong, the main ideas reflected in 
these codes are the primitive and tribal, and it is, in particular, 
difficult to imagine any race which had gone through the mill 
of Roman jurisprudence retaining the blood-feud and the 
composition for manslaughter. 

A few other indications there are which show the extent of 
Roman influence in Wales. Certain personal names, such as 
Ambrosius (Emrys), ^Eternus (Edern), Agricola (Aergol), 
Tacitus (Tegid), and Donatus (Dunod), became popular among 
the tribes, but this was chiefly, no doubt, as the result of the 
introduction of Christianity and therefore at the close of the 
period of Roman occupation. Two or three Roman sites, 
such as Caerwent and Caerfyrddin, are also known to have re- 
tained, after the withdrawal of the troops, the names which they 
had previously borne, but against these must be set a much 
larger number, such as Caerhun, Caerlleon, Caerllion, Caergybi, 
and Mur y Castell, of which the names are recent, or at any 
rate have not corne from Roman lips. Lastly, it has to be 
added that, while excavation has not as yet done much in 
Wales to bring to light actual vestiges of the British settle- 
ments of the Roman period, the scanty evidence which is 
afforded in no way runs counter to the view of the matter 
which has been put forward above. In 1849 Mr. W. Wynne 

124 Urk. Spr. 262, 137, 22, 77, 21 (see corrigenda), 169. For " brenin " see 
Celt. Br. (2), p. 282. 


Foulkes conducted a systematic exploration of the hill-fort- CHAP, 
resses which crown the long mountain rampart to the east of 
the Vale of Clwyd. There was clear evidence that they had 
been occupied both during and after the Roman period. On 
Moel Fenlli, Romano-British pottery was found in association 
with a stone knife ; a hoard of coins, ranging in date over the 
whole of the period of occupation, had also been exposed by a 
chance fire on the heathy hillside. 126 Moel y Gaer, too, 
yielded Roman pottery, which lay beneath the defences upon 
the original surface of the rock ; 126 pottery of the same type, 
with flint arrow-heads and vestiges of iron, came to light on 
Moel Arthur. 127 Roman coins and fragments of " Samian " 
ware were also found during the excavation of the ancient 
dwellings on Holyhead Island, which appear to have been in- 
habited by a mining population in a very rudimentary stage of 
culture. 128 

Roman civilisation, then, while it imported many new in- 
fluences into the old Celtic society, did not break up its essen- 
tial structure or sever its connection with the past. It left 
Wales richer in many respects, its parting gift of a new re- 
ligion (to be dealt with in the next chapter) being the greatest 
of all it bestowed, but the land remained a home of primitive 
ways and ideas, the dwelling-place of a people who, taken as a 
whole, had scarcely attained the level of culture of the Britons 
of the south-east at the time of the Roman conquest. 129 

NOTE TO CHAPTER III. i. Caratdcus. 

The popular form, Caractacus, comes from the early editions of Tacitus, in 
which the value of the Medicean MS., as the only independent authority for the 
text of Annals, xi.-xvi., was not properly recognised. In this MS. the name ap- 
pears almost always as Caratac(us), and this is the form which, according to the 
rules of Celtic philology, would yield the Welsh " Caradog". The name is to 
be connected with " car-u," to love; in Irish it is found as "Carthach," whence 
the modern McCarthy (Celt. Br. p. 284 ; Urk. Sfir. 71). 

Welsh tradition knows nothing of a Caradog ap Cynfelyn ll0 and has handed 

Arch. Camb. II. i. (1850), 82-7. Ibid. 174-81. n Ubid. 181-5. 

188 Cyttiau'r Gwyddelod, by the Hon. W. O. Stanley (London, 1871). 

139 Vinogradoff, in his Growth of the Manor (London, 1905), takes very 
definitely the view that the Romans did not attempt to uproot Celtic customs 
and institutions in Western Britain, and that even in the rest of the country Celtic 
speech and tradition survived. See bk. i. chap. 2. 

180 There is a " Caratauc map Cinbelin " in one of the pedigrees in Harl. MS. 
3859 (Cymr. ix. 176), but nothing seems to connect him in any way with the 
Caradog under discussion. 


CHAP, down no recollection of the gallant deeds of the Catuvellaunian prince. Its 
III. silence on this subject and on the no less stirring theme of the exploits of Boudica 
is clear proof that it does not really carry us beyond the period of the Roman 
occupation ; the so-called Historical Triads, while they may throw some gleams 
of light on the sixth and the seventh centuries, can do nothing to enlighten 
the darkness of the first. Even Geoffrey of Monmouth, ignorant as he was of 
the writings of Tacitus, is silent as to the fame of Caratacus, whose eventful 
career, had he known of it, would have provided him with the material for several 
eloquent chapters. Two Caradogs are commemorated in the Welsh mediaeval 
writings, but neither can be identified with the opponent of Ostorius. The one 
is Caradog ap Bran Fendigaid, a figure in the mythological group of which the 
doings are rehearsed in the Four Branches of the Mabinogi. When his father 
Bran the Blessed led an expedition to Ireland, he was left in charge of Britain 
as the chief of seven ministers or guardians (" kynweissat ") ; his father's cousin, 
Caswallon ap Beli, then came upon him, slew the other six under cover of a 
magic veil which concealed everything save the death-dealing weapon, and took 
the kingdom from him, so that he died of a broken heart (Mab. 35, 41 ; cf. W. 
People, pp. 40-1). In accordance with this, Caradog ap Bran appears in the 
Triads (though not in the oldest form of the triad see Mab. 302 and Myv. Arch. 
II. 12 (397)) as one of the Three Chief Ministers of the Isle of Britain (Myv. 
Arch. II. 4 (389), and IV. Anc. Bks. ii. 458). The other Caradog known to 
Welsh tradition is Caradog of the Stout Arm (Freichfras), a figure of the Ar- 
thurian cycle. He is represented as the son of Llyr Merini (Mab. 150, 261), as 
chief counsellor to Arthur, as a friend of St. Padarn (Cambro-Br. SS. 193) and 
as ancestor of the rulers of Brycheiniog (Breconsh. (2), pp. 28, 35) and of Morgan- 
nwg (Cymr. viii. 85). In spite of what is said by Camden (Britannia, 524) and 
Rhys (Arth. Legend, p. 172), he appears to be clearly distinguished both from 
the Caradog of the Mabinogi and the Caratacus of history. 




(In this section I owe much to the guidance afforded by Dr. Hodgkin in 
Italy and her Invaders.) 

THE fifth century beheld the overthrow of that marvellous CHAP. 
imperial system which had lasted so long that men had almost Iv - 
come to regard it as immortal. Unable to resist any longer 
the pressure of barbarian invasion, Rome bent her head to the 
storm, and, though Constantinople carried on for centuries the 
imperial traditions, the Teutonic flood submerged Western 
Europe and swept away for ever the symbols of Roman great- 
ness. Yet it must not be supposed that the causes of this great 
catastrophe were purely external. The barbarians of the North 
had for ages threatened the borders of the Empire and had 
occasionally broken through them in wild forays, but until the 
beginning of the fifth century they had won hardly any lasting 
successes ; some great captain or other had always arisen to 
vindicate the honour of Rome and drive them back into their 
forests. If they now succeeded, it was because the Empire had 
lost, through internal decay, its power of effective resistance. 
There was a general collapse of the arrangements for imperial 
defence, and this, it has been pointed out, 1 was the result of 
the economic situation. The ill-advised measures of the govern 
ment had reduced to misery and almost extinguished the 
middle class of the provinces, the smaller landowners who 
bore the main burden of taxation and whose industry and 
frugality made them the backbone of the Empire. Wealth 

1 Dill, Roman Society in the Last Century of the Empire (1898), pp. 204-34. 
According to Vinogradoff, Growth of the Manor, undue stress has been laid on 
this element in the problem, which still remains unsolved (pp. 67-8, 73-4). 


CHAP, there was still in abundance, but it was in the hands of a few. 

whose lands were tilled by servile labour and who contrived in 

a great measure to evade their obligations to the state. Thus 
the Empire fell as much, if not more, through its own weak- 
ness than through the might of the assailants. 

Britain, as an outlying member of the Roman State, was 
one of the first to feel the effects of this lowered vitality. At 
the end of the third century it had been for a time cut off from 
the rest of the Roman State through the achievements of 
Carausius, the Belgic adventurer 2 who first showed the insular 
strength of Britain and the ease with which an invader could 
be held at arm's length, when once the control of the Channel 
had been secured. The connection with Rome was, however, 
restored by Constantius, the father of Constantine the Great, 
and thenceforth maintained until the last quarter of the fourth 
century ; it is not until the age of Gratian and Theodosius I. 
is reached that plain signs appear of the coming dissolution. 
Maximus, a Spanish soldier of humble origin, who had held 
a minor post in the household of Theodosius when that 
emperor was with his father in Britain, 3 was stirred up by his 
late master's sudden rise to make a desperate bid for greatness 
himself. He found the troops which guarded Britain profoundly 
discontented with the rule of Gratian and had no difficulty in 
organising a movement for his own election as emperor. 4 
Crossing at the head of his army to Gaul he soon came face 
to face with Gratian, whom he caused to be slain on 25th 
August, 383. Gaul, Spain and Britain were now his, and for 

2 He was, according to Sextus Aurelius Victor, who flourished in the middle 
of the next century, " Menapiae civis " (De Casaribus, 39). As the contempor- 
ary panegyrist of Constantine calls Maximus an "alumnus" of Batavia (Pane- 
gyrici Latini, ed. Baehrens, p. 163), I see no reason why we should look further 
for his origin than the tribe of Menapii who dwelt in the region between the 
mouths of the Scheldt and the Meuse. See, on the other hand, W. People, p. 99. 
The name became popular in Britain and was borne by a contemporary of the 
usurper Constantine who issued a coin found not long ago at Richborough (A. J. 
Evans in Arch. Camb. V. v. pp. 138-63) and by a Christian who was buried in 
a cairn near Penmachno (Carnarvonshire) somewhat later in the fifth century 
(Lap. W. 175-6; Inscr. Chr. No. 136). 

3 The abuse of Ausonius (Ordo Urbium Nobilium, v. 70) and of Pacatus 
(Panegyrici Latini, ed. Baehrens, pp. 298-9) must rest on some fact of this kind. 

4 " Invitus. . . imperator creatus," says Orosius (Adversum Paganos, vii. 
34). But this was the emperor's standing excuse ; see the life of St. Martin by 
Sulpicius Severus, c. 20. The form Clemens Maximus is due to a misreading of 
a passage (ii. 49, ed. Halm) in the Chronicle of Sulpicius. 


several years he ruled them not unjustly, winning the favour CHAP, 
of St. Martin of Tours, whose prejudice against a usurper he 
contrived to overcome. But his ambition was not yet sated ; 
in 387 he attacked Italy, having collected for the purpose a 
great army of natives of Britain and Gaul. 5 At first successful, 
he was in 388 overthrown at Aquileia by Theodosius, who 
thus regained Western Europe for its legitimate rulers, Valen- 
tinian II. and himself. Such was the true career of the " Macsen 
Wledig " (Prince Maxen) of Welsh romance ; in the mediaeval 
accounts he is an emperor who has thirty-two crowned kings 
as his vassals, who journeys from Rome to Carnarvon to wed 
Elen, the daughter of a Welsh chief, and who recaptures the 
imperial city, taken by his enemies during his absence, with 
the aid of his wife's brave kinsmen. 6 The story, fantastic 
though it be, bears witness to the abiding impression made 
upon the people of Britain by the personality and the exploits 
of Maximus ; it is, moreover, noteworthy that history and 
fable alike speak of him as one who enlisted the youth of 
Britain in his military enterprises, thus weakening the province 
in its struggle against its barbarian foes. Gildas perhaps puts 
the matter too strongly when he alleges that Maximus despoiled 
the country of all its soldiers and military supplies, 7 but there 
can be no doubt that the strength of the garrison was reduced 
to carry out his ambitious schemes and that the way was thus 
prepared for the final separation of Britain from the Empire. 

A few years after the death of Maximus there was a 
temporary strengthening of the army of occupation, as the 
result of the energy and vigilance of Stilicho, the great general 
and minister of the Emperor Honorius. 8 But in 402 Stilicho 
seems to have withdrawn, for service against Alaric and the 
host of Goths who were invading Italy, the legion (probably the 
sixth) which was specially charged with the defence of the 

8 Sozomcn, Ecclesiastical History, vii. 13. Pacatus refers to the " satelli- 
tum Britannorum " of Maximus, who did his rough work in Gaul (Pan. Lat. p. 


6 " Breuddwyd Macsen Wledig" (Prince Maxen's Dream) is one of the 
stories included in the Mabinogion. For the Red Book text see Mab. 82-92. 

7 "Exin Britannia omni armato milite, militaribus copiis, rectoribus licet 
immanibus, ingenti iuventute spoliata, quae comitata vestigiis supra dicti tyranni 
(i.e., Maximi) domum nusquam ultra rediit " (c. 14). 

8 Claudian, On the First Consulship of Stilicho, ii. w. 250-5. 


CHAP, northern frontier of Britain. 9 It was then, in all probability, 
a sadly diminished body of soldiery which in 406 found its 
communications with Rome cut by the Teutonic occupation of 
a large part of Gaul, and in its alarm began to cast about for 
the means of ensuring its own safety. Three emperors, Marcus, 
Gratian and Constantine, were chosen in the space of a few 
months ; the first two lost life and office in quick succession, 
the third, although only a common soldier at the time of his 
elevation, was more fortunate he led his troops across the 
Channel and made for himself a position in Gaul which he 
maintained until 411. But the army which he carried off was 
the last which held the province for Rome ; when the in- 
habitants learnt that it was henceforth no part of Constantine's 
policy to defend Britain from barbarian attack, they ceased to 
look to the Empire for protection, made their own military 
arrangements and dismissed such civil officers of Rome as still 
remained in the island. 10 No doubt they still regarded them- 
selves as in theory members of the Empire, and the letter 
in which Honorius in 410 exhorted them to strenuous self- 
defence u should by no means be taken as a grant of independ- 
ence ; yet for all practical purposes Roman rule in Britain was 
at an end. 


By the withdrawal of the military protection of Rome, the 
province was left at the mercy of three groups of barbarian 
freebooters. The Saxons, a general name for the Low German 
plunderers who came from beyond the Frisian country, had 
menaced the south-eastern coast of Britain for more than a 
century, for they were among the foes with whom it was the 
special business of Carausius to deal when he took advantage of 

9 Claudian, Gothic War, w. 416-8. Hodgkin (Italy and her Invaders, vol. i. 
(2), 716) understands this passage of the Twentieth Legion, which is nowhere 
mentioned in the " Notitia," and was, he suggests, in transit when that docu- 
ment was compiled. 

10 Zosimus (vi. 5) is very explicit as to the steps taken by the Britons and is 
probably following here the earlier narrative of Olympiodorus. Coulanges takes 
the view that the " Roman officers " dislodged were merely those appointed by 
the usurper Constantine and that this was a preliminary to a return to loyalty. 
There is no hint in the narrative of anything of the kind. See V Invasion Ger- 
manique, Paris, 1891, pp. 6-7. 

11 Zosimus, vi. 10. 


the possession of a fleet to make himself supreme in the island. CHAP 
In order to check their piratical onslaughts, new military 
arrangements had been made in this quarter of Britain ; forts 
had been built and garrisoned on the threatened coast-line and 
a new officer, styled Count of the Saxon Shore in Britain, had 
been placed in command of the troops detailed for this service. 
The Picts first appear under this name in the years following 
the overthrow of Carausius and his successor, Allectus ; the 
term soon gained favour as a convenient general description 
of the unconquered tribes of Northern Britain, whose hostile 
attitude was the chief military problem which had to be faced 
by the governors of the province from the time of Agricola to 
that of the usurper Constantine. " Painted folk " they certainly 
were, tattooing their faces and bodies with the figures of birds 
and beasts, but it is not certain that this was the real origin of 
the name. 12 Against the inroads of the Picts the province had 
been fortified with especial care. Two great barriers had been 
raised at different times to keep them at arm's length ; the 
first, running from the Tyne to the Solway Firth, had been 
constructed in the time of the Emperor Hadrian ; the second, 
connecting the Forth and the Clyde, under his successor, 
Antoninus Pius. It had not been found practicable to hold the 
northern line of defence for more than a few years, 13 but the 
southern, of which the remains are still to be seen in great 
abundance, was one of the most important links in the im- 
perial frontier system and was guarded at all times by a very 
considerable force. 

The third quarter from which the province suffered attack 
was Ireland. It has already been suggested u that the danger 
of Irish invasion was one against which provision was made 
from the start, that this was, in fact, one reason for the estab- 
lishment of strong legionary stations at Chester and at Caerleon. 
But it is only at the end of the period of Roman occupation 
that evidence actually occurs of Irish inroads ; in 360 is found 
the first mention of the " Scotti," as acting with the Picts in 
certain raids in the region of the northern barriers. 16 The 

12 For another explanation see W. People, p. 79. 

13 This is the view of Prof. Haverfield (Social England, illustrated edition, 
i. p. 92). Cf. also Edinburgh Review, April, 1899, p. 376. 

14 P. 82. ls Ammianus Marcellinus, xx. i. 


CHAP. Scots, who were not for many ages to give to the land in which 
they afterwards settled its modern name of Scotland, were 
dwellers in the North of Ireland, and thus their association 
with the Picts was natural, for the passage from Ulster to 
Galloway, which is known to have been a Pictish country, 
was easily accomplished. So Gildas speaks of their attacking 
Britain from the north-west, while the Picts hailed from the 
north. 16 Partly, no doubt, they came by land, in which case 
the obstacle set in their path was the Wall of Hadrian ; but many 
crossed the Irish Sea in the rude skiffs they termed coracles, 
whose dark hulls, says Gildas, might be seen creeping across 
the glassy surface of the main like so many insects awakened 
from torpor by the heat of the noonday sun and making with 
one accord for some familiar haunt. 17 How the western coast 
was defended at this time is not at all certain ; the document 
which forms the chief authority for the organisation of the 
Empire at the beginning of the fifth century, the " Notitia 
Dignitatum," shows no legion at Chester or Caerleon, the 
Twentieth having disappeared and the Second being stationed 
at Rutupiae, our Richborough, at this time the chief British 
port on the Straits of Dover. 18 On the whole, it seems likely 
that the western system of frontier defence had been kept up 
to a comparatively late period, but had broken down about the 
time of Maximus. 

A theory has, indeed, been put forward that during the 
third and the fourth centuries of our era the western regions of 
Britain were not only constantly ravaged, but to a large extent 
settled by Irish sea-rovers. 19 In this way is explained the 
presence in Wales and the Cornish peninsula of a substantial 
Goidelic-speaking element in the population, as unmistakably 
evidenced by the inscriptions on their tombstones. A different 
explanation has, however, been adopted in the foregoing 
chapters, so far as Wales, at least, is concerned, namely, that 

16 " Scottorum a circione, Pictorum ab aquilone " (c. 14). 

17 C. 19. For Gildas see chap. v. 18 Not. Dig. 181. 

19 See a paper by Prof. Kuno Meyer in the Transactions of the Cymmrodorion 
Society for 1895-6 (pp. 55-86). The view there expressed is endorsed by Prof. 
Zimmer (The Celtic Church in Britain and Ireland, Eng. tr., 1902, p. 17) and 
M. Arbois de Jubainville (Rev. Celt, xviii. p. 355). Sir John Rhys replies briefly in 
W. People, pp. 82-3. 


the Irish thus commemorated by inscriptions in their own CHAP, 
tongue were the ancient inhabitants of the land, who had not 
been dislodged by the advancing tide of Brythonic invasion 
when the " Roman peace " put an end to tribal warfare. This 
explanation, though not without its difficulties, certainly gains 
force when regarded in the light of what is known as to the 
Roman occupation of Wales. It has already been shown that 
the road which led from Caerleon to Carmarthen was in regular 
use in the age of Diocletian (284-3 13), 20 an d, if evidence of the 
activity of the Romans in South Wales in the fourth century 
is of the scantiest, this is equally true of the province of 
Britain generally. There is, in short, no indication that the 
conquerors so far relaxed their hold upon the west in the last 
two centuries of their occupation of the island as to leave 
room for a real settlement, carried out on a large scale by 
Goidelic invaders. 

At the same time, it is not only possible, but likely that 
between the Goidels of Wales and those of the opposite coast 
there was regular intercourse. Roman coins have been found 
in great abundance on the east coast of Ireland, 21 bearing 
witness to a certain amount of commerce between the two 
countries. Along with trade there may well have been a 
certain amount of peaceful colonisation, leading to the forma- 
tion of ties of kindred between the subject and the independent 
Goidels. Wherever and whenever the Romans relaxed their 
hold, these ties would become important and would serve as a 
basis for the building up of native dynasties. It is thus one 
can most readily explain the interesting facts brought out by 
the upholders of the invasion theory that some of the Goidels 
of South Wales were known by the same tribal name (" Kin 
of Letan ") as the Ui Liathain, an Irish tribe settled between 
Cork and Lismore, 22 and that in the eighth century Tewdos 
ap Rhain, king of Dyfed, was claimed by the Deisi of our 
county Waterford as a descendant of one of their chieftains, 

80 P. 78. 

M G. T. Stokes, Ireland and the Celtic Church, second edition (1888), 
p. 16. 

M "Filii autem Liethan obtinuerunt in regione Demetorum et in aliis re- 
gionibus id est Guir (et) Cetgueli " (Hist. Britt. c. 14). Cormac in his Glossary 
also places a " Dind map Letani " among the Cornish Britons (s.v. mugeime). 

VOL. I. 7 


CHAP. Eochaid Allmuir, whose surname points him out as one who 
had sought his fortune beyond the sea. a3 

The task thrown upon the inhabitants of Britain by the 
withdrawal of the Roman forces was, it will be seen, one of 
grave difficulty. Nor is there anything to show that it was 
not manfully entered upon and vigorously carried out. The 
popular notion that, as soon as the legions departed, the 
Britons weakly resigned themselves to their fate and offered 
no serious opposition to their foes, rests entirely upon the 
language of Gildas, whose object it undoubtedly was to convey 
this impression. But Gildas writes more than a century after 
the departure of the Roman army and in such dense ignorance 
of the history of the island as to suppose that the Wall of 
Hadrian and the forts of the Saxon Shore were only built on 
the eve of the withdrawal, in order to help the distressed pro- 
vincials in the performance of their new duties. Moreover, he 
writes with a purpose, which is to convict the British nation 
(Briton though he is himself) of unfaithfulness to God, and to 
show that its calamities flow directly from its vices. In the 
earlier chapters of his work, misunderstanding a passage in the 
history of his chief authority, Paulus Orosius, he makes the 
truly astonishing statement that the conquest of Britain by the 
Romans was a bloodless one, threats and judgments serving 
instead of fire and sword, and the usual engines of war ! 24 
Having thus begun by painting the Britons as shifty cowards, 
he continues to represent them in that light as long as he can, 
until he reaches victories of which the memory is so vivid that 
he must perforce alter his tone. Little importance, therefore, 
attaches to the general outlines of Gildas's picture, though it is 

23 See the Irish tale edited by Prof. Kuno Meyer for vol. xiv. of the Cym- 
mrodor. The Welsh form of the pedigree is to be found in Harl. MS. 3859 
(Cymr. ix. 171) and Jesus College (Oxford) MS. 20 (Cymr. viii. 86). 

24 Orosius, following Suetonius (Claudius, 17), says (Adv. Pag. vii. 6) that 
the emperor Claudius "sine ullo proelio ac sanguine intra paucissimos dies 
plurimam insulae partem in deditionem recipit ". The ground had been so well 
prepared by Aulus Plautius that the emperor himself, when he came over, had 
nothing to do but receive the submission of foes already conquered. Gildas 
takes the passage as meaning that at no stage were the Romans resisted : " im- 
bellemque populum . . . non tarn ferro igne machinis, ut alias gentes, quam 
solis minis vel iudiciorum concussionibus . . . subiugavit " (c. 5). The " Par- 
thorum pacem " mentioned earlier in the chapter is, of course, that in the time 
of Augustus (20 B.C.), to which Orosius so often recurs (Adv. Pag. i. i ; iii. 8; 
vi. 21). 


likely enough that it preserves for us some valuable details, CHAP, 
such as the appeal for help in 446 (or soon afterwards) to 
Aetius, ruler of Northern Gaul in the name of Valentinian III. 86 
The facts are, that the Picts and the Scots were kept almost 
entirely out of the province and that the Saxons only effected 
a lodgment in it after a long struggle ; obscure as is the history 
of the period, it may be regarded as certain that the place of 
the Roman legions was taken by a fairly efficient fighting 

In its main features the new system was, no doubt, a con- 
tinuation of the old. The existing walls and forts gave the 
starting-point that was needed, and it was a simple matter to 
appoint new officers in place of those who had left. According 
to the "Notitia Dignitatum," 26 there were in Britain three 
important military officials in the closing years of the Roman 
occupation. The Count of Britain commanded forces which 
were not assigned to the defence of any particular district, but 
were available to meet any unexpected danger. The Count of 
the Saxon Shore had under him the Second Legion and the 
other troops which garrisoned the nine forts of the south-eastern 
coast. The Duke of the Britains was in charge of Hadrian's 
Wall, with the Sixth Legion and the garrisons of thirty-six forts 
at his command. No clear trace has yet come to light of a 
successor to the Count of Britain, though Sir John Rhys sug- 
gests that this was the real part played by Arthur, who is 
commonly termed " emperor " by Welsh tradition. 27 But the 
duties of the other two officers appear to have been carried on 
by certain Britons who were dignified with the title of " gwledig ". 
Derived from a Celtic root signifying " authority, lordship," 28 
this term in itself means no more than "prince, ruler," and is 
so employed in mediaeval Welsh literature. 29 It does not 
seem, however, to be applied to any historical figure of later 

9 Aetius might be addressed as " ter consul " not only in 446, but in any 
year thereafter until his fourth consulship and death in 454. 

Pp. 180, 182, 209. 27 W. People, p. 105. 

58 Celt. Br. (2), p. 67 ; Urk. Spr. 262. 

29 See, for instance, the proverb from the Red Book of Hergest printed by 
Dr. Gwenogvryn Evans in the Transactions of the Liverpool Eisteddfod of 1884 : 
" Gnawd gwin yn Haw wledig," i.e., " One expects wine in the hand of a prince ". 
Cf. also Myv. Arch. I. 192 (142), where the Deity is called " gwledig gwlad orfod," 
" lord of the realm of fate ". 


CHAP, date than the middle of the sixth century, and thus the con- 
jecture of Sir John Rhys, that it is, in the names in which 
it occurs, a Brythonic rendering of " dux " and " comes," has 
much to recommend it. 30 Of the four best-known wearers of 
the title, Macsen Wledig belongs, of course, to the Roman 
period ; Cunedda Wledig and Ceredig Wledig are connected 
with the North and appear to be guardians of the wall, while 
Emrys Wledig, the Ambrosius Aurelianus of Gildas, is the 
antagonist of the Saxons. Thus Cunedda and Ceredig may 
be regarded as Dukes of the Britains, while Emrys is a British 
Count of the Saxon Shore. 31 

It was inevitable that in the altered state of things the 
military element should outweigh and reduce to insignificance 
the civil. Rome had governed the province with the aid of 
two distinct sets of officers ; and her civil functionaries, re- 
sponsible as they were for justice and finance, were no less 
important than the captains of her troops. But, in ceasing to 
form part of a great imperial system, Britain lost the impulse 
which gave life and vigour to her civil institutions, and it was 
but natural that power should now fall entirely into the hands 
of the men who rendered the country the first and most neces- 
sary service by freeing it from its foes. One sign of this 
shifting of the centre of gravity is the disappearance of the 
cities as a political force in the island. City life, it would ap- 
pear, was at no time very highly developed in Roman Britain ; 32 
yet it was to the cities Honorius wrote when he bade the 
Britons defend themselves, and they might have been expected 
to play some part in the struggle with the barbarians. There 
is no evidence that they did so ; Gildas speaks in the next 
century of their past glories, the solid walls, the lofty towers, 

30 This view is expounded in Celt. Rr. (2), p. 104 and W. People, pp. 105-8. 
Skene had previously suggested that " gwledig " was the Welsh equivalent of 
"imperator " (7F. Anc. Bks. i. p. 48). Geoffrey of Monmouth seems to have been 
puzzled by the title in the case of Macsen, whom he styles " Maximianum senat- 
orem " (Hist. Reg. v. 9). 

31 A number of chiefs are styled " gwledig " in the Mabinogion (see Mab. 
342), but the only one of them (save Macsen) of whom anything is known is 
Amlodd Wledig, who is represented as the maternal grandfather of several not- 
able figures in British fable and histor , including Arthur and Illtud. See Mab. 
ioo ; Myv. Arch. II. 25 (416); Cambro-Br. SS. 158, and Myv. Arch. II. 289 

^Social England, illustrated ed., i. p. 83 (F. Haverfield). 


the well-built houses, 33 but in the ruin which marked the CHAP, 
barbarians' progress they fell, according to him, without re- 
covery, and in his own age still lay in hideous overthrow, 
desolate and untenanted. 34 Calpurnius, the father of St. Patrick, 
was a " decurio " or town councillor of some British city, but 
this did not prevent a band of pirate Scots from plundering his 
country estate and carrying off his young son into degrading 
captivity. 35 The case is typical and makes it easy to under- 
stand how in this crisis the cities lost such authority and 
prestige as they had, and how those who could command 
military force became the supreme arbiters of the destinies of 
the island. 

Of what origin the men were who now stepped to the 
front in Britain it is not very easy to say. Ambrosius 
Aurelianus, says Gildas, was the son of one who had worn 
the imperial purple, having, no doubt, been made emperor by 
the Britons ; he appears in the narrative as the only man of 
" Roman " race left in the island. 36 He belonged, it is likely, 
to a family founded by some Roman official. For Rome still 
exercised her majestic sway over the minds of the Britons, 
though she had ceased to command their obedience. St. 
Patrick, in the middle of this fifth century, and Gildas, a 
hundred years later, claim to be regarded as " cives," heirs 
and custodians of Roman civilisation, in opposition to the 
untamed barbarians. 37 Germanus of Auxerre, visiting Britain 
in 429, makes the acquaintance of a great man who wields 
" tribunician power ", 38 Even a thoroughly Celtic king like 
Voteporix, ruler of Dyfed in the days of Gildas, has graven on 

83 C. 3. In this chapter Gildas has, I believe, used an old description of the 
province, no longer applicable in respect of its account of the twenty-eight cities. 

34 " Sed ne nunc quidem, ut antea, civitates patriae inhabitantur ; sed desertae 
dirutaeque hactenus squalent " (c. 26). Mommsen must have overlooked this pas- 
sage when he hazarded the suggestion that the cities still survived, though ruin- 
ous, in the age of Gildas, and were centres, in the sixth century, of Latin speech 
(preface to Gildas, p. 10). 

85 See iii. of this chapter. * C. 25. 

w For St. Patrick, see his letter to the subjects of Coroticus (Ceredig Wledig 
of Strathclyde), addressed " non dico ciuibus meis atque ciuibus sanctorum Ro- 
manorum, sed ciuibus daemoniorum ob mala opera ipsorum, qui barbarorum ritu 
hostili in morte vivunt " (H. and St. ii. 314). For Gildas, see especially the end 
of c. 26 : " Quippe quid celabunt cives, quae non solum norunt sed exprobrant 
iam in circuitu nationes? " 

38 Bede, H.E. i. 18. 


CHAP, his tombstone no Celtic title, but the high-sounding style of 
" protector ". 39 These are facts which show that, when it 
was possible, the new rulers were glad to establish a connection 
with the old order of things and to represent themselves as 
lawful successors of the former officials. 

But it is fairly certain that it was not so much the official 
class as the wealthy landowners of Britain who supplied the 
new leaders of the British people. In Gaul the old landed 
families of the country, proud of their descent and tilling vast 
estates by the labour of a peasantry little better than slaves, 
form at this period the most prominent feature of provincial 
society, 40 and the conditions in the settled parts of Britain 
were probably much the same. Wealth and the habit of 
command would give members of this class the advantage 
necessary to raise them to power in the general struggle for 
ascendancy, and from it no doubt sprang the " tyranni " of the 
fifth century, men like Cunedda Wledig and Vortigern, who, 
whatever titles they may have assumed, were in point of fact 
the founders of Celtic dynasties. 


(On the subject of the early British Church there is a considerable literature. 
Nearly all the documentary evidence is given, in a most convenient form, in the 
first and second volumes of .Councils and Ecclesiastical Documents, for which 
Mr. Hadclan was responsible. The following works, dealing generally with the 
subject, have also been used: Usher, Britannicarum Ecclesiarum Antiquitates 
(Dublin, 1639) ; Warren, Liturgy and Ritual of the Celtic Church (Oxford, 
1881) ; Loofs, Antiques Britonum Scotorumque Ecclesice quales fuerint mores, etc., 
(Leipzig and London, 1882) ; Romilly Allen, Monumental History of the Early 
British Church (London, 1889). More recently, the whole ground has been tra- 
versed and new questions raised by Hugh Williams in the " Christian Church in 
Wales" (Transactions of Cymmrodorion Society for 1893-4), by Haverfield in the 
English Historical Review for July, 1896 (vol. xi. pp. 417 et seq.), and by Zimmer 
in his article on the Celtic Church in the Realencyclopddie fur protestantische 
Theologie und Kirche (vol. x.), separately issued in an English translation in 
1902 (London). For a searching review of the last named by Prof. Williams see 
the Zeitschriftfur Celtische Philologie, iv. pp. 527-74 (Halle, 1903).) 

No one has yet succeeded in penetrating the darkness 
which hangs around the first preaching of Christianity in 
Britain. Such was the fascination of the subject that legends 
were certain to grow up around it, and mediaeval writers do 

39 See pp. 115 and 132. 

40 Dill, Roman Society, p. 167; Edinburgh Review, April, 1899, p. 386. 


not fail to bring hither apostles and other New Testament CHAP, 
figures, such as Joseph of Arimathaea, as the first Christian 
missionaries. 41 Welshmen have been taught to believe that 
the father of the great Caratacus accompanied his son to Rome, 
there became a convert to the new faith, and, returning as a 
missionary to his native land, won immortal renown as " Bran 
the Blessed ", 42 A very old story, which gained credit at Rome 
as early as the end of the seventh century, tells how a pope 
of the second century was asked by one " Lucius, King of the 
Britains," to send a mission to this country, which became in 
consequence a Christian land. 43 Upon examination, these 
and similar stories vanish into thin air, and all that is certain 
is that at the beginning of the fourth century, when Christianity 
secured the protection and patronage of the Emperor Constan- 
tine the Great, it had already obtained a footing and made 
some progress in Britain. That the British martyrs honoured 
in later ages, Albanus of Verulamium, Aaron and Julius of 
Isca, 44 suffered in the great persecution which preceded this 
sudden change of fortune, may be reasonably supposed, though 
Gildas, who had heard something of their story, is not able 
to say so with certainty. 45 The first ascertained fact in the 
history of British Christianity is that in 314 the bishops of 
three British churches, those of York, London and (probably) 

41 H. and St. i. 22-4. 

42 The Bran story is only found in the Third and latest series of Triads 
(Myv. Arch. II. 61, 63 (402, 404)), and the related lolo MSS. (too, 115, 135, 147). 
It was unknown even to Theophilus Evans, as late as the appearance in 1740 of 
the second edition of Drych y Prif Oesoedd. Rees (Welsh SS. 77-81) showed 
that it could not be reconciled with what is known of Caratacus and his family. 
No doubt Rhys is right in supposing (Arth. Legend, pp. 172-3) that the mischief 
began with a misunderstanding of the name Bran Fendigaid (the Blessed). 

43 The Lucius story, in itself incredible, was not known, it would seem, to 
Gildas, Augustine of Canterbury or Aldhelm. It first appears in this country in 
the works of Bede (H.E. i. 4 and v. 24; see also De Tempoium Ratione in Opp. 
vi. 305), who, however, drew his information from a Roman source, the Liber 
Pontificalis. See H. and St. i. 25 ; Nenn. V. 140-6 ; Cymr. Trans. 1893 4 62-4 
(Hugh Williams). Harnack has recently shown how the story originated ; see 
Eng. Hist. Rev. xxii. 767-70. 

44 " Legionum (for legionis) urbis cives," says Gildas (c. 10). This might 
refer either to Deva (Caerlleon) or Isca (Caerliion), but the local tradition (see 
Lib. Lind. 225, " territorium sanctorum martirum iulii et aaron ") must be 
allowed some weight. 

48i( \. t conicimus " is the true reading (p. 31 of Mommsen's edition), " cog- 
noscimus" occurring only in the inferior (Cambridge) MS. Ff. I. 27. 


CHAP. Lincoln, 46 attended a council summoned to Aries in Southern 
Gaul to determine controversies which had arisen among the 
churches of Africa. It was as the result of the expansion 
of the new religion in an Empire in which communications 
were good and ideas rapidly circulated, and not in consequence 
of any notable missionary campaign, that the Christian faith 
was first planted in this island. 

During the next century, the last of the Roman occupation, 
the area covered by Christianity, which was now the imperial 
creed, was no doubt greatly extended, so that at the time of 
the severance from the Empire the whole province, so far as 
it was really Roman, may be regarded as Christian. Maximus, 
for instance, was not only a Christian, but took pains to put 
himself forward as a special champion of orthodoxy. It is 
true that archaeology furnishes very little evidence of the 
presence of Christians in the island during the Roman period, 
but it has been already remarked that it tells us little of any 
aspect of fourth-century life in Britain ; moreover, one very 
important bit of evidence has recently been unearthed at 
Silchester. Here, on the site of the little town of Calleva, a 
Roman settlement which was built within the circuit of the 
tribal fort of the Atrebates, the foundations were laid bare in 
1 892 of a building generally recognised to be a fourth-century 
Christian church. It measured 27 feet by 42, was approached 
at the eastern end by a porch, and had within the apse which 
rounded off the western end a mosaic panel which probably 
marked the position of a movable altar. In its design the 
structure closely resembles the early basilicas of Africa and 
Italy, and, as the earliest known place of Christian worship in 
Britain, it is of altogether unique interest. 47 Further light on 
the advance made by Christianity in this century is to be ob- 
tained from the accounts of the Council of Rimini (359), at 
which British bishops were present in such force as to make 
the action of three of their number who took a line of their own 
appear as that of an inconsiderable minority. 48 

46 The Corbey MS. has " colonia Londinensium " (H. and St. i. 7). This 
has been variously corrected, but the suggestion of Lingard, that it is an error 
for " Lindensium," appears the most reasonable. See Eng. Hist. Rev. xi. p. 419 

47 Ibid. xi. p. 424 ; Encyclopedia Britannica, loth edition, vol. xxxii. p. 627. 

48 Zimmer (Celtic Church, p. 4) shows how this is involved in the language 
of Sulpicius Severus (Chron. ii. 41, ed. Halm : " tres tantum ex Britannia") 


The question whether this growth of the new religion CHAP 
extended in the fourth century beyond the bounds of the area 
of which Roman civilisation was in full possession is a more 
difficult one, and has given rise to widely divergent views. " It 
is difficult to believe," says Prof. Hugh Williams, " that there 
were Christian churches in Wales before the beginning of the 
fifth century." 49 Prof. Zimmer, on the other hand, thinks it 
likely that " Christianity was gradually spread throughout 
Ireland in the fourth century by Irish-speaking Britons". 60 
On the whole, the balance of evidence seems to favour the 
former view. The army, it is said, was the part of the imperial 
system which longest retained the leaven of paganism ; nowhere 
was Christianity weaker than at the military centres. 61 Hence 
it -is not surprising that scarcely a trace of it is to be found at 
Isca or Deva or in the region of the Walls, and it will also 
follow that the Welsh tribes and those of the North were not 
likely to have heard much of the new creed from their nearest 
Roman neighbours. Testimony, at any rate, is so far wanting that 
any of the more primitive communities of the island had accepted 
Christianity before the closing years of the fourth century. 

About this time, however, a powerful impetus was given to 
missionary work by the extension to Britain of the monastic 
movement. Beginning in the deserts of Egypt, this passion 
for a life of bodily mortification and freedom from common 
joys and cares had travelled steadily westwards, and, in an age 
of violence and self-indulgence, had thrown its potent charm 
over many of the truest representatives of the Christian spirit. 
At the end of the fourth century the monastic idea had been 
popularised in Northern Gaul by St. Martin of Tours, and it is 
probable that before his death, which is placed about 400, 
disciples of his had reached Britain. One of their number, a 
Briton called Nyniaw, is said, indeed, to have undertaken about 
this time a mission to the Southern Picts and to have been 
buried in their midst, at Whithern in Galloway, in a stone church 
which he had dedicated to the memory of his revered master. 
The story, which has been accidentally preserved in the pages 
of Bede, 62 may be regarded as typical of the way in which the 

48 Cymr. Trans. 1893-4, 68. B0 Celtic Church, p. 26. 

81 So Prof. Haverfield in Eng. Hist. Rev. xi. pp. 419 (note 5), 423-4, 427. 
''-//.. iii. 4. It is not expressly stated, but is a fair inference, that 
Nyniaw was a disciple of St. Martin's. 


CHAP, new monastic enthusiasm proved a spur to missionary effort ; 
no doubt it was by men who had been trained in the same 
ascetic school, though not of necessity by the same teacher, 
that the gospel was now preached in the wilds of Cornwall, 
Wales and the North, and perhaps carried across the sea to 
some of the tribes of Southern Ireland. 53 

A vivid glimpse of British Christianity, as it appeared some 
thirty years later, is afforded by the narrative of the first visit 
of Germanus. 54 Pelagius, a British monk, 55 adopted at Rome 
about 400 the views as to original sin, free will, and Divine 
grace ever since connected with his name ; their promulgation 
gave rise to a furious controversy throughout the Christian 
Church, St. Augustine and St. Jerome being the leading 
champions of the orthodox or predestinarian theology. In 
418 Pelagianism was finally condemned, so far as the Western 
Empire was concerned, but the isolated position of Britain 
enabled it to find an asylum there, after it had been banished 
from Italy and Gaul. In 429 a synod of the bishops of Gaul 
resolved to send across the Channel two of the most distin- 
guished of their number, Germanus of Auxerre and Lupus of 
Troyes, in order to uproot this dreaded heresy. They spent 
several months in the island, and, according to the account of 
Constantius, who wrote a life of Germanus about forty or fifty 
years later, were entirely successful in their mission, in spite of 
the influential support which the Pelagians were able to com- 
mand. Christianity had by this time clearly won, in some form 
or other, the allegiance of the whole country, for heathenism 
makes no appearance in the story ; the conflict lies solely 
between heresy and the orthodox faith. The two envoys do 
not confine their activity to the churches ; everywhere, in town 

53 This seems to be involved in the reference of Prosper to " Scottos in 
Christum credentes," s.a. 431. 

54 Bede in this part of the Ecclesiastical History (i. 17-20) is simply a tran- 
scriber, the original source being the Life of Germanus by Constantius, a presbyter 
of Lyons, written about A.D. 480. 

55 For the British origin of Pelagius see H. and St. i. 15. The attempt of 
Zimmer (Pelagius in Irland, Berlin, 1901, pp. 18-21) to prove him an Irishman 
cannot be said to have been very successful (see Williams' review in the 
Zeitschrift fur Celtische Philologie, iv. pp. 531-34). As the name was widely 
diffused in the Empire, it is not probable it was a translation from the British ; 
but if so, the modern Welsh equivalent would be Morien (cf. lolo MSS. 42, 43) 
and not Morgan, as in the Welsh version of the Thirty-nine Articles (Celt. 
Heath, p. 229, note). 


and country alike, where they can find audience, they deliver CHAP, 
themselves of their message. The disputation between them 
and their adversaries is a public ceremony, of which men, 
women and children in great multitudes are spectators. News 
arriving during the bishops' stay of a barbarian invasion of 
the province, they enter the British camp and infuse a new 
spirit of confidence into the soldiery, so that the invaders, having 
been drawn into a narrow valley which is hemmed in by lofty 
hills, are scattered by the mere noise of the shouts of "Alleluia " 
sent ringing down from the o'erhanging rocks. The whole 
narrative is suggestive of a Christian atmosphere, nor does the 
statement that very many of the British soldiers sought baptism 
from Germanus and Lupus seriously affect this impression, for 
baptism was often deferred in this age by adherents of Chris- 
tianity until a late period of life. 

Something may be further learned of the British Christianity 
of this period from the early life of St. Patrick. Into the many 
vexed questions which cluster round the name of the so-called 
apostle of Ireland it is not necessary here to enter, since their 
bearing on the history of Wales is slight. But it is common 
ground among the investigators who have busied themselves 
with his career that the " Confession " and the " Letter to 
Coroticus " are two genuine works of his, which may be safely 
used as guides in threading the mazes of his history. 56 They 
bear, indeed, manifestly upon them, in their rough, uncouth 
Latinity, their abrupt transitions, and their passionate earnest- 
ness, the marks of their origin as the work, not of a scholar 
aiming at literary effect, but of a busy, zealous missionary 
whose purposes are all of the practical order. It would appear 
from them that the young Patricius came of a clerical family ; 
his father was a deacon, his grandfather (or great grandfather) 
a priest. Thus his ancestors had been Christians for several 
generations ; their names also suggest that they represented the 
Roman and not the Celtic element in the province. Patrick's 
father was, indeed, a " decurio " 57 or member of the " curia " 

88 See H. and St. (whose text I have used), ii. 296 ; Whitley Stokes, Tri- 
partite Life of St. Patrick (London, 1887), pref. xciii, c ; Zimmer, Celtic 
Church, pp. 27, 28 ; Eng. Hist. Rev. xvii. (1902), p. 235 (J. B. Bury). 

87 " Ingenuus sum secundum carnem, nam decurione patre nascor " (II. and 
St. ii. 316). 


CHAP, or town council of some British city ; he belonged to the class 
of smaller landholders and had an estate, worked by slaves, 
not far from the village or posting-house of Bannaventa, 58 
which stood on Watling Street, about 12 miles north of Tow- 
cester. Yet it would appear that the home language was not 
Latin, for in the " Confession " Patrick says that he could not 
write as others, who from earliest childhood had spoken the 
same language ; he was compelled to translate his utterances 
into a foreign tongue. 59 It is a reasonable supposition that 
he had been brought up to speak Brythonic. At the age of 
sixteen or thereabouts, he was carried off by a band of ravag- 
ing Scots to Ireland a fate, he says, which befell great numbers 
at that time. For six years he served a master whose home 
has been generally supposed to lie in the modern county of 
Antrim, not far from Slemish mountain. There, drinking the 
bitter cup of servitude and ignominy, he was deeply impressed 
by the religious truths to which he had paid little heed in the 
days of his boyish freedom and prosperity ; as he tended the 
cattle of Miliuc, he spent long hours in prayer, until at last 
a vision by night told him that the time of his deliverance 
had come. He escaped from captivity, made a journey of 
200 miles to a certain port, and there found a ship to take 
him to Britain. After many privations and strange adventures, 
he at last reached his home, little expecting to see Ireland again. 
The purpose of preaching the gospel to the people of that 
country, who were still almost entirely heathen, 60 seems to 
have gradually formed itself in his mind ; the claim of the 
benighted race upon one who had learnt their language and 
was familiar with their customs grew more and more insistent 
and, as was natural in one so impressionable and highly strung, 
took the shape of visions in which he heard the voices of those 
who dwelt beside the western sea calling him to their aid 
Many years passed by ere he found means to obey the 

58 This explanation of the " Bannauem taberniae " of the Book of Armagh 
(H. and St. ii. 296) was first put forward by Mr. E. W. B. Nicholson in the 
Academy for nth May, 1895. It is accepted by Dr. Zimmer (Celtic Church, p. 
43). For Bannaventa see I tin. Ant. 224, 229. 

59 H. and St. ii. 298. 

60 " Hiberione, qui numquam notitiam (Dei) habuerunt, nissi idola et 
inmunda vsque semper coluerunt, quomodo nuper facta est plebs Domini " 
(H. and St. ii. 308). 


summons, and in the meantime he did not repair the defects CHAP, 
of his early education. Thus, when at the age of forty-five or 
so, having been for some time a deacon, he pressed upon the 
church authorities of Britain the question of his consecration as 
bishop for Ireland, the opposition was considerable. In the 
" Confession " he bitterly complains of the treachery of one of 
his most intimate friends, who revealed in full council, when 
this matter was under discussion, an incident of the saint's 
boyhood which Patrick had confided to him, in penitential 
anguish of spirit, before entering on the diaconate. The affair 
was all the more discreditable, in that this very friend had a 
few days previously warmly congratulated Patrick upon his 
expected promotion. The root of the opposition was in all 
probability the question of training; throughout the "Con- 
fession " Patrick writes as though he had to justify his mission 
in the eyes of men who doubted his fitness for the work, 
themselves masters of rhetoric, full of wisdom and knowledge 
and learned in the law. The same attitude is taken up in the 
" Letter to Coroticus " ; " if my own people recognise me not, a 
prophet, be it remembered, has no honour in his own country ". 61 
There were also many to whom missionary effort made no 
appeal, such as the men whom Patrick describes in so lively a 
manner as jeering at him behind his back " Why should that 
fellow go and endanger his life among godless barbarians?" 62 
British Christianity in the first half of the fifth century was, 
therefore, well organised, cultured, and in close touch with the 
churches of Gaul. Patrick himself expresses the wish, could 
he safely leave the great work he is doing in Ireland, to visit 
the brethren in Gaul and meet the saints of that country face to 
face. 83 The conversion of the heathen was, perhaps, not being 
officially undertaken with the zeal that might have been ex- 
pected, but the contagion of monasticism was seizing hold of 
individuals and they, like Patrick and Nyniaw, were breaking 
through all obstacles and carrying unaided the torch of the new 

81 H. and St. ii. 316. 

82 " Iste quare se mittit in periculum inter hostes, qui Deum non nouerunt ? " 
(H. and St. ii. 310). 

83 H. and St. ii. 308-9. For the semicolon in the first line of p. 309, substi- 
tute a note of interrogation. The passage, as Prof. Williams points out (Zeit- 
schrift fur Celtische Philologie, iv. p. 555), affords no help towards determining 
the question whether St. Patrick ever actually visited Gaul. 


CH\P. faith into the darkness of barbarian lands. Such was the 
aspect of affairs when the success of the Saxon marauders in 
effecting a settlement in South-eastern Britain interposed a 
heathen barrier between the British churches and those of the 
nearest part of the Continent. Communication with Europe 
was not, indeed, rendered wholly impossible ; at the western 
end of the English Channel it was soon to grow more active 
than ever through the British settlement of the Armorican 
peninsula. But henceforward relations between British and 
continental Christianity were casual and unauthoritative, as is 
sufficiently shown by the fact that, when it is proposed by 
Augustine of Canterbury, a hundred and fifty years later, to 
re-establish them, the Celts are found to be still following rules 
for the computation of Easter which had been abandoned by 
Gaul about the time of the conquest of Kent. 64 


(This is a subject which Sir John Rhys has made peculiarly his own and 
treated fully in Celtic Britain, Celtic Folklore (Oxford, 1901), The Welsh People, 
and Arch. Camb. V. xii. 18-39, 264-302. The inscriptions, both Ogam and Latin, 
may be studied in Hiibner, Inscriptiones Britanniae Christianae (Berlin and 
London, 1876) ; Westwood, Lapidarium Wallice (Oxford, 1876-9) ; Rhys, 
Lectures on Welsh Philology, second edition (London, 1879) ; and the numerous 
notices in Archceologia Cambrensis, which has always extended to epigraphic 
studies a very liberal patronage.) 

The division of the responsibility for the defence of Britain 
between a Northern and a Southern " gwledig," the one dealing 
with the Picts and Scots, the other with the Saxons (whom it 
will henceforth be more convenient to call the English), had 
momentous consequences for the island. It led to the division 
of the Britons themselves into two sections, each with its own 
battle to fight and with little interest in that carried on by 
its neighbour. It has often been confidently stated that the 
natives of South-eastern Britain, when driven from their homes 
by the English invaders, found a refuge in large numbers 
in the mountains of Wales and thus became the ancestors of 
the Welsh people. 65 But for migration into Wales at this 

64 H. and St. i. 152. 

65 Early exponents of this view are Geoffrey of Monmouth (Hist. Reg. xi. 
10 : " secesserunt itaque Britonum reliquiae in occidentales regni partes, Cornu- 
biam videlicet in Wallias,") and William of Newburgh (Hist. Angl. ii. 5: "qui 
evadere potuerunt refugerunt in Wallias "). 


period from the East or from the South there is no evidence CHAP. 

* . 

whatever ; it is, on the contrary, remarkable how little genuine 
Welsh tradition has to tell of the great duel in the South-east 
between Celt and Saxon, while on the other hand it is circum- 
stantial in its account of the varying fortunes of the Britons of 
the North. The story of Vortigern is, of course, known in the 
West and the fame of Arthur penetrates there, but, save for a 
few great names like these, it would seem that the dwellers in 
Wales knew little of the course of the struggle wh ch was 
gradually making Britain England. Even their comrades on 
the southern side of the Bristol Channel, the men of Devon and 
Cornwall, they regarded as so divorced from them in interests 
as not to include them with themselves in the name of" Cymry " 
or " fellow-countrymen " 66 which they willingly extended to 
the Britons of the Solway and the Clyde. Thus the historian 
of Wales is not concerned with the progress of the English 
conquest until it has reached an advanced stage ; for him 
the importance of the fifth century lies in quite another direc- 

It was in this age that the Brythonic race secured a lasting 
supremacy throughout Wales, winning over the Goidelic tribes 
such victories as in time led to the total extinction of Goidelic 
speech on this side of St. George's Channel. Different opinions 
are held, as has already been pointed out, with regard to the 
origin of the Goidelic element in the population of fifth-cen- 
tury Wales, but there is no room for doubt as to its existence. 
Welsh popular tradition has always maintained that the 
" Gwyddelod " (Irish) preceded the Welsh in many parts of the 
country and has ascribed to them, under the name of " Cytiau 
Gwyddelod" (Irishmen's huts), the round stone dwellings of 
which the ruins were once so common on the bare slopes of the 
Welsh hills." 7 " Gryniau Gwyddelod " (Irishmen's ridges) are 
also shown, the supposed vestiges of a primitive agriculture. 
The tradition has left its mark in Welsh literature ; it appears 

88 For the history of this name see p. 164. 

67 Camden, in his account of Anglesey, refers to the fact that certain remains 
in the island were known as " Hibernicorum casulae ". Llwyd (Gibson, 677) and 
Rowlands (Mon. Ant. 27) were both familiar with the popular term " Cytiau 
Gwyddelod," though they deemed it a misnomer. Cytiau may still be seen near 
the South Stack, at Tre'r Ceiri, at Muriau'r Ore in Nant Gwynen, near Beddau 
Gwyr Ardudwy, and above Bwlch Gwernog in Nanmor. 


CHAP, in the lives of Welsh saints, 68 in the Triads, 69 and in the mem- 
oranda included in the " lolo MSS ". 70 Accordingly Dr. Basil 
Jones (afterwards bishop of St. David's) was able in 1850 to 
put forward in his Vestiges of the Gael in Gwynedd (pub- 
lished in 1851) a considerable amount of evidence in support 
of the view that the Gael, or as he is now more conveniently 
styled, the Goidel, was in possession of North Wales at the 
time of the collapse of Roman rule. But the decisive evidence 
of . Goidelic occupation (applying, however, rather to South 
than to North Wales) came to light after the publication of this 
book. It had long been known that in Ireland a mysterious 
system of writing, known as Ogam and consisting of groups of 
notches, had been in use in early times, but, the key to it hav- 
ing escaped the notice of scholars, ogam inscriptions remained 
undecipherable, until in 1848 Dr. Graves unfolded the mystery 
and showed how the alphabet was constructed. It was then dis- 
covered that a very large number of these inscriptions still 
survived, particularly in the counties of Kerry, Cork and Water- 
ford ; the language was invariably Irish, in an early stage of its 
history ; the purpose of the inscription was in each case to 
mark a burying-place. Some of the inscriptions appeared to 
commemorate Christians, though this could not be positively 
stated of the ogams as a whole. The ogam question took a 
new and interesting turn when inscriptions in this character be- 
gan to be discovered in Wales. The careful scrutiny of Edward 
Llwyd had in 1693 noted the ogams on a drawing of the Pool 
Park stone, but he offered no conjecture as to their purpose ; 
" as for y e stroaks on y e edges, I met with them on other tomb- 
stones, and I make not y e least question but this is also a 
tombstone". 71 Prof. Westwood in 1846 recognised as ogams 
the marks on the edges of the Kenfig stone. 72 It was not, 
however, until 1859, when a reading of the ogams on the 
St. Dogmael's stone was submitted to the Cardigan meeting of 
the Cambrian Archaeological Association, 73 that the Welsh in- 
scriptions of this class began to be systematically studied and 
the facts ascertained which show how firmly the Goidelic race 
was rooted in Wales at the epoch now under consideration. 

68 Cambro-Br. SS. 97, 101, 124. 

69 Third series, Nos. 8, 12 (Myv. Arch. II. 58, 59 (401)). 

70 Pp. 69, 71, 78-9, 81-2, 196-7. 71 Arch. Camb. I. iii. (1848), 310. 

. I. i. (1846), 182. ^Ibid. III. v. (1859), 345. 


Some thirty ogam inscriptions have come to light in CHAP, 
various parts of Wales. Of these two only, the Pool Park 
stone, which formerly stood on Bryn y Beddau, near Cloc- 
aenog and the stone lately discovered near Brynkir, Carnarvon- 
shire, 74 belong to North Wales, a circumstance which has 
its importance, though it does not mean that this part of the 
country had no Goidelic inhabitants. The rest are distributed 
as follows Brecknockshire, four ; Glamorganshire, two ; Car- 
diganshire, one ; Carmarthenshire, six ; Pembrokeshire, fif- 
teen. In every case in which the inscription is still legible, it is 
found to be sepulchral, giving the name of the deceased with 
(commonly) some indication of the parentage, while the occur- 
rence of the words " maqui " (son), " inigena " (daughter), 
" avi " 75 (grandson, descendant) makes it absolutely certain 
that the language in which they are written is Goidelic or old 
Irish. At the same time it is important to observe that twenty- 
two out of the thirty Welsh ogams are accompanied by in- 
scriptions in Latin capitals and in the Latin language, which 
appear in almost every case to be a reproduction of the Goi- 
delic epitaph in the more familiar tongue. In this respect the 
ogam inscriptions of Wales present as a group a great contrast 
to those of Ireland, for the Latin rendering is there a thing 
almost entirely unknown. While, therefore, the wide area over 
which these monuments are scattered makes it clear that 
Goidelic was a familiar tongue at the time they were erected 
throughout a large part of South Wales, it is equally certain 
that Latin had so far established itself as the literary language 
as to be used concurrently with the other in most of the ogam 
sepulchral inscriptions. 

These ogams, however, form only a small proportion of the 
inscribed stones found in Wales which are assigned to the period 
between the end of the Roman occupation and the beginning 
of the ninth century. No less than seventy Latin inscriptions 
of this age have at various times come to light, without reckon- 

74 Gibson, 686 ; Arch. Camb. VI. vii. (1907), 96. 

78 " Inigena " (in modern Irish, " inghean ") appears as the equivalent of 
" filia " in the Eglwys Cymun inscription (Arch. Camb. V. vi. (1889), 224-32; x. 
(1893), 285). " Avi," genitive of " avias, descendant, occurs, in the form 
" awi," on the Llanwinio stone (now at Middleton Hall, Llanarthney). See 
Inscr. Chr. No. 89 ; Lap. W. 91 ; W. Phil. 2), 280. " Maqui," genitive of 
" maquas," is fairly common. 

VOL. I. 8 


CHAP, ing the bilingual epitaphs mentioned above, and it becomes 
necessary to discuss their relation to the Goidelic question. 
The first point which claims notice is that the monoglot in- 
scriptions in Latin capitals are not separated from the Latin 
ones on bilingual stones by any well-marked characteristics, 
but appear to be of the same age and carved by the hands of 
the same race. At the east end of the ruined church of Llan- 
deilo, near Maenclochog, was discovered in 1889 a bilingual 
inscription to the memory of " Andagelli fili Caveti," but this 
man's brother (presumably), named " Coimagni fili Caveti " was 
commemorated by a Latin inscription only, which came to light 
at the same time, after having been long built up into the 
churchyard stile. 78 A third stone, now in the churchyard of 
Cenarth, but originally found in the same neighbourhood, 
records, in Latin capitals only, the name of a third member of 
this family, one " Curcagni fili Andagelli ". 7r Moreover, it is 
quite usual to find names which from their occurrence in ogams 
are known to be Goidelic also figuring in monoglot Latin 
epitaphs. Anglesey, where no ogams have as yet been dis- 
covered, furnishes two good instances in the " Cunogusi" of the 
Bodfeddan stone 78 and the " Maccudecceti " of that at Penrhos 
Lligwy. 79 

Hence, while the ogam inscriptions furnish incontestable 
evidence of the existence of a Goidelic population in the 
districts in which they are found, they afford no reason for 
limiting it to these parts of Wales. A comparison with Ireland 
will show that the fashion of using the ogam alphabet did not 
extend to every region settled by Goidels ; the character was 
in common use only in the south of the island, and north of 
Kildare a few sporadic specimens are all that have come to 
light. It was in some part of Munster, no doubt, that this 
kind of writing obtained its first foothold in Ireland, and for 
some reason or other it had no great vogue outside that pro- 
vince. Similarly in Wales, the ogam stronghold is the region 

Arch. Camb. V. vi. (1889), 304-13. 

77 Westw. 86 ; W. Phil, (a), 388. For the situation of the stone in 1743 
see Arch. Camb. V. xiii. (1896), 134. 

78 W. Phil. (2), 363 ; Lap. W. 192. 

Inscr. Chr. No. 154; Lap. W. 189; W.Phil. (2), 361; Mon. Ant. 156; 
Celt. Br. (2), p. 230. 


of Dyfed, where, no doubt, the character made its first appear- CHAP, 
ance in Britain, and the fact that elsewhere, and particularly in 
North Wales, it is sparsely represented only means that the 
Goidels of these districts did not take kindly to it. It would, 
in fact, appear that the really significant distinction is not that 
between the ogam-yielding districts and the rest of the country, 
but that which marks off the regions producing inscribed stones 
with rude Latin capitals from those where they are practically 
unknown. A line drawn along the course of the Clwyd, of 
the upper Dee and of the Mawddach would in North Wales 
mark off the district to the north-west in which these inscriptions 
are fairly common from that to the south-east in which they 
hardly ever occur. The corresponding line in South Wales 
would run from the mouth of the Aeron to the great bend of 
the Wye at Glasbury. In the angle between these two lines 
lies the region known as Powys, the country of the Ordovices, 
which has been treated above as a Brythonic region ; the con- 
clusion, therefore, seems irresistible that, as suggested by Sir 
John Rhys, 80 the custom of erecting over the grave of a notable 
man or woman a standing stone with the name of the deceased 
inscribed upon it was characteristically Goidelic and had little 
or no currency among the Brythonic tribes. Inscribed stones 
which commemorate Brythons are no doubt to be found, such 
as the well-known Porius stone at Penystryd near Trawsfynydd, 
but in the genuinely Brythonic parts of Wales their erection 
was not a customary way of paying honour to the dead. 

It should be said that no serious question can arise as to 
the date of these inscriptions. While the earlier of them are 
shown by the comparatively pure style of lettering to be very 
little later than the end of the Roman occupation, there is 
from this period onward a gradual deflection from the severity 
of Latin capitals until a stage is reached, about the beginning 
of the ninth century, when the whole inscription is in " minus- 
cules " or small letters of the Hiberno-Saxon type. At least 
two of the stones bear the names of historical persons who 
can be identified, namely, of King Voteporix of Dyfed 81 and 

80 Celt. Br. (2), pp. 248-50. 

81 Arch. Camb. V. xii. (1895), 303-13 ; xiii. (1896), 107 ; xv. (1898), 274. See 
also the Academy for nth January, 1896 (p. 35), and W. People, p. 503. 



CHAP. King Cadfan of Gwynedd, 82 who flourished in the sixth and 
the seventh century respectively. As a class they belong 
undoubtedly to Christian times. The majority of them have 
been found on or near ecclesiastical sites ; very many have a 
cross incised on the stone, in addition to the inscription. At 
Cefn Amwlch in Lleyn two priests are commemorated, one 
being the Hynog (" Senacus " on the stone) whose name is 
still preserved in that of the neighbouring hill called " Bryn 
Hynog ", 83 Even as regards the ogams, which it is not un- 
natural to regard as of pagan origin, the connection with 
Christianity is perfectly clear. More than half the Welsh 
ogams have been found in close association with churches or 
ancient chapels ; several have incised crosses, though in one or 
two cases these may be the work of later hands. 

During the fifth and sixth centuries the process was going 
on which gradually drove Goidelic speech altogether out of 
Wales and made the country what it has remained, save for 
the progress of English, to this day, a Brythonic-speaking land. 
Unfortunately, the story of the change has not been recorded 
by any contemporary writer ; such details as have survived are 
embedded in traditional narratives, of which the authority is 
not too high. Nevertheless, it does not seem impossible to 
construct with their aid a fairly consistent account of the 
Brythonic conquest. The primary authority is the portion of 
the " Historia Brittonum," usually ascribed to Nennius, which, 
occurring in two MSS. only, is known as the " Saxon Genea- 
logies " and is now recognised by scholars as a little Northum- 
brian or Cumbrian tract written at the end of the seventh 
century and tacked on to the " Historia " proper. 84 In this it 
is said that the great-grandfather of Maelgwn Gwynedd was 
"Cunedag," who with his eight sons came from the North, 
from Manaw Gododin, and drove out the Scots from Gwynedd 
with very great slaughter, so that they never returned. 85 This 

82 Mow. Ant. 156; Inscr. Chr. No. 149; Lap. W. 190-1 ; W. Phil. (2), 364; 
Celt. Br. (2), pp. 127-8. 

83 Inscr. Chr. Nos. 144 and 145 ; Lap. W. 177-8 ; W. Phil. (2), 365-6. The 
stones came from the site of an ancient chapel (Arch. Camb. V. xiii. 138). 

84 Cymr. xi. 134-8 (E. Phillimore) ; Nenn. V. 74-81 ; Mommsen, Pref. to 
Hist. Britt. (p. 117). Mommsen holds that the whole "Historia" may be of 
this date. 

85 Hist. Britt. 205 (c. 62). 


was 146 years before the reign of Maelgwn, a chronological CHAP. 
indication which is at once precise and vague, but which, taken 
in conjunction with the statement that Cunedda was Maelgwn's 
great-grandfather, may be regarded as assigning him to the 
beginning of the fifth century and the earliest years of the 
period of British independence. The genealogies of Welsh 
princes in Harleian MS. 3859, which were put together not 
later than the middle of the tenth century, supply the links 
which connect Maelgwn and Cunedda and furnish the latter 
with a long pedigree. 86 Maelgwn was the son of Cadwallon 
La whir (the Long-handed), the son of Einion Yrth (the Im- 
petuous ?), 87 the son of Cunedda ; as to the ancestors ascribed 
to Cunedda, most of them may be disregarded, but the names 
of his father, grandfather and great-grandfather, namely, Edeyrn, 
Padarn Beisrudd (of the Red Robe), and Tegid, wear a historical 
aspect. In the same MS. is to be found an interesting account 
of the sons of Cunedda. Their names are given, in tenth- 
century orthography, as Osmail, Rumaun, Dunaut, Ceretic, 
Abloyc, Enniaun Girt, Docmail, and Etern ; a ninth is added, 
called Typiaun, who is said to have died in Manaw Gododin 
before the great expedition and to have been represented in 
the division of the Brythonic spoils by his son Meriaun. The 
boundary of the nine, it is stated, was on the one hand the 
Dyfrdwy (Dee) and on the other the Teifi, and they held very 
many regions in Western Britain. 88 Which these regions were, 
though the MS. does not expressly mention them, may be 
inferred from the bounds given and from certain pedigrees 
which it contains, pedigrees of princes who traced their descent 
to sons of Cunedda and were clearly rulers of cantrefs or larger 
districts in North-west Wales. They were the districts which 
still preserved the names of Cunedda's sons and grandson, 
" Osmeliaun," which has not so far been identified, Rhufoniog, 
Dunoding, 89 Ceredigion, Aflogion (a cymwd of Lleyn), 90 Dog- 

86 Cymr. ix. 170. For the date of the MS. see Mr. Phillimore's introduc- 

87 Dafydd ap Gwilym speaks of " Brychan Yrth, breichiau nerthawg " 
(Works, ed. Cynddelw, 1873, 108). Cf. Myv Arch. I. 259 (187) : " A gwyr gyrth 
am byrth yn burthyaw gonvlad ". 

88 Cymr. ix. 182. 8 Mab. 73. 

90 This appears in the various lists as Gaualogion, Ganelogyon, Is Clogyon, 
etc. But the true form seems to be preserved in Jesus Coll. MS. 20, viz., 
" aphlocyawn " (Cymr. viii. 85). 


CHAP, feiling (the cantref of Dyffryn Clwyd), 91 Edeyrnion and 
Meirionydd. Other authorities add nothing of value to the 
story of the Brythonic conquest ; the statement of the " His- 
toria Brittonum " that Cunedda and his sons drove the Ui 
Liathain out of South-west Wales appears to be, as Prof. 
Zimmer holds, 92 an unwarranted extension of the original 
narrative, though, strangely enough, it is only in the region 
of Cydweli (expressly mentioned in the " Historia ") that the 
name of the Brythonic leader survives in " Allt Cunedda " 
(Cunedda's Hill). As to the poem in the "Book of Taliesin" 
which deals with Cunedda, 93 it probably contains, as Sir John 
Rhys has pointed out, 94 some remnants of the primitive legend, 
but the use of the false form " Cunedda/j" which restores a 
final " f " (= v) which never existed, shows that the writer is a 
bard of the Middle Ages, handling material of which he has 
not entire mastery. 

Cunedda, it may safely be inferred from the names of his 
immediate ancestors, Eternus, Paternus and Tacitus, came of 
a family which, whatever its origin, had been for some time 
Roman and not Celtic in its manner of life and its traditions. 
Very probably the "red robe" of his grandfather Paternus 
was a robe of state, marking out its wearer as a high Roman 
official. Cunedda himself bore a Celtic name, which may 
indicate that about A.D. 400 the Celtic element in the 
province was beginning to earn for itself a quite unusual re- 
cognition, as the power of Rome visibly declined. Yet, if we 
may trust the account given above, he gave Latin names to 
three of his sons, Romanus (Rhufon), Donatus (Dunod) and 
Eternus (Edeyrn), and there was also a Marianus (Meirion) 
among his grandsons. His home lay near the Firth of Forth, 
for Manaw Gododin, so called to distinguish it from Ynys 
Manaw or the Isle of Man, was on the northern border of the 
country of the Votadini, where Slamannan in Stirlingshire 
still keeps the name alive. His title of " gwledig," it has al- 
ready been suggested, points to him as the successor of some 
Roman general, and, if he really hailed from the North, no office 
would seem to be more appropriate to his position than that 

91 Owen, Pemb. i. 201, note by E. Phillimore. 92 Nenn. V. 92-3. 

93 IV. Anc. Bks. ii. pp. 200-2; Myv. Arch. I. 71 (60-1). 
9t Celt. Br. (2), pp. 118-9. 


of defender of the northern or the southern wall. It is fairly CHAP. 


certain that he was a Christian, for not only does tradition 
connect him and his family with important missionary work 
in Wales, but the names Donatus and Marianus are of dis- 
tinctively Christian origin. 

Whether the men who gave their names to the districts 
lying between the Teifi and the Dee were really the sons of 
Cunedda and not rather his followers and lieutenants may be 
open to doubt, but that they were the actual founders of 
Brythonic chieftainships in this region can hardly be questioned. 
Names like Ceredigion (Caraticiana), Rhufoniog (Romaniaca), 
Meirionydd (Mariania), require a Ceredig, a Rhufon, a Meirion 
to make them intelligible and the Cunedda legend supplies the 
simplest and most reasonable explanation of their origin. 
Moreover, the legend in its oldest form establishes the sons of 
Cunedda in the precise part of Wales where history is prepared 
to find them. None are assigned to the isle of Anglesey or 
the opposite coast of Carnarvonshire, the Mon and Arfon of 
Welsh writers, and this harmonises with the tradition that it 
was in a later age that the Goidels were overcome in these, 
their latest strongholds. On the other hand, none are alleged 
to have made any conquests in Powys, which is thoroughly 
consistent with the view that the men of this region were al- 
ready Brythons and not likely, therefore, to have had anything 
to do with Cunedda's enterprise save as allies. 

An interesting question is raised by this last consideration, 
namely, how far Cunedda drew support from the Brythonic 
element in Wales itself. Tradition is silent on the point, but 
Sir John Rhys has adduced several reasons for thinking that 
the Ordovices of Powys were sharers in his undertaking and 
benefited largely by it. 95 Rhyd Orddwy, near Rhyl, Cantref 
Orddwyf, an old name of Meirionydd, and Dinorwig, which 
is found as Dinor^/wig, seem to mark their progress into 
Goidelic country, while at Penbryn, in Cardiganshire, is an in- 
scribed stone of the fifth century set up in memory of " Cor- 
balengi Ordous," an Ordovician settler, it may well be believed, 
who came into this district in the train of Ceredig. 96 A Bry- 
thonic immigration on a large scale is indeed required by the 

96 Celt. Br. (2), p. 302. 

88 Gibson, 648 ; Inscr. Chr. No. 115 ; Lap. W. 146 ; W. Phil. (2), 379. 


CHAP, facts of the case, for it is not at all likely, despite the talk of 
"expulsion" in the traditional accounts, that there was any 
great displacement of the older population, and thus only the 
arrival of a very numerous Brythonic colony would establish 
that balance in favour of the language which enabled it ere 
long to sweep Goidelic from the field. 

According to tradition, the reduction of the Goidels of 
Gwynedd was not completed by Cunedda and his sons. One 
of the eight, Einion Yrth, did not, it may be noticed, give his 
name to any cantref or cymwd, and the history of his de- 
scendants suggests that this was because he remained in the 
fighting line, waging war until the end of his life against the 
Goidels of Mon and Arfon. Be this as it may, it is to his son, 
Cadwallon the Long-handed, that the credit is given by tradi- 
tion of finally securing Brythonic ascendancy in North Wales 
by defeating the Irish in a great battle fought at Cerryg y 
Gwyddyl, near Trefdraeth in Anglesey. This fight is first 
mentioned in the Triad which speaks of the Three Fettered 
Warbands of the Isle of Britain ; one of their number was the 
warband of Cadwallon Lawhir, who, when fighting with 
" Serygei " the Irishman at Cerryg y Gwyddyl in Anglesey, 
tied to their feet the fetterlocks of their horses and so made 
flight impossible, leaving victory and death as sole alterna- 
tives. 97 Other accounts speak of the battle as fought at Llan 
y Gwyddyl, 98 but this was probably another name for the 
same place. 99 Henceforth, the Brythons were supreme in 
North Wales ; Cadwallon's son Maelgwn seems to have had 
no difficulty with the newly conquered Goidels on the con- 
trary, Gwynedd was the centre of his power. How long 
Goidelic speech lingered on in this district it is not easy to say, 
but it is not probable that it disappeared at once. The 

97 Triads, first series, No. 49 = second series, No. 40 (Myv. Arch. II. 12, 16 
(39 1 ) 397) ! Cymr. vii. 129 ; Mab. 305). Geoffrey of Monmouth makes " Cad- 
uallo Leuirh rex Venedotorum " one of the vassal princes of Arthur (Hist. Reg. ix. 
12). Cf. Bruts, 200. Caswallon is only found in late authorities, e.g., lolo MSS. 

98 lolo MSS. 78-9. 

99 So lolo MSS. 123. A stone circle in the parish of Towyn is (or was) 
known by the name of " Eglwys y Gwyddel " (Arch. Camb. IV. v. (1874), 242). 
Hence " Llan y Gwyddyl " and " Cerrig y Gwyddyl " are probably both popular 
descriptions of some monument of the kind. 


" Maccudecceti " inscription, for instance, belongs to the sixth CHAP, 
rather than to the fifth century. 100 

Even less is known of the progress of the Brythons in the 
South than in the North. Save for the occupation of Cere- 
digion, the activity of the house of Cunedda appears to have 
been confined to North Wales, and it is only possible to con- 
jecture how the Brython gradually edged out the Goidel in the 
Vale of Towy, in Dyfed and in the land of the Silures. The 
Brythonic colonisers of the South were probably the Ordovices 
of our Radnorshire, whence they pressed up the vale of the 
Irfon into the region of Llandovery, Llandeilo and Llanelly. 
A movement such as this might explain the undoubted fact 
that few Latin inscriptions of this period have been discovered 
between the Towy and the Tawe, and also the story as to the 
conquests of Cunedda in Gower and Kidwelly. The Goidels of 
South Wales would thus be split up into two divisions, those 
of Dyfed and those of the Silurian country, and in each of these 
two districts there is evidence that the ancient language lingered 
for many generations. In Dyfed it was in use in the middle of 
the sixth century, as is shown by the tombstone of Voteporix 
which stood in Castell Dwyran churchyard, on the borders of 
Pembrokeshire and Carmarthenshire. 101 St. David, if we may 
believe his legend, found about the same time a " Scottus," that 
is, a Goidel, named Baia or Boia in possession of the land on 
which was afterwards established the saint's famous monastery 
of Mynyw. 102 In the other region, that of Morgannwg, evi- 
dence is likewise forthcoming of the late survival of Goidelic ; 
in a rude Latin inscription at Merthyr Mawr, which is shown 
by its lettering to be not earlier than the seventh century, re- 
ference is made by the man who set up the memorial to his 
" scitlivissi," a word which Sir John Rhys regards as without 
doubt Goidelic for " scout, newsbearer ". 103 

Thus it is likely that the older Celtic tongue, ever threatened 
by the victorious progress of Brythonic, nevertheless maintained 
a precarious hold upon certain parts of Wales until the eve of 
the Danish invasions. When it finally went, it left several 

100 Hiibner places it in his second class (sixth and seventh centuries). See 
Pref. to Inscr. Chr. xx. 

101 See pp. 115, 132. loa Cambro-Br. SS. 124. 

103 Inscr. Chr. No. 67; Lap. IV. 15-7; Arch. Camb.V.xvi. (1899), 158-63. 


CHAP, legacies behind it The syntax of the old language, which 
was itself an inheritance from pre-Aryan times, profoundly 
modified that of the new, giving modern Welsh those syntac- 
tical peculiarities which distinguish it so sharply from most 
other languages of the Aryan family. Goidelic loan-words 
remained in the Brythonic vocabulary, and many place-names, 
such as Cenarth ( = Brythonic Pennarth) and Cwm Llwch (from 
Goidelic "loch," a lake), continued to wear their Irish dress. 
The folk-lore and legends of the conquered race also lived on, 
arid thus it was that Welsh children came to hear of the great 
lawgiver, Dyfnwal Moelmud, of the wizard pilot and merchant, 
Manawyddan son of Llyr, and of the wily enchanter, Gwydion 
son of Don. 

NOTE TO CHAPTER IV. iv. The Historical Triads. 

The form of expression known as the Triad, in which objects are grouped 
together in threes, with a heading indicating the point of likeness (e.g., " Three 
things not easily restrained, the flow of a torrent, the flight of an arrow, and the 
tongue of a fool") is characteristically Welsh and is found in the earliest Welsh 
literature. It has been supposed to be of Druidic origin, and passages from 
Pomponius Mela and Diogenes Laertius have been quoted as containing triads 
which were current in the Druidic schools (Edw. Davies, Celtic Researches, 
London, 1804, pp. 150-1). But the passage cited from Mela (iii. 2, " ut forent 
. . . manes ") is not a triad at all ; it simply states the Druidic doctrine of the 
transmigration of souls. That from Diogenes (prooem. 5) is triadic in arrange- 
ment; it does not, however, record an actual saying of the Druids, but is a sum- 
mary of teaching common to this order and to the gymnosophists of India. The 
triad may more reasonably be connected with that reverence for the Trinity to 
which Giraldus bears testimony (vi. 203 ; Descr. i. 18) and which, according to 
him, was the source of the Welsh custom of sitting to a meal in groups of three. 
Whatever may have been its origin, it was a rhetorical form which flourished 
exceedingly in Welsh soil. Triads are found in the oldest Welsh MS., the Black 
Book of Carmarthen (fo. 14 a, b), and in the ancient versions of the Welsh Laws 
(see the " Three Stays of Blood" in LL. i. 456, 784 ; ii. 768, 890). References 
to well-known triads occur also in the " four branches " of the Mabinogion (42, 
43, 44, 49, 80, 94) and in the poems of the mediaeval bards ; see, for instance, the 
reference to the " Three Open-handed Chiefs " in Myv. Arch. I. 293 (208). As 
time went on, the mass of triadic literature grew more and more bulky, and it 
ultimately became substantial enough to occupy a considerable proportion of the 
third volume of the Myvyrian Archaiology. 

The so-called " Historical " Triads, or Triads of the Isle of Britain, have to 
do with the personages of early British history, of whom they tell a good deal 
which is not to be found in other sources. The question of their historical value 
is, therefore, of some importance. Two mistakes have commonly been made by 
those who have drawn upon the Triads for historical material ; the worthless 
third series has been treated as of equal value with the first and the second, and 
it has been forgotten that the triad, as we have it, is not a literary record, but 
merely preserves an oral tradition at a certain stage in its transmission. The 
three series referred to are the three printed in the second volume of the Myvyrian 


Archaiology ; it will be seen, if they are carefully examined, that their compilers CHAP. 
had in the main the same triads before them, but each collection represents the IV. 
work of a separate editor. Series i. (II. i-ig, upper text (388-394)) contains 
ninety-two triads; it is not clearly shown by the Myvyrian editors how they 
arrived at their text, but the fact that Peniarth MS. 45, formerly Hengwrt MS. 
536, a MS. of the end of the thirteenth century (Evans, Rep. i. pp. 379-80), con- 
tains Nos. 7-45 of this series in the Myvyrian order (printed in IV. Anc. Bks. ii- 
pp. 456-64) makes it fairly certain that it represents in the main an old collection. 
Stephens boldly assigned it to the fifteenth or sixteenth century (Lit. Kym. (2), 494), 
but the reference to " ystoria y Greal " (No. 61) on which he relied is 'probably a 
late addition. Series ii. (II. 1-22, lower text (395-399)) is taken from the Red Book 
of Hergest ; a more accurate text will be found iq Mab. 297-308. The Red Book 
was written about the year 1400 (Evans, Rep. ii. p. i), but there is evidence which 
carries this second series somewhat further back, for Nos. 13-60 appear in Peniarth 
MS. 12, formerly Hengwrt MS. 202, of the fourteenth century, and it would 
seem that the MS., when complete, contained the whole series and that it was 
from this source that the series was copied by the scribe who included it in the 
Red Book (Cymr. vii. 89-90, 95-100; Evans, Rep. i. pp. 305, 324). It should be 
added that Stephens, when he sought (Lit. Kym. (2), 494) to date this series with 
the aid of a supposed reference to Owain Gwynedd (ob. 1 170), overlooked the obvi- 
ous fact that " owein gwyned " (Mab. 302) is a misreading of " o vein gwyned " 
(Cymr. vii. 98). 

Considerable use has been made in these pages of the triads of the first and 
second series, in the belief that they have handed down genuine old traditions 
and thus contain historically valuable matter, though with a mythical admixture. 
The third series (Myv. Arch. II. 57-75 (400-11)) is of an entirely different char- 
acter. Nothing at all resembling it is to be found in any ancient MS., and the 
history of the Myvyrian copy does not seem to go much further back than 1601. 
This is much the fullest of the three series, not only in the number of triads (126) 
which it contains, but also in its explanatory detail. The added matter is, how- 
ever, of a very suspicious kind; it introduces a number of personages, such as 
Hu the Mighty, Dyfnwal Moelmud, and Plenydd, who are not mentioned in the 
other two series and of whom some are certainly mythical. An examination of 
this series, which has been largely used by writers on Welsh history from the 
time of Dr. Owen Pughe, leaves little doubt that it is the product of the unhappy 
activity of the Glamorgan school of bards and antiquaries of the sixteenth century 
(Cymr. xi. 126). For the historian it is almost wholly useless. 




(The material for this section has been chiefly derived from Gildas, for 
whom see the next section, with the appended note.) 

CHAP. BETWEEN the first settlement of the English in Britain and 
v< the mission of Augustine lie 150 years which are not a whit 
less important in the early history of Wales than in that of 
England. During this period the Welsh tribes cast off all 
traces of heathenism and of political subjection and, in common 
with the more civilised Britons of the East, become well organ- 
ised Christian communities, ruled by powerful monarchs, 
ministered to by a learned clergy, led to battle by champions 
whose renown has not yet faded. This is the age of Maelgwn 
Gwynedd, of Gildas and David, above all, of Arthur and his 
knights of the Round Table. Did we know the true story of 
these years, it would beyond a doubt be found to be no less 
stirring and romantic than that which the deft hand of fable 
has woven around the dim figures of that far distant day. But 
the period is one for which we have, with one important 
exception, no contemporary evidence ; were it not for Gildas, 
it would be involved in all but total darkness. There is, indeed, 
a mass of traditionary material, embodied in the Triads, the 
lives of the Welsh saints, the Mabinogion, the narrative of 
Geoffrey of Monmouth, and the works of the older poets, which 
probably has, to a far greater extent than is usually supposed, 
a substantial historical basis. But the clue to the interpretation 
of this evidence, the touchstone which will enable the inquirer 
to distinguish fact from fiction, history from legend, in this 
wonderful medley has not yet been discovered, and accordingly 
but sparing use has been made in these pages of tradition, in 



spite of the temptations which it spreads in the path of him CHAP, 
who would write a connected history of the Welsh people. 

In the South and in the East the great struggle was going 
on which had as its final issue the making of England, the 
creation of a group of Teutonic states in which hardly a trace 
remained of the language, the religion, and the institutions 
of fifth-century Britain. It has been already remarked * that 
the interest of the Welsh tribes in this struggle seems to have 
been but languid and that they handed down few traditions of 
it. The fame of Arthur, who vindicated the honour of the 
British arms against the heathen invader in many victories, 
stood high among them, but they seem to have had little 
detailed knowledge of his achievements ; he was " the Emperor 
Arthur," leader of a puissant band of warriors, but in the Welsh 
romances which recount his great doings and those of his men, 
the " Saeson," strangely enough, are not even mentioned. 2 
One is not disposed, on this account merely, to banish him to 
the realm of myth, for there are many circumstances which 
tend to establish his real historical existence. The name, it 
has been pointed out, is probably of Roman origin and derived 
from that of the Artorian clan ; 3 moreover, Gildas expressly 
says that the battle of Badon Hill, fought about the year 500, 
was a decisive victory for the Britons, giving them immunity 
from hostile attack for a generation, 4 and such a victory could 
surely only have been won under the leadership of a great 
captain. If Gildas refrains from giving the name of the 
victorious commander, that is of a piece with the reticence 
shown in this respect throughout his book ; it is with him the 
exception to introduce a name, even where the narrative 
obviously calls for it. 6 But, while there can be little doubt 

1 P. in. a See the index to Mob. 

3 Celt. Br. (2), p. 237. For the name of Artorius see Tacitus, Annals, xv. 71 ; 
Juvenal, iii. 29; Social England, illustrated edition, i. p. 98. Many members of 
the gens are included in Pauly-Wissowa. 

4 Cap. 26. " Mons Badonicus " is still unidentified. The view of Guest 
(Origines Celtica, ii. pp. 187-9) that it was Badbury in Dorset has been very 
generally accepted, but Stevenson has shown (Eng. Hist. Rev. xvii. (1902), pp. 
633-4) that the names are not, as was supposed, identical. 

8 See c. 23, where Vortigern is merely called " superbo " and " infausto 
tyranno " ; c. 28, " immundae leanae Damnoniae," " sancti abbatis," and " legi- 
tima uxore"; c. 31, " propria coniuge"; c. 33, " avunclum regem " ; c. 36, 
" magistrum elegantem ". 


CHAP, that a great warrior named Arthur led the Britons to victory 
about the beginning of the sixth century, the vagueness of 
Welsh tradition 6 leaves the historian with no means of pinning 
him down to any particular part of the country and furnishes 
only negative evidence, for it may fairly be inferred from it 
that Wales, at any rate, was not the theatre of his deeds of 
prowess. On the whole, it is most likely that his wars were 
waged in the South and East of the island, and it may be of 
some significance in this connection that, according to the notice 
in the Historia Brtttonum, four of his battles were fought on 
the river "Dubglas" in the region of "Linnuis," i.e., Lindsey. 7 
The districts which were under the sway of the " gwledig " 
of the North-west do not seem to have tasted the bitterness of 
this strife until far on in the sixth century. In the old age of St. 
Patrick, about the year 450, a glimpse is afforded of the doings 
of Ceredig Wledig, the " Coroticus " of his letter of denuncia- 
tion, who bore rule, it would seem, in the valley of the Clyde. 8 
He is a powerful prince, who not only successfully raids the 
Irish coast and carries the natives off as slaves, but is also on 
good terms with his neighbours, the Picts and the Scots, to 
whom he is able advantageously to dispose of his ill-gotten 
gains. He is also a Christian ruler, having holy men about 
him and rendering alms to the Church ; in his attack upon him 
for shameless disregard of the rights of Irish converts, St. 

6 The Historia Brittonum, in which appears the oldest account of Arthur, 
gives the names of twelve battles which he fought, winding up with " bellum in 
monte Badonis " (c. 56). So far as the places can be identified, they belong to 
the North and thus lend support to the theory of a Northern Arthur. But, to 
judge from the inclusion of " in urbe legionis " (the battle of Chester, fought 
about 615), the list is a miscellaneous one of famous battles of the period and 
furnishes no clue to the real field of Arthur's operations. Some of the Welsh 
romances give him a court at Caerleon (Mai. 162, 215, 244) ; older traditions 
connect him with Celli Wig in Cornwall (Mab. 133, 135, 141, 301). 

7 Lindens-es would yield Linnuis in old Welsh. 

8 For the letter to Coroticus see H. and St. ii. 314-9. This prince was 
long supposed to be Ceredig ap Cunedda, lord of Ceredigion (Welsh SS. p. 135 ; 
Todd, St. Patrick, p. 352; H. and St. ii. 314). But attention was called to the 
fact that Muirchu Maccu-Machtheni, a seventh-century writer on the life of St. 
Patrick, speaks of Coroticus as " Coirthech regem Aloo," i.e., Ail Cluade or Dum- 
barton (Stokes, Tripartite Life, pp. 271, 498), which is a much more likely home 
than South Wales for an associate of Picts and Scots. " Ceretic guletic " ap- 
pears in the pedigree of the princes of Strathclyde at a point consistent with 
the floruit here suggested, viz., A.D. 450 (Cymr. ix. 173). The phrase " quern 
ego ex infantia docui " in the letter appears to point to the end of St. Patrick's 
career as the time of its composition. 


Patrick can appeal to the Christian sentiment of his court and CHAP, 
remind the warriors of Coroticus, whose conduct has put them 
outside the pale of Roman citizenship and made them the 
comrades of demons and fellow to the Pict and the Scot, that 
the glory of an earthly realm may vanish as a cloud, or as 
smoke that is scattered by the wind. The vices of Coroticus 
and his followers are those of the prosperous and strong, of 
men who hold a responsible position, but despise restraint ; 
there is no hint that the British community had anything to 
fear from without, or that heathen peoples not so well in hand 
as the Picts and the Scots were surging at the gates. 

A like impression of masterful strength, unbridled and un- 
dismayed, is conveyed by the picture painted by Gildas of the 
British rulers of his day. No doubt, the barbarian ravages, 
with their sequel, the heathen occupation of a large part of 
Britain, fill a large place in his narrative, and it is from him that 
modern historians have chiefly derived their conception of the 
conquest as a process of savage extermination. But, lurid as 
is the landscape when he sets himself to depict the devouring 
tide of conquest, it is, it should be remembered, a story of the 
past, of what befell Britain before he was born. That siege of 
Badon Hill, which happened in the year of his birth (and he 
wrote at the age of forty-three), saw the last and decisive 
victory of the British arms ; wars with external foes have now, 
he says, ceased ; the time is one of prosperity and skies are 
serene. 9 Unless we are entirely to discard the evidence sup- 
plied by Gildas, it must be believed that the first stage of the 
English conquest and occupation came to an end about the 
beginning of the sixth century, to be followed by a truce of 
half a century, and that it was not until 550 or thereabouts 
that a second onward movement began, threatening this time 
the Britons of the North and West and reducing them ere 
long within very narrow limits. The rulers whom Gildas ad- 

9 C. 26. I follow Zimmer (Nenn. V. 100) and Mommsen (pref. to Gildas, 8) 
in understanding Gildas to say that the year of Badon Hill and that of his 
own birth was the forty-fourth, reckoning backwards, from that in which he 
wrote. As he wrote before the death of Maelgwn in or about 547, the year 
referred to cannot have been much later than 504, while the fact that Gildas 
lived until about 570 makes it probable it was not much earlier. The entry in 
the chronicle in Harl. MS. 3859 ascribing "bellum badonis " to the year 516 
(Cymr. xi. 154) is of no particular authority, and with it goes the common 
ascription of Gildas's work to the year 560. 


CHAP, dresses are men of assured position, harassed by no external 
difficulties. Law and order are maintained in their territories ; 
criminals are brought before the seat of judgment, thieves 
hunted out and imprisoned, judicial oaths administered. 10 They 
have their troops, with which they wage war against each other, 
their bards who celebrate their praises and clergy who are 
much dependent on their favour. 

Of the princes specially named and singled out for attack 
by Gildas, the most notable is Maglocunus, 11 in whom it is easy 
to recognise the Maelgwn Gwynedd of Welsh tradition. Had 
our knowledge of this vigorous ruler been drawn solely from 
this latter source, he would, without a doubt, have been treated, 
like Arthur, as a purely mythical figure and perhaps would 
have been refined into a solar deity. But in the pages of the 
De Excidio he is unmistakable flesh and blood. Taller in 
stature than most of the chieftains of Britain and therefore well 
deserving his title of " Maelgwn Hir " (the Tall), 12 he overtops 
them also in power. This position he has won by deeds of 
violence ; many a neighbouring prince has been sacrificed to his 
ambition, and his reign began, when he was but a raw youth, 
with the overthrow of his maternal uncle, 13 whose brave troops 
he defeated and whose crown he usurped. His later years have 
been stained by crimes no less heinous, the murder of his wife 
and of a nephew, whose faithless spouse he was thus enabled 
to marry. Yet, heavy as is the catalogue of misdeeds laid to 
his charge, he is not without a certain tincture of nobleness. 
He is a liberal giver, and no common tyrant would, in the 
heyday of his greatness, have laid aside his royal dignity and 
have withdrawn, as Maelgwn did, to the austere seclusion of a 
monastic cell. All Christians must, it is true, deplore his sad 
relapse into a life no less worldly and sin-ridden than before, 
but the very making of the experiment proves him a prince of 
no ordinary mould. 

10 C. 27. 

11 Cc. 33-6. The mediaeval form is always Maelgwn, not Maelgwn. 
Maelgwn is nowhere styled " gwledig," yet it is reasonable to assume that his 
power rested on the continuance of some office of the kind. 

My v. Arch. I. 476(318). 

13 As the name of Maelgwn's mother is unknown, his " avunculus " cannot 
be identified. Some have thought he might be Arthur (Sharon Turner, History 
of the Anglo-Saxons, bk. iii. c. 3 ; Arth. Legend, 8). 


Such was Maelgwn as he appeared to a contemporary who CHAP, 
lashed his vices with no friendly hand, yet did not allow himself, v> 
as he did in attacking others, to sink into the language of con- 
tempt. The details of his career which have been handed 
down by tradition seem to harmonise well with the picture 
drawn by Gildas. He was the son of Cadwallon Lawhir, who 
finally overcame the Goidels of North Wales. 14 With this 
people, therefore, he is not represented as fighting ; on the 
contrary, Gwynedd, in its Latinised form Venedotia, is the 
centre and stronghold of his power, 15 and his designation of 
"island dragon" is naturally explained with reference to his 
secure possession of Anglesey. 16 His especial home seems to 
have been the Creuddyn peninsula, where on the rock of 
Degannwy, the ancient hold of the Decanti, was the "court of 
Rhos " in which, according to the popular saying, he slept to 
awake no more. 17 His name is still preserved in the neighbour- 
hood in that of Bryn Maelgwn, a height not far form De- "* 
gannwy. 18 Tradition agrees with Gildas in explaining 
Maelgwn's advent to power as the result of a competition, but 
describes it as one not of arms, but of constructive skill ; the 
contest came off, we are told, on the sands in the estuary of 
the Dovey, since known as "Traeth Maelgwn" (Maelgwn's 
Strand), and Maelgwn owed his victory to the cunning artificer 
who provided him with a floating chair, since he alone was 
able to hold his ground against the incoming tide, before which 
his rivals had to flee. 19 What the germ of truth may be in 
this story it is hard to tell, but it would seem, at any rate, to 
embody the view that the son of Cadwallon was a masterful 
man and one of many resources. 

14 See the pedigrees in Harl. MS. 3859 (Cymr. ix. 170). 
15 " Mailcunus magnus rex apud Brittones regnabat, id est in regione 
Guenedotae " (Hist. Britt. c. 62). 

16 " Insularis draco " (Gildas, c. 33). This is the view of Zimmer (Nenn. 
V. 101) ; Rhys takes the island to be Great Britain (IV. People, p. 106). 

17 The passages which connect Maelgwn with Degannwy are mostly late 
(Bruts, 234 ; Myv. Arch. II. 359 (547) ; ibid. (597) ; " Hanes Taliesin," as 
printed in Lady Charlotte Guest's edition of the Afabinogion). But the place 
was an " arx " of some consequence as early as the beginning of the ninth 
century (Cymr. ix. 163, 164), so that the connection is a likely one. For the 
original form of the proverb about Maelgwn's long sleep see Ann. C. s.a. 547. 

18 The name is at least as old as Pennant's time (iii. 145). 

19 LL. ii. 48-50 (Peniarth MS. 32, of about 1380, and another). Cf. lofa 
MSS. 73-4. 

VOL. I. 9 


CHAP. Tradition has something to say of the relations of Maelgwn 
both with the Christianity and the bardism of his time. He 
plays a prominent part in several of the lives of the saints, 
which exhibit him as in frequent conflict with the religious of 
his day, but as invariably brought to reason and repentance. 
St. Brynach in Dyfed, St. Cadog in Gwynllwg, St. Cybi in 
Anglesey, St. Padarn in Ceredigion, 20 and St. Tydecho in 
Powys" all had encounters with Maelgwn, which ended in 
their obtaining from him substantial privileges for their monastic 
foundations. St. Cybi received from him the royal "caer" or 
"castellum" in Holyhead Island, which has ever since borne 
the name of Caer Gybi, and had the legend of St. Deiniol, the 
founder of Bangor in Arfon, survived, some similar statement 
would doubtless be found in it explaining how the name of 
Maelgwn Gwynedd came to be connected with the origin of that 
monastery also. 22 All is in keeping with the " largior in dando, 
profusior in peccato " of Gildas, save that in the stories told by 
the writers of the legends, the order of the epithets is reversed 
and the almsgiving is the atonement offered for the crimes. 
Gildas and tradition are, again, in accord in representing 
Maelgwn as surrounded by a troop of sycophant bards. In the 
passage which describes, 23 with relentless invective, the king's 
return from the pure life of a monk to his former dignity and 
his ancient sins, it is said that he no longer hears the sweet 
strains of ecclesiastical melody and the praises of God sung 
by the tuneful voices of His servants, but worthless laudation 
of himself, as the rascally, lying quacks who serve him spit out 
their bacchanalian ravings. Though the fact is not expressly 
stated, the nature of the antithesis makes it very likely that 
the praises of Maelgwn were, like those of the Most High, set 
to music, and that the men so scornfully treated by Gildas were 
the bards of the royal court. Tradition has a good deal to 
say of these bards, who to the number of twenty-four surrounded 
Maelgwn on state occasions at Degannwy, but were confounded 
by the superior skill, in magic no less than in the more legiti- 

Cambro-Br. SS. 10-12, 50-6, 186-7, 191-2. 21 Camb. Reg. ii. 376. 

22 That Maelgwn founded Bangor appears to rest at present on the author- 
ity of the C.C.C.C. MS. of John Ross (Joannis Rossi, Historia Regum Anglia, 
Oxoniae, 1716, p. 65). Ross had visited Anglesey and Carnarvonshire in search 
of material (ibid. p. 54). 

23 C. 34 (end). 


mate aspects of the poet's art, of the great Taliesin, who de- CHAP, 
livered his master Elphin from the prison of thirteen locks in 
which Maelgwn had immured him. 24 Whatever view may be 
taken as to the real nature of the conflict so described, it is 
clear that Maelgwn was traditionally regarded as a patron of 
native poesy and music. 

This busy, restless career was cut short by the plague which 
devastated Europe in the middle of the sixth century. 25 Accord- 
ing to the chronicle preserved in Harleian MS. 3859, the year 
was 547, 26 but better authority is needed ere so precise a date 
can be without question accepted ; it is enough to say that 
probably it is not far wrong. The ancient proverb already 
referred to, viz., " Hir hun Faelgwn yn Llys Rhos" (the long 
sleep of Maelgwn in the court of Rhos"), suggests that he died 
at Degannwy, but a more picturesque account says that, when 
the Yellow Plague was roaming through the land in the guise 
of a loathly monster, he took refuge from it in a church not 
far from his court. "And Maelgwn Gwynedd beheld the 
Yellow Plague through the keyhole in the church door and 
forthwith died." 27 As to what it looked like, authorities are 
not agreed ; in the Story of Taliesin, it is described as a 
strange beast, with yellow eyes, teeth and hair; but in the 
life of St. Teilo contained in the Liber Landavensis it is 
a column of vapour rising from earth to heaven and sweeping 
along the ground as a shower of rain sweeps along the low- 
lying valleys. Terror gave it many shapes, and it was none the 
less mysterious that it had laid low the mighty prince who 
feared no face of man. 

Four other British princes are attacked by name in 
the De Excidio Constantine 30 is called " the tyrannical 

34 " Hanes Taliesin ; " lolo MSS. 73. 

25 Bury, History of the Later Roman Empire, i. p. 401. Cymr. ix. 155. 

87 lolo MSS. 78, where the church is said to be Eglwys Rhos. Other ac- 
counts (Britts, 234 ; Myv. Arch. II. 359 (547) ; ibid. (597)) merely say it was near 
Degannwy. The "eglwys y brodyr " of Jesus Coll. MS. 19 (Myv. Arch. II. 359 
(471)) is perhaps explained by the appropriation of Eglwys Rhos to the Cistercian 
abbey of Conway (after 1284, Maenan). 

K Lib. Land. 107. 

"For various views as to the five kings of Gildas see Usher, Antiquitates 
(1639), 536-47 ; Celt. Br. (2), pp. 122-5 ! Nenn. V. 306-7 ; Gildas, ed. Hugh 
Williams (1899), pp. 69-83. The four beasts (leo, pardus, ursus, draco) appear to 
be taken from the Apocalypse (xiii. 2), where the dragon is supreme. 

** C. 28. 





whelp of the unclean lioness of Damnonia," so that it is 
probable he bore rule in the Devonian peninsula. " Aurelius 
Caninus," 31 left solitary by the death of all his kin, is not 
connected with any particular British region, but it is very 
likely that we have in him the degenerate offspring, elsewhere 
referred to by Gildas, 32 of the great Ambrosius Aurelius or 
Emrys Wledig, and in this case his kingdom must be looked 
for in the neighbourhood of the English settlements rather 
than in Wales. The two remaining kings may be confidently 
placed on Welsh soil. " Vortiporius" 33 is described as the 
tyrant of the Demetae, or men of Dyfed, the bad son of a good 
father, a man who has grown grey in the devil's service and 
who shows no sign of repentance as he draws near his end. 
The pedigree of the royal line of Dyfed as recorded in the 
eighth century included a " Guortepir," who was the son of 
" Aircol," or Agricola, the Long-handed, 34 and it is remarkable 
that, while in the Liber Landavensis several gifts of land to 
St. Teilo are associated with Aircol Lawhir, 35 none such are 
credited to his son. Even more interesting is the fact that in 
1895 the tombstone of the object of Gildas's invective was dis- 
covered in the heart of his kingdom of Dyfed. 36 It is a rude 
standing-stone or " maenhir," with an inscription in Latin 
capitals running across horizontally and another in ogam 
characters along one of the edges. In the former, which is 
headed by a wheel cross betokening that the dead man was a 
Christian, the legend runs " Memoria Voteporigis protictoris " ; 
the latter has nothing but the name, in its Goidelic form, 
" Votecorigas ". The stone has latterly stood in the grounds 
of Gwarmacwydd House, but it originally came from the church- 
yard of Castell Dwyran, which may have been a chapel at- 
tached to a manor of the rulers of Dyfed. It will be observed 

31 C. 30. "Canine" is the reading of the Cottonian MS. Yet the name 
may really have been " Cuna(g)nus " (Cynan), as suggested by the Avranches 
MS. (see p. 43 of Mommsen's Gildas), and the other form have been coined by 
the author as a grim joke at the expense of Aurelius. Cf. his treatment of the 
name of " Cuneglasus ". 

32 C. 25 (end). 33 C. 31. 

34 Harl. MS. 3859 (Cymr. ix. 171) ; Jesus Coll. MS. 20 (Cymr. viii. 86) ; see 
also the Irish forms in Cymr. xiv. 112. 

3& Lib. Land. 125-30. 

36 For references seep. 115. There is a good illustration of the stone in 
Social England, vol. i. (1901), p. 173. 


that the stone, in addition to giving this prince the title of CHAP. 
" protector," bestowed by the Romans in the declining years of 
their Empire upon notable barbarian leaders and no doubt borne 
by Voteporix hereditarily, is at variance with Gildas as to the 
spelling of the name ; in this matter its authority is more 
weighty than that of the man of letters, who may easily have 
been misled by the idea of an analogy with such forms as 
Vortigernus and the like. 

" Cuneglasus " 37 was also, no doubt, one of the rulers of 
Wales, for the pedigrees 38 include a certain "Cinglas," in 
modern Welsh Cynlas, who was a son of Owain the White- 
toothed, son of Einion, son of Cunedda, and therefore belonged 
to the same generation as Maelgwn. There is nowhere any 
indication of the seat of his power, but the probabilities are in 
favour of some part of North Wales lying between Powys and 
the region directly ruled over by Maelgwn. The ancient vill of 
Cynlas 39 in Penllyn may commemorate his name ; his brothers 
Engan and Seiriol founded churches in North-west Wales, 40 
and the Caradog King of Gwynedd who died in 798 appears to 
have been his descendant. 41 For Cynlas, Gildas has a special 
hatred and contempt, explained, it may be, by the fact that 
this king alone of the five is accused of active hostility to the 
Church, being not only a contemner of God, but also an 
oppressor of His clergy. 42 His Celtic name is held up to scorn 
and parodied in Latin as "Grey Butcher". He has trans- 
gressed, not only after the fashion of his fellow-kings, by 
warring against his Christian neighbours and by violating the 
sanctity of the marriage tie, but also by direct attacks upon the 
" saints " (*>., monks), whose sighs and groans will one day 
recoil upon him to his undoing. 

87 C. 32. 

38 Harl. MS. 3859 (Cymr. ix. 172). 

39 Until lately a township of the parish of Llandderfel, Merionethshire, now 
in that of Llanfor in the same county. 

40 At Llanengan in Lleyn and Penmon in Anglesey (Myv. II. 23 (415)). 

41 C/. the entry under the year 798 in Harl. MS. 3859 with pedigree No. iii. 
in the same MS. (Cymr. ix. 163, 172). 

4a " Dei contemptor sortisque eius depressor " (c. 32). 

1 34 tfIS TOR Y OF WALES. 



(The authorities mentioned in connection with Chapter IV. iii. continue to 
be important for the sixth century, and to them must be added Mommsen's edi- 
tion of Gildas, my obligations to which are very great. I have also used with 
profit Williams's edition of the same author. Other sources, such as Rees's 
Welsh Saints, are mentioned in the notes.) 

It is now time to turn to the one figure of sixth-century 
Wales who, revealed by his writings, stands out clearly among 
the somewhat elusive shapes of that romantic age. Gildas did 
not (save as an author) greatly impress his British contempor- 
aries, if one may judge from the neglect of his name and 
achievements by the voice of Welsh tradition, 43 but he has the 
advantage of men who in his own day far outshone him in 
renown in that his " little homily," 44 as he modestly calls it, is 
the only literary effort of his age and country which has come 
down to us, and thus while Arthur, Dewi, Illtud, Dyfrig and> to 
a less degree, Maelgwn are but shadows, Gildas is a man of 
whom it may be said that we have seen his face and heard him 

Little, indeed, is known of him outside his own work. Two 
lives have been handed down which profess to tell his history, 45 
but so widely do they differ from each other in the tale they 
unfold that many of the older school of critics imagined there 
must have been two men of the name, viz., the author of the 
De Excidio, born in the early part of the sixth century, and 
a Gildas of the North, who belonged to an older generation and 
had dealings with the great Arthur. 46 It is now recognised 
that "Gildas Albanius," as portrayed in the life ascribed to 
Caradog of Llancarfan, is a figment of the monks of Glaston- 
bury, who desired to connect so famous a Briton with the early 
history of their house and had a life written to order for the 

43 No Welsh church is dedicated to Gildas. His name occurs several times 
in the Mabinogion (107, 160, 258), but no legend is told of him. In the older 
" Bonedd y Saint" he appears only as the father of saints (Myv. Arch. II. 25 

44 " Admonitiuncula " (c. i). 

45 For an account of the lives, MSS. and editions of Gildas see note ap- 
pended to this chapter. 

46 This view was adopted from Bale and Pits by Usher (Antiquitates, 1639, 
441-2), and subsequently by Stevenson (Pref. to ed. of Gildas, vi.). Schoell (de 
ecclesiasticae Britonum Scotorumque historiae fontibus, Berlin, 1851) showed 
how baseless it was. 


purpose. 47 Caradog's work (if it be really his) 48 contains some CHAP, 
interesting Arthurian matter, put together before all stories 
about the British king were made to match the portrait painted 
by Geoffrey of Monmouth, but of the true history of Gildas 
it hardly preserves a trace. The other life, written by a monk 
of Rhuis in Southern Brittany, is far less open to suspicion ; its 
historical setting is not, for the most part, inconsistent with 
what is known of Gildas, and it is from some such foundation 
as Rhuis, traditionally connected with his name, that genuine 
records of his life and work might be expected to make their 
appearance. Still, its evidence, as that of a late source, can 
only be received after careful scrutiny and by way of supple- 
ment to the knowledge afforded by the De Excidio itself. 

Gildas 49 was a native of Britain. Though he nowhere ex- 
pressly calls himself a Briton, 60 the whole spirit of his treatise 
is that of one who was addressing his fellow-countrymen. In 
his preface he apologises for the uncompromising severity of 
his tone ; let no one think him a misanthropist or a self- 
righteous despiser of his kind ; he has a true pity for the woes 
and misfortunes of the country and would rejoice to see them 
remedied. 51 He descants with genuine pleasure upon the 
natural beauties of Britain, where, to quote familiar lines, 

Every prospect pleases 
And only man is vile. 82 

He claims a share in the martyrs of Britain, " lamps of exceed- 
ing brightness set alight for us, lest Britain should be involved 
in the thick darkness of pitchy night". 53 According to the 
Rhuis life, he came from the fertile land of " Arecluta," which 
lay (as its name implies) along the river Clyde ; this is a 
statement there is no ground for disputing, and " Arecluta," in 

47 Williams, Gildas, pp. 390-1. 

48 Some doubt is raised by expressions which suggest that the author was 
not a Welshman, e.g., "balnea . . . quod diligebatur asua gente maxime" (c. 3) 
and '' Walenses indigenae " (c. 8). 

49 The name is not of Latin origin and is not otherwise known among the 
Brythonic peoples. It appears in the forms Gilda (Mab.), Gildus (Bede), Gillas 
(Tighernach) and Giltas (Columbanus). Possibly it is Goidelic; cf. Mab. no, 
where mention is made of a " Gilla goes hyd " (Gilla with the shanks of a hart) 
who was the best jumper in Ireland. 

"According to Mommsen (pp. q-io), Gildas's "patria" (cc. 1,2, 18, etc.) 
means no more than " provincia ". " Mea patria " is certainly not found. 
81 C. i. C. 2. C. 10. 


CHAP, later Welsh Arglud, has been very reasonably identified with 
the rich alluvial tract to the south of the Clyde between 
Greenock and Glasgow. 54 Thus Gildas was a North Briton, 
but when it is further stated that his father " Caunus " was 
king of this region, one is inclined to pause and remember how 
many of the saints of this age are alleged to have been of royal 
blood. Addicted as Gildas was to rhetorical depreciation of 
himself and his merits, he would scarcely, in addressing 
Maelgwn, have conceded that he was " of the meanest rank," 55 
had his birth been as good as that of the redoubtable lord of 
Gwynedd himself. Nor does his attitude towards the British 
chieftains who built up their power on the ruins of the Roman 
state, men whom he calls tyrants and kings not anointed of 
God, 56 favour the suggestion that he drew his own origin from 
this class. 

He was born, on his own testimony, in the year of the 
battle of Badon Hill, i.e., in or about A.D. 5OO. 57 Christianity 
had by this time firmly rooted itself in the valley of the Clyde, 
for Ceredig Wledig, who had ruled over this region some fifty 
years earlier, was a professed, if a somewhat rudimentary Chris- 
tian. Gildas, it is to be noted, knows nothing of any survival 
of paganism among those whom he recognises as his fellow- 
countrymen : he has seen in the ruins of the deserted temples 
monstrous effigies of gods worshipped aforetime ; he has heard 
that mountains, hills and rivers were once invoked as divine ; 
but all this is ancient history, which he will not, he says, rake 
up against the men of Britain. 58 He was brought up, it is clear, 
in a purely Christian atmosphere. 

No one will doubt that he received an excellent education. 
Chafe as we may at the endless intricacies and involutions of 
his style, insipid as we may find his sugared rhetoric, he was 
beyond question a master of the Latin language, and the gulf 

54 IV. Anc. Bks. i. pp. 173-4; Cymr. xi. 75. Ar^cluta, with the connecting 
vowel (cf. " Aremoricis gentibus" in Orosius, Adv. Pag. vi. n),must come from 
an almost contemporary source. The district is still known for its heavy crops 
of wheat. 

85 " Licet vilissimae qualitatis simus " (c. 36). 

56 < Ungebantur reges non per deum, sed qui ceteris crudeliores exstar- 
ent " (c. 21). 

57 See p. 125 and note 9. 

58 C. 4. The argument at the end of c. 38 implies that the men attacked 
by Gildas were not open to the charge of idolatry. 


which separates him, as a writer of Latin, from later British CHAP, 
historians such as "Nennius" is immense. He commands a 
wide vocabulary, is entirely at home in Latin syntax, and is 
familiar with the Latin poets, notably Vergil. 59 Of the Scrip- 
tures he had made an exhaustive study, and every reader of the 
De Excidio remembers how large a part of that work is 
made up of quotations, carefully marshalled, from holy writ. 
The school in which Gildas acquired this knowledge was, ac- 
cording to the Rhuis life, that of " Hildutus," the St. Illtud of 
Welsh tradition under whose care, it is said, were many lads 
of good family, receiving a training not only in divine but also 
in secular learning. 60 Something will be said in a later section 
of the work and influence of Illtud ; here it is enough to note 
that Gildas received an education which linked him closely with 
the past, and that, while in effect one of the pioneers of a new 
era, he drew his inspiration from the waning civilisation of 
Rome and by temper and disposition was always of the old 
order and not of the new. The literary culture of Rome, ere 
it faded from Britain, cast over the young Gildas a spell which 
bound him firmly to the end. 

It was, no doubt, in the school of Illtud that he came under 
another spell, that of monasticism. He became not only a 
cleric, but also a monk, 61 full of pride in his order and of con- 
fidence in it as the sole salvation of the world . The monks 
are in his pages the true " saints " of God ; 62 their convents are 
cool caverns and safe hiding-places where the good may find 
shelter from the attacks of the Evil One. As yet, the monastic 
element in the British Church is not in the ascendant ; the men 
whom Gildas singles out for praise are " very few," 63 a minority 
in the Church, whom he leaves out of account, save for one or 
two apologetic allusions, 64 in his comprehensive indictment of 
the British nation. The Church is not governed by monks ; 
one reference there is to an abbot, 65 but the ecclesiastics who fill 

89 Pref. to Mommsen's ed., p. 6. 90 Ed. Mommsen, p. 92. 

81 " Clericorum in nostro quoque ordine " (c. 65) is so interpreted by 
Williams (p. 160). 

88 See cc. 28 (sanctorum choris), 34 (sanctorum speluncas). 

""Faucis et valde paucis " (c. 26). C/. " paucissimos bonos pastores " 
of c. no. 

94 See c. 65 (ab his veniam impertiri, etc.). 

95 " Sub sancti abbatis amphibalo " (c. 28). 


CHAP, the scene are bishops and priests, whose power arises out of 
V ' their office and has nothing monastic about it. Thus the party 
to which Gildas attached himself, though it threw its subtle 
charm for a time over the great Maelgwn, was not the dominant 
one in the island, and it was from no vantage-ground, save that 
of sincerity and zeal, that he hurled at his contemporaries the 
ringing challenge of the De Excidio. 

The work had been long in preparation when at the age of 
forty-three he gave it to the world. Ten years had gone by 
from the time when its design had first taken shape in his mind 
and still he shrank from its publication, until at last he was 
overcome by the entreaties of his brother-monks and a sense of 
what was due from him as a witness for God's truth. He had 
not the official standing of a bishop 66 or the authority of a 
teacher of wide renown ; 67 Britain, he pleads, has its recognised 
watchmen and rulers, and it scarce becomes the mere hand and 
foot, the humble instruments, to usurp the office of the warning 
lips and the vigilant eye. But the need of a clear testimony 
was too great to allow of any further postponement ; that very 
year 68 deeds had been done which must be exposed in their 
true villainy ; and thus, as Gildas puts it, the long-standing debt 
was finally paid. 69 

No hint is given in the De Excidio as to the place of its 
composition. The view that Gildas was at the time in Brittany 
and launched it across the Channel 70 rests on no better 
authority than the Rhuis life, 71 where it was almost inevitable 
that such a statement should be found, and it is rendered very 
unlikely by the use of the word " transmarine " to describe the 
continental sources used by the author in default of native 
authorities. 72 If the work, as seems most probable, was written 
in Britain, it surely made its appearance in a British state of 
which the ruler had not the honour to figure in its pages, for 
it can scarcely be believed that, as Zimmer has suggested, 73 

66 The Annals of Inisf alien seem to style Gildas "episcopus," but in the 
Stowe Missal he appears, with St. Columba, among the " sacerdotes " and not 
the " episcopi " (Warren, Liturgy and Ritual of the Celtic Church, p. 240), and 
this is doubtless the sounder tradition. 

67 He says he is not " conspicuo ac summo doctori " (c. i). 

68 "Hoc anno " (c. 28). B9 C. i. 

Mon. Hist. Br., pref., 60, note 2. 71 Ed. Mommsen, p. 97. 

7211 Transmarina relatione " (c. 4, end). Nenn. V. 308. 


it was composed by Gildas in some monastery in the heart of CHAP, 
that south-western corner of the island of which the great men 
are all held up to public reprobation. It has been pointed out 
that, so far as can be seen, not one of the five kings pilloried by 
Gildas belongs to the British country which at that time 
stretched from Chester to Dumbarton, a country on which the 
English invader had as yet made little impression. 74 May not 
his work have been written here, under the protection of a king 
who, while too bad for a formal blessing, was too good for 
cursing in the heroic vein which was thought necessary to meet 
the case of the hardened sinners of the South ? 

It was no part of the purpose of Gildas to write a history 
of Britain. His aim is that of the preacher and reformer, and 
the whole of his treatise, which he occasionally styles a 
" letter," 76 is meant, in its narrative no less than its admonitory 
portions, to subserve the one moral end. If a good deal of 
history is introduced, this is because history, as handled by 
Gildas, has its lessons to teach. He is a disciple of Paulus 
Orosius, the Spanish presbyter who rather more than a century 
before had set the fashion of writing history with a homiletic 
purpose, and, like Orosius, he has for his theme " the lusts 
and the retributions of sinful men, the conflicts of the world 
and the judgments of God ", 76 He would have it understood 
that the miseries of Britain in bygone days, miseries which he 
paints in the darkest colours, were the direct result of the 
wickedness and perversity of the natives of the island, and that, 
though a season of prosperity and peace has now succeeded, 
continuance in evil-doing will bring back the old calamities. 
Thus Gildas is of the order of prophets and not of historians, 
and what he says must be viewed in relation to the ethical 
purpose which constantly swayed him. 

In one respect he was ill fitted to do justice to the new 
world which was being formed around him. It was a world in 
which the barbarian, the Celtic element, was daily growing more 
powerful, and Gildas was Roman to the finger-tips. His 
favourite name for his fellow-countrymen is " cives," citizens ; 77 

74 Nenn. V. 308. 

78 See cc. i (in hac epistola), 93 (huic epistolae). In c. 37 the Cottonian 
MS. has " flebilis haec querulaque malorum huius aevi historia". 
76 Adv. Pag. vii. ^3. 77 See cc. 4, 19, 26, 28, 32. 


OHAP. with them he contrasts the nations round about, the Saxon 


foe, the fierce hordes of Picts and Scots, shaggy-haired and of 
indecent nakedness, 78 barbarians, in short, with whom he re- 
cognises no bond of union. The Latin language is his tongue ; 79 
for Celtic names he has nothing but contempt. Throughout 
the story of Rome's dealings with the island a story often 
grotesquely remote from the real facts the men of Rome are 
the powerful lords and protectors, the men of Britain cowardly 
slaves who rebel when their masters' backs are turned, but 
who cannot stand in their own strength. No one is praised by 
name after the departure of the Romans save Ambrosius 
Aurelianus (Emrys Wledig), sole survivor in the catastrophes of 
the fifth century of the Roman and imperial race. 80 

Gildas cannot have had many who sympathised with him 
in this attitude and, as an upholder of legitimism against the 
ever-thickening growth of British chieftaincies, must have been 
indeed a voice crying in the wilderness. Was he isolated, too, 
in his onslaught upon the lives of the men who ruled at this 
time in Church and State ? That there were men of influence 
who did not share his passionate conviction that Britain was 
rushing headlong to ruin seems open to no doubt. He blames 
such for their indifference to the immorality of the age, though 
he has no charge to make against them personally. 81 There 
are false prophets abroad, he intimates, doctors filled with a 
spirit of contrariness, whose words are smoother than butter 
and who cry " Peace, peace ! " when there is no peace. 82 They 
were opportunists and friends of compromise, who condoned 
the vices of monarchs and gladly accepted their gifts to the 
Church gifts which, in the opinion of Gildas, were no substi- 
tute for repentance and reform. 83 The British Church was by 
this time a well-developed organisation. Its bishops and priests 
were numerous ; their offices were valuable and worth taking 
much trouble to secure. 84 In such a society there would be 
many, even among the reputable and diligent members, who 
would place loyalty to the institution before fidelity to a lofty 

c. 19. 

79 " Tribus, ut lingua eius exprimitur, cyulis, nostra longis navibus " (c. 
23). The Avranches MS. reads ' Latina," which is, no doubt, the right inter- 
pretation, for, if Gildas had meant to use the Brythonic loan-word " longa " 
(Welsh " Hong "), he would not have added " navibus ". 

80 C. 25. 81 Cc. 69, no. 82 C. 40. 83 C. 42. 84 Cc. 66, 67. 


and perhaps unattainable standard ; Gildas the iconoclast CHAP, 
would be in their eyes a dangerous fanatic and they would do 
their utmost to counteract his influence. 

But it is not at all likely that they succeeded. The evils 
against which he so strenuously inveighed were real evils, the 
natural result of a long period of wealth and abundance in the 
history of a little community cut off from the rest of Europe. 
As to the crimes so definitely laid to the charge of the five 
princes there could, at any rate, be no doubt, and there is every 
reason to suppose that the earnest and fiery eloquence of the 
De Excidio did arouse an echo in many hearts. This may 
be inferred from the esteem in which " Gildas the Wise " 85 was 
held by later generations and from the remarkable progress of 
monasticism in the second half of the sixth century. He was, 
in fact, justified by the event. He foretold calamities as the 
inevitable issue of the licence and presumption of his age, and 
two calamities of the first magnitude ere long befell the British 
race the yellow plague, spreading terror and ruin as it went, 
and the renewal of barbarian aggression, putting an end to the 
long truce between Briton and Saxon and reopening the con- 
flict for the mastery of the island. Though the evidence for 
the progress of this conflict comes chiefly from traditional 
sources and is, therefore, a somewhat uncertain guide, it seems 
possible to trace, in the Saxon Chronicle, for instance, the rise 
of a forward movement on the part of the English about the 
year 55O 86 which would afford a vivid contrast to the " present 
security" described by Gildas, and, coupled with the devasta- 
tions of a great pestilence, would not fail to turn men's minds 
to thoughts of repentance and atonement. In the general 
humiliation Gildas would be recognised as a true seer, and the 
party to which he was attached, the monastic and ascetic one, 
would gain the upper hand in the British Church. 

If it may be assumed that the De Excidio was written 
not long after 540, Gildas lived for a quarter of a century after- 
wards to wield great influence in the Celtic Church. The date 
of his death cannot be precisely given, but it was not far from 

"The title "Sapiens" is given to Gildas by Ann. C. MS. C., the Cam- 
bridge MS. Ff. i. 27, and the Book of Leinster (Stokes, Tripartite Life of St. 
Patrick, p. 514). 

89 Mak. Eng. pp. 91-2. 


CHAP, the year 57o. 87 He spent this closing period of his life, it 
would seem, in organising and directing the powerful monastic 
movement which had now laid hold upon the Celtic commun- 
ities. In Wales and Cornwall, it is true, there is little evidence 
of influence directly exercised by him ; the offence given by his 
attacks upon the monarchs of those regions was not so easily 
forgotten. But elsewhere, his authority was high. In the 
North, his account of the English settlement was used in the 
next century by the author of the Saxon Genealogies and 
in -the eighth by the Venerable Bede. Both lives, discordant 
though they are in most other respects, agree that he under- 
took a mission tour to Ireland, which the Rhuis life connects 
with the reign of Ainmire mac Setna, king of Erin about 
568. 88 Here monasticism had already found a congenial soil ; 
Clonmacnois had been founded by St. Ciaran, Clonard by the 
elder Finnian, and Clonfert by the elder Brendan. But the 
advent of so notable a champion of the " perfect way " must 
have been a source of much encouragement to those of the 
same school of thought, and it is this, no doubt, which lies be- 
hind the somewhat mysterious statement made in an eighth- 
century Irish tract that the second order of Irish " saints," be- 
longing to the latter half of the sixth century, received their 
ritual from Bishop David and Gildas and a third Briton. 89 
That Gildas was held in high esteem in his lifetime among the 
monks of Ireland is made certain by the fact, recorded by 
Columbanus before the sixth century was out, that " Ven- 
nianus," probably Finnian of Moville, consulted the famous 
author as to the treatment of those inmates of monasteries who 
left their abbots without permission to take up the life of the 
anchorite. 90 Gildas, it is said, replied in a very charming letter, 
and some of the fragments bearing his name which have been 
preserved in Irish collections refer, in fact, to this very subject. 
These fragments show, not only how the words of the> British re- 
former were valued, but also that, like many another reformer, 
he spent the evening of his days in checking and keeping with- 

87 This is the date implied in the entry in Harl. MS. 3859 (Cymr. ix. 155). 
Cf. the Annals of Tighernach (Rev. Celt. xvii. (1896), p. 149), Ann. Ult. s.a. 569 
(repeated s.a. 576), Ann. Inisf. s.a. 570. 

88 Ed. Mommsen, p. 94. 89 H. and St. ii. pt. 2, 293. 
90 Gildas, ed. Mommsen, p. 21 ; ed. Williams, pp. 256-7, 415. 


in bounds the tendency and spirit which he had done so much CHAP, 
to rouse by the eloquent outpourings of his youth. He has to 
deal, no longer with sloth and indifference, but with misplaced 
energy and misguided zeal. 91 

There need be no hesitation about accepting the statement 
of the Rhuis life that he died in Brittany. It is confirmed by 
the existence of monasteries at Rhuis and elsewhere claiming 
him as their founder and patron, 92 by the occurrence of his name 
in an early Breton litany, 93 and by the Breton origin of one of 
the two old MSS. of the De Excidio^ Thus Gildas ended 
his days in a species of exile, dwelling among the descendants 
of those harassed Britons whom he had described as seeking 
homes across the sea with loud lamentations, and chanting 
beneath the bellying sails their piteous refrain : " Thou hast 
given us as sheep appointed for meat and hast scattered us 
among the heathen ". With some reason might he have an- 
ticipated another champion of righteousness in the well-known 
words : " I have loved the law of God and hated iniquity ; there- 
fore I die in exile ". 


Gildas does not mention the name of a single ecclesiastic of 
his own time, and even leaves in obscurity the " very few " whom 
alone he regards as true sons of the Church. It is necessary, 
therefore, to turn to other sources in order to learn who were in 
this century the leading figures in the British Church, and this 
later evidence, dealing as it does with men who have been for 
ages the theme of religious legend, has to be used with great 

In the early part of the century the figure of Illtud 95 seems 

91 For the fragments of Gildas see H. and St. i. 108-12; Mommsen, pp. 
12, 86-8; Williams, pp. 255-71. 

93 A cape at the mouth of the Loire is known by his name. 

93 H. and St. ii. pt. i, 82. 

94 The Avranches MS. formerly belonged to the abbey of Mont St. Michel 
(Mommsen, p. 14). 

95 The life of Illtud printed in Cambro-Br. SS. 158-82 (from Cott. MS. Vesp. 
A. xiv.), though put together about 1 100, at Llanilltud Fawr itself, is not consistent 
with the old authorities, making the saint die, for instance, at Dol in Brittany. 
Possibly, as Prof. Hugh Williams has suggested (Cymr. Trans. 1893-4, IIO ) 
two men of the same name have been confused, the one an Armorican, the 
other a denizen of Greater Britain. 


CHAP, to stand out with some distinctness. The earliest extant life of 
a British saint, that of Samson of Dol, which belongs in all 
probability to the seventh century, 96 represents " Eltutus " 97 as 
the head of a great monastery, which was also a school of the 
highest reputation, for the abbot was "of all the Britons best 
skilled in Holy Scripture, both the Old Testament and the New, 
as well as in every kind of learning, such as geometry, rhetoric, 
grammar, arithmetic and the knowledge of all arts ; in divination, 
too, he was well proven and he had foreknowledge of the future ". 
To this school young Samson was sent that he might be trained 
for a position in the Church ; according to later accounts, he 
had as fellow-pupils Gildas, Paul Aurelian, and even Dewi, 
though the evidence for the last of these three names is not con- 
vincing. 98 The situation of Illtud's monastery is not indicated 
(though the writer of Samson's life had been there in quest of 
material), 99 but it seems certain it was in South Wales, and no 
site has better claims than Llanilltud Fawr, the Lantwit Major 
of to-day. 100 Wherever it stood, the school seems to have wielded 

96 Prnted by Mabillon in Acta sanctorum ordinis S. Benedicts, i. 165 
(Paris, 1668) and by Sollerius (to whose text my references are made) in the great 
Acta Sanctorum, July (zS), vi. 573-91 (Venice. 1749). The internal evidence 
points unmistakably to a date in the century following Samson's death (Cymr. 
xi. 127 ; / iverpool W. Nat. Trans. 1896-7, zoo), which is also suggested by such 
a form as " Tigerinoma'e," occurring in the dedication, pp. 5730, 5876. All 
the other lives are, as Sollerius says, derived from this, and I cannot except that 
printed by Plaine, from Andeg. MS. 719, in Analecta Bollandiana, vi. (1887), 77- 
150, which appears to me in many passages to introduce confusion into the 
Bollandist text. The reference to Germanus (d. 448) ns Illtud's teacher (p. 
5756) I regard as an early interpolation ; it raises insurmountable difficulties of 
chrono ogy. 

97 P. 5756. " Eltutus " yields the Elltud of Llanelltud, Merionethshire. The 
/lltud of the South seems to have been a Goidelic form (Urk. Spr. 41). 

98 Life of Paul Aurelian (founder of S. Pol de L6on in Brittany), as printed 
by Plaine in Analecta Bollandiana, i. (1882), 209-58 and Cuissard in Rev. Celt. 
v. pp. 417-58 ; Rhuis life ot Gildas. The former bears date 884. For Dewi, see 

99 " In cujus magnifico monasterio ego fui " (5756). 

'00 The later writers have confused two places which the author of the life 
of Samson keep^ quite distinct, viz., the " monasterium " of Illtud and the " in- 
sula" (also a monastery; cf. Cymr. Trans. 1893-4, JI 3)> founded " non longe ab 
hoc monasterio " by the priest-monk Piro (5786). The latter may well have been 
Caldy Island (the " Enis Pir " of Gir. Camb. vi. 92 (Itin. i. 12)), where remains 
have been found of an early settlement of the kind (Lap. W. 106-8 ; Arch. Camb. 
V. xiii. (18961, 98-103); the former might be Llanilltud Fawr, if "non longe" 
were taken loosely as a traveller's casual estimate. In the life of Paul, Illtud's 
school is represented as being at the place " quern nunc Iltuti monasterium 


great influence upon British society. It trained not only Gildas, CHAP. 
but, as has been indicated above, other young men of energy and 
ability who took up the monastic idea ; nay, even Maelgwn 
Gwynedd himself, according to a likely interpretation of a 
passage in the De Excidio was one of Illtud's pupils, a 
fact which would help to explain his brief divergence into the 
life of the cloister. This was, in fact, the permanent part of 
the great abbot's work ; he gave new life and vigour to mon- 
asticism, but he founded no British school of learning. Tales of 
his austere seclusion at Oystermouth 102 and at Llanamwlch 103 
were reverently heard by later generations, but his fame as a 
scholar almost entirely perished. He spent his whole life in 
Britain and probably died about 54O. 104 

Of the men of the next generation, the contemporaries of 
Gildas, none has so well attested a history as Samson. 105 He 
was not of royal birth, though later ages connected him with 
Uthr Bendragon and Emyr Llydaw, 106 but belonged to the 
class which stood next in order, the royal courtiers and 
servants whose pride and boast it was to be the foster fathers 
of kings. 107 His father, Ammon, was a man of Dyfed 
(Demetiana patria), his mother, Anna, came from another 
South Welsh district, viz., Gwent (de Ventia). It was not at 
first intended to make him a cleric, for his father held the 
order in no little contempt, 108 but the mother's influence was 
used on the side of the Church and at the age of five Samson 
entered the school of Illtud. Here he made great progress in 

dicunt" tod also on an island " Pyrus nomine Demetarum patriae in finibus sita ". 
The Rhuis life of Gildas does not mention Piro, but he says that Illtud and his 
flock dwelt in a barren island, " quae insula usque in hodiernum diem Lanna 
Hilduti vocitatur " (ed. Mommsen, p. 93). 

101 " Habueris praeceptorem paene totius Britanniae magistrum elegantem " 
(c. 36.) See Cymr. Trans. 1893-4, IO 9- 

loa Nennius, c. 71 (Mirabilia). For the identification with Oystermouth 
see Arch. Camb. IV. xi. (1880), 155. 

103 Gir. Camb. vi. 28 (Itin. i. 2) ; Breconsh. (2), 452. 

104 For the Illtud dedications, which are chiefly in Glamorganshire, and for 
modern accretions to his legend, see Welsh SS. pp. 178-81. 

108 For the Bollandist and other lives see note 96. 

108 lolo MSS. 107, in, 152. The "annun du " (= Antonius the Black) of 
Myv. Arch. II. 24 (415) was probably a different person from Samson's father. 

107 . Altrices regum " (5740). For the custom see LL. i. 788 ; Gir. Camb. 
vi. 2H, 225 (Descr. ii. 4, 9) ; Bruts, 279. 

IDS 4i u t p Ote qui semper minister terreni regis fuisset " (5750). 

VOL. I. 10 


CHAP, learning, was ordained deacon and afterwards priest by Bishop 
Dubricius, and seemed to be marked out by his ability and 
sanctity as Illtud's successor. But Illtud had a nephew, 
advanced to the priesthood, who looked upon the abbacy as 
family property and his lawful inheritance, and with the aid of 
his brother, who was at the time cook, sought to rid himself of 
his rival by means of poison. The attempt failed, but it con- 
vinced Samson that the monastery could no longer be his 
home. He therefore transferred himself to the abbey of Piro, 
where by the favour of Dubricius he became, first, cook or 
steward (pistor) and then, on the death of Piro, abbot During 
his tenure of the latter office he visited Ireland, having been 
persuaded to undertake the journey by certain Irish scholars of 
distinction who were passing through the country on their way 
back from a pilgrimage to Rome. 109 A true son of his age, 
Samson found his hunger for the things of the spirit only 
partially satisfied by life in a monastic community ; after ruling 
the house for a year and a half, he resolved to quit it for the 
solitude of a hermitage. Taking with him three companions, 
he finds a spot to his mind on the banks of the Severn ; 
the three live in a deserted fort, where they build a church to 
which Samson resorts on Sundays, but he himself withdraws to 
a cave in the trackless forest and his hiding-place is long un- 
known. He is, however, discovered and against his will made 
abbot of one of the famous houses of Britain. The day comes 
round when the bishops of Britain annually meet in synod to 
raise, according to their custom, three of their clergy to the epis- 
copal dignity. Two candidates are in readiness, but who the third 
is to be remains uncertain, until Dubricius designates Samson for 
the honour. It is now revealed to the new bishop that he must 
become " peregrinus," must leave his native country and spend 
the rest of his life in service across the seas. With the journey 
through Cornwall and the voyage to Brittany, Samson's career 
in Britain closes ; on the other side of the channel he appears 
as the founder of Dol and other monasteries and the successful 
champion of one Breton count against another. The life brings 
him into association with Childebert, king of Paris from 511 

109 Quidam peritissimi Scoti de Roma venientes " (5820). That they had 
acquired their learning (didicisse) in Rome is an embellishment of the later writer 
(Anal. Boll. vi. 101). 


to 558, and, in harmony with this, he is found signing, as CHAP. 
" Samson peccator episcopus," the decrees of the council held at 
Paris in 555 or 557. 110 

A prominent part was played in the early history of Samson 
by Dubricius. He moves mysteriously, indeed, across the 
stage, appearing with the authority of a bishop or overseer in 
various South Welsh monasteries, but with no hint furnished 
of the place from which or the sphere within which he exercised 
his sway. Yet he is clearly a genuine sixth-century ecclesiastic 
and may not be dismissed as one of the many unsubstantial 
shapes which were drawn into the vortex of the great Arthurian 
legend. He is in that company by the deliberate design of 
Geoffrey of Monmouth, who also invented for him his arch- 
bishopric of Caerleon on Usk, so that there might be a fitting 
dignitary to preside over the crowning of the puissant king. 111 
The purely native tradition knows nothing of Dyfrig in this 
connection ; the tales it has to tell are of his marvellous birth, 
of the fame of his school at " Henllan on the Wye," moved 
after seven years to " Mochros" in the same region, and of his 
retreat at the end of his life as an anchorite to that favourite 
haunt of British saints and goal of British pilgrims, Bardsey 
Island. 112 An attempt was made in the twelfth century to con- 
nect Dubricius with the see of Llandaff; his relics were trans- 
lated from Bardsey in May, 1 1 20, by Bishop Urban, and 
installed with much ceremony in the South Welsh cathedral. 113 
But the chief churches dedicated to him 114 are to be found 

110 H. and St. ii. 75. There are no Welsh dedications to Samson, but he 
is commemorated in Cornwall (Southill), Guernsey and the Scilly Isles. His 
archbishopric of York is due to Geoffrey of Monmouth (Hist. Reg. viii. 12) ; in 
the pages of Giraldus Cambrensis, also, he plays an impossible part (H. and St. 
i. 149). 

111 Hist. Reg. viii. 12; ix. i, 4, 12, 13, 15. The mention in ix. 12 of the 
" trium metropolitanarum sedium archipraesules, Londiniensis videlicet, Ebo- 
racensis, necnon ex urbe Legionum Dubricius " suggests that Geoffrey had seen 
somewhere the names of the British bishops present at the Council of Aries 
(H. and St. i. 7) and anticipated the conjecture of Stillingfleet (Origines Brit- 
annicee, ed. Pantin, 1842, i. p. 115). 

112 Life in Lib. Land. 78-84. The life by Benedict, a monk of St. Peter's, 
Gloucester (printed in Anglia Sacra, ii. 654-61), is later and adopts the fables 
of Geoffrey. 

118 Lib. Land. 84-6. 

114 Llanfrother (now extinct), with its chapels of Hentland and Ballingham ; 
Whitchurch by Monmouth. 



CHAP, within the limits of Archenfield in Herefordshire, the Welsh 

Erging, making it likely that this, rather than any other dis- 
trict in South Wales, was the special scene of his activity. 
What is certainly known of Dyfrig is that he was a bishop who 
used his influence on behalf of monasticism, and must there- 
fore have been, in sympathy and aims, of the party of 
Gildas. 115 

Leaving the little group of men associated with the career 
of Samson, the student finds himself with scarcely any sure 
guide to the history of the great mass of Welsh " saints," n6 the 
founders in the sixth and seventh centuries of the principal 
parish churches of Wales. As to nearly all of them, it may be 
said that nothing has been handed down by tradition save their 
parentage, the names of the churches dedicated to their memory 
and presumed to have been founded by them, and the dates of 
their festivals, the feast day marking in each case the anniver- 
sary of the saint's happy translation to a better world. 117 Thus, 
to take an instance, Tydecho is said to have been the son of 
Anhun Ddu, son of Emyr of Brittany, to have founded the 
churches of Llanymawddwy (with its chapels of Mallwyd and 
Garth beibio) and Cemais and to be commemorated on December 
1 7th. 118 There is no reason why particulars of this kind should 

115 The obit given in Harl. MS. 3859, which is equivalent to A.D. 612 (Cymr. 
ix. 156), is pretty certainly fifty or sixty years too late. 

116 Of the " saints " of Wales, St. David is the only one canonised (about 
1120) by the Roman Church. The term was used by the Celtic Church in the 
sense of Gildas (see note 62 above), i.e., monk. 

117 The oldest form of " Bonedd y Saint " (Genealogies of the Saints) is that 
printed (from a Hafod MS.) in Myv. Arch. II. 23-5 (415-6). This is found in 
various MSS. of the thirteenth and fourteenth centuries, such as Peniarth MSS. 
16 and 45 (Evans, Rep. i. pp. 339, 379). The documents printed in lolo MSS. 100- 
153 are greatly inferior in value, having been compiled at a much later period by 
ignorant and reckless antiquaries (see the criticisms of Phillimore in Bye-Gones, 
1890, pp. 448-9, 482-5, 532-6). In his Essay on the Welsh Saints (London, 1836), 
Rice Rees makes the fullest use of the material supplied by the " Bonedd " MSS., 
the dedications and the feast days of Welsh churches and similar data, and 
though the conclusions adopted (e.g., as to the archbishopric of Caerleon, p. 173) 
will in many cases not stand the test of modern criticism, the book still remains 
most useful for purposes of reference. The ground is now being worked afresh 
in the admirably full and judicious Lives of the British Saints (S. Baring-Gould 
and J. Fisher), of which two volumes (A-E) have been issued. 

118 My v. Arch. II. 24 (415) ; Welsh SS. 218. There was a legend or life of 
St. Tydecho, now lost, which brought him into conflict with Maelgwn Gwynedd ; 
it is only known from a metrical version of it by Dafydd Llwyd of Mathafarn (fl. 
1480), printed in Camb.Reg. ii. 375-7. 


not be in the main authentic, for it is certain that they were CHAP. 
kept on record from a very early period. " Many times," says 
the author of the life of Samson, " have I heard read at St. 
Samson's altar, when mass was sung, the names of both his 
parents." 119 The founding of monasteries and churches appears 
in the same life as an ordinary incident in the career of a 
monastic devotee, 120 and it would seem extremely probable 
that the memory of the "saint" who set apart the site of a 
monastery for religious uses, perhaps after a severe course of 
fasting such as is described by Bede, 121 would be carefully pre- 
served on the spot Thus a form like " Llandysilio" would be 
explained as " Tysilio's monastery," 122 and its extension over 
the whole of Wales would create no difficulty, when it was borne 
in mind that the sixth-century " saint " was habitually a migrant, 
regarding the call to pilgrimage and travel in distant lands as 
a high spiritual distinction. The solemn observance of the day 
of a holy man's death, also, was a custom inherited by the 
Celtic Church from primitive times, 123 and there is every reason 
to think that the dates connected with the names of the patron 
saints of Welsh churches are, for the most part, genuine anni- 

But, though these brief notices are no doubt trustworthy 
in the main, they are embodied in documents so recent that it 
is not safe to use them for the purpose of detailed historical 
reconstruction. Many errors there must be, the result of a 
careless transcription, of confusion between persons of the same 
or of similar names, of attempts to give a favourite saint a little 
added height and dignity. As long as means do not exist of 

120 1< Confirmatis itaque in bonis operibus his omnibus atque ad monasteria 
fundanda suggestis " (5806) ; Samson's mother expresses the hope that very soon 
he will be able, as bishop, to consecrate " monasteria quae nobis suggeris fundanda 
et ecclesias construendas ". 

i H.E. iii. 23. 

121 The original meaning of "llan " (Celtic "landa") is an open space or 
cleared enclosure, as in " gwinllan," " ydlan " (Urk. Spr. 239). It acquired in 
British speech the specialised sense of " monastery " (Life of Gildas, ed. Momm- 
sen, c. 27 ; Life of S. Pol de Leon, Rev. Celt. v. p. 440), and, in Wales, owing to 
the prevalence of churches founded by monks, came at last to mean "church " 

188 The custom is referred to in the early life of Samson, " magnifica ilia ac 
sancta annualis solennitas . . . imminet " (5876-5880) ; indeed from this point 
the life is really a saint's day homily. Cf. Adamnan's Life of St. Columba, ii. 45. 


CHAP, tracking these errors, it is best to regard this whole group of 
facts as bearing abundant witness to the vigour and activity of 
the British monastic movement of the age of Gildas, and not 
to seek to erect upon it any more elaborate historical super- 

It might have been expected that the history of some at 
least of these notable pioneers of the faith would have been 
illustrated by the inscriptions of the period, which is fruitful in 
monuments of the kind and these of a distinctively Christian 
order. But, in point' of fact, only a few fitful gleams of light are 
to be derived from that quarter. In one case only has time 
unearthed what would seem to be the original tombstone of a 
founder in the spot which still bears his name. The stone which 
records the burial of "beatus Saturninus" and "sua sa(ncta) 
coniux," who had no doubt also embraced the monastic life, 
was found in Llansadwrn churchyard, not far from Beau- 
maris, in the early part of the eighteenth century, 124 and, as it 
is certainly not later than 550, may be taken to belong to the 
" Sadwrn " to whom the foundation of the church is ascribed. 125 
In another case, the saint's tombstone came to light some miles 
away from the church which bears his name, but this need cause 
no difficulty, even if it were certain that the stone was in its 
original position. The "Vendesetli" of this Llannor inscrip- 
tion undoubtedly represents the form which Gwynhoedl would 
assume in the fifth and sixth centuries, and there is no difficulty 
in supposing it to have been meant to commemorate the 
founder of Llangwynodl, an ancient church in the same region 
of Lleyn. 126 In like manner, it may be presumed that the 
stone found at Tyddyn Holland, near Llandudno, which is now 
read " Sanctanus sacerdus," marked the grave of Sannan, the 
founder of Llansannan in Denbighshire. 127 It remains to 

124 Inscr. Chr. No. 153 ; Lap. W. 188 ; W. Phil. (2), 363 ; Arch. Camb. 
V. xiii. (1896), 139. 

125 Rhys calls attention (Cymr. xviii. 32-3) to the difference between Sadwrn 
(Saturnus) and Sadyrnin {Saturninus). But, according to Carlisle, Top. Diet., 
the annual fair of Llansadwrn, Carmarthenshire, was held on the same day 
(Oct. 5) as that of Llansadyrnin, which suggests that the two names were used 
interchangeably. The longer was perhaps an affectionate diminutive. 

126 Inscr. Chr. No. 139; Lap. W. 180 ; W. Phil. (2), 366-7; Arch. Camb. 
I. ii. (1847), 201-3; IV. viii. (1877), 141-4; Urk. Spr. 265. 

127 Laf. W. 182 ; W. Phil. (2), 370-2 ; Arch. Camb. V. xiii. (1896), 138 ; 
xiv. (1897), 140-2. Rees (Welsh SS. 240) confounds Sannan, whose day was 


mention the case of Paulinus, which is by no means so clear. CHAP. 
In the life of St. David by Rhygyfarch it is said that the saint 
received instruction for many years from " Paulinus scriba," 
who led a holy life in a certain unnamed " insula " and had 
around him many disciples. 128 Paulinus, in short, was head of 
a monastic school of the same type as Illtud's. No hint, 
however, is given as to where it stood. That it was situated at 
Whitland or " Y Ty gwyn ar Daf" (The White House on the 
Taf) in Dyfed is pure conjecture, resting on no ancient au- 
thority ; 129 the foundations associated with the name of Paul- 
inus are Llangors church in Breconshire 13 and the chapel of 
Ystradffin, anciently known as " Capel Peulin," on the borders 
of Cardiganshire and Carmarthenshire. 131 If regard be now 
had to the early inscriptions bearing the name, they will be 
found to be three. One on a stone now in Margam church is 
the epitaph of" Cantusus," set up (so the legend is construed) by 
his father " Paulinus " ; 132 another at Llandysilio in Dyfed marks 
the resting-place of "Clotorigi " (Clodri) son of "Paulini" ; 133 
the third, found at Maes Llanwrthwl and now kept at Dolau 
Cothi, Carmarthenshire, is more elaborate and may be thus 
rendered " Keeper of the faith and constant lover of his 
country, Paulinus lies here : he was a most devoted follower of 
the right ", 134 It is possible that this last inscription, with its 

I3th June (see Mostyn MS. 88) and who was connected with St. Winifred (Penn. 
ii. 179), with S. Senan of Scattery Island, Clare, who was commemorated on 8th 

128 Cambro-Br. SS. 122. See also p. 137 ; Lib. Land. gg. 

149 Whitland is a township in the parish of Llangan and, until the founda- 
tion of the Cistercian abbey, had no ecclesiastical associations. 

130 See Arch. Camb. IV. xiv. (1883), 44-5, 144, 146, 153, 154, for twelfth- 
century references to Llangors as the church of St. Paulinus. 

131 It is " Capel Pylyn " in Speed's map of Carmarthenshire, and is reason- 
ably identified by Rees (Welsh SS. 187) with the chapel of St. Paulinus mentioned 
as subject to the abbey church in a Strata Florida document of 1339 (Sir. Flor. 
App. li.). 

Inscr. Chr. No. 77; Lap. W. 38-9; Arch. Camb. V. xvi. (1899), 145-6. 
The stone was formerly at Port Talbot and was a Roman milestone before it was 
converted into a gravestone. 

I33 lnscr. Chr. No. 97; Lap. W. 111-2; W. Phil. (2), 397-8. In Celtic 
Folklore, p. 535, Rhys seeks to connect the " marinilatio " of this stone with 
the church of Paulinus at Llangors. 

184 Gibson, 624 ; Inscr. Chr. No. 82 ; Lap. W. 79-81 ; W. Phil. (2), 392. 
When first described by Llwyd, the stone was a footbridge at " Pant y Polion," 
which is, I am informed still the name of a field on Bron Deilo farm, opposite 
Maes Llanwrthwl. 


CHAP, emphatic terms of praise, may record the virtues of the Paulinus 
of tradition, but, in view of the fact that the name was in 
common use at this period, even this must not be regarded 
as certain, and, in general, it may be said that the history of 
this " saint " cannot be eked out by means of the " Paulinus " 

In the case of the more famous of the Welsh "saints," 
further evidence is available in the shape of regular lives, 
recording for the edification of the faithful the great deeds and 
marvellous experiences of men who appear in a variety of 
trying situations as the constant favourites of Heaven. Lives 
of this kind no doubt contain in most cases a substantial 
nucleus of truth and may be used, if due caution be exercised, 
for historical purposes. But, as they are in their very nature 
panegyrics of particular saints, in which everything is subordin- 
ated to the enhancement of the hero's glory, the element of 
exaggeration (not to speak of the avowedly miraculous) is from 
the beginning not absent and it grows more and more pro- 
nounced as the story is told and retold in successive ages. 
Hence it is much against the lives of the great monastic 
founders of Wales Dewi, Padarn, Teilo, Cadog, Illtud and 
Cybi that not one of them can in its present form be assigned 
to an earlier date than the era of the Norman Conquest, that is 
to say, five hundred years after the period with which these 
documents are concerned. 135 It would be folly for this reason 
to question the existence of Dewi and his companion saints, or 
to deny that they founded many of the churches which bear 
their name, but it must be held that we know little of their real 
history. Great as is the fame of Dewi Sant, who has for ages 
been recognised as the patron saint of Wales, for the historian 
he is but a lay figure compared to Gildas. 

The life of Dewi, 136 better known as St. David, is, it should 

135 Most of these lives were edited, with English translations, by W. J. Rees 
for the Welsh MSS. Society in Lives of the Cambro-British Saints (Lland very, 
1853). It is to. be regretted that the text of this volume, in the words of Rhys 
(W. Phil. (2), 425), "teems with inaccuracies"; for a long list of important cor- 
rections see Cymr. xiii. 76-96. The principal MS. source used for this edition 
was Cottonian MS. Vespasian A. xiv. (ff. 13-94), written about 1200; the lives 
of St. David, St. Cadog, St. Gwynllyw, and St. Illtud appear, however, to have 
been composed about a century earlier (Cymr. xi. 127-9). 

138 Dewi represents the popular and Dafydd the learned form assumed by 
David in Welsh (Loth, Mots Latins, 160). 


be said, in all likelihood the oldest and most trustworthy of the CHAP, 
group. It was written about 1090 by Rhygyfarch, son of Bishop 
Sulien of St. David's, who belonged to a family of scholars and 
had access to such records as the cathedral could supply. 187 
It certainly seems, in its account of the early monastic discipline 
of St. David's, to embody ancient materials and to describe 
institutions which had long fallen into abeyance. Thus, in spite 
of its late date and general legendary air, this life perhaps merits 
closer attention than most compositions of the class to which 
it belongs. St. David, according to this narrative, was the son 
of Sanctus, a king of Ceredigion, and Nonnita, a nun of Dyfed ; 
he was born in the latter region, at the spot on the north shore 
of St. Bride's Bay now marked by St. Nonn's Chapel, 138 when 
King Tryffin and his sons ruled over Dyfed. The date thus 
indicated, viz., about 52O,' 39 is probably not far wrong; as to 
the parentage, all that can be said is that the reverence shown 
towards the memory of Nonnita, or, to use the better-known 
form, Non, not alone in Wales, but also in Cornwall, Devon and 
Brittany, seems to give her an assured historical position, though 
the idea that she was a nun at the time of St. David's birth may 
fairly be set down as due to a misunderstanding of her name. 140 
In later life she probably became, like Samson's relatives, a 
convert to the monastic life, and hence the churches dedicated 
to her at Llannon H1 and Llanerchaeron in Cardiganshire, 
Llannon in Carmarthenshire, Alternon in Cornwall, Bradstone 
in Devon and Dirinon in Brittany. 142 Sanctus or Sant is a much 

137 Cambro-Br. SS. 117-43. For Rhygyfarch (" michi autem qui Ricemar- 
chus nominor '') and his connections, see chap. xii. extra note. On this life de- 
pend that composed by Giraldus Cambrensis (iii. 377-404) and the Welsh 
" buchedd," for the earliest text of which see Llyfr yr Ancr, 105-18. 

138 Fenton (2), 63 ; Jones and Freem. 227, 243 ; Arch. Camb. V. xv. (1898), 

139 The pedigree of the royal house of Dyfed makes " Triphun " the grand- 
father of " Guortepir," the " Vortipori " of Gildas (Cymr. ix. 171). 

140 Several instances of the personal name " Nonnita " occur (W. Phil. (2), 
404), one in a Cornish inscription of about this period (Inscr. Chr. No. 10 ; Arch. 
Camb. V. xii. (1895), 54). Gir.-ldus had the " sanctam monialem " of Rhygy- 
farch before him, but thought it desirable to suppress it. No importance need be 
attached to statements providing Non with local connections (Myv. Arch. II. 23 
(415). 37-8 (123); lolo MSS. 82-3, 101, no, 124). 

141 A chapel (now ruined) under Llansantffraid (Arch. Camb. V. xiv. (1897), 

14a "Buhez Santes Nonn," a Breton miracle play, was found in a MS. of 
about 1400 at Dirinon and printed at Paris in 1837 ; the text was re-edited by 


CHAP, more shadowy personage, and better evidence than we have at 
present is needed to prove that St. David was of royal blood 
and grandson of Ceredig ap Cunedda Wledig. 143 

The saint, it is further said, was baptised by St. Ailbe of 
Emly 144 (who died about 530) and spent his earliest years at 
Henfynyw, 145 Cardiganshire, in a school apparently taught by 
one " Guistilianus," a bishop and his cousin on the father's side. 146 
After taking priest's orders, it is added, he went to the school 
of Paulinus, already mentioned, and was there for very many 
years. It would seem as if in these passages two different 
accounts of St. David's education had been combined, more 
especially as the ninth-century life of Paul Aurelian offers a 
third, mentioning " Sanctum Devium " among the famous pupils 
of Illtud. 147 Rhygyfarch's next excursion is into the region of 
pure legend, when he tells how his hero founded, before he made 
his home in Mynyw, twelve famous monasteries, among them 
being the well-known English foundations, Repton, Croyland, 
Leominster and Bath ! He is on surer ground when he brings 
the saint back to Mynyw, in its Latin form, Menevia, and re- 
counts the struggles he had with the chieftain of the district, one 
" Baia Scottus," whose fort is still shown at " Clegyr Foia " 
(Boia's Rocks), 148 ere he was allowed to settle peaceably in 
"Glyn Rhosin," the little valley ever since inseparably associ- 

E. Ernault in Rev. Celt. viii. pp. 230-301, 405-91. The story is taken entirely from 
Rhygyfarch's life, save for some additions from Geoffrey of Monmouth. 

143 Sanctus is " Sant " in the older Welsh authorities (Myv. Arch. II. 23 
(415); Llyfr yr Ancr, 106), the form "Sandde" being an eccentricity of the 
lolo MSS. (82, 101, no, 124). He does not appear in the pedigree of the royal 
house of Ceredigion, which is carried to Ceredig through a son " lusay " 
(Cymr. ix. 181). In the " Buhez," Ceredig is named as the saint's father. 

144 " Helue Menevensium (read, Muminensium) episcopo." He probably 
owes his place in the legend to the existence near St. David's of a Llaneilfyw or 
St. Elvie's. 

145 1 follow the explanation of Giraldus (iii. 384), that the " Vetus Rubus " 
of Rhygyfarch is a translation of Henfynyw by some one who thought, as 
Giraldus did himself, that Mynyw was from the Irish muine, a bramble. 

146 Possibly the " Justinianus " whose astonishing legend is given in Cap- 
grave's Nova Legenda Anglia (ed. Horstman, 1901, ii. pp. 93-5) and who was 
commemorated at Llanstinan and at Capel Stinan, near St. David's (Fenton 
(2), 64 ; Jones and Freem. 224). 

147 Rev. Celt. v. p. 421. 

148 Excavations conducted by Mr. Baring-Gould in 1902 showed that the 
spot had been occupied by a people in the Early Iron stage of civilisation (Arch. 
Camb. VI. iii. (1903), i-n). It need not be supposed that Boia and his clan 
were much above this level. 


ated with his name. That he found a Goidel in possession CHAP. 
in this remote corner of Dyfed is most likely U9 and opposition 
to a monastic settlement may well have come from a self-willed 
landowner who did not at all repudiate the name of Christian. 

After the settlement in St. David's, the life suggests that, 
save for one important exception, the saint's wanderings came 
to an end and that henceforth his energies were devoted to the 
work of organising and controlling the monastic community he 
had got together in this remote angle of Britain. The exception 
is the journey to Jerusalem which, according to the legend, he 
made in company with Padarn and Teilo and which had as its 
issue his consecration as " archiepiscopus " by the Patriarch, but 
this passage in the life is without doubt purely mythical, the 
intention being to show that St. David was not beholden for his 
episcopal authority to any ecclesiastic in the West. Though 
he was reverenced in later ages in Devonshire, 150 in Cornwall, 151 
and in Brittany, 152 and was in close intercourse with the mon- 
astic founders of Ireland, with men like St. Maedhogof Ferns 153 
and St. Senan of Clare, 154 there is nothing to show that he 
travelled in those regions ; the influence he wielded was as 
abbot of the monastery of Mynyw or Tyddewi and the founder 
of daughter monasteries in other parts of South Wales. 

The type of monasticism which prevailed at St. David's was 
of the most rigorous kind. The abbot himself was known as 
" Dewi Ddyfrwr," 155 David the Waterdrinker, and for ages his 
successors were under a solemn obligation to abstain from 
meat. 156 In the picture drawn by Rhygyfarch of the ancient 
but vanished order of the fraternity, monks are seen yoked to 
the plough in place of oxen, others dig and hoe the ground 

149 P. 121. 1BO At Thelbridge and at Ashprington. 

181 At Davidstow. m At S. Divy, near Landerneau. 

188 The "Aidus" of the life in Cambro-Br. SS. 232-50. In Rhygyfarch's 
text he appears as " Aidanus " (130) and " Maidoc" (133), the latter being an 
affectionate derivative from the former name (Arch. Camb. V. xii. (1895), 36-7). 
Ferns, the seat of his monastery, is " Guernin " in the one passage and 
" aquilento," i.e., marsh (absurdly rendered " north " on p. 436 1), in the other. 

l **Acta Sanctorum, March, vol. i. (8th day), 7720 (Venice, 1735). 

168 Life of Paul Aurelian (Rev. Celt. v. p. 421), Rhygyfarch (118), Giraldus 
Cambrensis (iii. 379), LL. ii. 318, lolo MSS. 300, 301. Lib. Land. 128 suggests 
a somewhat different interpretation, and Mr. Phillimore regards the original 
meaning of the epithet as an open question (Owen, Pemb. 206-7). 

188 Gir. Camb. vi. 104 (Itin. ii. 2). 


CHAP, in religious silence, others carry saws for the felling of timber. 
No one may claim anything as his own, not even the sorry 
skins he wears ; the careless use of such a phrase as " my book " 
is an offence to be expiated by severe penance. It would 
appear certain, therefore, that St. David's was a monastery of 
that stricter pattern which caused Gildas, if the extracts cited 
under his name on this subject are really his, 157 so much anxiety 
in his later years. Monks, he complains, are forsaking their old 
allegiance in many monasteries of the ancient and less exacting 
type in order to join communities having a more rigorous ideal, 
in which the eating of meat, the drinking of all beverages save 
water, the use of horses and carriages are abjured. Bread is 
eaten by measure ; oxen are discarded so that the zeal of the 
brethren may show itself in the drawing of ploughs, and mean- 
while, such is the burden of his lament, there is a notable 
falling off in Christian charity and a dangerous uprising of the 
pharisaic spirit. If Gildas and David did, as is dimly suggested 
by these extracts, champion opposing schools of ascetic thought, 
it is certain that the school of David left the deeper impress 
upon the Welsh mind ; as Christianity had in the first instance 
made good its foothold in Wales through the powerful appeal 
of monasticism, so now Celtic enthusiasm cast its suffrage in 
favour of the more uncompromising exponents of the monastic 

That St. David was a bishop 158 may be without hesitation 
believed, and no doubt every abbot who succeeded him in the 
headship of the monastic community of Mynyw held the same 
office. He may even, by contrast with men of inferior fame 
and influence, have been known as " archiepiscopus " or " chief 
bishop". 159 But there is no warrant for supposing that he 
sought to wield any authority as metropolitan, or archbishop 
in its later sense, over the British Church, either from St. David's 
or any other centre. Geoffrey of Monmouth, having made 
Dubricius Archbishop of Caerleon, is naturally led to indicate 
David as his successor in that dignity ; 16 Rhygyfarch is less 

157 Printed by H. and St. (i. 108-13), Mommsen (pp. 86-8) and Williams 
(pp. 256-70), who are all disposed to regard them as genuine. 

158 He is styled " Dauid episcopus moni iudeorum " in Harl. MS. 3859 
(Cymr. ix. 156). 

159 For this use of the title see Cymr. Trans. 1893-4, 131. 

160 Hist. Reg. ix. 15 ; xi. 3. 


concerned about ecclesiastical order, and, while claiming for his CHAP, 
hero the position of metropolitan, represents the office as con- 
ferred upon him by general acclamation after his wonderful 
preaching to the multitudes assembled at the Synod of Brefi. 
But in the days of St. David the system of dioceses, under 
which each bishop has assigned to him for his exclusive rule 
a definite extent of territory, was clearly not known in Wales, 
and still less could there be an archbishop recognised as supreme 
from end to end of that country. 

Nothing could well be more legendary than the account of 
the Synod of Brefi as it has come down to us. The Pelagian 
heresy has revived (though Gildas, it should be noted, has not 
a word to say about a relapse which, had it really occurred, 
must have deeply moved him), and 1 1 8 bishops meet at Brefi 
in Ceredigion to proclaim the true faith to the immense throng 
there gathered together. But not one of them can make 
himself heard, and matters are at a standstill until it is suggested 
that Daniel and Dubricius shall fetch the Saint of Mynyw. 
David is with difficulty persuaded to return with them to the 
synod and leaves them on the way to raise from the dead the 
only son of a widow. But, when at last he reaches Brefi, he is 
master of the situation ; the ground beneath his feet rises into 
a little hill (on which now stands the church of Llanddewi), 
and his voice, resonant as the sound of a trumpet, reaches the 
farthest limits of the assembly. Fantastic as is this story, the 
Synod of Brefi may not itself be mythical ; it is known from 
Bede m and from the life of St. Samson 162 that the bishops of 
the Britons were accustomed to meet in such gatherings, and, 
if there was no occasion to discuss Pelagianism, there were 
other burning questions, notably those connected with the 
monastic life, which made counsel and joint action very neces- 

St. David died, as is well known, on the 1st of March; 
the year of his death cannot, unhappily, be fixed with the same 
precision. Probably that intended by the Irish Chronicon 
Scotorum, viz., 588, is not far from the mark, 163 and he would 

181 H.E. ii. 2. IBS P. 5836. 

193 The entry is simply " Dauid Cille Muine". So Tighernach (Rev. 
Celt. xvii. p. 158) has " Dabid Cille Muni " ; Harl. MS. 3859 also leaves out 
the " obiit ". Thus all three notices clearly come from a common, probably 


CHAP, thus be a contemporary of Gildas, but somewhat younger. 
The fame he acquired in later ages, culminating in his adoption 
by Welshmen generally as the patron saint of their country, is 
to be explained by the very large number of churches (Rees 
reckons fifty-three 164 ) in all parts of South Wales which were 
regarded as under his protection. In many cases the dedica- 
tion is probably not one which dates from St. David's own day, 
but, when all deductions have been made, it still remains true 
that, within the limits of South Wales, he had an exceptional 
position in respect of the many important churches, like 
Llangyfelach in Gower, Llanarthneu on the Towy, 165 Glascwm 
in Elfael, and Dewchurch Magna in Herefordshire, which 
owned allegiance to his memory. The only saint who seems 
to have been regarded as in any way his rival was Cadfael, 
familiarly known as Cadog and Catwg, the founder of Llancar- 
fan in Morgannwg, and a saint of great reputation in that 
quarter of Wales. 168 In the life of Cadog, written about 1075 
by one Lifric, who was son to Bishop Herwald of Llandaff and 
" magister " of Llancarfan, 167 it is said that David summoned 
the Synod of Brefi while Cadog was away on pilgrimage and 
that his reluctance to take precedence of so distinguished a 
saint was only overcome by the strict injunctions of an angel. 
When Cadog returned, no one for a while ventured to tell him 
what had occurred, and, when at last the news was broken to 
him, he gave vent to most unsaintly anger, which was only 
appeased by another angelic interposition. It is noteworthy in 
this connection that none of the ancient churches of the king- 

an Irish, source, and the later date implied in the Welsh chronicle, viz., A.D. 601, 
is perhaps due to an attempt to connect the saint with the " Sinodus urbis 
legion " ascribed to that year (Cymr. xi. 156). If the statement of Rhygyfarch, 
that he died on a Tuesday, were to be accepted, the year might be given with 
some confidence as 589. 

164 Welsh SS. 45. 

165 Llanarthneu appears in a poem by Gwynfardd Brycheiniog, a contem- 
porary of the Lord Rhys ap Gruffydd, as " llan adneu " (Myv. Arch. I. 271 
(194) ; Lit. Kym. (2), 155). It is, therefore, no doubt, the " Depositi Mona- 
sterium" of Cambro-Br. SS. 117, " adneu" being the Welsh legal term for a 
deposit (LL. i. 244, 258, 484 ; Wotton, Glossary, s.v.). At the same time, this 
current derivation was probably wrong, Arthneu being really a proper name ; cf. 
the " Ian hardneu " of Lib. Land. 279. 

166 His life, from Vesp. A. xiv., is printed in Cambro-Br. SS. 22-96. For 
the name Cadfael (in old Welsh, Catmail), see pp. 25, 28. 

167 Cambro-Br. SS. 80 ; Lib. Land. 271, 273, 274. 


dom of Glywysing, lying between the Tawe and the Usk, are CHAP, 
dedicated to St. David ; this was the special domain of Cadog, 
who was sprung from the princely house of the district, his 
father Gwynllyw (commemorated at St. Woollo's, Newport) 
being the ruler who gave his name to the ancient cantref of 
Gwynllwg. 168 Thus there would appear to be some ground 
for the belief that Dewi and Cadog were not close fellow- 
workers, while both tradition and the evidence of dedications 
go to show that Dewi worked harmoniously with the other two 
monastic pioneers of the South, with Padarn or Paternus of 
Llanbadarn Fawr in Ceredigion and with Eliud, familiarly 
known as Teilo, the patron of many a Llandeilo in South 
Wales. 199 But, if there was jealousy in Glamorgan of the fame 
of. the Demetian saint, the rivalry of Cadog never became 
elsewhere serious, for his renown always remained a strictly 
local affair. 

NOTE TO CHAPTER V. i. Harleian MS. 3859. 

This MS., which at one time belonged to the abbey of Montauban in the 
South of France, includes, among other writings unconnected with Wales, a 
copy of the Historia Brittonum attributed to Nennius (ff. 1746-1896), followed 
by a set of Latin annals, a number of genealogies of Welsh and Cumbrian 
princes, a list of British cities, and the tract De Mirabilibus Britannia (ff. 1896- 
198). Various estimates have been given of the age of this little collection ; 
Petrie assigns it to the tenth century (Man. Hist. Br. 68), Hardy to the eleventh 
(Catalogue of Materials for British History, i., No. 778), Maunde Thompson to 
the beginning of the twelfth (Cymr. ix. 145-6). The point is not, however, of 
the first importance, since it has been clearly shown that the Harleian MS. is 
not an original in respect of these documents, but is derived from an older MS. 
written in the Hiberno-Saxon hand. The parent MS. may be confidently as- 
signed to the end of the tenth century, in view of the fact that the principal 
pedigree given is that of Owain ap Hywel Dda, who died in 988, and that the 
annals come to an end in his reign. This conclusion is also supported by the 
spelling of the Welsh names, which are uniformly in the Old- Welsh form. 

The MS., therefore, represents a tenth-century contribution to the history 
of Wales, and the information given in the annals and the genealogies is of high 
value. Both have been printed, with scrupulous care, by Phillimore in the ninth 
volume of the Cymmrodor (152-83), where also will be found a full introduction 
(141-51). Less satisfactory is the edition of the annals in the volume entitled 
Annales Cambria, which appeared in the Rolls series in 1860 under the editor- 
ship of John Williams Ab Ithel. Harl. MS. 3859 is MS. A. of this edition, in 
which it is awkwardly combined with MSS. B. and C. of the thirteenth century. 

1M Cymr. vii. 118-9; >" 4- J - The oldest known form of the name is 
given in Harl. MS. 3859, viz., Guinnliguiauc (Cymr. ix. 167), for Gwynllywiog. 

16 For the lives of Padarn and Teilo see Cambro-Br. SS. 188-97 and 
Lib. Land. 97-117 respectively. Both contain friendly references to St. David, 


CHAP. The annals begin with a year which appears to be A.D. 444 and run on, so 

V. far as the marking of the years is concerned, to 977, though the last event re- 
corded is under 954. The year of our Lord is nowhere given, the only addition 
to the successive entries of " an(nus) " being the figure set against every tenth 
year which gives the number of that year, counted from the first (see the fac- 
simile in the Rolls edition). As a record of events set down from year to year, 
these annals obviously belong to St. David's ; after 809 no bishops are men- 
tioned save those of that see, while of them there is a fairly complete account. 
It is known from the story of Asser that learning was not without its devotees 
at St. David's during this age. But, while the later entries are undoubtedly of 
contemporary authority, the earlier ones have no claim to be put on the same 
footing. It has often been pointed out that the basis of the chronicle was 
furnished by some Irish source, also used by Tighernach and the Irish annalists 
who followed him (Mon. Hist. Br. 92-3 ; Ann. C. xv. xvi. ; Cymr. xi. 139). 
Nor is it unlikely that, as suggested by Skene (Chronicles of the Picts and Scots, 
1867, xxviii.), it was this Irish document which provided the starting-point of 
A.D. 444, otherwise not easily explained, for the Annals of Ulster assign to this 
year the foundation of Armagh and the beginning, therefore, of organised 
Christianity in Northern Ireland. But, besides this foreign source, the earlier 
annals depend to a great extent on the Historia Brittomim, and probably upon 
other similar traditional accounts of the doings of the Britons. In the eighth 
century the notices are predominantly northern, and are, for the most part, such 
as would naturally occur in a Strathclyde chronicle. It is not until the age 
of Offa is reached that the entries begin to wear the aspect of a contemporary 
record kept in South Wales, and it is from this point only that the chronicle 
takes the position of a historical authority of the first class. 

NOTE TO CHAPTER V. ii. MSS. and Editions of Gildas. 

Gildas's work was not popular in the Middle Ages (" raro invenitur," says 
William of Newburgh, who had lighted upon a copy) and few MSS. are available 
for the determination of the text. Of those extant the most important is Cot- 
tonian MS. Vitellius A. vi., which belonged to the abbey of St. Augustine, 
Canterbury, and was written in the eleventh century. It was used by the early 
editors of Gildas, but suffered severely in the Cottonian fire of 1731 and was 
thus neglected in the nineteenth century, until Mommsen showed in 1894 that 
much of it was still legible and of great-value for purposes of textual criticism. 
In the same edition another important MS. was used, viz., that now kept in the 
public library of Avranches (No. 162), which came originally from the neigh- 
bouring abbey of Mont St. Michel. The readings of this MS. had not been 
given in any previous edition ; it is of the end of the twelfth century, and, though 
corrupt and interpolated, supplies a useful check upon the Cottonian MS. 
There are two MSS. of Gildas in the Cambridge Public Library; Dd. i. 17 
came from Glastonbury and belongs to about 1400 ; it is clearly a copy of the 
Cottonian MS. and is valuable chiefly as evidence of what that MS. once con- 
tained. Ff. i. 27 was the property of the Cistercian abbey of Sawley in York- 
shire and is a thirteenth-century MS. It is a copy of a MS. written by one 
Cormac, in which great liberties were taken with the original. 

The first editor of Gildas was the Italian Polydore Vergil, who used the 
Cottonian MS. and one akin to that of Avranches. This edition appeared in 
1525. In 1568 John Josselin, Latin secretary to Archbishop Parker, published a 
second, based upon the Cottonian MS. and its Cambridge derivative, Dd. i. 17. 
Thomas Gale next included Gildas in his Scriptores XV. (Oxford, 1691), and for 
the first time made use of the older Cambridge MS., Ff. i. 27, by no means to the 


advantage of the text. Nearly a century and a half went by ere another attempt CHAP, 
was made to give to the world a text of Gildas, and the edition of Joseph Steven- V. 
son, published by the English Historical Society in 1838, was in some respects 
the least satisfactory of all. Stevenson relied on the Cambridge MSS. and the 
early editions, and appears to have taken it for granted that no better authorities 
were to be found ; the Cottonian MS. he does not even mention. Gildas finds 
a place in the huge volume entitled Monumenta Historica Britannica (1848) and 
also, with the omission of the historical portion, in the first volume of Councils 
and Ecclesiastical Documents (London, 1869), but both Mr. Petrie and Mr. 
Haddan, while aware of the importance of the Cottonian MS., believed it to be 
lost and brought no fresh material to the study of the text. In 1894, 17 however, 
appeared the Monumenta Germania Historica edition, under the superintendence 
of Mommsen ; for this full use was made both of the Cottonian and the Avranches 
MSS. and textual problems were for the first time adequately discussed in a de- 
tailed introduction. In the edition issued by Hugh Williams for the Cymmro- 
dorion Society (1899-1901), the text of Mommsen is reprinted with an English 
translation and notes. 

Gale was the first to divide the work into " Historia " (cc. 1-26) and " Epis- 
tbla" (cc. 27-110), the source of this error (in which he was followed by Steven- 
son) being the fact that the MS. to which he pinned his faith, viz., Ff. i. 27, ends 
with c. 26. In no other MS. is there any break at this point, and the monk of 
Rhuis, in his life of Gildas (p. 97), joins the end of c. 26 to his citation of c. 27 
in the most natural manner possible. What the original title of the work was it 
is difficult to determine ; in the Cottonian MS. it was apparently described as 
De Excidio Britannia. 

The authenticity of the De Excidio as a real production of the early sixth 
century is no longer seriously questioned. The MSS. are all of late date, but 
the extensive use of the work by Bede, who mentions " Gildus " by name (H.E. 
i. 22), makes it impossible to suppose it of later date than A.D. 700, and the efforts 
of Thomas Wright (Biographia Britannica, i. 115-35) ar *d A. Anscombe (Academy, 
1895) to nn d a place for it, either as a whole or in part, in the seventh century have 
been quite unsuccessful. 

The Rhuis life was first printed from a defective Fleury MS. in Dubois" 
Floriacensis Bibliotheca (1605), 429-63, whence it was taken by Bolland for the 
Acta Sanctorum (Jan. (29), II. 958-67). Mabillon in Acta Sanctorum Ordinis 
S. Benedicti (i. 138-52) used a Rhuis MS. to supply some of the defects of the 
Fleury MS., but was not able to add 'the missing close. Mommsen edits (pp. 
91-106) from the printed sources. In c. 34 reference is made to an event stated 
to have happened in 1008, but it is not unlikely that, as Williams suggests 
(Gildas, p. 318), these later chapters are an addition to the original life, furnishing 
no clue as to its date. The Glastonbury life was first printed by Stevenson in 
his edition of Gildas, from Burney MS. 310 (British Museum), written in the 
priory of Finchale, near Durham, in 1381, and a MS. copied from this in the 
sixteenth century (Royal MS. 13 B. vii.). Phillimore pointed out in 1890 (Cymr. 
xi. 79) how the most valuable MS. of all, C.C.C.C. MS. 139 (which is of the 
twelfth century), had been neglected, and cited some of its readings. Mommsen's 
text (pp. 107-10) is based on this and the Burney MS. 

170 This is the date of the separate issue of the " Gildas and Nennius " fasci- 
culus the whole volume (torn, xiii.) is dated 1898. 

VOL. I. 1 1 



(For this period I have made much use of Mr. Plummer's editions of Bede's 
Ecclesiastical History and of the Anglo-Saxon Chronicle. Celtic Britain is still 
indispensable. Mr. Skene devoted much labour to the elucidation of the history 
of this period and embodied the fruit of his researches in The Four Ancient Books 
of Wales and Celtic Scotland. But his analysis of the material he used was not 
sufficiently searching, and many of his conclusions rest in consequence on a very 
insecure basis.) 


CHAP. THE first event on record in the history of the struggle between 
the English and the Britons who were under the sway of the 
northern gwledig is the foundation by Ida of the kingdom of 
Bernicia. Bede gives the year as S47, 1 a date which he no 
doubt inferred from the particulars given in the old list, kept 
from an early period, of Northumbrian kings and the years of 
their rule. 2 Whether Ida founded an entirely new settlement or 
was raised to the dignity of king by a body of Anglian warriors 
already established on the Northumbrian coast is a point as to 
which information is wanting, but it is likely that any colony 
previously settled in this region was insignificant, and that it 
was from the middle of the sixth century that the Britons who 

1 H.E. v. 24. 

2 The references to the treatment of the " infaustus annus," 633-4, m H,E. iii. 
i, 9, afford absolute proof that there was such a list in existence in Bede's time 
and that he regarded it as a trustworthy record. In point of fact, the Moore 
MS. of H.E. has appended to it (see Man. Hist. Br. 290) a table of this kind 
which appears to have been entered in the MS. in 737, two years only after the 
historian's death. It gives the length of Ida's reign, with Bede, as twelve years, 
and the data it furnishes, assuming that ^Ethelfrith died in 616, would fix the 
accession of Ida at 547. The same list appears, with the accidental omission of 
the one-year King Glappa, in the Saxon Genealogies (Hist. Britt. c. 63) ; 
Simeon of Durham has it also (Hist. Reg. 12), but with a number of errors of 
transcription. The list in Fl. Wig. i. 6 differs in important respects ; it appears 
to me to have been altered so as to bring it into conformity with the Anglo-Saxon 
Chronicle, s.a. 588. 


dwelt on the banks of the Forth, the Tweed and the Tyne had CHAP. 


reason to be genuinely alarmed at the progress of the Teutonic 
invader. The centre of Ida's power is indicated by the fact 
that tradition ascribes to him the foundation of Bamborough, 8 
a coast castle between Berwick and Alnwick, which long 
remained an important royal stronghold. The people over 
whom he bore rule called themselves Baernice, 4 a name which 
was possibly derived from that of the Brigantes who anciently 
occupied this region 5 and was transferred into Welsh in the 
forms Byrneich and Bryneich. 8 

During the second half of the sixth century war must have 
been all but incessant between the Bernician interlopers and 
the older inhabitants of these northern lands. No details of 
the conflict have, however, been handed down, save the brief 
notices contained in the little tract which a Briton of the 
district put together at the end of the seventh century and 
which now appears, in four MSS., at the end of the Historia 
Brittonum. In this document (commonly known as the 
Saxon Genealogies) 1 the names are mentioned of five kings 
of Bernicia who succeeded Ida and preceded ^thelfrith, and of 
two of them it is said that British chieftains waged war against 
them. These were Theoderic, a son of Ida, who reigned from 
572 to 579 "against him Urien and his sons valiantly did 
battle" and Hussa, who was king from 585 to 592 "against 
him the four kings, Urien, Rhydderch the Aged, Gwallog and 
Morgan, fought ". 8 In 592 a grandson of Ida, ^Ethelfrith 
son of jEthelric, obtained the crown, a man whom Bede terms 
a very Saul for plundering his enemies ; no English leader, he 
says, made himself master of more British land, either by 
driving out the Britons or reducing them to servitude. 9 Thus 

3 A.S. Chr. MS. E. s.a. 547. Cf. also the difficult passage in the Saxon 
Genealogies (Hist. Britt. c. 61, end), which I read: " (i)unxit (as in the "Nen- 
nian " recension) Dinguayr(u)i guurth Berneich," i.e., he joined Dinguarwy, or 
whatever the form should be, to Bernicia. So Zimmer, Nenn. V. 307 ; on p. 80 
he has another suggestion. 

* A.S. Chr. MS. E. s.a. 634. B Celt. Br. (2), pp. 113-4. 

8 Byrneich is the older form ; see Mommsen's edition of the Hist. Britt. pp. 
201 (Berneich, Bernech), 204 (Birneich), 205 (Berneich, Birnech) ; Bruts, 101 
(Byrneirch), 102 (Byrneich), 103 (Byrneich). 

7 See note on p. 116. 

8 Hist. Britt. c. 63. The dates are calculated from the list of kings referred 
to above. 

//.. i. 34. 

II * 


CHAP, at the end of the century Bernicia had become, out of small 
beginnings, a most formidable power. 

It seems most probable that it was this struggle, continued 
by the Britons with varying but on the whole decidedly 
adverse fortune until the middle of the seventh century, which 
created the national name of Cymry. 10 Hitherto, the general 
designation of the race holding power between the Severn and 
the Forth had been " Britons," in modern Welsh " Brython " ; 
this distinguished them alike from the Goidels, Scotti, or Irish 
on the one hand and from the English on the other. But in 
the extremity which now beset them, in face of the resistless 
advance of the English, not in Bernicia alone, but in the 
kingdom of Deira also, to the south of the Tees, and through- 
out Mid Britain, Brython and Goidel would seem, along the 
whole of this line, to have agreed to cast aside race distinctions 
and to recognise only the common name of " Combroges " or 
" fellow-countrymen," fighting for freedom under the authority 
of one gwledig. Only thus does it seem possible to explain 
the late appearance of the name in history, its extension to the 
North, where it is still preserved in Cumberland, the " land of 
the Cymry," and its failure to reach the Britons of Devon, 
Cornwall and Brittany, between whom and the more northerly 
Britons there was kinship, but, so far as can be seen, no 
martial alliance. 

The leaders of the Cymry in this long and obstinate 
conflict were the men known to tradition as " GwyY y Gogledd," 
the Men of the North. 11 No single figure towers above the 
rest, as in the age of Arthur or of Maelgwn Gwynedd, though 
it may be presumed that the office of gwledig was still in 
existence ; the Britons were clearly much divided and fought 
under the leadership of local chieftains. Four of these are 
mentioned in the Saxon Genealogies as opponents of Hussa, 
viz., Urien ap Cynfarch, Gwallog ap Llyenog, Morgan (ap 
Coleddog ?) and Rhydderch ap Tudwal. Others whose names 

10 The view here adopted is that of Rhys (Celt. Br. (2), pp. 115-6, 139 ; W. 
People, p. 26). For the derivation and application of the name, see the note ap- 
pended to this chapter. 

11 Peniarth MS. 45 (formerly Hengwrt MS. 536) gives " Bonedd Gwyr y 
Gogledd" (Pedigrees of the Men of the North), as printed in IV. Anc. Bks. ii. 
p. 454. But these pedigrees are to be found in a much more trustworthy form in 
Harl. MS. 3859, as printed by Phillimore in Cymr. ix. 169-82. 


have been handed down are Llywarch ab Elidyr (surnamed the CHAP. 
Aged), Clydno Eiddin, Gwrgi and Peredur his brother, Gwen- 
ddoleu ap Ceidio. Of most of these little can be said save that 
they seem to be real historical personages, but of one or two a 
few facts are authentically recorded. The author of the Saxon 
Genealogies had for some reason or other a special interest in 
Urien, 12 whom he describes as the most brilliant war-leader on 
the British side ; he besieged the English for three days and 
three nights in the island of Lindisfarne, 13 but, in spite of his 
services to the cause of British freedom, fell a victim to the 
jealousy of Morgan, who contrived his death in the course of 
the expedition. Urien continued to be for many centuries, 
under the name of Urien of Rheged, a figure of importance in 
Welsh legend, but his exploits against the Irish of South Wales, 
the fame of his son Owain, lord of three hundred ravens, and of 
his daughter Morfudd, beloved of Cynon ap Clydno Eiddin, and 
his death at the hands of Llofan Llawddifro, all belong to the 
realm of fable rather than of history. 14 Rhydderch ap Tudwal, 
known to the older writers as Rhydderch the Aged, 15 but in 
later times as Rhydderch the Open-handed (Hael), 16 is another 
British leader who may be said to be something more than a 
name for the historian. In Adamnan's life of St. Columba, 17 
written at the end of the seventh century, he is spoken of as a 
friend of that saint, who ruled at the Rock of Clyde, the Allt 
Glud of the Britons, now known as Dumbarton. 18 He had 
many enemies, of whom he stood in daily fear, and in his 

12 In addition to the account of Urien himself, there are references to a son 
Rhun (c. 63) and a great grand-daughter Rhiein(melth) (c. 57). 

13 " Insula Metcaud ". C/. " Inis Metgoit " of Tighernach (Rev. Celt. xvii. 
p. 182). 

14 See Arth. Legend, cap. xi., where perhaps overmuch emphasis is placed 
on the mythological element. 

18 " Riderch hen " in the pedigree (Cymr. ix. 173) as in the narrative 
(c. 63). 

18 He appears, with Nudd and Mordaf, as one of " Tri hael Ynys Prydain " 
(Triad i. 8 = ii. 32 = iii. 30). 

"I. 15. 

18 The old- Welsh form, Alt Glut, in which "Alt" = hill, height, is found 
in the Namur MS. of Bede (H.E. i. i) and in Harl. MS. 3859 (Cymr. ix. 166) ; 
it seems to be preserved also in the name of the Arthurian warrior " tarawc allt 
clwyt " (Mab. 138). More common, however, is the Irish Ail (= rock, the 
" Petra " of Adamnan and H.E. i. 12) Cluaithe, whence the ordinary Alclud. 
Dumbarton is Dunbretan, the fort of the Britons. 


CHAP, anxiety sent a private message to the prophet of lona to know 
whether he was destined to fall by their hands. Columba's 
answer was that he would escape all their wiles and die in his 
own house, reposing on his couch of feathers, a prophecy 
which, according to Adamnan, was literally fulfilled. Rhy- 
dderch and Columba being on these terms, it is easy to credit 
what Jocelin of Furness, writing at the end of the twelfth 
century, says of the relations between the ruler of Dumbarton 
and another saint of the period, viz., Kentigern or Cyndeyrn, the 
founder of Glasgow. 19 The hostility of " Morken," who may be 
the "Morcant" or Morgan of the Saxon Genealogies, had 
driven Kentigern from the banks of the Clyde to North Wales, 
where he founded the monastery of Llanelwy or St. Asaph. 
But when Rhydderch obtained supreme power in the North, 
there was a change of attitude towards Christianity, or rather, 
it may be supposed, towards monasticism ; the new ruler was 
in full sympathy with the aims of Kentigern and invited him to 
return to his old field of operations. There was a memorable 
meeting at Hoddam (in Dumfriesshire), where the saint for 
a time took up his abode ; in the course of a few years the 
monastic centre of Glasgow was founded, not far from Rhy- 
dderch's stronghold of Dumbarton. It has been suggested that 
Rhydderch attained his commanding position as the result of 
the battle of Arderydd, fought at Arthuret, near Carlisle, about 
575, and that he then defeated a semi-pagan host and achieved 
a Christian victory ; 20 but in the course of ages, so thick a 
legendary haze has gathered round the history of this famous 
encounter 21 that one may not venture to say more of it than 

19 Life as edited by A. P. Forbes in The Historians of Scotland (vol. v. 
Edinburgh, 1874), cc. 23, 24, 29-33. 

20 IV. Anc. Bks. i. pp. 66, 175. Harl. MS. 3859 assigns " bellum armterid " 
to the year 573 (Cymr. ix. 155), but its dates ought not hereabouts to be accepted 
without question. All the known forms of the name, including " Armterid " 
and the " Arywderit " ofBlk. Bk. zb, 316, imply the spelling Ardeiydd, and Skene 
had no ground for using the form Ardderyd, save the wish to support his view that 
the site of this battle was in the neighbourhood of Arthuret, a little north of 
Carlisle. Arthuret was, however, Ar^uret in the thirteenth century (Tax. Nich. 
319) and the identification, which rests on a passage in the fifteenth-century 
edition of the chronicle of Fordun (ed. Goodall, Edinburgh, 1759, i. 135-6), 
mentioning a battle " cunctis in hac patria constitutis satis noto, quod erat in 
campo inter Lidel et Carwanolow situate," needs no such illicit garnishing. 

21 For the legends about Rhydderch and Arderydd and various interpretations 
of them, see the mediaeval Latin poem entitled " Vita Merlini " (Die Sagen von 


that it was a triumph won by Rhydderch over Gwenddoleu ap CHAP. 
Ceidio. aa Possibly it was the real turning-point of his career ; 
his power, at any rate, was by some means or other firmly 
established, for the family of Rhydderch, in a collateral branch, 
furnished kings of Cumbria or Strathclyde for many genera- 
tions. 88 

How the Cymry fared in their struggle with the Anglian 
kings of Deira and with the chieftains who won Mid Britain 
for the Teutonic race it is beyond the power of the historian to 
say, for we have no trustworthy information on the subject 
until Bede raises the curtain upon the doings of ^Ethelfrith at 
the beginning of the seventh century. 24 The most prominent of 
the princes of the Southern Cymry in this age was Rhun, son 
of Maelgwn Gwynedd, who succeeded to his father's authority 
in North Wales about 550, but tradition has nothing to say of 
any fighting between him and the English. Yet it leaves in 
the mind a fairly definite impression of the man as one of the 
notable figures of the second half of this century. 26 In the 
mediaeval romance known as The Dream of Rhonabwy, Arthur 
and his knights are represented as receiving from their enemy, 
Osla of the Broad Knife, a request for a six weeks' truce. The 
king gathers his counsellors together to consider the proposal 
and thereupon the whole group makes its way to the spot 
where a great, tall man, with red-brown, curly hair, sits apart. 
Rhonabwy in his vision asks the reason of this extraordinary 

Merlin, San-marte, Halle, 1853) ; the Welsh poems printed in IV. Anc. Bks. ii. pp. 
3-5, 18-28, from the Black Book of Carmarthen; Davies's Mythology and Rites 
of the British Druids (London, 1809), pp. 469-74; Lit. Kym. (2), 198-267; Diet. 
Nat. Biog. s.v. Rhydderch Hael. 

23 The name of Gwenddoleu is perpetuated in Carwhinelow, near Arthuret, 
anciently Caerwyndlo (Bye-Gones, ist Oct. 1890, p. 483) and Carwanolow (note 
20 above). 

83 See Harl. MS. 3859, Pedigrees V. and VI. (Cymr. ix. 172-3)- 

84 The elaborate speculations of Dr. Guest, as contained in Origines Celtica 
(London, 1883), have been accepted to a surprising extent by Stubbs, Freeman, 
Green and other writers of sound judgment. But they rest on the flimsiest foun- 
dations, and in particular take no account of the established rules of Celtic or 
Teutonic philology. See W. H. Stevenson in Engl. Hist. Rev. xvii. (1902), pp. 

25 For Rhun see Harl. MS. 3859, Pedigree I. (Cymr. ix. 170) ; Jesus Coll. 
MS. 20 (Cymr. viii. 87) ; Mab. 159, 160 ; LL. i. 104 ; Triad i. 9 = ii. 8 = iii. 25 
and i. 22 = ii. 43 = iii. 28 ; Diet. Nat. Biog. s.v. Rowlands (Man. Ant. p. 148) 
makes him give his name to Caerhun on the Conway, which is perhaps rightly 
so explained. 


CHAP, proceeding, and is told that this is Rhun ap Maelgwn Gwynedd, 
whose privilege it is that all shall come and take counsel of 
him, not he of others. Rhun and Arthur did not, of course, 
live in the same age ; yet the privileged position assigned to 
Rhun in the story may well be an echo of a real predominance 
held by him as gwledig in succession to his father. The refer- 
ence to his stature seems, also, a historical touch : for the 
poets call him " Rhun Hir " 26 (it will not be forgotten that 
Maelgwn o'ertopped his fellow-princes) and there is a triad, the 
Three Golden-shackled Men of Britain, which makes him out 
to have been of so gigantic a frame that special arrangements 
were necessary for getting him seated on a horse's back. 

One incident only in Rhun's career is recorded, and this 
brings him into relation with the Men of the North. Accord- 
ing to the story preserved in the oldest copy of the Venedo- 
tian version of the Welsh Laws, 27 Elidyr the Bounteous was 
slain, while on a visit to North Wales, by the men of Arfon at 
Aber Meweddus, not far from Clynnog. 28 This brought down 
upon the district a punitive expedition from the North, led by 
Rhydderch Hael, Clydno Eiddin and other chieftains of the 
Cumbrian region, who gave Arfon to fire and sword. How 
Rhun comported himself during the progress of this avenging 
onslaught is not stated, but, when it was over, he organised a 
great counter expedition, which carried the arms of Gwynedd 
as far as the Forth. It was to supply historical grounds for 
the claim of Arfon to lead the van in the Venedotian host that 
this narrative was first put on record, and the tale is duly told 
how the battle stood still on the banks of the northern river 
until a message from Gwynedd ended all strife by an authorita- 
tive decision in favour of the men of Arfon. Nevertheless, in 
spite of some legendary features, the tradition probably rests 
on a basis of fact, and shows, what one is very ready to be- 

My v. Arch. I. 189(140). 

27 Peniarth MS. 29 (Black Book of Chirk), referred to as A. in Owen's edition 
of the Laws. It is of the early part of the thirteenth century. The section 
" Breiniau Arfon " is also found in E., which is a transcript of A. See LL. i. 

28 Eben Fardd drew attention in Cyff Betmo (Tremadog, 1863) to the 
fact that a brook which runs into the Desach from Bron yr Erw is called " yr 
afon Wefus " (p. 66). A link between this and the " Mewedus " of MS. E. of the 
Laws (Camb. Reg. ii. 308) is supplied by " Moweddus," included in a list of the 
possessions of Clynnog church in Rec. Cam. 257. 


lieve, that the struggle with the English was not too absorbing CHAP, 
to allow room for internal conflict. 

There is excellent evidence that these encounters, whether 
with the foreign foe or among the Cymry theinselves, were the 
theme of poets who sang their glories, while they were still 
recent, in the Brythonic tongue. It has been already shown 
that the Welsh tribes retained, in spite of the Roman occupa- 
tion, their native system of poetry and music, owing its origin, 
most likely, to the Druidic discipline. 29 Thus there is no 
difficulty in accepting the statement of the Saxon Genealogies 
that in the age of Ida, i.e., about 550, Talhaearn the "Father 
of Fantasy " was famous in poesy, and that Aneirin, Taliesin, 
Blwchfardd, and Cian, known as the Wheat Singer, were all 
at the same time renowned composers of British verse. 30 
It is true that of the five names thus signalised only the first 
two were known to Welshmen of the Middle Ages, and even 
the fame of Aneirin was not as widely diffused as might 
have been expected, 31 but this tells in favour of the notice 
as really ancient and not the product of the Middle Ages, when 
Welshmen had different ideas as to the great singers of the 
sixth century. At that period the " Cynfeirdd " or " Primi- 
tive Poets " of the Welsh people were understood to be Aneirin, 
Taliesin, Merlin (in Welsh Myrddin) and Llywarch the Aged ; 
Talhaearn, Cian and Blwchfardd had been forgotten, 32 and two 

89 See p. 86. 

30 Hist. Britt. c. 62 : " Tune Talhaern Tataguen in poemate claruit, et Neirin 
et Taliessin et Bluchbard et Cian, qui vocatur Gueinth (read Guenith) Guaut, simul 
uno tempore in poemate Britannico claruerunt ". For explanations of this passage 
see Nenn. V. 103, and Cymr. xi. 135. Three MSS. read " Tatawguen," whence 
the ridiculous " Talhaiarn Tad Tangwn " of lolo MSS. 77 ; cf. 79 and 128. 

31 This is suggested not only by the " et Neirin," for an original aneirin, of 
Harl. MS. 3859, but also by the gross blunder made in " Llyfr Gwyn Rhydderch " 
and afterwards in the Red Book of Hergest in copying Aneurin's title of " mech- 
deyrn beirdd," i.e., prince of bards, which is taken to be "merch Teyrnbeirdd," 
daughter of Teyrnbeirdd! See Cymr. vii. 98-9. The mediaeval form is always 

32 Talhaearn and Cian are both mentioned in the poem called " Angar 
Cyfyndawt " (Myv. Arch. I. 34-5 (35) ; IV. Anc. Bks. ii. p. 130, 131), but it may be 
conjectured that the poet was drawing upon Nennius. Llanfair Talhaearn in Den- 
bighshire is an ancient chapel of the cathedral of St. Asaph, dedicated to the 
Virgin, and owes its name to its position in the township of Talhaearn (Thomas, 
St. Asaph. 386). In the Black Book of Carmarthen the title " Tad awen " is 
given to one " tedei " (fo. 320), from whom is derived the " Tydain Tad awen " 
of the Third Series of Triads (Nos. 57, 92, 93) and Davies's Celtic Researches 
(159, 160) and Mythology (193, 526). 


CHAP, other names had taken their place in the roll of great bards of 
the sixth century, of which it may be said that Myrddin has a 
suspiciously mythical air and that Llywarch Hen was a chieftain 
of whose devotion to bardism there is no satisfactory evidence. 
To what extent the poems ascribed in mediaeval MSS. to 
Taliesin and Aneirin may be regarded as the work of those 
poets is a question which has occupied critics for a century 
and has not yet received a final solution. 33 It is indeed certain 
that not one of these poems has come down in a sixth-century 
dress, and certain, also, that many of them were written in the 
Middle Ages with an eye to political conflicts then raging and 
were merely assigned to an ancient bard, as were the Myrddin 
poems, for stage effect. But it still remains doubtful whether 
some of these warlike strains, in which the setting is Cumbrian 
rather than Welsh, primitive rather than mediaeval, may not 
embody fragments of the older music, fitted to the diction of a 
later age. This is a possibility which has to be seriously con- 
sidered in the case of the principal work attributed to Aneirin, 
the "Gododin". 34 Though it no doubt contains many late 
additions, its principal theme, the ill fortune of the Brython on 
the field of Catraeth, seems to belong to early Cymric history 
and to have nothing to do with the Wales of the Middle Ages. 
Catraeth has not been satisfactorily identified, 35 but Gododin 
or Gododdin has been generally taken to be the country of the 
Otadeni or Votadini, placed by Ptolemy between the Tyne and 
the Forth. 36 The difficulty of interpreting the poem also 
deepens the impression of antiquity it conveys ; it wears the 
aspect of a genuine relic of a long-forgotten strife, a massive 
boulder left high on its rocky perch by an icy stream which has 
long since melted away. 

The men went to Catraeth ; merry was the host. 
The grey mead was their drink and their poison too. 

33 The chief works dealing with the question have been Sharon Turner's 
Vindication of the Genuineness of the Ancient British Poems (London, 1803), 
Stephens' Literature of the Kymry (Llandovery, 1849), Nash's Taliesin (London, 
1858), and Skene's Four Ancient Books of Wales (Edinburgh, 1868). 

34 For the Gododin see Myv. Arch. I. 1-16 (1-20) ; IV. Anc. Bks. i. pp. 374- 
409 ; ii. pp. 62-92, and the editions by John Williams Ab Ithel (Llandovery, 1852) 
and Thomas Stephens (published in parts by the Cymrodorion Society, 1888). 

35 For the philological objections to Catterick see Arth. Legend, 240-1. 

36 W. People, p. 98. 


Such was the melancholy story a tale of valour and high daring CHAP, 
brought to the biting of the dust by lack of self-governance, a 
tale the pitiful recollection of which, though it told of humiliation 
and defeat, the Cymry diligently kept alive through the ages. 


" He who acts as guide to the barbarians, let him do penance 
for thirteen years, that is, if there does not ensue a slaughter of 
Christian folk and the shedding of blood and lamentable cap- 
tivity. When these follow, let the man abandon his arms and 
spend the rest of his life in penance." 87 Such was the temper 
in which the Cymry waged their conflict with the English ; the 
war against the heathen invader was to be a crusade, in which 
no relations were to be permitted with the enemy, and what 
in other wars was common desertion became treason to the 
Christian faith. Briton and Englishman were forbidden to have 
dealings with each other, and the mutual suspicion and hostility 
thus engendered brought it about that no effort was made by 
the older inhabitants of the island even to convert the new- 
comers to their own religion. In the Life of Beuno, a saint who 
belongs to the beginning of the seventh century, it is said that, 
after having been for some time settled at Berriew, near the 
Severn, Beuno one day heard an Englishman's voice on the 
further side of the river egging on his dogs to the chase of a 
hare. Whereupon he turned where he stood and without delay 
went back to his followers, bidding them make up their baggage 
and prepare for instant removal. " For," said he, " the kinsmen 
of yonder strange-tongued man whose voice I heard across the 
river setting on his dogs will obtain possession of this place, and 
it will be theirs, and they will hold it in ownership." 38 The 
life is late and the incident may be fictitious, but the spirit 
breathed in these words was most certainly that which possessed 
the British Church in St. Beuno's day. 39 Where the English- 
man planted his foot, it was held that there was no place for 
missionary effort. 

37 This is one of the canons attributed to the synod of the "Grove of Vic- 
tory " (H. and St. i. 118). 

36 Llyfryr Ancr, 120-1; Cambro-Br. SS. 14-5. " Buchedd Beuno" is as- 
cribed to the thirteenth century (Cymr. xi. 129). 

39 Bede complains of it (H.E. i. 22). 


CHAP. This was the posture of affairs when in 597 a Roman mis- 
VL sion despatched by Pope Gregory I., under the leadership of 
Augustine, head of the convent of St. Andrew's in the imperial 
city, landed in Kent and secured the adhesion to the Christian 
faith of its king, Ethelbert, and the Kentish people. Ethelbert 
was not only king of Kent, but had also made himself overlord 
of all the English kingdoms south of the Humber ; hence, his 
change of religion at once affected a very wide area, quite apart 
from, its importance as the first step in the process of winning 
England for Christianity. After the preliminary difficulties had 
been overcome and the work of the mission set on a firm basis, 
Augustine was certain to avail himself of the far-reaching 
authority of the Kentish king to get into touch with the 
Christians of the unconquered West, and the question of the 
relations which were to exist between him and the British 
Church would call for immediate settlement. 

For 150 years there had been no intercourse between the 
Christians of the British Isles and those of the Continent, 
with the exception of that carried on through Brittany, 
the British colonists of which appear to have had almost 
as little to do with their Prankish neighbours as had the 
insular Britons with their English foes. The state of things 
which now arose was the fruit of this isolation ; the Celtic 
Churches had not shared in the general movement of Western 
or Latin Christianity, but had travelled in a path of their own 
making. Conservative in some respects, they had innovated in 
others, and their successful breaking of new ground, their con- 
version of the heathen Picts, the fame of their seats of learning, 
had given them confidence in themselves and made them little 
disposed to give up at the bidding of an outside authority 
customs and institutions which had come down from the early 
ages of Christianity. Hence the situation was one which re- 
quired careful handling, and it was unfortunate for the cause 
which Gregory had at heart that his representative had none of 
the gifts of a diplomatist, but relied on the authority bestowed 
upon him by the pope and thus assumed from the first that 
supremacy which he should have attained by policy and self- 
restraint. Gregory himself, it should be said, showed little 
appreciation of the true nature of the problem when he handed 
over to Augustine all the bishops of Britain, " so that thou mayst 


teach the unlearned, fortify the weak by thy exhortations, and CHAP, 
by the exercise of thy authority reduce the perverse to obedi- 
ence ". 40 Ecclesiastics who had long been accustomed to the 
fullest independence were not likely to be won over by this rough 
and ready adaptation to their case of the methods of the school- 

There was no insurmountable barrier, it would seem, between 
Augustine and the British bishops. 41 No theological differences 
parted the Roman from the Celtic Church, for the notion that 
the latter was the home of a kind of primitive Protestantism, 
of apostolic purity and simplicity, is without any historical basis. 
Gildas shows clearly enough that the Church to which he be- 
longed held the ideas current at Rome in his day as to the 
sacrifice of the eucharist and the privileged position of the 
priest. 42 The Roman missionaries knew of nothing against the 
Christians of Britain before they landed in the island, but on 
the contrary held them in high esteem for their reputed holiness 
of life, 43 nor is it to be supposed that Augustine would have 
asked them to join him in preaching the gospel to the English 
if he had not known them to be, from the Roman point of view, 
of unquestionable orthodoxy. 44 It was, no doubt, the case that 
they had not been used to acknowledge any special authority 
over other churches as vested in the Bishop of Rome ; in 
the eye of Gildas, 45 every bishop sits in the chair of St. Peter 
and has entrusted to him the keys of the kingdom of heaven. 
Yet this was due to Celtic isolation and not to any anti-Roman 
feeling; the Irish missionary Columbanus, sturdy champion 
though he was of Celtic independence in matters ecclesiastical, 
nevertheless says of the pope "By reason of Christ's twin 
apostles (Peter and Paul), you hold an all but celestial position, 
and Rome is the head of the world's Churches, if exception be 
made of the singular privilege enjoyed by the place of Our 

40 H.E. i. 27 (Plummer, p. 53). 

41 On the whole question see H. and St. i. 152-5 ; Warren, Liturgy and 
Ritual of the Celtic Church, pp. 63-82 ; Zimmer, Celtic Church in Wales and 
Ireland, pp. 60-1 ; Cymr. Trans. 1893-4, 99-i*. I 3-7- 

Williams, Gildas, p. 159. 

43 See the letter of Laurentius and his comrades in H.E. ii. 4. 

44 The views of F. V. Conybeare on this subject (Cymr. Trans. 1897-8, 84- 
117) have been shown to be baseless by H. Williams (Zeit. Celt. Ph. iv. 541-5). 

48 C. 66 (sedem Petri apostoli inmundis pedibus usurpantes). Cf. c. 109. 


CHAP. Lord's resurrection (Jerusalem) ", 46 When this much was con- 
ceded, it was but a short step to the acknowledgment of such 
claims as were put forward by Rome at this early stage in the 
history of the papal power. 

The only extant account of the conferences which took place, 
probably in 602 or 603, between Augustine and the leading 
clergy of Southern Britain, is that contained in the Ecclesi- 
astical History of the Venerable Bede. 4T As a witness, Bede 
labours under the disadvantage of being a warm partisan of the 
Roman against the Celtic party, the breach being still unhealed 
when he wrote in the early part of the following century. But, 
prejudiced as he was, he had the instincts of a historian, and 
his narrative allows us to see pretty plainly the point of view 
of the opposite party. If he tells the story of the blind man 
whom the Britons could not cure, but who forthwith received 
his sight from the Roman envoy, he has also preserved the 
much more interesting anecdote, which he got, no doubt, from 
a British source, of the hermit who, when consulted by his 
fellow-Britons as to the line of action they should follow, bade 
them take their cue from Augustine's own bearing and deport- 
ment if he paid them the courteous attention of rising on their 
approach, let him be submissively heard as a true servant of 
Christ ; if, on the other hand, he kept his seat, in arrogant 
assumption of superiority, let them have nothing to do with 
him. Augustine, according to the story, did not rise, and from 
that moment the spirit of discord and suspicion had the upper 
hand. Whatever measure of historical truth may lie in this 
story, it has undoubtedly symbolic value ; it contains the British 
justification of the refusal to work with the new Archbishop of 
Canterbury, and in recording it Bede gives proof of his honesty 
and diligence as a chronicler of the past, his willingness to 
make use of any material that lay ready to his hand. 

The first conference, arranged with the help of King Ethel- 
bert, took place not far from the estuary of the Severn at a 
spot (probably the modern Aust) which long bore, in memory 

46 Letter to Boniface IV. in Monumenta Germania Historica (new series), 
Epistolae, torn. iii. pp. 174-5. Cf. Warren, Liturgy and Ritual of the Celtic 
Church, pp. 38-9. 

47 H.E. ii. 2. For the date see H. and St. iii. 4, and Plummer's Bede, ii. 


of the event, the name of Augustine's Oak. 48 Certain bishops CHAP, 
and divines attended on behalf of the Britons, and, having 
listened to the demands of Augustine, asked that the matter 
should be adjourned for the consideration of a fuller and more 
representative assembly. This was duly arranged; to the 
second conference there came seven British bishops and many 
learned men, especially, says Bede, from the great monastery 
of Bangor, presided over at that time by abbot Dunod. The 
reference is undoubtedly to Bangor on the Dee, where a monas- 
tery had been established in the middle of the sixth century by 
Deiniol or Daniel, the founder of the original Bangor on the 
Menai Straits. 49 It seems likely that the conference was held 
not far off, and the " Sinodus urbis legion " assigned by 
Harleian MS. 3859 to the year 60 1 60 may well refer to some 
clerical assembly got together at Chester in connection with 
these negotiations. It would be of the utmost interest to know 
who were the leading figures on the British side, but the 
attempt even to fix the sees of the seven bishops who met 
Augustine is a hopeless one, 51 and all that is certainly known 
is that the conferences were a failure. According to Bede's 
account, Augustine expressed his readiness to overlook many 
peculiarities of the Celtic Churches which were contrary to uni- 
versal Christian custom, if only they would in three respects 
make a fresh beginning ; let them adopt the Roman calculations 
for fixing the Easter of each year, "complete" the ordinance 
of baptism as was done at Rome, and join him in preaching the 
gospel to the English, and all would be well. But to none of 
these things would the Britons consent, nor, adds the narrative, 
would they accept Augustine as archbishop. 

Throughout the period of severance between the Celtic and 
the Roman Churches great stress was laid on the divergence with 

48 Augustinaes Ac " is not found out of Bede and no identification can 
therefore be confidently put forward. But it is in favour of Aust, the Austreclive 
of Domesday (i. 1646, 2), that it was known in Welsh as Penrhyn Awstin (Triad 
i. 30 = ii. 56) and that it appears as "act Austin" in the charter of 692 or 693 
which bestowed it on the see of Worcester (Cod. Dipl. i. 35). Green's objection 
(Mak. Eng, p. 224) disappears if we take Bede to be speaking of the boundaries 
of his own day. 

49 See note appended to this chapter. 

50 Cymr. ix. 156. The ascription of the death of Gregory to the same year 
is possibly an inference from Bede's " interea " at the beginning of H.E. ii. 2. 

81 H. and St. i. 148 and iii. 41 ; Plummer's Bede, ii. p. 75. 


CHAP, regard to Easter. Like the question of the proper form oi 
clerical tonsure, which Augustine does not seem to have raised, 
though it was keenly debated between the two parties in later 
times, 52 it forced itself on men's notice as a visible sign of dis- 
cord, since the sudden transition from the gloom of Holy Week 
to the rejoicings of the Day of Resurrection was an event in 
the Christian year to catch the attention of the most careless, 
and to see one Christian still keeping the Lenten fast while 
another by his side was in the midst of the Easter revels 53 
brought out in the clearest fashion how far they were from 
dwelling as brethren in unity together. Nevertheless, the 
divergence was one which did not issue from any theological 
principle, but was due entirely, as has already been suggested, 
to the long separation between the Celtic Christians and those 
of the Continent. From an early period it had been agreed to 
celebrate our Lord's Resurrection, not, like His Nativity, on a 
fixed date, but at the season of the Jewish Passover, historically 
so closely connected with it. The Passover has always been 
observed on the day of full moon (known as the fourteenth) of 
the first month of the Jewish ecclesiastical year, that month 
being the one which coincides with the spring equinox. When 
it was further decided that the Easter festival should always be 
held on Sunday, itself a weekly celebration of the Resurrection, 
and not on the day which happened to be that of the paschal 
full moon, the elements of a complicated problem had been got 
together, and for many centuries no agreement as to a uniform 
system seemed possible. 54 The Celtic Churches accepted the 
principles stated above ; their Easter was not a fixed date, nor 
did it ever fall on any other day than a Sunday. But the rules 
which they followed in determining which Sunday should be 
Easter Day in any particular year were those which had been 
current in Rome at the beginning of the fourth century, while 
in the meantime at Rome itself changes had been successively 
adopted which had produced an entirely new system. In this 
way it came about that, as the result of one rule, the Celtic 

52 Plummer's Bede, ii. pp. 353-4. 

53 This occurred in the royal household of Northumbria under Oswy (H.E. 
iii. 25). 

54 For a full and lucid account of the technical questions involved see A. 
Giry, Manuel de Diplomatique (Paris, 1894), pp. 141-54. The Celts, as Zimmer 
says (Celtic Church, pp. no-i), followed the older " supputatio Romana". 


Easter often anticipated the Roman by a week, while occasion- CHAP, 
ally, through the operation of another, it would fall no less than 
four weeks later. Among the Celts 25th March was the earli- 
est possible Easter Day, 2 1 st April the latest, while at Rome 
the range of oscillation was from 22nd March to 25th April. As 
a result of these conflicting calculations, it was the exception 
for the two Easters to coincide. 66 

There is considerable doubt as to the meaning of Augus- 
tine's second demand, with reference to the " completion " of 
baptism. It is most often understood as implying the absence 
of the rite of confirmation 56 or some defect in the manner in 
which confirmation was carried out. 57 As, however, this par- 
ticular element of discord is not elsewhere touched upon, it was 
clearly not one of the first magnitude. The same cannot be 
said of the British refusal to accept Augustine's third condition, 
namely, that Briton and Roman should join hands in the 
great undertaking of the evangelisation of the English king- 
doms. It is, unfortunately, nowhere stated in express terms 
why this task was declined, so that the grounds of the refusal 
can only . be conjectured. Race hatred, the fruit of a century 
and a half of race conflict, will no doubt supply a partial ex- 
planation, as it explains the fact that nothing had yet been 
done by the Britons themselves in this direction. But the con- 
viction is forced upon one that this was not the sole reason, 
and it is a fair inference from the narrative of Bede that the 
claim of Augustine to exercise ecclesiastical supremacy over 
the whole island was the real stumbling-block. Where pride 
of race told was in indignation at the thought that the British 
Churches, the origin of which lay far back in a distant past, were 
to be disposed and ordered at the will of a mere missioner to 
the English, living among these Christians of yesterday and 
making their interests at all times his first consideration. The 
Britons may even have pictured to themselves the spectacle, 
which was actually witnessed in little more than fifty years, of 
an Englishman seated in the chair which it was sought to invest 
with such dignity and authority. 

The breach having been once made, there is no doubt that 
the Britons did their best to keep it open. They made use 

55 See the tables of Giry, pp. 187-8, 212-3. 

88 Plummer's Bede, ii. pp. 75-6. " Trans. Cymr. 1893-4, 103-6. 

VOL. I. 12 


CHAP, against those who conformed to the Roman system of that 
weapon of excommunication which had already been found by 
the Celts a far too handy resource in their domestic disputes 
about monasticism. 58 "It is to this day," writes Bede in 731, 
" the fashion among the Britons to reckon the faith and religion 
of Englishmen as naught and to hold no more converse with 
them than with the heathen." 59 This had begun as early as 
the time of Laurentius, the next successor of Augustine at 
Canterbury, who complains that one Dagan, an Irish bishop, 
had not only refused to eat at his table, but would not take his 
food anywhere under the roof which sheltered him. 60 In the 
days of Aldhelm, who died in 709, the Britons would make no 
use of pots and pans which had served for a Saxon meal until 
they had been thoroughly cleansed and scoured. 61 It was a 
losing battle which the Celtic Churches were fighting, since they 
could not hope to maintain their traditions, which were tradi- 
tions merely and represented no great principle, against the 
growing influence of Rome, yet it was fought none the less 
bitterly for that reason. 


At the beginning of the seventh century the English attack 
upon the Cymry became, in the hands of yEthelfrith, a most 
threatening movement For the first twelve years of his reign, 
viz., from 592 to 604, he was, if the author of the Saxon Gene- 
alogies is to be trusted, 62 king of Bernicia only, his southern 
border being the Tees, or possibly the Tyne ; 63 in the latter 
year he annexed the neighbouring English kingdom of the 
Deiri (known in Welsh literature as Deifr), 64 expelling from 

58 See the fragments ascribed to Gildas (H. and St. i. 108-12). 

59 H.E. ii. 20. Ibid. 4. 61 H. and St. iii. 271. 

62 " Eadfered Flesaurs (the epithet has not been explained) regnavit duodecim 
annis in Berneich et alios duodecim in Deur ; viginti quatuor annis (so H.E. i. 
34) inter duo regna regnavit " (Hist. Britt. c. 63). It is true that the acceptance 
of this chronology involves the abandonment of the usual view, based on entries 
in the Anglo-Saxon Chronicle, that Deira was annexed by Ethelric, ^Ethelfrith's 
father, on the death of JElle as early as 588. But a much better authority than 
the Chronicle, -viz., Bede's short chronicle (Man. Hist. Br. 96), speaks of the Eng- 
lish north of the Humber as under the rule of Kings /Elle and ./Ethelfrith in the 
days when Kent was receiving the gospel, so that there is no difficulty in accept- 
ing the statement that the former lived until 604. 

63 For the limits of Bernicia southward see Plummer's Bede, ii. p. 120. 

64 Rhys (Celt. Br. (2), p. 291) connects Deifr with the Welsh " deifr," 


it the young Edwin, son of its former king, ylle. He now CHAP. 
held an exceedingly strong position, and, though the predomi- 
nance of Ethelbert prevented him from exercising much influ- 
ence south of the Humber, north of that river he was without 
a rival. In 603 he had been attacked by Aidan, king of the 
important Irish or " Scottish " colony which had established 
itself in Argyll, with a large army which probably included 
a contingent of Cymry from the region of the Clyde ; but Aidan 
sustained a crushing defeat at Degsastan (The Stone of Degsa), 86 
probably Dawston at the head of Liddesdale. atJ " From that 
day to this," says Bede, " no king of the Scots dwelling in 
Britain has dared to take the field against the English race." 
It was probably the distinction he achieved through this victory 
which emboldened ^Ethelfrith soon after to lay his hands upon 
Deira, and it resulted from this further acquisition that he was 
brought into touch with the southern Cymry, with the men of 
Gwynedd and of Powys and the dwellers along the Mersey and 
the Ribble. 67 Thereupon opened the second stage in the 
relations between the English and the Cymry that which led 
in the space of about fifty years to the final separation of the 
northern from the southern section of the defeated nation and 
thus set Wales and Strathclyde travelling their several ways. 
By what route ^Ethelfrith pushed forward to Chester there 
is nothing to show. But it was near this city that, about the 
year 6i5, 88 he won his most famous victory over the Britons. 
Bede gives in his Ecclesiastical History 69 some account of 

waters, and assumes an original Debria or Dobria. It is strange, if this be so, 
that the labial is not represented in the early forms Deur (Sax. Gen.), Deiri and 
Deri (Bede). 

88 H.E. i. 34. IV. Anc. Bks. i. 177. 

87 In lolo MSS. 86, " Teyrnllwg " is said to be the ancient name of the 
region between the Dee and the Cumbrian Derwent. But this is probably a mere 
inference from the name Cadell Deyrnllwg, which in its oldest form, as given in 
Hist. Britt. c. 35 (ed. Mommsen, p. 176), is " Catell Durnluc" (the Blackhanded). 
See Cymr. vii. 119. 

88 Bede gives no date, but says incidentally that Augustine (ob. 604 or 605) 
had died " multo ante tempore ". This rules out the conjectures of the Anglo- 
Saxon Chronicles (605, 607) and points to the period 610-6. Harl. MS. 3859 
assigns " Gueith cair legion " to 613 (Cymr. ix. 156), which is also the year im- 
plied in Ann. Ult. s.a. 612. The notice in Tighernach is as follows : " K. ui 
(which, if correct, would be 6n or 616) . . Cath (the battle of) Caire Legion ubi 
sancti (i.e., the monks) occissi sunt et cecidit Solon mac Conain rex Bretanorurr 
et Cetula rex cecidit. Etalfraidh uictor erat, qui post statim obit " (Rev. Celt. 
xvii. pp. 170-1). 69 H.E. ii. 2. 

12 * 


CHAP, the battle ; unhappily, he was only interested in what may be 
regarded as an accident of the struggle. It was his purpose to 
show that on this occasion a prophecy uttered many years be- 
fore by Augustine was fulfilled, to the effect that, if the British 
clergy would not join him in preaching to the heathen English, 
they would assuredly be the victims of their barbaric rage. 
Accordingly he tells the familiar story of the appearance on 
the battle-field, after a three days' fast, of many hundreds of 
monks from the not far distant monastery of Bangor, who were 
stationed in what was supposed to be a place of safety under the 
protection of one Brochwel, 70 and from that post of vantage 
began to implore the blessing of God upon the arms of their 
fellow-countrymen. ^Ethelfrith asked the -meaning of this 
strange spectacle, and, on being told, bade his troops forthwith 
carry by storm the citadel of prayer ; " for," said he, " if they 
cry to their God against us, they fight against us as surely as 
do those who bear weapons ". The command was obeyed ; 
Brochwel and his men fled at the first onset, and no less than 
twelve hundred " saints " are said to have been put to the 
sword. Such was the massacre of the monks of Bangor a 
piteous tragedy, yet one which beyond a doubt had often been 
paralleled in the relentless warfare between the pagan invader 
and his Christian foe. As to the main issues of the battle, 
Bede has nothing to say, except that yEthelfrith's decisive vic- 
tory was not won without considerable loss on his own side. 71 
This disposes of the view, which is also, it may be added, in- 
consistent with the later history of Cheshire, 72 that the battle 
of Chester was at once followed by a Northumbrian occupation 
of the plain around the city. 73 Genuine as the victory was, 
it was pretty certainly not one to have results of this kind : 
^Ethelfrith withdrew from the district, and shortly after, in 616, 

70 " Brocmail " can hardly be Brochwel Ysgythrog, ruler of Powys, for his 
grandson, Selyf ap Cynan, was slain in this very battle. Nor is it likely he is 
the " Brocmail " of the year 662 in Harl. MS. 3859 (Cymr. ix. 158). The name 
was, in fact, a very common one; see Cymr. ix. 177, 178, 179, 181, 182, and, for 
the early form " Brohomagli," Inscr. Chr. No. 158, Lap. W. 202, W. Phil. (2), 


71 " Non sine magno sui exercitus damno." 

72 Cymr. x. 23 (A. N. Palmer). The dialect of Cheshire is Mercian in its 
affinities, not Northumbrian ; see Darlington, Folkspecch of South Cheshire 
(Eng. Dialect Soc., London, 1887). 

73 Mak. Eng. pp. 242-5 and map. 


met his death on the banks of the Idle in battle against Raed- CHAP. 
wald of East Anglia. 74 

So far as can be seen, the leader of the Cymry in the battle 
of Chester was Selyf (or Solomon), son of Cynan of the White 
Car son of Brochwel of the Tusks, who, as representative of the 
ancient line of kings of Powys, was the natural defender of 
the valley of the Dee. 75 The Welsh and the Irish notices of the 
battle name him as the most notable among the slain, and one 
of the Triads reckons him among the British heroes who were 
avenged from their graves ; 7<t this may be a reference to the 
mythical British victory, which, according to Geoffrey of Mon- 
mouth, almost immediately wiped out the disgrace of the day of 
Chester, 77 or it may be merely an allusion to the fact that 
^Ethelfrith lived but a short time to enjoy his triumph. There 
is no evidence that the forces of Gwynedd took any part in the 
battle, for, though the chronicle in Harl. MS. 3859 assigns to 
the same year the " falling asleep " of lago (Jacob), son of Beli 78 
and grandson of Rhun ap Maelgwn, 79 it does not connect the 
event with its notice of " Gwaith Caerlleon " ; moreover, " dor- 
mitatio " is almost always used of the death of an ecclesiastic 
and suggests that lago, if at any time king of Gwynedd, had by 
this time resigned that office and withdrawn to the quiet of a 
monastery. 80 Thus the ruler of Mon and Arfon and the in- 
heritor of the claims of Maelgwn Gwynedd at the time was 
probably Cadfan, who appears in the pedigrees as the son of 

74 H.E. ii. 12. As to the date, the year 616 appears to be the one which is 
required by the figures of Bede ; see H.E. i. 34 ; ii. 14 ; and especially ii. 20. It 
is also the year implied in the old list of Northumbrian kings (Mon. Hist. Br. 
go). That MS. E. of the Anglo-Saxon Chronicle has 617 is of no importance, in 
the light of the evidence drawn from older and better authorities. 

76 For the pedigree of Selyf see Cymr. ix. 179 (Harl. MS. 3859) ; viii. 87 
(Jesus Coll. MS. 2o(Evans, Rep. ii. p. 31)). " Garwyn" appears in the Old- Welsh 
form ' Carguinn " in Cambro-Br. SS. 79, as corrected in Cymr. xiii. 80; cf. 
also the "garrvin" of Blk. Bk. fo. 146. " Ysgythrog" is from " ysgythr," a 
tusk, for which see LL. i. 312, Mab. 122, 135. 

79 Triad i. 65 = ii. 30 = iii. 76. 

77 Hist. Reg. xi. 13. Palmer thinks the second battle may be historical and 
may underlie the " Gwaith Perllan Fangor " of Triad i. 66-7 = ii. 38 (Cymr. x. 


78 Cymr. ix. 156. Ibid. 170. 

80 Triad i. 39 = ii. 29 = iii. 48 says, indeed, that one of his own men cleft 
his skull with an axe, but this may be due to confusion with lago ab Idwal, who 
was killed " a suis " in 1039. lago ap Beli was reputed a benefactor of the 
cathedral church of Bangor (Browne Willis, Survey, 184). 


CHAP. lago ap Beli. 81 Nothing is known of this king from any ancient 
' literary source, 82 but the caprice of time, which has overwhelmed 
so many other memorials of vastly greater interest, has spared 
us his tombstone. 83 After long serving as the lintel of the 
south door, it is now within the church of Llangadwaladr in 
Anglesey, and its inscription reads " Catamanus rex sapien- 
tisimus opinatisimus (most renowned) omnium regum," language 
which is reasonably interpreted to mean that he claimed, as 
hereditary " gwledig," a primacy among the chieftains of the 
Cymry. 84 As the foundation of the church is traditionally 
attributed to Cadfan's grandson, Cadwaladr, 85 the inscription may 
not actually date from the year of the king's burial, but the 
form of the name 86 and the characters employed point pretty 
clearly to the seventh century. 

If Cadfan is to be reckoned among the obscurer personal- 
ities of Welsh history, his son and successor Cadwallon 87 holds 
a place in the forefront of those who have earned the grateful 
remembrance of the Welsh race by vigorous championship of 
the national cause. The memory of his great duel with Edwin 
of Northumbria, carried on with marked ill fortune for many 
years, but ending in the defeat and death of the English king, 
deeply impressed itself on the minds of his fellow-countrymen, 
so that Edwin became the typical English antagonist, 88 and every 
bold defender of the freedom of Wales was hailed as a new 
Cadwallon. 89 This was, indeed, the last great struggle between 

81 Cymr. ix. 170. Though this is fairly good authority, the number of 
names (five) in the pedigree between Maelgwn (ob. circa 550) and Cadwaladr 
(ob. 664) makes one a little sceptical. 

82 Geoffrey of Monmouth (Hist. Reg. xi. 13 ; xii. i) is here, as elsewhere, a 
mere romancer. 

83 Mow. Ant. 156(157); Arch. Camb. I. i. (1846), 165-7; Inscr. Chr. No. 
149; W. Phil, (2), 160, 364; Lap. W. 190-1. 

s *Celt. Br. (2), pp. 127-8. 85 Myv. Arch. II. 33 (421). 

86 Brythonic " Catamanus," for an earlier " Catwmanus," is on its way to 
become the old Welsh " Catman ". 

87 For the pedigree see Cymr. ix. 170. Bede's form " Caedualla " is due to 
the influence of the name of Ceadwalla of Wessex. 

88 See Myv. Arch. I. 194 (143), where a twelfth-century militant bard says 

Gwalchmai y'm gelwir, gal Edwin ac Eingl. 

(" Gwalchmai am I called, a foe to Edwin and every Englishman.") 
Cf. Elk. Bk. fo. 240 (IV. Anc. Bks. ii. 17) a poem of about the same date. 

89 See the poem from the Red Book of Hergest in Myv. Arch. I. 121-2 (96- 
7) ; IV. Anc. Bks. ii. 277-9, which has no relation to the battles fought by the 
Cadwallon ap Cadfan of history and must commemorate the deeds of some 
mediaeval prince. 


Briton and Englishman for supremacy in the island, and the CHAP, 
overthrow of Edwin for a brief space raised hopes that Britain 
might yet be snatched from the grasp of the Teutonic conqueror. 
The fall of Cadwallon a year later scattered these hopes to the 
winds, and, though the contest with Northumbria was not 
abandoned, it had henceforth little prospect of success, and on 
the death of Penda, who furnished it with Mercian support, it 
came suddenly to an end. 

The death of ^thelfrith in 616 had brought about a dynastic 
revolution in Northumbria. Power was seized by Edwin, the 
representative of the royal line dislodged from Deira in 604, 
and thus, while the two kingdoms still remained under one head, 
it was the southern and not the northern realm which now 
wielded supremacy. Edwin had spent his youth he was now 
thirty-one in exile, fleeing from court to court to escape the 
ruthless enmity of ^Ethelfrith ; during his wanderings he had 
lived in East Anglia, and also, if we may accept the evidence 
of the Triads, in Gwynedd, for his name is included in a trio of 
" Three Chief Oppressors of M6n, nurtured within the island ". 90 
He soon showed the energy and resolution to be expected from 
one who had been trained in the austere school of adversity. 
Attacking the British kingdom of Elmet, or Elfed, as it would 
now be written, which lay around our Leeds, he completely 
subdued it and drove King Ceredig from his throne. 91 By this 
conquest the chief barrier which parted Deira from the Irish 
Sea was removed, and very shortly afterwards Edwin must have 
effected that breach between the Cymry of the North and those 
of Wales which the battle of Chester foreshadowed, but did not 
actually bring about. His relations with the other English 
kingdoms, over all of which save Kent he established his ascend- 
ancy, his acceptance of Christianity at the instance of a Roman 
missionary from Canterbury, and his assumption of something 
like imperial state invest his short reign with great interest, but 
for the historian of Wales his most notable achievement was 
his conquest of what Bede calls the Mevanian islands, lying 

90 I. 81 = ii. 56 (end). Geoffrey of Monmouth's account (Hist. Reg. xii. i) 
of the nurture of Edwin and Cadwallon together no doubt rests on this tradition. 

91 Hist. Britt. c. 63; Mak. Eng. pp. 253-7. F r Ceredig see H.E. iv. 21 
(" rege Brettonum Cerdice ") ; he may be the " Ceretic " of Harl. MS. 3859, 5.0. 
616 (Cymr. ix. 157), but in that case the year is most probably wrong. 


CHAP, between Britain and Ireland, 92 in other words, of Anglesey and 
Man. Such a conquest implies the equipment of a fleet, which 
was probably fitted out at Chester ; 93 not only was this necessary 
for any operations against the Isle of Man, but it was repeatedly 
shown under the Norman and early Plantagenet kings that 
nothing could be effected against Anglesey without naval assist- 
ance, so strong were the natural defences of the island on the 
landward side. With a fleet in possession of the Irish Sea, 
troops flushed with victory over the Britons of Elmet, and 
borders secured from attack by the greatness of his name, Edwin 
entered upon what was the first English invasion of Wales with 
notable advantages in his favour. 

It may be judged from Cadwallon's later history that there 
was no lack of spirit in the defence. But tradition has nothing 
to say of his share in the campaign, and only commemorates 
the valour of one Belyn of Lleyn, who is described as fighting 
Edwin with his " teulu " or warband at " Bryn Edwin " in 
Rhos, 94 and also at Erethlyn, near Eglwysfach, in the same 
region. 95 Rhos, lying as it does between the Clwyd and the 
Con way, was just the region in which to oppose the progress 
of any expedition making its way to Anglesey, and the Hill of 
Edwin may be that known at present under the slightly altered 
form of Bryn yr Odyn, not far from Llanelian. The brave 
stand of Belyn and his doughty followers was, however, made in 
vain, for the struggle closed, after having been waged apparently 
for some years, with the retreat of Cadwallon to the little island 
of Priestholm or Ynys Lannog, off the coast of Anglesey, where 

92 H.E. ii. 5, g. The description given of the two islands in c. g shows 
clearly that Anglesey is meant to be one, but Bede had no warrant for extending 
the name " Mevania " to this island, which is always Mona and Mon in the 
ancient writers. As for " Meuania," it is found in certain authors, e.g., Orosius 
(Adv. Pag. I. ii. 82), as a name of the Isle of Man, being probably a mis ead- 
ing of " Menauia," which again is the Celtic " Manavia " or Manaw (Celt. 
Heath, pp. 663-4). 

93 Mak. Eng. p. 257. 

94 Triad i. 49 = ii. 40 = iii. 27. That Belyn belongs to history and not 
merely to fable is shown by the " Belin moritur " of Harl. MS. 3859, s.a. 627 
(Cymr. ix. 157). 

98 Triad ii. 19, in which " yn " must be read before Erethlyn (Erythlyn in 
the Red Book). The name appears in the old one-inch Ordnance map as 
" Hiraethlyn " ; for the better form " Pennant Ereithlyn " see Thomas, St. 
Asaph, p. 538. 

he was hemmed in by Edwin's fleet. 96 It was now, in all CHAP. 


probability, that the flight to Ireland took place, 97 which is 
vouched for by tradition 98 and which must have made Edwin's 
triumph for the moment complete. 99 

But ere long the wheel of fortune took a sudden turn. 
Cadwallon, on his return to Wales, entered into an alliance with 
Penda, who had stepped forward as the leader of the Mercians, 
the English settlers in the basin of the upper Trent. 100 The 
motives of Penda are not difficult to discern ; he was a pagan 
and remained until his death the chief upholder of heathenism 
among the English ; he resented as a Mercian the ascendancy 
of Northumbria. There is more to wonder at in the attitude 
of Cadwallon, for never before, so far as can be seen, had Briton 
and Englishman made common cause in any quarrel that had 
arisen in the island. The British king had, however, realised, 
in the light of recent events, that his first concern must be to 

99 " Obsessio Catguollaun regis in insula glannauc " Had. .MS. 3859, s.a. 
629 (Cymr. ix. 157). As the fall of Edwin is assigned to 630, the year is prob- 
ably 632. 

97 Celt. Br. (2), p. 131. 

"Triad i. 34, which says, however, that he was an exile for seven years. 
Geoff. Mon., in deference to the tradition, takes Cadwallon to Ireland (Hist. Reg. 
xii. 4), but soon moves him on to Brittany (" Armoricam," whence the " Ar- 
mowicam " of Reginald's life of St. Oswald Sim. Dun. i. 345 cf. 350), so that 
he may be restored by Breton help. A poem in the Book of Taliesin (IV. Anc. 
Bks. ii. p. 206 ; Myv. Arch. I. 74 (62)) has a reference to the story 
Pan dyfu gatwallawn Dros eigyawn iwerdon. 
(" When Cadwallon returns o'er the Irish sea ") 

but the poet is no doubt using the old tradition in the interests of some prince of 
his own day. 

99 Wm. Malm, was the first to suggest that Anglesey was so called in con- 
sequence of Edwin's reduction of it, the name being really " isle of Angles or 
English " (Gesta Regum, ed. Hardy, i. 69). There are many objections to this 
derivation. One fails to see why the very brief occupation under Edwin should 
have led the English ever afterwards to speak of this as by pre-eminence their 
island. Further, the name does not make its appearance until the eleventh 
century ; Alfred's version of H.E. ii. 5, 9, has " Monige " for the Mevanian islands 
(Plummer's Bede, ii. p. 94) and MSS. C. and D. of the Anglo-Saxon Chronicle 
have s.a. luoo the same form for Anglesey (Mon. Hist. Br. 407). " Angles ege " 
is first found in MS. E. s.a. 1098. Philological difficulties have also been pointed 
out (Academy, 2nd June, 1894, p. 458) and altogether, the derivation proposed by 
Dr. H. Bradley and supported by Mr. W. H. Stevenson (Descr. Pemb. i. 322-3) 
is much to be preferred. Starting from the form Ongulsey found in the Orkneyinga 
Saga (Rolls ed., p. 70), they connect the name with the Norse ongull, a fiord, 
and interpret as " the island of the strait ". 

100 H.E. ii. 20. According to some Welsh pedigrees, Cadwallon married 
Penda's sister (Bye-Gones, ist Oct. 1890, p. 480). 


CHAP, break the power of Edwin, and that in the furtherance of this 
purpose he must not be over scrupulous as to the means he 
employed He had no reason to dread the triumph of Mercia, 
a state as yet in its infancy, and Penda's religion was no 
stumbling-block to one who did not regard the English Chris- 
tians, followers of the ways of Rome, as brethren in the faith. 
In 633 Cadwallon and Penda met Edwin in battle, defeated and 
slew him, and for a time had Northumbria at their feet. The 
scene of this memorable encounter cannot, unfortunately, be 
fixed with certainty. Bede gives the name of the place as 
Haethfelth and conveys the impression that it was in Deira, or 
not far distant from its borders ; 101 accordingly, Hatfield, near 
Doncaster, has been a popular identification. 102 The Saxon 
Genealogies, on the other hand, speak of this battle as " bellum 
Meicen," 103 and later Welsh traditions connect it with the 
Meigen which is known to have lain somewhere near the 
Breiddin, on the borders of Montgomeryshire and Shropshire. 104 
On the whole, it is most likely that the battle was fought in the 
north, at a spot known to the Cymry as Meigen, and that this 
was subsequently confused in Welsh literature with the more 
familiar Meigen on the confines of Powys. 

During the year which followed the overthrow of Edwin, 
Northumbria was entirely at the mercy of the British king 
and his Mercian ally. 105 It was treated as a conquered country 
and pitilessly ravaged ; no respect was paid by either of the 
two victors to the nascent Christianity of the district. Cadwallon 
was the dominant spirit, and it is clear that his policy was 

101 See the reference in H.E. ii. 14 (end) to ravages which followed the 
fight. The Berne MS. of Hist. Reg. xii. 8 has, it should be noted, "hedfeld " 
and not " Hevenfeld," as in Giles's edition. Geoffrey knows only the English 
names of the battle-fields of this period. 

102 Smith in Mon. Hist. Br. 171 ; Mak. Eng. pp. 269-71 ; Plummer's Bede, 
ii. pp. 115-6. 

103 Hist. Britt. c. 61, which is the source of the entry in the chronicle in 
Harl. MS. 3859, s.a. 630. For an explanation of the " Meiceren " of the Rolls 
editor (Annales Cambria, ed. J. Williams Ab Ithel, 1860, p. 7), see Cymr. ix. 
157 ; xi. 147. 

104 Triad i. 63 = ii. 15 suggests that Meigen was in Powys, and the poem 
styled " Marwnad Cadwallon" associates it with the Severn and Dygen (IV. 
Anc. Bks. ii. p. 277; Myv. Arch. I. 121 (97)), the latter being in full Dygen 
Freiddin (Myv. Arch. I. 193 (142)). According to lolo MSS. 18, there was a 
place of the name (" Meigen Cil Ceincoed "), on the banks of the Rhymni. 

105 H.E. ii. 20; iii. i. 


purely destructive ; his aim was not to subjugate Northumbria, CHAP, 
but to ruin it. On the death of Edwin, Bernicia and Deira 
had again become separate kingdoms ; a son of yEthelfrith's, 
Eanfrith, had seized the crown of the northern, a cousin of 
Edwin's, Osric, that of the southern state. Cadwallon set him- 
self to make an end of both rulers. He defeated and slew 
Osric in the summer of 634, finding his opportunity when that 
prince endeavoured to shut him and his army within the circuit 
of a walled town ; 106 in the autumn of the same year he lured 
Eanfrith to his camp and despatched him also. With a very 
large army, to which had flocked, no doubt, the Cymry alike 
of the north and of the south, in eager expectation of a final 
triumph over the Saxon foe, he planted himself in the heart of 
Northumbria, on the hills that slope northwards towards Hex- 
ham and the Tyne. 

A second son of ^Ethelfrith's, named Oswald, had, however, 
at once stepped into the place of his brother Eanfrith, and, with 
a small body of troops on which he could thoroughly rely, 
marched south to contest with Cadwallon the supremacy of 
Bernicia. The two brothers had spent the reign of Edwin in 
exile among the Scots of Argyll and had there been imbued 
with the principles of the Christian faith, which they received, 
of course, in their Celtic form. Eanfrith's creed was, it would 
seem, but lightly held, for he abandoned it on his accession ; 
Oswald, on the other hand, had embraced the new religion 
with all the earnestness of a singularly noble character, and 
the tale was long told how, on the day before the battle which 
he fought with Cadwallon, he had set up a rough wooden cross, 
the first ever seen among the English of those parts, and had 
knelt at its foot with his soldiery to pray for victory in the 
coming struggle. The spot bore the name of Heavenfield ; 
it lay close to the Roman Wall, not far from the point where 
this is cut by the North Tyne, and the devotion of later ages, 
which held the place to be one of the most sacred in North- 
umbria, raised there the chapel of St. Oswald's, which marks 
it to this day. 107 The following night Oswald's army resumed 

106 Often supposed to be York, but Bede obviously did not know its name. 

107 H.E. iii. 2. There is nothing in the narrative of Bede to suggest that 
there was any fighting at " Hefenfelth " or St. Oswald's. On the contrary, he 
states explicitly that Cadwallon and his army were overwhelmed at " Denises- 


CHAP, its march and at daybreak surprised the host of Cadwallon, as 
it lay encamped some 10 miles further south, on the banks of 
a stream now known as Rowley Water. In spite of the dis- 
parity of numbers, the rout of the British was complete and 
the death of Cadwallon made it irretrievable defeat. 108 

The year of Cadwallon's ascendancy in the North showed 
that, though the courage of the Cymry ran high, they did not 
possess the secret of rule. They failed to follow up their 
victory in the field by any measures which might incline the 
defeated Northumbrians to accept their overlordship, and 
accordingly the question whether they would reconquer the 
island or be driven into the highland regions of the West was 
really settled by the events of this year. Some time went by 
ere the contest with Northumbria was abandoned, but its issue 
was henceforth certain. 

Oswald established himself as king of the whole of North- 
umbria, and, according to the testimony of Bede, 109 was not 
inferior in the extent of his power to Edwin himself. But no 
such struggle with the Cymry as marked his predecessor's reign 
is coupled either by history or by tradition with his name, and 
nothing is known of his relations with the successor of Cad- 
wallon, if indeed any Welsh prince was able at this time to 
assert his claim to the office of " gwledig ". His chief enemy 
was Penda of Mercia, and it was Penda who, in 642, when he 
had ruled over the Northumbrians for some eight years, at- 
tacked and slew him in the battle of Maserfeld. 110 There is 

burna," which has been shown to be several miles to the south (Plummer's Bede, 
ii. pp. 122-3). This harmonises well with the fact, to which Bede and an equally 
good authority, Adamnan (Vita Sancti Columbtz, i. i), bear witness, that Os- 
wald, with a much smaller army, won his victory over the thousands of Cad- 
wallon by means of a night march, followed by an attack at dawn. It was, says 
Adamnan, " felix etfacilis . . . victoria". 

108 In the Saxon Genealogies the battle is called " bellum Catscaul " (Hist. 
Britt. c. 64), whence the " cawtscaul " of Harl. MS. 3859,5.0. 631 (Cymr. ix. 
157). Skene's derivation of the name from " cad ys guaul," the battle at the 
Wall (Celtic Scotland, 1876, i. p. 245), is quite impossible, not to speak of the high 
probability that the battle was not fought at the Wall at all. The recent habit 
of dating it 635, instead of 634, is due to a misunderstanding of Bede (Plummer's 
Bede, ii. p. 121). 

109 H.E. ii. 5 (" sextus Oswald . . . hisdem finibus regnum tenuit "). 

110 Ibid. ii. 9-13 ; for the date see v. 24. The Sax. Gen. calls this " bellum 
Cocboy" (Hist. Britt. c. 65), a name which cannot be used either for or against 
the Oswestry identification. 


no evidence that the Cymry had any part or lot in this battle, CHAP, 
though it would seem probable that it was fought not far from 
their borders. For, while no ancient authority furnishes any 
hint as to the situation of the battle-field, a local tradition, 111 
which was in existence at the time of the Norman Conquest, 
fixed it at Oswestry, the Oswald's Tree (in Welsh " Croes 
Oswallt," i.e., Oswald's Cross), from which the place derived its 
name, being taken to be the wooden post or stock on which 
was set by Penda's orders the head of the fallen king. When 
Oswald came to be regarded as an English saint and martyr, 
a church was raised on the spot to his memory ; hard by may 
still be seen Oswald's Well, once highly esteemed for its healing 
virtues and at that time overshadowed by Oswald's Ash, of 
which remarkable legends were also told. On the whole, 
there is much in favour of this identification ; it is implied in 
Bede's account of the miracles which signalised the spot that 
it lay in a wild region sometimes visited by British wayfarers, 112 
and there can be no difficulty in imagining the Mercians as 
having pushed so far west, 113 for Penda's name is preserved in 
that of Llannerch Panna, near Ellesmere, 114 and perhaps in 
that of Pontesbury between Shrewsbury and Montgomery. 116 
The victory of Maserfeld made Penda for many years the 
chief power in Southern Britain. Northumbria became again 
a divided kingdom, and the Mercian leader was able not only 
to control affairs in Deira, but also to harass by constant 
plundering expeditions the furthest limits of Bernicia, ruled 
over by Oswald's brother, Oswy. 116 So far as can be judged, 
he had at all times the support of the Cymry, and about 645 

111 See the life of Oswald by Reginald of Durham, printed in part by T. 
Arnold in the Rolls edition of Simeon of Durham (vol. i. 1882), and especially 
PP- 35i 35 6 -7- The earliest mention of Oswestry by that name is in Earl Hugh's 
charter to Shrewsbury abbey (Mon. Angl. iii. 520 " Oswaldestre " ; in the foun- 
dation charter it is " ecclesiam sancti Oswaldi," ibid.). 

na H.E. iii. 10 (" quidam de natione Brettonum . . . iter faciens iuxta ipsum 
locum "). 

113 Green, who supposed that Shrewsbury did not become English until the 
time of Offa (Mak. Eng. pp. 419-20), was thereby led to set aside the view that 
Maserfeld was in North-west Shropshire. 

114 This name, which may be a translation of the English Penley (Bye-Gones, 
ist Oct. 1890, p. 480), appears in Peniarth MS. 176 in the more regular form of 
Llanerch Banna (Evans, Rep. i. pt. 2, p. 979). For Panta, the Celtic spelling of 
Penda, see Hist. Britt. cc. 60, 64 ; Rev. Celt. xvii. pp. 181, 185, 194. 

118 Pantesberie in Domesd. i. 2556, i. 118 H.E. iii. 16, 17. 


CHAP, a struggle, noticed only by the Irish annalists, 117 took place 
between Oswy and the Britons of Strathclyde or of Wales, in 
which it is highly probable that Penda played a secret, if not 
an open part. Gwynedd was ruled over at this time by 
Cadafael son of Cynfedw, who was not of the stock of Maelgwn, 
but is ranked by the Triads among the Three Peasant Kings 
of the Isle of Britain. 118 Cadwallon had, indeed, left a son 
Cadwaladr, but he was probably at the time of his father's 
death of tender years, and would seem to have had to wait for 
his crown 'for a considerable period. Other Welsh leaders of 
the period are not known to history, but there is little doubt 
that they acted with Penda, and that most of them were in the 
army which in 655 ll9 marched upon Bernicia with the inten- 
tion of overwhelming Oswy. It was the crisis of that king's 
career ; he was shut up in the strong fortress of ludeu, which 
lay somewhere near the Firth of Forth, 120 and was forced to 
deliver up to Penda the treasures of the royal hoard, the heir- 
looms he had received from his ancestors and the rich spoils 
of many a victory over the Britons. These the Mercian king 
distributed, with the pride of a conqueror, as largess to his 
followers, and the delight of the Cymry at recovering their 
ancient possessions made memorable for years to come the 
" Restoration of ludeu ", 121 The army then returned in triumph, 
and it was probably as the serried hosts were passing through 
Deira, in the careless mood of men who had achieved their 
purpose, that Oswy burst upon them and in the battle of 
Winwaed Field 122 routed the great confederacy, slew the 

117 Tighernach in Rev. Celt. xvii. p. 186 (Cath Ossu inter eum et Britones) ; 
Ann. Ult. s.a. 641. 

118 I. 76 = ii. 59 = iii. 26. III. 48 seeks to make him the murderer of lago 
ap Beli ; it is not a very likely story. 

us Bede gives the date in H.E. v. 24. 

120 Celt. Br. (2), pp. 133, 151, 268 ; W. People, pp. 115-6. The difficulty in 
identifying the " urbem quae vocatur ludeu " of the Sax. Gen. (Hist. Britt. c. 
64, end) with the " urbem Giudi " of H.E. i. 12, which is almost certainly Inch- 
keith, might be met by supposing that Bede had confused an " urbs Giudi " and 
an " insnla Giudi ". 

121 1 follow the Sax. Gen. here rather than Bede, because the phrase " At- 
bret ludeu " must have had its origin in an actual restoration. "Atbret" is now 
" edfryd " ; cf. Gr. Celt. (2), p. 900. 

122 Not yet identified, though, with Bede's words before one (" hoc autem 
bellum rex Osuiu in regione Loidis . . . confecit "), it is difficult to avoid placing 
it in the West Riding. The suggestion that here (H.E. iii. 24) " Loidis" means 


implacable enemy of his house, and finally freed Northumbria CHAP, 
from the domination of Penda and his British allies. Cadafael VI< 
escaped destruction by making for Gwynedd with all his men 
the night before the encounter, which led the wits of the day 
to affix a new epithet to his name, viz., Cadafael Cadomedd, 
" The Battle-seizer who battle declines ". m But among the 
thirty noble leaders who fell around Penda there must have 
been no small number of Britons, the last of their race seriously 
to contest with the English the supremacy of the isle of 

Oswy's victory enabled him to reach a height of power and 
influence attained by no earlier English king. He was recog- 
nised alike by Saxon, Angle, Briton, Pict and Scot as the 
supreme ruler of Britain, and after his death in 671 a good 
part of his authority was retained by his son Egfrith. This 
final victory of Northumbria over the Cymry put an end to the 
existence of the latter as a united force and irrevocably divided 
the Cumbrians from the Welsh. Thus the year 655 forms an 
epoch of great importance in the history of the Welsh people ; 
it closes the period of definition, during which they were 
gradually marked off from the other inhabitants of these 
islands and constituted a separate people ; it brings upon the 
stage a nation, isolated and self-contained, dependent hence- 
forth upon its own resources for its development 

NOTE TO CHAPTER VI. i. The Name " Cymry ". 

Zeuss first proposed in 1853 the " Combroges " derivation and cited the 
similar form " Allobroges," explained by an early commentator on Juvenal, 
Satires, viii. 234, as meaning " men of another land" (Grammatica Celtica, first 
ed., p. 226). It has now been generally adopted ; see Urk. Spr. 221 ; W. People, 
p. 26. Among obsolete derivations mention may be made of the following. 
Geoffrey of Monmouth, in order to explain Cambria, the Latinised form of 
Cymru, invented a Kamber, son of Brutus (Hist. Reg. ii. i), who does not appear 
in the original Brutus legend, as given in various forms in the Historia Brit- 
tonum. Theophilus Evans (Drych y Prif Oesoedd, ed. 1740 and reprint of 1902, 
p. 7) in 1715 derived Cymry from Gomer, son of Japheth, who had from the 
time of Josephus (Ant. Jud. I. vi. i) been regarded as the ancestor of the Gauls 
(Hist. Britt. c. 18 : " primus, Gomer, a quo Galli "). This derivation soon ac- 

Lothian seems far-fetched ; it would be unlike Bede not to warn his readers that 
the name is not the same as that which he mentions in ii. 14, where the reference 
is unmistakably to a region in Deira. 

Jas Hist. Britt. c. 65. The Sax. Gen. call the battle " strages Gai campi " 
(probably translating some such form as " Maes Gai "). 


CHAP, quired great popularity among Welshmen, who have not yet given up the habit 
VI. of speaking of themselves as " hil Gomer " and of their language as "yr Omer- 
aeg ". A more specious explanation was that which connected Cymry and 
Cimbri, the name of a tribe now definitely assigned to the Teutonic family, 
though long supposed to be Celtic (Rev. Celt. xix. p. 74). This was first put for- 
ward in the sixteenth century by Humphrey Llwyd (Commentariolum, ed. Moses 
Williams, London, 1731, pp. 65-6) and caught the fancy of many capable his- 
torians, including the penetrating critic, Thomas Stephens (Lit. Kym. first ed., 
1849, Pref.). John Walters proposed in his Dissertation on the Welsh Language 
(1771) a native derivation, from "cyn" = first, original and "bro" = country, 
giving the word the sense of " aborigines " (so W. O. Pughe's Dictionary, s.v. 
Cymmro), but, as is pointed out by Zeuss (Gr. Celt. (2), p. 207), this combination 
yields, not " Cymro," but " Cynfro ". For the distinction between the two 
prefixes " kom, kon " and "kentu," see Urk. Spr. 86 and 77. It ought, per- 
haps, to be added that the ever delightful George Borrow makes an original 
contribution of his own to the discussion of this question, which shall be given in 
his words: "The original home of the Cumro was Southern Hindustan, the 
extreme point of which, Cape Comorin, derived from him its name " (Wild 
Wales, appended note). 

The main facts as to the early use of the name are given by Phillimore in 
Cymr. xi. 97-100. Its first appearances are in connection with the history of 
Cumbria, not of Wales, and it was only very slowly adopted by Welsh writers 
as a substitute for the ancient " Brython," " Brittones ". Thus it is unknown, not 
only to Gildas, Bede and Nennius, but also to Asser, the Liber Landavensis, 
and the compilers of the older parts of the Annales Cambria. But there is one 
important exception ; " Cymro," " Cymraes " and " Cymry " are of fairly common 
occurrence in the Welsh Laws, and, although the MSS. of these are of compara- 
tive late date, the terminology they employ is no doubt ancient. See LL. i. 96, 
98, 152, 206, 208, 508, 530, 646, 694, 750 ; ii. 94, 100, 114. In these passages, how- 
ever, the " Cymro " appears to be, not so much a member of a particular race, as 
one holding a definite legal status. He is usually distinguished from the " alltud " 
or landless man, and occasionally from the " caeth " or slave. He is, in fact, the 
" treftadog " or " priodor," whether bond or free, the man who has landed 
property or expectations. It is possible, therefore, that the term had a legal 
before it had a historical application and the use of '' Cymry " in the sense of 
"co-proprietors" may have prepared the way for its adoption as a badge of 
national union. 

NOTE TO CHAPTER VI. ii. Bangor. 

The idea that " bangor " denotes in Welsh a primitive type of monastery 
is due to the Glamorgan school of antiquaries, one of whom expressly says (lolo 
MSS. 114) : "the bangors preceded the monasteries and afterwards disappeared, 
with the exception of those which became monasteries ". The theory took its 
rise from the fact that some half-dozen Celtic ecclesiastical sites actually bear 
the name; to these the writers of the lolo MSS. added another dozen, such as 
Bangor Illtud (Llanilltud Fawr) and Bangor Dathan (Caerwent), which were 
entirely of their own invention. It was a further help to this explanation that 
the name Bangor had long been explained as derived from " ban," high, conspicu- 
ous, lofty, and " cor," a choir (see Davies, Diet. s.v. ban). But it has lately been 
pointed out that, where " bangor " occurs as a common noun, it has a very dif- 
ferent meaning, viz., the binding part of a wattle fence (Evans, Diet, s.v.), and 
the suggestion has been put forward that the first monastic site so called took 
its name from the wattle enclosure surrounding it, while the other monasteries 


were called Bangor in honour of the first and most famous (A. N. Palmer in Cymr. CHAP, 
x. 16-7; cf. Cymr. xi. 83-4). If this be so, there can be little doubt that the VI. 
original Bangor was that on the Menai Straits, known to this day among Welsh- 
men as " Bangor Fator (the Great) yn Arfon," and to be identified with the 
" Bennchor moer in Britannia " the burning of which is mentioned by Ann. Ult. 
s.a. 631 (really 634). This monastery is commonly said to have been founded in 
the early part of the sixth century, so that the famous Bangor of Belfast Lough, 
established by Com gall about 558 (Chron. Scot, and Ann. Ult. s.a.) may well 
have been modelled upon it. Nor is there any difficulty in regarding Bangor 
Iscoed, which is Bede's "Bancor" and " Bancornaburg," as an offshoot also ; 
the true tradition appears to make Deiniol, the founder of episcopal Bangor, its 
patron saint (see below). Of the other places bearing the name little need be 
said. Bangor on the Teifi is an ancient church dedicated to St. David (Myv. 
Arch. I. 271 (194)), but there is nothing to show that it was ever the seat of an 
important monastery. Capel Bangor near Aberystwyth was not an ecclesiastical 
site until 1839 ; it takes its name from the adjacent farm of Maes Bangor. 

The " Bancornaburg " of H.E. is, of course, an English derivative, explained 
as meaning " the stronghold of the men of Bangor " (Plummer's Bede, ii. p. 75). 
The existence of the form and Bede's use of the present tense (" vocatur ") are 
not favourable to the view which is sometimes expressed (Cymr. x. 15) that the 
monastery did not survive the famous massacre of ^Ethelfrith's day. It has not 
been observed that the name was still current, in the form " Bankeburw," at the 
end of the thirteenth century (Tax. Nich. 248). 

" Dinoot," which represents the old Welsh Dunot (with the narrow ), from 
the Latin Donatus (Celt. Br. (2), p. 304), is given by Bede as the name of the 
abbot of Bangor-on-Dee in Augustine's time. But a little consideration will show 
that this cannot be the father of Deiniol and the son of Pabo Post Prydain men- 
tioned in the old genealogies (Cymr. ix. 174 ; Myv. Arch. II. 23 (415)) and in the 
Triads (i. n = ii. 31 = iii. 71). For this Dunodis everywhere represented, not 
as a " saint," but as a mighty warrior ; he belongs, moreover, to the beginning 
and not to the end of the sixth century. There is, so far as I know, only one 
passage which can be cited in support of the ascription of this church to Dunod 
ap Pabo, viz., lolo MSS. 105, and elsewhere in these notices (113, 127, 129) the 
view is taken that it was Deiniol who had the chief share in the matter of its 
foundation. That this is the sounder tradition is shown by the fact that two of 
the ancient chapels of Bangor, viz., Marchwiel and Worthenbury, are dedicated 
to Daniel, that there was in the parish a " Daniel's Well " (Cymr. x. 19) and that 
the parish wake or annual festival followed St. Deiniol's day, formerly nth Sep- 
tember (lolo MSS. 152 ; Evans, Rep. i. p. 17), but now, as the result of the change 
of style, 22nd September (Arch. Camb. IV. vii. (1876), 297; Thomas, St. Asaph, 
p. 799). It is, of course, quite possible, and even likely, that Bede's " Dinoot " was 
of the family of Deiniol. 

VOL. I. 13 




CHAP. THOUGH there is little contemporary evidence as to what took 
VII> place in Wales during the two hundred years which followed the 
battle of Winwaed Field, the indirect evidence leaves no doubt 
as to the political condition of the country. Fromi the royal 
genealogies preserved in Harleian MS. 3859 and from the 
Welsh laws, no less than from the scanty notices of the 
annalists, it may with confidence be inferred that the Welsh 
were during this period under the rule of minor chieftains, 
"kings" of districts which were often of less extent than a 
modern Welsh county, and that, if any one of these claimed, by 
right of ancestral dignity, pre-eminence among his fellows, the 
utmost to which he could attain was an honorary primacy, 
carrying with it no important practical consequences. 1 The 
ambitious hope of recovering Britain from English domination 
had for ever faded and with it had ended the mission of the 
gwledig ; henceforth, the conflict with the English would be a 
border warfare, waged against the kings and ealdormen of 
Mercia, in countless skirmishes and border forays, under many 
local leaders, and centuries would go by ere the spectacle 
would again be beheld of a great national movement led by a 
prince whose authority was recognised by the whole of Wales. 
In this and the following two chapters the history of this 
period of subdivision and local independence will be treated, 
not chronologically, for any attempt to weave the scattered 
strands into a thread of continuous narrative must prove a 
failure, but topically, each branch of the subject being taken 
separately. The establishment of a border between the English 

1 See p. 231 below. 


and the Welsh peoples, the progress of the Welsh Church, the CHAP, 
geographical and territorial divisions of Wales, the social life 
and the characteristic institutions of the Welsh will be in turn 
discussed, and the way will thus be prepared for the study of 
mediaeval Wales, when a certain measure of stability has been 
reached in these matters and the main interest lies once more 
in the action of individual princes and of their opponents. 

Little is known of the process by which the boundary 
between the English and the Welsh was evolved. No record 
has been preserved of the English conquest of Cheshire, Shrop- 
shire or Herefordshire, and one can but conjecture the course 
of events in this region during the seventh and the eighth 
centuries. On the whole, it appears likely that it was the 
earlier and not the later of these two centuries which witnessed 
the triumphs of Mercia along the border, and that the great 
age of territorial expansion was that of Penda (d. 655) and his 
energetic son, Wulfhere (d. 675). Chester and its neighbour- 
hood, though not occupied by ./Ethelfrith as the result of his 
famous victory, 2 probably fell into Mercian hands not long 
afterwards ; this may well have been one effect of the fall of 
Cadwallon. If it may be supposed that St. Werburh had a 
convent here before the translation of her relics to the place 
from Hanbury in 874, it was thoroughly English as early as 
680, for she was a daughter of King Wulfhere. 3 To this it 
may be added that the fact that Bede gives the monastery of 
Bangor Iscoed an English name, viz., Bancornaburg, 4 implies 
that, when he wrote in 731, the English border was not far 
from the Dee. In Shropshire the evidence is in the same 
direction. It has already been suggested that Penda's name is 
preserved in that of Llannerch Panna, near Ellesmere, and 
that the battle of Maserfeld was fought at Oswestry. 5 In the 
next generation, Wulfhere gave his name to a Wulfheresford in 
the hundred of Mersete, 6 which was known to the Welsh as 

a See p. 180. 

3 Fl. Wig. i. 32, 265. The date of the translation is from Higden, who was 
a monk of St. Werburgh's (Polychr. vi. 126). 

* H.E. ii. 2. P. 189. 

6 " In Merset hund. Rogerius comes tenet Wlferesforde. Rex Edwardus 
tenuit " (Domesd. Shrops. i. 2596, 2). No identification is suggested by Eyton 
(Shrops. xi. 43). 



CHAP. " Rhyd Wilfre ar Efyrnwy," 7 and must, therefore, have been 
close to Llanymynech or Melverley. Shrewsbury itself is not 
mentioned until a comparatively late date, 8 but the nunnery at 
Wenlock, not far to the south-east, was founded by St. Milburh, 
who was a cousin of St. Werburh, 9 so that this district must 
have been in English hands before the end of the seventh 
century and was not won for Mercia, as has been sometimes 
supposed, 10 by the victorious sword of Offa. That English 
Herefordshire was also conquered about the middle of the 
seventh century (if not earlier) does not admit of any doubt. 
This was the region of the Hecana or Magesaetas, who formed 
a separate kingdom under Merewald, a brother of Wulfhere, 11 
and from about 680 a separate diocese, with Hereford as the 
seat of the bishop. 12 The Liber Landavensis, compiled from 
old records at Llandaff in the twelfth century, places as early 
as the age of Oudoceus, i.e., the first half of the seventh century, 
the overthrow of the Britons in the triangle formed by the 
Dore, the Worm and the Wye, 13 and it is indeed evident that, 
in the interests of the security of Hereford, this tract of country 
must have been seized about the time of the foundation of the 
city. The general effect of the evidence, therefore, is to make 

7 Mab. 144. 

8 The earliest reference to Shrewsbury (if the document be genuine) is to be 
found in Ethelred's charter to Wenlock, done in 901 "in ciuitate scrobbensis " 
(Cod. Dipt. ii. 137). It next appears in the Anglo-Saxon Chronicle, MS. F. s.a. 

9 F1. Wig. i. 33, 265; Wm. Malm. G.R. 78, 267 (i. no, 369-70). 

10 Green (Mak. Eng. pp. 419-20 ; cf. also maps on pp. 395 and 429) adopted 
too readily the statement of Welsh writers that Offa conquered the country 
between the Wye and the Severn and thus brought about the transference of the 
capital of Powys from Pengwern (i.e., Shrewsbury) to Mathrafal on the Vyrnwy. 
The first to put forward this view was David Powel in his Historic of Cambria 
(1584), and it does not appear that he had any warrant for doing so (see p. 17 of 
reprint of 1811). There was an ancient tradition that Shrewsbury, which has 
been long known to Welshmen as " Amwythig," had once borne the name of 
" Pengwern " and that on the site of St. Chad's Church had stood the palace of 
Brochwel Ysgythrog, prince of Powys (Gir. Camb. vi. 81 (I tin. i. 10), 169 (Descr. 
I. 4); Historia Monacellae in Arch. Camb. I. iii. (1848), 139). But nowhere in 
the older sources is it suggested that the ruin of Pengwern was brought about by 
Offa. Rhys thinks (Celt. Br. (2), 141) that " Scrobbesbyrig " is a translation of 
" Pengwern". 

11 Fl. Wig. i. 265. 

12 Ibid. 41, 238; H.E. v. 23 (Ualchstod), with Plummer's notes (ii. 341; 
cf. also 222). 

13 Lib. Land. 133-4. 


it fairly certain that at the beginning of the eighth century, the CHAP, 
age of the greatness of Mercia, that state had already reached, 
in the main, its westernmost limits, and that the work of Offa, 
important though it was, lay rather in the direction of definition 
than of conquest. 

During the years 7 16 to 757 Mercia was ruled by Ethel bald, 
who was the principal English monarch of his day, holding 
Wessex for the greater part of his reign in subjection and tower- 
ing above Northumbria and its feeble line of kings. There is, 
however, no record of any warfare carried on by him against 
the Welsh, 14 and all that is known of the border conflict which 
no doubt went on incessantly during this period is that in or 
about 722 the Welsh won two victories in South Wales, the 
one at a " Pencon " or " Pencoed," not yet identified, and the 
other at Garth Maelog, which was probably the place of that 
name near Llanbister in Radnorshire. 16 It is under Ethelbald's 
successor Offa, who was king of Mercia from 757 to 796, that 
the struggle between Welsh and English again emerges into 
the light of history, just at the stage when the final limit is 
being set to the westward progress of English colonisation. 

Offa was one of the most powerful kings of the early 
English period, formidable in Kent, Wessex and East Anglia, 
and dealt with as an equal by his mighty neighbour, Charles 
the Great. Thus it is not at all surprising that he should have 
shown vigour and resourcefulness in his treatment of the Welsh, 
to whom his realm lay open from the Severn to the Dee. A 
battle of Hereford between Welsh and English is recorded 
under the year 760 ; 16 whether he was concerned in this it is 

14 By the " Wealas " of A.S. Chr. 743 I understand, as generally in this 
part of the Chronicle, the West Welsh, or men of Cornwall and Devon. Ethel- 
bald came to the aid of Cuthred as his overlord. 

""Bellum hehil apud cornuenses. gueith gart mailauc. cat pencon. apud 
dexterales brittones. et brittones uictores fuerunt in istis tribus bellis" (Harl. 
MS. 3859 in Cymr. ix. 160). The present Caerfaelog, or Cyfaelog, close to the 
village of Llanbister, was formerly " Gardd (for Garth) Vaelog " (Dwnn i. 266). 
Another possible identification is with Garth Mailwg, near Llantrisant, Gla- 
morganshire ; there is, however, some evidence that this is properly Garth Mt'lwg 
(Lib. Land. 384). I know no reason why Garth Maelog should be placed, 
as is done by Powel (p. 12), in North Wales. 

18 " Bellum inter brittones et saxones, id est gueith hirford " (Harl. MS. 
3859 in Cymr. ix. 161). The Dyfnwal ab Tewdwr mentioned in the same annal 
was a prince of Strathclyde ; see the pedigree on p. 172. 


CHAP, not possible to say, but his name is expressly coupled by 
VIL Harleian MS. 3859 with two raids which were made upon 
Welsh territory in 778 and 784. 17 These attacks seem to have 
been made on so large a scale as to attract the special notice 
of the chronicler, but they were probably not intended as part 
of any scheme for the conquest of Wales, for the enterprise 
particularly connected with the name of Offa is the boundary 
dyke which he caused to be raised along the Welsh border, and 
the rearing of which at enormous cost must be looked upon as 
a deliberate closing of the era of conquest. 18 Attempts have, 
indeed, been made to discredit the traditional account of the 
origin of this great earthwork and to show that it is much 
older than the age of Offa. But they have been signally un- 
successful. It is true that the older English and Welsh 
chronicles, for the most part, have nothing to say of the digg- 
ing of the dyke, but the testimony of Asser, 19 a Welshman 
familiar with England who wrote less than a century after Offa's 
death, outweighs the silence of the other sources and makes it 
all but absolutely certain that the popular name of the dyke 
preserves the true account of its origin. It has been pointed 
out that the form and disposition of this entrenchment prove 
it to have been cast up by an Eastern folk for protection 
against dwellers in the West ; the ditch or fosse is always on 
the western side, and wherever the line of a cliff or escarpment 
is followed, the face of this is always to the west. That the 
dyke is also post-Roman is clear from the discovery in it at 
the Ffrith, near Hope, of relics of a Roman settlement which 
were disturbed when it was erected, and when the theory that 
it was prehistoric has thus been disposed of, there seems no 

17 Cymr. ix. 162. The entry in Ann. C. MS. C. s. a. 795, " Vastatio Rienuch 
ab Offa " no doubt refers to an expedition into Dyfed (Rheinwg), which antici- 
pated that of Cenwulf in 818. 

18 A. N. Palmer has admirably discussed the leading problems connected 
with the dyke in Cymr. xii. 65-86. See also Arch. Camb. II. i. (1850), 72-3 ; 
III. ii. (1856), 1-23 ; III. vi. (1860), 37 ; IV. vi. (1875), 275-81, and Stevenson's 
Asser, pp. 204-5. 

19 " Rex nomine Offa qui vallum magnum inter Britanniam (his usual name 
for Wales) atque Merciam de mari usque ad mare facere imperavit" (c. 14). 
The passage was copied by Sim. Dun. ii. 66, and a little later by the author of 
the twelfth-century life of St. Oswald (Sim. Dun. i. 353). There is no allusion 
to the dyke in Harl. MS.385g, Ann. C., or the older Bruts ; what is said on the 
subject by the Gwentian Brut (Myv. Arch. II. 473-4 (686) ) should be entirely dis- 


reason for depriving Offa of the credit of the undertaking, CHAP, 
which has been ascribed to him by English and Welsh tradition 
alike for the past ten centuries. 

According to Asser, the dyke ran from sea to sea, and in 
order to be a complete boundary, it was no doubt requisite 
that it should do so. Both the northern and the southern end 
are, however, difficult to trace. No vestiges of the dyke have 
been found to the south of Bridge Sellers on the Wye, 20 about 
6 miles above Hereford, unless the entrenchments which line 
the east bank of that river between Monmouth and Chepstow 
are to be regarded as part of the great work. From the 
fourteenth century to the time of Pennant it was supposed that 
the northern end lay at Basingwerk, near Holywell, 21 but the 
earthworks here, though locally known as " Clawdd Offa," 
were shown in the first edition of the Tour of 1773 22 to be part 
of Wat's Dyke, which, lying a few miles to the east of Offa's 
Dyke, runs parallel to it as far as the borders of Montgomery- 
shire. 23 The western dyke probably touched the sea not far 
from Prestatyn, 24 but its course, except for a short length near 
Newmarket, is quite uncertain through the greater part of 
Flintshire. Near Treuddyn Church its traceable course begins ; 25 
the line of the " vallum " may thence be followed without 
serious interruption through Adwy'r Clawdd (The Gap in the 
Dyke), Ruabon, Chirk Castle Park, Selatyn, Llanymynech and 
Llandysilio to the Severn. Here intervenes a break of about 
5 miles and it is reasonable to suppose that the river itself 
was treated for this distance as the boundary. At Buttington 
the dyke re-appears and thence runs southward through Forden, 
Lymore near Montgomery, Mainstone, Knighton (the Welsh 

* Arch&ological Survey of Herefordshire (1896), p. 7. 

31 Higden (Polychr. ii. 34) was the first to place the end of the dyke between 
Basingwerk and Coleshill (collem carbonum). He was followed by Gutyn Owain 
in the Book of Basingwerk (B.T. MS. E. p. 8), Humphrey Llwyd (Comment. 
(2), 64), and Edward Llwyd (Gibson, 587). 

B Penn. i. 31. 

23 For Watt's Dyke, or " Clawdd Wad," see Penn. i. 349-50. Palmer thinks 
it may also have been thrown up by Offa (Cymr. xii. 75). The notion that the 
space between the two dykes was neutral ground cannot be traced further back 
than Churchyard's Worthiness of Wales (1587). 

M Arch. Camb. III. iv. (1858), 335-42 (Guest); Cymr. xii. 79-80. 

K For the course of the dyke see the old maps (one inch to the mile) of the 
Ordnance Survey, Sheets 79, 74, 60, 56; Penn. i. 350-2 ; Radnorsh. (2), pp. 123-4 
Archceological Survey of Hertfordshire, p. 7. 


CHAP, name of which is " Tref y Clawdd," the Town on the Dyke), and 
' Discoed to Knill, near Kington in Herefordshire. Hencefor- 
ward its course is broken ; a portion has been traced near 
Lynhales and another on the north side of the Wye near 
Bridge Sellers, but beyond the latter point, as has been already 
stated, no sign of it can be perceived ; for reasons which can 
only be conjectured, Offa did not think it necessary to place 
any barrier between himself and the men of Ewias and Erging. 26 
It is obvious that a power which possessed no standing army 
could have made no other use of this dyke than as a boundary, 
the violation of which it visited with penalties, and such, it 
must be supposed, was the purpose of Offa in erecting it In 
the twelfth century it was believed that by an ordinance of 
the builder of the dyke every Welshman found with a weapon 
on the eastern side of it had his right hand forthwith cut off. 27 
Whether such a decree was ever issued by Offa or not, it may 
safely be said that the main intention of the dyke was to mark 
definitely the frontier between the two races and so to signify 
to the Welsh on the one hand, how far they might come, and 
on the other, that no further aggression at their expense was 
intended. What was English was to remain so, but no more 
Welsh " trefs " were to be turned into English " hams " and 
" tons ". It is probable that for many years the border had 
not sensibly advanced ; while the plains had been won with 
comparative ease, the tide of invasion had washed in vain 
against the immovable ramparts of the Welsh mountains, and 
on the lower slopes of these the line was now drawn which 
was to separate English and Welsh for centuries and indeed 
separates them at many points to this day. 

A study of the older place names along theidyke brings out 
clearly the fact that it was a real national border line. To the 
east of it the village names are of English origin, Suttons, 
Astons, Actons, Middletons, Newtons of the ordinary type ; 28 
to the west, the names are, save for some exceptions to be 

26 The ordinary view, that the Wye served as the boundary from Bridge 
Sellers to Monmouth, seems to me to leave the Welsh much too near Hereford 
for that city's safety. 

27 John of Salisbury (Polycraticus, vi. 6) ascribes the law to Harold, Walter 
Map (De Nugis,n. 17, p. 86) to Offa, but with the substitution of "foot" for 
" right hand". 

28 This was noticed by Humphrey Llwyd (Comment. (2), 64-5). 


presently noted, Welsh in formation, from Rhuddlan and CHAP. 


Diserth in the north to Bleddfa and Llangynllo in the south. 
In Flintshire, it is true that the facts have been somewhat 
obscured by the process of reconquest carried on by the Welsh 
in later years ; yet the help of Domesday is hardly needed to 
enable one to see Preston in Prestatyn, Westbury in Gwespyr, 
Merton in Mertyn, Bishopstree in Bistre, 29 and thus to bring 
to life again the English settlement which once occupied the 
region of Englefield or Tegeingl almost as far as the Clwyd. In 
the neighbourhood of Wrexham, Chirk and Oswestry, the dyke 
still forms the dividing line between the two peoples, 'as it does 
in the Welshman's popular phrase, when he speaks of England 
as "y tu draw i Glawdd Offa" (the other side of Offa's Dyke). 
In Montgomeryshire and Radnorshire the English village 
names extend at certain points a mile or two to the west of the 
dyke; cases in point are Leighton, Forden, Hopton, Water- 
dine, Pilleth, Cascob and Radnor. 30 These are ancient Eng- 
lish settlements, for they are mentioned in Domesday, and it 
must be understood that hereabouts, even after the making of the 
dyke, the process of English colonisation for a little time went on 
without check. The portion of the dyke just north of the Wye 
seems to have become obsolete as a frontier not long after its 
erection, as the result of the settlement of Eardisley and the 
surrounding villages, which threw the Welsh back upon the 
outlying ridges of Radnor Forest. 

By the action of Offa the border between Welsh and Eng- 
lish was thus to a large extent fixed. The border warfare was, 
however, not brought to an end, nor yet the forays into the 
heart of Wales by means of which Mercia at this time demon- 
strated her strength and kept the Welsh in awe. In 796, the 
year of the death of Offa, a battle was fought at Rhuddlan, in 
which, it may be conjectured, the English sought to defend 
their new frontier in Tegeingl. 31 Under Offa's successor, Cen- 

89 The Domesd. forms are Prestetone (2690, 2), Wesb(er)ie (ibid.), Meretone 
(2690, i) and Biscopestreu (2690, 2). 

30 Trans. Cymr. 1899-1900, 123-4. 

sl "Bellum rud glann " (Harl. MS. 3859 in Cymr. ix. 163). The entry 
" Caratauc rex guenedote apud saxones iugulatur " comes two years later (798) in 
all the old authorities, and it was not until the time of Powel (p. 17) that the two 
notices were merged and the foundation laid for the popular account of the de- 
feat of Caradog in the battle of Rhuddlan Marsh (Morfa Rhuddlan). 


CHAP, wulf, the English in 816 harried Rhufoniog, which lay west ot 
VI1 ' the Clwyd, and the region of Snowdon itself; in 818 they 
penetrated into Dyfed. Cenwulf died in 821 at Basingwerk, 32 
perhaps in the course of a new campaign against the Welsh, 
nor did his death, though it marks the end of the period of 
Mercian greatness, lead to a cessation of the attacks upon the 
men of the west, for in 822 the Welsh fortress of Degannwy, 
once the secure stronghold of Maelgwn Gwynedd, was destroyed 
and the realm of Powys was overrun. Shortly afterwards, 
however, the supremacy of Britain passed from Mercia to 
Wessex, and, when Mercia further began to feel the weight of 
the Norseman's sword, the Welsh were able to breathe more 


While the kings of Mercia were confining the Welsh 
people within limits which grew daily narrower, the Welsh 
Church was also being more and more cut off from the rest of 
the Christian world. Little by little the Celtic communities 
outside the borders of Wales adopted the Roman Easter and 
so abandoned the original Celtic position, that the insular tradi- 
tions ought to be maintained against any innovation hailing 
from the Continent. 33 First, the Southern Irish yielded to 
Rome during the papacy of Honorius I. (625-38); next, 
the Northumbrian Church, an offshoot of lona, declared for the 
See of Saint Peter at the famous Synod of Whitby in 664. 
Through the influence of Adamnan, abbot of lona and bio- 
grapher of St. Columba, the Northern Irish followed suit at the 
very end of the seventh century. lona itself was for some 
years obdurate, but after seeing its Pictish branches forced to 
accept the Roman customs, gave up the struggle in 718. It 
was probably about this time that the Britons of Strathclyde 
gave way, and about 705 a large section, if not the whole, of 
the men of Devon and Cornwall were won over to the Roman 
cause by Abbot Aldhelm of Malmesbury. In the middle, of the 
eighth century it is most likely that the people of Wales were 

32 " Transit el liu de Basewerce," says Geoffrey Gaimar (v. 2240), a late au- 
thority, but here quoting, no doubt, some lost record. 

33 Zimmer, Celtic Church (1902), pp. 77-86 ; H. and St. ii. 6; i. 673. 


the only considerable community of Christians in these islands CHAP, 
who maintained the old attitude of isolation from Rome. 

This position of solitary protest could not be long retained, 
and accordingly it is not surprising to find the Welsh in their 
turn forced to give way in the early part of the reign of Ofifa. 
The submission to Rome was all the more thorough inasmuch 
as it was not, apparently, brought about by foreign arms but 
by a peaceful revolution within the country itself. In the year 
768, says the sole authentic record of the event, " Easter was 
altered among the Britons, the reform being the work of that 
man of God, Elbodugus ", 34 According to the untrustworthy 
Gwentian Brut, 35 South Wales did not yield without a conflict, 
but it is obvious that the particulars which it gives have been 
invented in order to bring out the independence of that part 
of the country and its unwillingness to follow the North in 
basely truckling to Rome. History knows nothing of any 
struggle of the kind, and the character of the prime mover in 
the matter suggests that the change was really due to the 
feeling of the abler and more spiritual leaders of the Welsh 
Church that they were, by a meaningless conservatism, cutting 
themselves off from the religious life of Christendom. Elfodd 
(for such would be the modern Welsh form of the name of 
Elbodug) 36 is a somewhat shadowy figure, but it may be in- 
ferred from the epithet " homine Dei " that he was a monk ; 37 
tradition makes him a member of the monastic community 
of Caer Gybi or Holyhead. 38 He must have embarked upon 
his movement of reform at a comparatively early age, for he 
lived more than forty years after its successful completion, to 
die in 809 with the title of "chief bishop in the land ot 
Gwynedd". 39 This greatness, however, he attained in later 

14 " Pasca commutatur apud brittones emendante elbodugo homine dei " 
(Had. MS. 3859 in Cymr. ix. 162). 

38 S.a. 755.777,809. 

38 The "Elbodugo," "Elbodg" of Harl. MS. 3859 and the " Elvodugi," 
"Elbobdus"of the MSS. of Nennius (ed. Mommsen, pp. 143, 207) represent 
the old Welsh Elbodug and Elbodu, to be compared with Arthbodu (Lib. Land. 
80) and Gurbodu (ibid. 230). Cf. Gr. Celt. (2), p. 22; W. Phil. (2), 386. 

37 Trans. Cymr. 1893-4, I2 9- c f- the Welsh " meudwy," a hermit, which is 
for " meu duiu," servus dei (Urk. Spr. 198). 

38 Myv. Arch. II. 42 (425). Other evidence seems to point to a connection 
with Abergele (Thomas,, p. 351). 

19 " Elbod(u)g archi episcopus (in) guenedote regione migrauit ad dominum " 
(Harl. MS. 3859 in Cymr. ix. 163). 


CHAP, life ; his victory, there is reason to think, was not the victory 
of the prelate or man of affairs, but of the scholar and stu- 
dent. Nennius in the next generation introduces himself 
as author with the proud title of " disciple of Elfodd," 40 whom 
he elsewhere styles " most saintly of bishops " ; 41 incidentally, 
he allows us to see that his master was a student of the 
works of Bede. 42 Slight as are these indications, they seem 
to show that it was the learning and devotion of Elfodd 
which won this battle for him, and no authority which he 
wielded as metropolitan or as bishop of Bangor, titles, be it 
observed, one is not warranted in attaching to his name. 43 

Thus the schism between the Christians of Wales and those 
of the rest of Western Europe came to an end, for the sub- 
mission in respect of the observance of Easter undoubtedly 
carried with it submission in regard to other points of difference 
so far as these were considered to be a breach of Catholic unity. 
Welshmen came, in common with other dwellers in the Western 
world, to regard Rome as the centre of the religious world, and 
Welsh princes and prelates adopted the fashionable habit of 
pilgrimage to the holy city which held the, bones of St. Peter 
and St. Paul. 44 Nevertheless, the Welsh Church still retained 
in many respects the marks of its Celtic and monastic origin ; 
acknowledgment of the supremacy of Rome by no means 
implied at this period the acceptance of a uniform system of 
worship and church organisation, so that there was still room 
left in Wales for the growth of distinctive features of church life. 
It will now be convenient to take a brief survey of that life as 
it manifested itself in the centuries preceding the Norman 

40 " Ego Nennius Elvodugi discipulus" (ed. Mommsen, p. 143). 
41 "Elbobdus episcoporum sanctissimus " (ed. Mommsen, p. 207). 

42 It was Elfodd and another bishop, one " Renchidus," who pointed out to 
him that Edwin of Northumbria was really baptised by Paulinus of York, and 
not, as the old British record alleged, by Rhun ab Urien. Nennius cut the knot 
by treating the two men as one and the same. 

43 " Archi episcopus " was at this time a title of honour merely and did not 
necessarily imply metropolitan authority (Trans. Cymr. 1893-4, I 3 1 )- Elfodd is 
styled bishop of Bangor by late writers only (Gwentian Brut s.a. 755 ; lolo MSS. 
117, 127), who could not imagine an Archbishop of Gwynedd seated at any other 

44 Recorded instances are those of Cyngen of Powys in 854, Hywel (of what 
line is uncertain) in 886, Hywel Dda in 928, and Bishop Joseph of Llandaff in 


The salient fact in the history of the Welsh Church at this CHAP. 


time is that the principal churches, those having ancient tradi- 
tions and a position of honour and prestige, were in the hands 
of communities of clergy which in origin, whatever they may 
have in time become, were monasteries. 45 Primitive Welsh law 
divides churches into two classes, vtz., " mother churches " and 
those of less consideration. 48 The former are treated as always 
having an abbot (abad), who should be a cleric and lettered 
(dwyfol lythyrwr), with a community or " clas " of canons 
(cynonwyr), including at least one priest (offeiriad). In the 
smaller churches there are no abbots or canons, but merely 
parsons and priests. The "clas" was an important and re- 
sponsible body ; it received the half of all payments made to the 
church, 47 succeeded to the movable property of the abbot when 
he died, 48 and decided finally all disputes arising among its 
members. 49 Though the " claswyr " are not styled monks, but 
canons, the title of their chief officer, the abbot, and the manner 
in which they consumed in common the revenues of the Church, 
afford strong evidence that the " clas " was at first a monastery, 
smaller, no doubt, than the great monastic establishments of 
the sixth and seventh centuries, but of the same general type 
and in many cases, for instance at Llantwit and St. David's, 
carrying on the traditions of the age of the saints. 50 

This view of the organisation of the Welsh Church in the 
early Middle Ages does not rest for support solely upon the 
statements of the Welsh Laws ; it is confirmed by many inci- 

45 The otherwise admirable discussion of this subject in A. N. Palmer's essay 
on "The Portionary Churches of Medieval North Wales " (Arch. Camb. V. iii. 
(1886), 175-209) would have gained greatly in point if the evidence for the 
monastic origin of mother churches had been fully appreciated. 

46 LL. i. 78-80 (Yen.) ; 432-4 (Dim.) ; ii. 842 (Lat. B.) 

47 LL. i. 434 ( 3) ; ii. 842 ( 6). Ibid. ii. 10 ( 27). 4 Ibid. ( 28). 
50 For other references to the "clas," see LL. i. 106 (clas Bancor a rey 

Beuno) ; ii. 63 (yclaswyr ar personeit ; kanys vynt yssyd berchenogyon ar yr 
eglvys) ; Buchedd Gr. ap Cynan in Arch. Camb. III. xii. (1866) 42 (ar escop ae 
athraon a holl clas er arglwyd dewi). " Monastica classis " is found in Rhygy- 
farch's life of St. David (Cambro-Br. SS. 127), a phrase which suggests a deriva- 
tion from " classis" = " corpus, collegium" (Ducange s.v.). The word occurs 
in place names ; Higher and Lower Clas are two hamlets of the parish of 
Llangyfelach, an ancient Dewi church ; Clas Garmon is a township of the parish 
of St. Harmon's ; Treclas contains the parish church of Llanarthne, for which 
see note 165 to chap. v. The head of the body of clergy at Caergybi (Holyhead) 
was styled " penclas " (Penn. iii. 73). 


CHAP, dental references to the churches themselves to be found in the 
literature of the period. The Liber Landavensis, for instance, 
bears witness to the existence of many monastic churches in 
South Wales ; the head of the church of Llancarfan (anciently 
Nant Carfan) is described as " abbas Sancti Catoci " 51 and 
" abbas Carbani vallis " ; 52 at Llantwit Major there is " abbas 
Sancti Ilduti," 53 at Llandough, near Cardiff, " abbas Docguinni". 84 
Abbots also appear at Caerwent, Moccas, Garway, Welsh 
Bicknor, Llandogo and Dewchurch, and, if, as is most likely, 
" princeps " was but an alternative title, Bishopston in Gower 
and Penally may be added to the list. 55 In 1188 the church 
of Llanbadarn Fawr, near Aberystwyth, had an abbot, though 
in this case, as in others in Wales at this period, the title was 
held by a layman, who, having first got himself recognised as 
the advocate or guardian of the shrine, had afterwards ap- 
propriated its landed endowment. 56 Nor was the case different 
in North Wales. In 1147 there was an abbot of Towyn in 
Meirionydd, 57 while in 856 the death is recorded of a " princeps " 
of Abergele. 58 Llandinam had its abbot in the middle of the 
twelfth century, 59 and as late as the fifteenth the memory sur- 
vived of the abbot and " claswyr " of Llanynys. 60 The churches 

51 Lib. Land. 140, 143, 144 (Catmaili, the older form, occurs on p. 131). 

02 Ibid, passim. For the form Nant Carfan, see also the colophon to Cara- 
dog's Life of Gildas (ed. Mommsen, p. no), and the life of Gwynllyw in 
Cambro-British Saints (p. 149). Geoffrey of Monmouth seems to be responsible 
for Llan Carfan (Hist. Reg. xii. 20). 

58 Lib. Land, passim. " Abbas lannildut " occurs once (145). 

84 Ibid, passim. For the identification see Mar gam Abb. 3. 

65 Ibid. 222 (guentonie urbis), 164 (mochros), 166 (lann guruoe), 164 (lann 
garthbenni), 223 (lann enniaun, id est lann oudocui), 164 (lann deui), 145 (lann 
cynuur ; cf. 239, in monasterio sancti cinuuri, id est lann berugall), 149 (aluni 

56 Gir. Camb. vi. 120-1 (Itin. ii. 4). 

57 " Moruran abat y ty gwyn " (Bruts, 315 ; B.T. 174) has been generally 
taken for an abbot of Whitland, but the part played by him is altogether unsuited 
to a Cistercian monk and the difficulty vanishes when it is remembered that 
Cynfael, the scene of the incident, is close to Towyn Meirionydd. 

58 " Jonathan princeps opergelei moritur " (Harl. MS. 3859 in Cymr. xi. 
165). For the ancient importance of Abergele church see Thomas, St. Asaph, 
pp. 350-2. 

59 " Dolfin abbas Llandinaw " (the correct form of the name) is among the 
witnesses to a Trefeglwys charter granted by Madog ap Maredudd, who died in 
1160 (Arch. Camb. III. vi. (1860) 331). 

60 According to a petition of Griffin Young to Pope Boniface IX. (Papal 
Letters, iv. 349), the revenues of Llanynys (in Dyffryn Clwyd) were anciently 
divided into twenty- four portions called " claswriaiethe," instituted for the main- 


mentioned in this list are among the oldest and most important CHAP, 
in the country, so that it is plainly not with monasteries in the n> 
ordinary sense of the term that we have to do, but with the 
general framework of church organisation. 81 

In the case of the four cathedral churches, the title of abbot 
was from the first merged in that of bishop, always regarded 
in Wales as the more honourable, notwithstanding that in 
Ireland and Scotland matters were usually reversed. Yet 
evidence is not wanting that they, too, were served by com- 
munities of the same pattern. When Bishop Bernard of St. 
David's (1115-1148) came into possession of his see, he found 
there a body of " claswyr," who regarded the cathedral revenue 
as a common stock for their support and had not divided it, as 
was usual elsewhere, into canonries or prebends for the main- 
tenance of a fixed number of clergy. 82 They were continuing 
the traditions of the " monasterium " of Asser's day. Llandaff 
is also termed a " monasterium " ; 68 it rejoiced, indeed, in the 
name of archmonastery, which was probably meant to emphasise 
its supremacy as the mother house over the other convents 
founded in honour of St. Teilo. 84 Were there any ancient 
accounts of Bangor and St. Asaph, they would probably tell the 
same tale. It is, indeed, likely that chance alone determined 
which of the many monasteries founded in the sixth century 
should permanently become episcopal sees. At the beginning 
of the tenth century seven important churches in the kingdom 
of Dyfed were traditionally known as " esgoptai " or episcopal 
houses. 86 Mynyw headed the list and still retained its bishop. 

tenance of twenty-four perpetual portionaries called " abbatathelaswyr " (*'.., 
abad a chlaswyr). To one of these, the portion of David the priest, was assigned 
the cure of souls. I am indebted to Mr. J. R. Gabriel for calling my attention 
to this document. 

61 The fact that " llan," meaning at first a monastery, came to be applied to 
all churches alike is further proof of the monastic origin of the older Welsh 
Churches. See note 122, chap. v. 

68 Gir. Camb. iii. 153 (Men. Eccl. ii). The form " glaswir " for " claswyr " 
is, of course, inaccurate, but the substitution of g for c in this word seems to have 
been not uncommon. Thus Jocelin of Furness explains Glasgu as " cara 
familia" (Life of St. Kentigern, ed. Forbes, p. 182) and "Y Klas ar Wy " (Pen. 
MS. 147), which was Clastbirig in the twelfth century (Fl. Wig. s.a. 1056 in 
Mon. Hist. Br. 608), has since become Glasbury. 

83 Lib. Land. 144, 214. 84 Ibid. 74, 75, 129; Cymr. xi. 131-2. 

65 LL. i. 556-8 (Dim.) ; ii. 790-1 (Lat. A.), 869 (Lat. B.). The section on 
" saith esgopty Dyfed " is clearly primitive and in all probability formed part of 
the original " law of Hywel ". 


CHAP. But the other six houses, viz., " Llan Ysmael " (St. Ishmael's on 
VIL Milford Haven), "Llan Degeman" (Rhoscrowther), "Llan 
Usyllt"(St. Issell's near Tenby), " Llan Deulyddog " (Carmar- 
then), and a " Llan Deilo " and a " Llan Geneu " which cannot 
be identified, were presided over by simple abbots, and only in 
the higher status and privilege of these dignitaries was there 
any substantial recognition of the former standing of these 
churches as episcopal sees. 66 Llanbadarn Fawr in Ceredigion 
is another ancient church which early mediaeval tradition alleged 
to have been at one time the seat of a bishop, 67 and it can hardly 
be doubted that other Welsh monastic centres were also ruled 
by bishop-abbots until the time came when the ideas of 
Christendpm as to the necessity of parcelling countries into 
dioceses, with one bishop for each of these divisions, were 
accepted by the Welsh and the four principal houses of St. 
David's, Llandaff, Bangor and St. Asaph were recognised as 
supreme in their own quarter of Wales. 68 

Scarcely any direct evidence is to be got as to the life of 
the Welsh monastic communities in the days of their early 
zeal and activity. But the indications are that it did not differ 
greatly from the manner of life led by the monks of Ireland 
and Scotland during the same period, so that what is known 
from the life of St. Samson 69 and the early British peniten- 

66 For the situation of the seven churches see Aneurin Owen's notes (LL. 
i. 559); Fenton (2), 218-9; Descr. Pemb. i. 296, 304, 307, 308, 310. There is 
nothing to indicate which Llandeilo was meant, and several churches with this 
dedication, e.g., Llanddowror and Llandeilo Abercywyn, lie between St. Issell's 
and Carmarthen. No Llangeneu is now to be traced in Dyfed. Lat. A. clumsily 
introduces " Egluyss Hwadeyn," i.e., Llawhaden, into the list, thus betraying, 
as in 13 (p. 791), the hand of a St. David's editor. 

67 Gir. Camb. (vi. 121-2 (Itin. ii. 4)) expressly states, what is implied in the 
" metropolis alta" and"antestes . . . Paternus" of the poem of leuan ap Sulien 
(H. and St. i. 665), that Llanbadarn was once " cathedralis ". It was no doubt 
on the strength of this tradition that Geoffrey of Monmouth made Cynog 
" Lampaternensis ecclesiae antistes " (Hist. Reg. xi. 3). While the tradition 
itself is of historical value, no importance should be attached to the notice in Gw. 
Brut, s.a. 720, implying there was a bishopric of Llanbadarn at that date, and it 
is very unlikely, too, that the " Idnert " stone found at Llanddewi Brefi com- 
memorates, as Edward Llwyd suggested, the murdered bishop of Gerald's 
account (Gibson, 644 ; Lap. W. 140 ; Inscr. Chr. No. 120). 

68 Some evidence of the monastic origin of Bangor and St. Asaph is afforded 
by the statement of Gir. Camb. (vi. 170 (Descr. i. 4)) that the former was 
under the patronage " Danielis abbatis " and by the reference to St. Kentigern's 
foundation of a monastery in " Llyfr Coch Asaph " (Thomas, St. Asaph, p. 179). 

69 Cited from the Bollandist text ; see note 96 to chap. v. 


tials 70 may be pieced out with the aid of Bede and of Adamnan, CHAP, 
the biographer of St. Columba. It is certain, in the first place, 
that the monks did not inhabit a single building, but lived in 
separate huts or cells, which were surrounded by a wall or 
rampart, after the pattern of that which girt the various 
buildings of a royal court or " llys ". n This was the " llan," or 
enclosure ; within it were also the church, the abbot's cell, the 
" hospice " for the entertainment of visitors, and such necessary 
outhouses as the kiln in which corn was dried to fit it for the 
mill. None of these buildings were of stone, any more than 
those set up within the precincts of a " llys," which are known 
to have been of timber and wattle. The very church was of 
wood, for, according to Bede, stone churches were almost 
unknown among the Britons. 72 It is not surprising, under 
these conditions, that not a vestige remains of the buildings of 
any early Welsh monastery and that no part of any Welsh 
church now standing is anterior in date to the coming of the 
Normans. 73 

Entrance into the community was through the monastic 
vow, which was taken in the church on bended knee 74 and was 
known as the vow of perfection. 75 The man thus became one 
of the " brethren," a body of comrades among whom there was 
complete equality, but who owed unquestioning obedience in 
all things to their spiritual father, the abbot. 76 There were 
often other officers who relieved the abbot of some of his 

70 Cited from H. and St i. 113-20. The MS. (Paris 3182) comes from 
Fe'camp Abbey and ultimately from Brittany. It is of the eleventh century 
(Bradshaw, Collected Papers, 1889, pp. 473-4). 

71 See, especially, V.S. Colttmb., with its references to the "ecclesia" or 
"oratorium" (i. 8, 32; iii. 23), the " hospitium " (i. 31, 32; ii. 39), the abbot's 
cell (i. 35 ; Hi. 21, 22), and the "canaba " (i. 45). In i. 3 mention is made of the 
" vallum monasterii " at Durrow. There was a hospice in the monastery of 
Piro (Ada SS. July, vi. 5796). 

78 " Ecclesiam de lapide, insolito Brettonibus more" (H.E. iii. 4). In -Lib. 
Land. 277, the erection of a wooden church about 1060 at Llangarran in Archen- 
field is mentioned, and it would seem likely that churches generally in that 
district were of light construction, in view of the large number which Bishop 
Herwald is said to have consecrated during the fifty years of his episcopate. 

73 So, substantially, Allen in Monumental History of the British Church 
(1889), p. 43. 

74 V.S. Columb. i. 32. 
H. and St. i. n8(9). 

78 " Offensus quis ab aliquo debet hoc indicere abati, non tamen accusantis 
sed medentis affectu, et abas decernat" (H. and St. i. 115 18). 
VOL. I. 14 


CHAP, duties, a priest who offered the eucharistic sacrifice, 77 a cook or 
steward who saw to the food supply, 78 occasionally a teacher 79 
and a scribe. 80 But none of these trenched in any way upon 
the authority of the abbot. Election by the monks appears to 
have been the regular mode of appointing an abbot in the 
larger monasteries, but considerable influence was exercised by 
the bishops of the district, 81 and another important factor made 
itself felt at an early period, namely, the force of blood re- 
lationship. Although the inmate of a monastery was regarded 
as having divorced himself from all family ties, it was neverthe- 
less the fact that abbots sought to secure for their own relatives 
the succession to their offices. One instance of the kind has 
already been given ; it will be remembered that Illtud had a 
nephew who confidently expected to succeed to the position 
held by his uncle and who was filled with jealousy when he 
saw that the dazzling virtues of Samson might bring about the 
ruin of his hopes. 82 This incident reveals clearly the exact 
situation ; there was no rule of hereditary succession, but merely 
a presumption in favour of relatives, which would not stand 
against the claims of a really brilliant outside candidate. 
Nevertheless, it is worthy of note that, of the eight abbots who 
ruled over lona and its daughter houses in the seventh century, 
all save one are known to have been relatives, near or distant, 
of St. Columba. 83 In the dependent monasteries, the heads of 

77 See Lib. Land, for " presbiter Catoci " at Llancarfan (268, 272, 273), 
"sacerdos" and "presbiter S. Ilduti " at Llantwit (257, 272), " sacerdos " and 
" presbiter Docunni " at Llandough (249, 258, 268, 272), " presbiteri tathiu " at 
Caerwent (270). In the Book of St. Chad (Lib. Land. pref. xlvi) is " sacerdos 
teiliav ". 

78 When Samson is made "pistor" of Piro's monastery, he is thus ad- 
dressed by Dubricius : " omnia bona quae in hac cella, Deo donante, abundant ad 
dispensandum tibi praecipio " (5816). Elsewhere in the life the " pistor " and 
the " oeconomus " are distinguished (577-8). 

79 Lib. Land, has "doctor Catoci" (273), "magister sancti catoci" (271, 
274). Cf. Cambro-Br. SS. 82-3, for the constitution of the "clas " at Llantwit 

80 " Dissaith scriptor " (Lib. Land. 224) recalls the " scribe " of Irish mon- 
asticism, for whom see Warren, Liturgy and Ritual of the Celtic Church, p. 18. 

81 Samson was elected abbot in place of Piro by the suffrages of the monks 
("consilio facto . . . omnes voluerunt "), but Dubricius was the moving spirit 

82 577^- Compare the story told in the same life of the desire of Dubricius to 
hand on his office to an unworthy favourite (probably a relative called Morinus 


83 See the genealogical table in Fowler's edition of V.S. Columb. 


which were appointed, by the abbot of the parent community, CHAP, 
there was still further scope for the exercise of family partiality ; 
the appointment by St. Columba of a maternal uncle as head 
of " Hinba insula " 84 and by St. Samson of a paternal uncle as 
head of the monastery ceded to him in Ireland 85 are instances 
of what was no doubt a common practice. The inherent right 
of noble blood to rule was in this age not questioned, either in 
Church or State, and when a family of good birth had once 
given an abbot to a particular monastery, their interest in it 
and influence over it tended to become permanent. 

The monks were at first entirely supported out of the lands 
immediately attached to the monastery, which they tilled with 
their own hands. Other endowments were regarded with 
suspicion, as likely to lower the high spirit of monastic self- 
denial, 89 and though they inevitably came, as the monastery 
grew in reputation, until each important " llan " had scores of 
rent-yielding estates in different parts of the country, this was 
a departure from primitive ways and ideas. All the early 
authorities agree in representing the monks as engaged in 
agriculture and the care of cattle. " He who breaks his hoe," 
runs one of the provisions of the monastic code ascribed to 
Gildas, " where it had no fracture before, let him pay for the 
damage by extra work or keep a special fast." 87 They went 
out to their various tasks, assigned to them by the abbot or his 
representative, in the morning and spent a long day in the 
fields, to return at nightfall for the one set meal of the day. 88 
The fare was of the plainest, its main constituents 'being bread, 
butter, cheese, eggs, milk and vegetables, but meat and beer 
were not altogether excluded, save in the monasteries bound 

V.S. Columb. i. 45. 

88 5826. " Avunculus " is used for " patruus ". 

88 Bede says of the Irish monks who settled in Northumbria under Oswald 
and Oswy that none of them accepted " territoria ac possessiones ad construenda 
monasteria, nisi a potestatibus saeculi coactus " (II. E. iii. 26). 

87 H. and St. i. 115, 26. 

88 The " discipuli " of Piro went out in the morning " ad opus exercendum " 
(5796). Cf. V.S. Columb. i. 37 (fratres, post messionis opera, vespere ad 
monasterium redeuntes) ; iii. 12 (dum fratres, se calceantes, mane ad diversa 
monasterii opera ire praepararent) ; iii. 23 (ad visitandos operarios fratres . . . 
in occidua insulae louae laborantes parte). The monks of Bangor Iscoed lived 
" de labore manuum suarum " (Bede, H.E. ii. 2). H. and St. i. 114, 15, shows 
that the "cena " covered a good deal more than half the daily ration of food. 



CHAP, by rules of special severity. 89 Drunkenness, though esteemed 
a serious offence, was not altogether unknown. 90 

Manual labour was, of course, not the sole occupation of the 
monks. Much time was given to the study of the scriptures, 91 
and writing absorbed the energies of those brethren who had 
special qualifications for this work. 92 There was each day and 
night a regular succession of services in the church, when 
portions of the Psalter were chanted in Latin, the universal 
language of the Western Church. 93 On Sundays and saints' 
days there was also a celebration of holy communion, while the 
holiday character of the day was further marked by a general 
cessation of labour and a more generous diet. 94 Wednesdays 
and Fridays, on the other hand, were days of fasting 95 and the 
season of Lent was always observed with great rigour. 96 The 
saints' days observed included not only the festivals recognised 
by Christendom at large, but also the anniversaries of the great 
figures of the Celtic Church, whose death-days were treated as 
heavenly birthdays and made an occasion for feasting and 
rejoicing. 97 

Hard work and assiduous devotion were not the only means 
of moral discipline provided by this ancient monastic system. 
There were also severe penalties for all lapses from the monastic 
standard of purity and simplicity. The monastic virtues were 
humility, readiness to obey, almsgiving, and above all, chastity, 
and any breach of these obligations was punished in proportion 
to the gravity of the offence. In the more serious cases a period 
of " penitence " was prescribed, during which the offender was 
excluded from communion, placed upon fasting diet, and 
required constantly to ask pardon for his sin. Penitents were 
sometimes grouped together in particular monasteries, where 
they were sent to work out their discipline ; often, however, they 

89 H. and St. i. 113, i ; 115, 22 ; 119, n. 

90 For legislation dealing with it see H. and St. i. 114, 10 ; 118, 2 ; 119, 
3 4 ' f r tne actual case of Piro see the life of St. Samson, 5820. 

91 V.S. Columb. i. 24. 

92 Columba was a notable scribe (ii. 8, g, 16 ; Hi. 15). 

93 H. and St. i. 115, 19; Williams, Gildas, p. 282. 

94 V.S. Columb. ii. i, 44; iii. 12, 17. 
s *Ibid. i. 26; Bede, H.E. iii. 5. 

96 Dubricius spent every Lent in the monastery of Piro (5816). 
97 V.S. Columb. ii. 45 ; iii. ii ; Life of St. Samson. 


were banished to distant countries, the enforced pilgrimage being CHAP, 
added as a further item to the sum of penance. 98 

In their relations with the outside world the early Celtic 
monks resembled the friar rather than the cloistered monk of 
mediaeval times. Custom did not limit them to the precincts of 
their monastery, but, subject to the authority of their abbot," 
allowed them to wander hither and thither on errands appropriate 
to their calling. In the early period, when heathenism had not 
completely relaxed its hold upon the country, they went about 
preaching, 100 and thus the monastery came to be regarded, not 
merely as a home for pious recluses, but also as the natural 
ecclesiastical centre of the district in which it stood. Monks 
were often charged with special commissions which took them 
far from their place of settlement, or were suffered as " peregrini " 
to undertake long journeys on which they had set their hearts. 101 
A monk might even be sent forth, with a number of companions, 
as bees from an old hive, to found a new community, and, in 
the days when ascetic fervour was at its height, it was always 
the ambition of such a band to discover some desert island or 
dell in the heart of the forest in which they might live the life 
of self-denial. 102 Thus it was that the isles of the Atlantic 
coast' in Ireland and Scotland came to be "isles of the saints," 
that the islands of Priestholm, Bardsey, St. Tudwal, Ramsey, 
Caldy and Barry on the Welsh seaboard also offered asylum to 
men weary of the turmoil of the world, and that on the main- 
land many a monastic " Diserth " or " desert " arose in what 
was once a forest solitude. 103 

It must, of course, be understood that the picture of Welsh 
monastic life which has just been drawn will only hold good 

98 For the penitential system see H. and St. i. 113-20; Williams, Gildas, 
pp. 272-5. 

89 V.S. Columb. i. 6. 10 Bede, H.E. in. 5, 26 ; Life of St. Samson. 

101 The " peregrinus " appears in Bede (H.E. iii. 19), the Life of St. Samson 
(5826, 5840), and V.S. Colnmb. ii. 39 ; iii. 7. The British bishop Marcus in the 
ninth century spent the end of a long and busy life in voluntary exile at the 
abbey of St. Me'dard, nearSoissons (Mommsen's Nennius, pp. 120, 172). 

108 V.S. Columb. i. 6, 20; ii. 42. 

103 For this use of " desertum " and its Welsh and Irish equivalents see 
Cymr. Trans. 1893-4, IX 3 > Descr. Pemb. i. 260. Three churches in Wales bore 
the name, viz., " Y Ddiserth yn Nhegeingl " (Bruts, 369), " Y Ddiserth yn Elfael " 
(Pen. MS. 147) and " Y Ddiserth yn Rhos " (Evans, Rep. i. p. 971), now Llansant- 
ffraid Glan Comvay. 


CHAP, for the early monks, those who founded the Welsh Church in 

the sixth, seventh and, one may perhaps add, eighth centuries. 

As the primitive ideals lost their fascination, as the monasteries 
gathered wealth and the monk's life became one of ease, main- 
tained at the expense of the labour of others, degeneration set 
in, and the havoc caused by the Danish attacks of the ninth 
century added another element of disorder. How rapidly a mon- 
astic system might absorb, like some subtle poison in the veins, 
the secular spirit is shown by Bede's letter to Archbishop Egbert 
of York, in which he complains that within thirty years the 
monasteries of Northumbria, once famous for their unworldly 
purity, had to a large exent become secular institutions, in 
which no monastic rule was observed. The two most important 
changes which affected the Welsh communities were the growth 
of territorial endowment and the abandonment of the celibate 
life. The first process must have begun early; it was the 
natural way in which a prince or wealthy landowner paid honour 
to the memory of the local saint or atoned for some injury 
done to the Church. On the margins of the Book of St. Chad, 
a MS. of the Gospels which was at Llandaff during the ninth 
century and part of the tenth, are several entries of that period 
typical of the method by which the greater churches grew rich. 104 
" This writing sheweth," runs one, " that Rhys and the tribe of 
Grethi have given to God and St. Eliud (i.e., Teilo) Trefwyddog 
... its render is forty loaves and a ram in the summer, in the 
winter forty loaves, a sow and forty sheaves (of oats). . . . He 
that shall keep this compact shall be blessed, he that shall break 
it shall be cursed, of God." 105 Such were the gifts which in the 
course of three or four centuries made St. David's, Bangor, 
Llandaff, St. Asaph, Clynnog, Llancarfan, and other churches the 
centres of groups of manors or hamlets of rent-paying serfs. 
The abbot was no longer an apostle, a worker of miracles, a 
terror to evil-doers, but simply a mighty landowner. The vow 

104 The marginalia of the Book of St. Chad are given in facsimile, with a full 
discussion of their meaning, in Lib. Land. pref. xlii-xlviii and the accompanying 
plates. See also Bradshaw, Collected Papers, 1889, pp. 458-61. 

105 Mannuclenn " is not otherwise known, but, judging from the ordinary 
"dawn bwyd" of a South-Welsh servile " tref " as given in the Laws (LL. i. 
532-4, 770), I regard " sheaf" or "handful" (from "manucla" = manua, a 
bundle) as a far more likely explanation than the " sucking pig" of Lib. Land. 


of poverty, though still so far observed that the monk had no CHAP. 
property of his own, had lost most of its meaning through the 
great accession of riches to the community. As for the vow of 
celibacy, that probably went more gradually, as the result of 
the position held by the monks in so many cases as clergy in 
the chief or mother church of the locality. Although the monk 
took a vow which put marriage out of the question, no such 
solemn obligation was entered into by the cleric, and the 
objection to a married clergy, though very general in the Middle 
Ages and enforced by reformer after reformer, did not rest upon 
any fundamental law of the Church ; it was partly a matter of 
sentiment, arising out of the belief that the monastic ideal was the 
higher one, and partly a matter of policy, based on well-founded 
dread of the rise of a system of hereditary succession. Whether 
in any particular age and country the clergy were actually 
married men and founded families depended upon the local 
public opinion, and in Wales it seems clear that in and after the 
ninth century clerical marriage and family property in church 
offices were pretty firmly established. Among the clergy who 
witness the grant of freedom to Bleiddud son of Sulien and his 
heirs entered in the Book of St. Chad is " Cuhelyn son of the 
bishop," i.e., of the Bishop Nobis who then presided over 
Llandaff, 106 and in the same century another Bishop Nobis, 
seated at St. David's, was followed in the see by a relative 
named Asser, who became famous as the biographer of King 
Alfred. 107 At the era of the Norman Conquest the system had 
reached its height. The sons and grandsons of Sulien, who was 
Bishop of St. David's from 1073 to IO 78 and again from 1080 to 
1085, were the leading clergy of the diocese for the best part of 
a century, and when Giraldus Cambrensis says that in Wales sons 
regularly succeed their fathers in church livings, 108 his statement 
is confirmed by the evidence of the Liber Landavensis that this 
was quite the rule in the district of Archenfield, then altogether 
Welsh, at the end of the eleventh century. 109 While there are 
no documents which make it possible to trace the gradual dis- 

106 Lib. Land. xlvi. 

107 " Expulsione illorum antistitum qui in eo (i.e., St. David's) praeessent, 
sicut et Nobis archiepiscopum (for this title see note 43), propinquum meum, et 
me expulit " (c. 79). 

108 Works, iii. 130 ; cf. vi. 214 (Descr. ii. 6). 10B Lib. Land. 275-7. 


CHAP, appearance of the celibate ideal, it is indubitable that a change 
of this kind was brought about, were there no evidence other 
than that afforded by St. David's, where in the sixth century 
Dewi preached and practised a pitiless austerity and in the 
thirteenth century the canons lived with their wives, as Giraldus 
complains, under the very shadow of the cathedral. 110 

Yet there were, let it not be forgotten, certain communities 
which kept the primitive discipline as late as the time of 
Giraldus, favoured by their isolation from the world. One 
such occupied Priestholm, 111 off the easternmost point of 
Anglesey, a body of hermits who still followed the custom of 
maintaining themselves by the labour of their own hands and 
whose rigorous asceticism was such that no woman was ever 
allowed to- enter the island. The soil was hallowed by the 
bones of departed anchorites without number, and the very 
animals of the place, it was believed, were in league with the 
powers above, for any quarrels which ruffled the surface of this 
peaceful and sheltered community were forthwith avenged by 
inroads made on its supplies by the little field-mice of the 
island. Bardsey, the " Ynys Enlli " of the Welsh, was held 
by a group of hermits of the same pattern, 112 and had, in con- 
sequence, a high reputation throughout Wales for sanctity. 
The dearest wish of the Welsh warrior or poet, as he approached 
the end of his stormy career, was to be buried in " the beauteous 
isle of Mary," where the heaving ocean made a girdle round 
the churchyard and where he might share the sleep of twenty 

no w or k Si iii. 128-9. In South Wales the law recognised as legitimate the 
sons of clerics who had not taken priest's orders, but the son of a priest was " be- 
gotten against law " (LL. i. 444 (Dim.) ; 760 (Gw.) ; ii. 857 (Lat. B.)). 

m Gir. Camb. vi. 131 (I tin. ii. 7). The name Priestholm (Priests' isle) is of 
Scandinavian origin. In Welsh the island was known originally as Ynys Lannog 
(" insula glannauc " in Harl. MS. 3859 Cymr. ix. 157), from the mythical 
Glannog, father of Helig Foel and grandfather of certain local saints (Bonedd y 
Saint in Myv. Arch. II. 24, 30, 45 (416, 419, 426)). This is the name which Gir. 
Camb. gives as " Enislannach " and, in an unfortunate excursion into the fields 
of Welsh etymology, explains, by reference to " llan," as " insula ecclesiastica ". 
Ynys being a feminine noun, the disappearance of the g, which led him into this 
pitfall, is quite regular for Welsh of this period. " Ynys Seiriol," due to the 
close connection of the island with Seiriol's church of Penmon, appears to be 
much later, while " Puffin Island " is a modern tourists' designation. 

112 Gir. Camb. vi. 124 (I tin. ii. 6). Rhys derives the Welsh name from an 
original Ynys Fenlli and connects it with Benlli Gawr (Arth. Legend, 354). 


thousand saints, " the pure-souled dwellers of Enlli "."* Nor CHAP, 
were these stricter communities confined to the islands of the 
coast ; there was one at Beddgelert,* in the rocky heart of 
Snowdon, a body of clergy who led a celibate life and were 
given to self-denial and the practice, as became their situation 
on the confines of Eifionydd, Ardudwy, Arfon and Arllechwedd, 
of hospitality. 114 

These groups of celibates or " meudwyaid " 115 appear to 
have been the only monasteries in the strict sense which sur- 
vived among the Welsh at the era of the Norman Conquest. 
There, were no Benedictine houses until the Norman conquerors 
of South Wales brought into the country this type of 
monastic foundation, and upon the native Welsh, it will be 
seen in a later chapter, no impression was made by any move- 
ment of this kind until the middle of the twelfth century, the 
great age of Cistercian expansion. Thus the men under vows 
(diofrydogion) mentioned in the Welsh Laws, who have abjured 
women, the eating of flesh, and riding on horses, men of 
peculiar sanctity whose concurrence gives special virtue to a 
judicial oath, are probably the members of these celibate com- 
munities, 110 together with another class which was to be found 
at all times in Wales from the introduction of Christianity to 
the age of the Protestant Reformation, namely, solitary hermits, 
dwellers in isolated cells. This type remained unchanged, 
however much communities might alter, throughout the ages ; 
the holy man consulted by the British clergy before they went 
to their second conference with Augustine m is represented six 
centuries later by Wechelen, the hermit of Llowes in Elfael, who 
was regarded by the whole countryside as a prophet and a 

113 See the " Deathbed of Meilyr the Poet " (Marw Ysgafyn Veilyr Brydyt 
" ysgafyn " is explained in Mots Latins, 215) in Myv. Arch. I. 192-3 (142). The 
translation in Lit. Kym. (2), 13-5, gives the general sense, but is not to be 
trusted in details. For an important early account of Enlli see Lib. Land. 1-5 
(story of Elgar). 

114 Gir. Camb. iv. 167. The place is not named, but is clearly Bedd- 

118 For " meudwy" see note 37 above. 

118 LL. i. 408, 594 (Dim.), 688, 750 (Gw.) ; ii. 769, 794, 803 (Lat. A.), 836, 
850 (Lat. B.) all South-Welsh texts; the inclusion of " gwyr diofrydog" in 
a " rhaith " was not apparently usual in Gwnyedd. Some of the texts have 
" lliein " for " cic ". 

117 Bede, H.E. ii. 2. 


CHAP, healer of the sick and whose guidance and ghostly counsel 
v were sought, poor and illiterate though he was, by so consider- 
able a person as Giraldus Cambrensis. 118 The " gwr ystafellog " 
(chambered man) appears in the Laws, in virtue of the pay- 
ment due on his death to the lord of the district in which he 
lived, 119 and from the same source it is known that there were 
women anchorites, singular among all Welshwomen in that 
they also paid this " ebediw " or heriot, which was otherwise 
only exacted from men. 120 This was because other women 
paid an " amobyr " or marriage-fee, which would never be- 
come due in the case of a woman vowed to life-long seclusion. 
The " mother churches " of monastic origin were well dis- 
tributed over the whole of Wales, and 'their position in the 
earlier ages of Welsh Christianity is accurately defined in the 
name which has been given them of " missionary churches " 
Every important " clas " had its " out-stations," to use the 
language of modern missions, for each of which one or more 
of the members was responsible and which became in time a 
chapel of the! mother church. Thus the " clas " of Cybi at 
Holyhead had chapels at Bodedern, Llandrygarn and Bodwrog, 
while in Powys traces of the supremacy of Meifod as the old 
ecclesiastical metropolis of the district in which it stood are 
found as far afield as Llanfair Caereinion and Alberbury. In 
this way may be explained the rise of a very large number of 
the churches which ultimately, through the assignment to them 
of the tithes of certain definite areas around them, became the 
parish churches of to-day. They were at first the mission 
stations of the mother churches. But a number of churches 
remain which cannot be accounted for in this way, and these, 
it is reasonable to suppose, were erected by the efforts of laymen, 
who wished to make provision for their own spiritual needs. 
Churches built in this way seem to have been far fewer in Wales 

us Works, i. 89-93 (De Rebus, iii. 2) ; i. 175 (Invect. vi. 20). 
U9 LL. i. 492, 24 (Dim.), 686, n (Gw.) ; ii. 12 (from a Venedotian 
source), 797 (Lat. A.), 885 (Lat. B.) all fix his " ebediw " at 24d. 

120 LL. i. 96, 52 (Yen.), 492, 25 (Dim.), 686, 12 (Gw.) ; ii. 12, 797 
(Lat. A.), 885, which give the payment variously as I2d. and i6d. Aneurin Owen, 
following Moses Williams (Wotton, 585), translates " ystafellog " as " cottar " 
(i. 687), but the " sanctimonialis " of Lat. A. and B. is conclusive as to the mean- 

121 Palmer in Arch. Camb. V. iii. (1886). 

THE AGE OF I SOL A T10N. * 1 9 

than in England, possibly because building in wood was so CHAP, 
much simpler and less costly that there was here no occasion to 
invoke the aid of wealth when a new sanctuary was to be raised. 
But the "king's chapels," the churches built for the use of the lord 
of territory and his train at the principal courts of his kingdom, 
were certainly of this type. According to what seems to be 
the old law, the chapel of the court, in which the men of the 
court worshipped, 122 was one of the buildings which the king's 
villeins were to put up and keep in repair ; 123 it had a per- 
manent chaplain, who lived in the " maerdref," the hamlet 
attached to the court, 124 but its chief officer was the " offeiriad 
teulu," the king's priest, who followed the king as he travelled 
from court to court, acted as his chaplain and secretary, and 
received the bulk of the religious offerings of the king and his 
courtiers. 125 He was the real parson of the royal chapels and 
his office was naturally in the gift of the king. 126 


(The chief authority upon Celtic art in Christian times is the late J. Romilly 
Allen, whose book on the subject, dealing also with the earlier art of the pagan 
epoch, appeared in 1904. Use has also been made in this and other chapters of 
the handbook written by Mr. Allen for the Society for Promoting Christian 
Knowledge, entitled The Monumental History of the Early British Church 
(London, 1889). The stones are fully described in Lap. W. discoveries made 
after 1879 are dealt with in Arch. Camb. See also articles by Mr. Allen in Arch. 
Camb. V. x. (1893), 17-24; xvi. (1899), 1-69.) 

What has been said of the ruling tendencies in the Welsh 
Church during the centuries which followed the age of Gildas 
and Dewi suggests that, though much quiet and enduring 

1H LL. ii. 68, 69. 

123 The "cappel " appears in Dim. (LL. i. 486) and Gw. (i. 772), and was 
found, there is reason to think, in the copies of the laws used in the compil- 
ation of Lat. A. and B. see ii. 785, 828. In Ven. (i. 78) the list of nine buildings 
is otherwise made up, but this may well be due to the fact that in the thirteenth 
century royal chapels were no longer built in the old haphazard fashion. 

124 The " offeiriad teulu," like other members of the court, had no house 
there; he lodged with "caplan y dref" (LL. i. 358, 634), ' sui capellani " (ii. 
755), or " cappellani sub eo servientis " (ii. 819, 897). In Ven. (i. 16, 52) this 
officer is styled " clochydd," i.e., bellringer or sexton. 

185 Ven. i. 16-18 ; Dim. i. 364 ; Gw. i. 638. 

IK Ni ddyly esgob bersoni neb ar sapelau y brenin heb ei ganiad " (A 
bishop ought not to appoint any one to the king's chapels without his consent) 
Ven. i. 18, MSS. A., E. MSS. B., D. read differently, but assert the same 


CHAP, work was done, there was a loss of originality and a declension 
VIL from early spiritual ideals. This impression is not removed if 
one looks at the intellectual output of the period, the per- 
manent records of its activity in the spheres of literature and 
art. It was not a great epoch or an epoch of great men ; 
though in touch with Ireland, 127 the home at this time of a 
valuable culture of native growth, Wales was but feebly moved 
by this life across the sea and rather shared the backwardness 
of the rest of Britain and of Western Europe. 

Of the architectural work of the period it is impossible to 
judge, for it has been already stated that nothing remains of 
any church built in Wales before the Norman Conquest 128 
Even if it be true that all churches of the time were not of 
wood, and occasionally one was built of stone, it seems clear 
that the conditions were such that architecture had as an art no 
scope for development. 129 The only memorials in stone which 
have survived as evidences of the artistic culture of the period 
and these are fairly abundant are the carved stone crosses 
and inscribed stones which stood in the graveyards and on 
other consecrated spots connected with the early churches. In 
respect of these a direct succession may be observed from the 
standing stones bearing inscriptions in Latin capitals or ogam 
characters, or both, which were discussed in Chapter IV. of this 
work. At first the stones are without ornament, save an oc- 
casional incised cross, and the later date is inferred merely from 
the form of the letters, which are no longer capitals but of the 
type known as minuscule. There is a stone of this description 
in the churchyard of Llanwnnws, Cardiganshire, which is be- 
lieved to be of the early part of the ninth century and bears a 
Latin legend that may be thus rendered : " Whosoever may 
decipher this name, let him utter a benediction on the spirit 
of Hiroidil son of Carotinn ". 13 Still older is the inscription 
on the stone at Caldy Island ; to the old ogam epitaph was 

127 The British bishop Marcus, to whom the copyist of the Vatican MS. 
erroneously ascribes the Historia Brittonum, was educated in Ireland in the 
early part of the ninth century (Mommsen's Nennius, pp. 120, 172). 

128 P. 209. 129 Allen, Celtic Art, p. 179. 

130 it Quicunque explicaumt hoc nomen det benedixionem pro anima hiroidil 
filius carotinn " (Arch. Camb. V. xiv. (1897), 156-8). See also Inscr. Chr. No. 
122; Lap. W. 144; Arch. Camb. IV. v. (1874), 245-6, and V. xiii. (1896), 135. 
The ' ' nomen " is the xps which is placed at the head of the cross. 


added about A.D. 750 the following: "And I have marked it CHAP 
with the sign of the cross. I beseech all that pass hereby VII. ' 
to pray for the soul of Cadwgan." 131 Meanwhile the art of 
illuminating manuscripts had reached a high pitch of perfection 
in Ireland, and a special Irish style of decoration had been 
evolved, in which the curves and spirals of the earlier Celtic 
world were grafted upon the interlacing ribbon and key-pattern 
work of Christian art and results achieved of wonderful com- 
plexity and artistic merit. Nothing more beautiful of its kind 
can be imagined than the illuminated border and initial work 
in the Book of Kells, a copy of the gospels prepared in the 
first half of the eighth century for the use of one of the most 
famous of the monasteries of St. Columba. In the course of 
time the principles of design followed in the decoration of 
books were applied to the ornamentation of stone monuments, 
and the sculptor's art produced the graceful stone crosses, 
covered with intricate ornament, of which many examples are 
to be found in Scotland and Ireland. The ninth century is 
believed to have been the age in which the highest degree of 
skill was attained by the Celts in craftmanship of this kind. 132 
In Wales the same causes were at work, evolving a native art 
of sculpture out of primitive and Christian decorative elements, 
but with less felicitous results. A good deal of carving in the 
Celtic style was at this period executed in the country ; nearly 
a hundred stones if fragments be taken into the reckoning 
with Celtic ornament upon them have come to light in such 
ancient centres of church life as Penmon, Corwen, Meifod, St. 
David's, Margam, Llantwit Major, Penally, Nevern and Llanba- 
darn Fawr. But it is agreed that the Welsh crosses are inferior 
in design and workmanship to the Irish ones ; spirals are almost 
wholly absent, there is little figure sculpture, and there is less 
grace of form. The new style of decorating tombstones was 
adopted in Wales, but the artists and gravers had not the Irish 

Among the more notable of the Welsh crosses may be 
mentioned the two at Nevern and Carew, the former 12 and 

131 " Et singno crucis in illam fingsi : rogo omnibus ammulantibus ibi exorent 
pro anima catuoconi " (Arch. Cam. V. xiii. (1896), 98-103). See also Inscr. Chr. 
No. 94; Lap. W. 106-8; Fenton (2), 251. 

132 Celtic Art, pp. 286-8. 


CHAP, the latter 14 feet high. The Carew cross bears, as was often 
the case with these elaborately sculptured monuments, the 
name of the artist, who appears to have been one Maredudd of 
Rheged. 133 An important group of these carved tombstones 
belongs to Glamorgan, where over thirty stones showing Celtic 
ornament have at various times been discovered. The Gla- 
morganshire crosses, of which twelve belong to Margam and 
Llantwit Major, are often of remarkable form, having a short, 
broad shaft and a round head, and bear singular inscriptions, 
such as this on a stone at Llantwit Major : " In the name of 
the Supreme God begins the cross of the Saviour which Abbot 
Samson prepared for his soul and for the soul of King Ithel 
and Arthfael and Tecan ", 134 It is noteworthy that the inscrip- 
tions on these crosses, when they extend beyond the mention 
of a name, are (with one exception) l35 written in Latin and 
not, as in Ireland, in the vernacular speech, which may be 
regarded as showing that such culture as was maintained in the 
Welsh Church ran on the traditional Roman lines, and was not, 
like the Irish, a native growth, drawing its inspiration from 
popular sources. The custom of setting up tombstones with 
minuscule inscriptions and Celtic ornament lasted as late as 
the end of the eleventh century and the advent of the 
Normans, for in 1891 there was discovered in St. David's 
Cathedral a memorial of this kind to Hedd and Isaac, sons 
of Bishop Abraham, who presided over the see from 1078 to 
io8o. 136 

Beyond the carved and inscribed stones little remains to 
tell of the degree of progress attained by the Welsh in artistic 
performance at this time. There is no evidence that they had 
the Irish skill in illuminating, for, though the Book of St. Chad 
was for a considerable time at Llandaff and was then known 

133 Inscr. Chr. No. 96; Lap. W. 119; Arch. Camb. V. xii. (1895), 

134 K j n nom ine di summi incipit crux salvatoris quae preparauit Samsoni apati 
pro anima sua et pro anima iuthahelo rex et artmali et teca(n) " (Inscr. Chr. No. 
62; Lap. W. 12 ; Arch. Camb. V. xvi. (1899), 147-150). 

135 This is the well-known stone preserved in Towyn Church, the inscription 
on which has never been satisfactorily explained and is possibly not genuine. 
See Inscr. Chr. No. 126 ; Lap. W. 158 ; Arch. Camb. V. xiv. (1897), 142-6 

136 Arch. Camb. V. ix. (1892) 78 ; x. (1893), 281. 


as " Efengyl Teilo," m it came there by purchase 188 and may CHAP. 
well have been produced by Irish art. Ancient bells, of the ' 
quadrangular Celtic form, have been found in two or three 
places in Wales, but only in one case was there any ornament 189 
Learning did not wholly die in the monasteries of Wales. 
It is known, from the existence of MSS. of Ovid, of the poet 
Juvencus, and of Martianus Capella, in which many of the 
words are interpreted in old Welsh of the ninth century, that 
the studies of the monks were not confined to the text of the 
Scriptures, but ranged over a fairly wide field. 140 Nevertheless, 
the bad Latin of the inscriptions shows that composition in 
that language was at a low ebb, and there are in fact but two 
works written by Welshmen between 600 and 1050 which have 
survived, neither of them of much literary merit, though they 
are full of interest for the historian. They are the Historia 
Brittonum usually connected with the name of Nennius, and 
Asser's De Rebus Gestis ^Elfredi. Many problems, into which 
it is not possible here to enter, cluster round the former of 
these two works ; the date at which it was first put together, 
its original author, the relations to each other of the various 
MSS., are still under discussion. 141 But some facts stand out 
with sufficient clearness and enable an estimate to be formed 
of the significance of the work as an index to Welsh culture. 
About the year 800 a little collection of tracts on the history 

187 " Sit maledictus a deo et a teiliav, in cuius euangelio scriptum est " (Lib. 
Land. xlvi.). 

138 Emit gelhi filius arihtiud hoc euangelium de cingal . . . et dedit pro 
animasua istum euangelium deo et sancti teliaui super altare " (Lib. Land, xliii.). 
It is probably of the eighth century (Celtic Art, p. 175). 

139 Arch. Camb. IV. ii. 271-5 ; Celtic Art, pp. 196-201. The famous bell of 
St. David, called " Bangu " and kept at Glasgwm, was probably of this type (Gir. 
Camb. vi. 18 (Itin. i. i)). 

140 For the old Welsh glosses see Bradshaw, Collected Papers, pp. 281-5, 
453-88 ; IV. Anc. Bks. ii. pp. i, 2, 311-4 ; Gr. Celt. (2), 1054-1063; Arch. Camb. 
IV. iv. (1873), 1-21. 

141 Earlier editions of the Historia Brittonum (Gale, Scriptores XV. 93- 
139, 1691; Gunn, 1819; Stevenson, 1838; Petrie, Man. Hist. Br. pp. 47-82) 
have been superseded by that of Mommsen (Monumenta Germaniae Historica : 
Chronica Minora saec. iv. v. vi. vii. Berlin, 1894), in which there is a complete 
critical apparatus, and use is made of an important MS. not known to the earlier 
editors, viz., the Chartres MS. of about A.D. 900. In Nennius Vindicates (Berlin, 
1893) Zimmer had for the first time shown the true relation of Nennius to the 
work so commonly cited under his name, and in Momm sen's edition the 
" Nennian " passages are printed separately. 


CHAP, and geography of Britain was known in Wales in a number of 
IL copies ; it included the Saxon Genealogies, which has been 
treated in the earlier chapters of this book as a < valuable author- 
ity for the sixth and seventh centuries, 142 and the rest of the 
collection may possibly be quite as old. 143 One copy, from 
which is derived nearly all the extant MSS., was transcribed 
by a man who had a special interest in the ruling family of 
Buellt and Gwerthrynion, districts bordering on the upper 
waters of the Wye ; 144 a somewhat later copyist wrote in the 
fourth year of Merfyn Frych, ruler of Gwynedd, i.e., about 829. 145 
Another copy, compiled about 800, was the work of one 
" Nennius," or, to use the Welsh form, Nyniaw, who calls him- 
self a disciple of Bishop Elfodd and freely edits the original 
" volume of Britain," as he terms it, which he has in his hands. 146 
It is beyond dispute that the original author of the Historia 
Brittonum and its editor Nennius were very unskilful writers 
of Latin and had a very limited knowledge of the general 
course of history. The gulf between them and Gildas as Latin- 
ists is immense ; fault may no doubt be found with the over- 
elaborate style of the older writer, but it was at any rate the 
fruit of profound study. The matter, too, of the Historia 
reveals to us a community in which folk-lore takes the place of 
learning. Use is made of some of the historians whose works 

142 See note 84 to chap. iv. 

143 This is Mommsen's view (Introd. 113, 117), as against Zimmer, who holds 
that the Historia was written about 800 (Nenn. V. 10). 

144 All MSS. save the Chartres MS., which is an imperfect copy, have the 
passage (c. 49), "Fernmail ipse est qui regit modo in regionibus duabus Buelt et 
Guorthigirniaun," and the descent of Ffernfael from Vortigern. According to 
Jesus Coll. MS. 20, Ffernfael's cousin Brawstudd was married to Arthfael of 
Morgannwg (see Pedigrees IX. and XIV. in Cymr. viii. 85, 86), which would 
indicate that this king flourished about 800 a conclusion also to be drawn from 
the number of generations between him and Vortigern. 

145 The reference to the " annum quartum Mermini regis " (c. 16) is not found 
in the Chartres MS. or in the Irish version, and does not belong, therefore, to the 
original text of the Historia, or, apparently, to the Nennian edition. The 
year is no doubt the same as that given in many MSS. as A.D. 831 (Mommsen, 
p. 146), for it is not likely that Merfyn Frych obtained the throne of Gwynedd 
until the death, recorded by Harl. MS. 3859 under the year 825 (Cymr. ix. 164), 
of Hywel ap Rhodri, the last male representative of the line of Cunedda. 831 
may easily, in fact, be an error for 829 (observe that the year of the passion is 
given as 796), the precise year required if Merfyn succeeded in 825. 

146 As a disciple of Elfodd, Nennius must have lived about 800, and the same 
conclusion is suggested by the fact that the Historia was known under his 
name to the Irish scholar Cormac (836-908). 


were current in ecclesiastical circles, of Jerome, of Prosper and CHAP, 
of Isidore, 147 but inferences are drawn from the writings of 
these reverend fathers which would have greatly surprised them. 
Because the chronicle of Jerome used for the earlier part of the 
Roman period is without the computation of the year by means 
of its consuls, while the continuation of Prosper has them, it is 
gravely said that in the time of the Emperor Maximus " there 
began to be consuls at Rome and never afterwards were they 
called Caesars " ! 148 An old list of seven emperors who visited 
Britain is taken, and it is assumed that save for the visits of 
these seven the Britons were independent of Rome ; then the 
writer adds that, while he has only found the names of seven 
" in the ancient traditions of our elders," the Romans speak of 
two more; these, however, are Septimius Severus and Con- 
staritius, whom he has already, without realising the fact, 
included in his list 149 

Nennius, it has already been said, styles himself a disciple 
of Elfodd, from which it may be inferred that he. flourished 
about the year 800 and belonged to the party which desired a 
closer connection between the Welsh Church and the rest of 
Christendom. 160 No copy of the Historia exactly represent- 
ing his edition of the work seems to have been preserved, but 
it has been reconstructed in its main features by the skill of 
Zimmer and Mommsen. In his preface Nennius represents 
himself, as was the manner of his kind, as an original compiler 
rather than a mere editor ; " I have," he says, " gathered to- 
gether all I could find not only in the Roman annals, but also 
in the chronicles of the holy fathers Hieronymus, Eusebius, 
Isidorus and Prosper, and in the annals of the Irish and the 
English and in our own ancient traditions ", 161 Elsewhere he 
lays aside the pretence of originality and speaks with the voice 
of the mere copyist : " But since my master, priest Beulan, 
thinks the Saxon and other genealogies useless, I have refrained 
from copying them, but I have written out the ' Cities ' and 
the ' Marvels of Britain,' as other scribes have done before 

147 Mommsen, Introd. 114-5. 148 C. 26 (Mommsen, p. 166). 

149 For the original list see the Chartres MS. (Mommsen, p. 163, note 4). 

180 See section ii. of this chapter. 

181 From the shorter preface. The longer is only found in one late MS. 
and is spurious (Mommsen, Introd. 126). 

VOL. I. 15 


CHAP, me". 152 His purpose was to produce a new and enlarged 
edition of the Historia, but he found himself sadly destitute 
of material. The doctors of Britain, he complains, have kept 
no records of the history of the race ; incessant war and pesti- 
lence have dulled the senses of the Britons, so that they have 
ceased to care about the memorials of their past 153 It must be 
confessed that Nennius did little to improve the text of the 
Historia. He had learnt from readers of Bede that to 
Paulinus, and not Rhun ab Urien, was due the credit of the 
conversion of Edwin of Northumbria, 154 and he knew that 
Septimius Severus died at York. 155 But it is an unhappy guess 
which identifies the mythical Ceredig, Hengist's British inter- 
preter, with the historical king of that name who ruled the 
northern El fed in the days of Edwin, 156 nor is it of any great 
advantage to have the lineage of Brutus, the progenitor of all 
Britons, traced to the accursed Ham instead of the more re- 
putable Japheth. 167 Genuine as was the patriotism of the 
disciple of Elfodd, he had undertaken a task which was beyond 
the compass of his narrow powers. 

A hundred years later another Welsh scholar appears in 
the person of Bishop Asser of St. David's. 158 A relative of 
Bishop Nobis, brought up in the famous monastery of Dewi and 
gradually promoted from the grade of scholar, through that of 
priest, to the highest dignity the place could offer him, 159 Asser 

152 Mommsen, p. 207. 

153 Preface. It is remarkable, as Zimmer points out (Nenn. V. 133), that 
the expression " hebitudo gentis Britanniae " is echoed in the early ninth century 
MS. (so Bradshaw) which speaks of the alphabet invented by " Nemniuus" in 
answer to the taunts of an English scholar, who said the Welsh had no native 
alphabet; this the Welshman did " ut uituperationem et hebitudinem deieceret 
gentis suae " (Gr. Celt. (2), 1059). 

154 See note 42 above. 155 Mommsen, p. 165. 
158 Ibid. p. 178. 1B7 Ibid. p. 151. 

158 I have, of course, used the recent edition of the life of Alfred by 
W. H. Stevenson (Oxford, 1904), in which the many difficult questions 
raised by the text of this work are for the first time fully and scientifically 

159 ju a tam sanc t a loca, in quibus nutritus et doctus ac coronatus (i.e., 
tonsured) fueram atque ad ultimum ordinatus" (Stev. p. 64). I understand the 
last phrase to refer to his elevation to the episcopate ; for, despite Stubbs (Reg. 
Sacr. (2), 217), it would seem clear that only a bishop of St. David's could speak 
of " omnia quae in sinistrali et occidentali Sabrinae parte habebam " and 
couple himself with Nobis as one of the " antistites " of the " parochia Sancti 
Degui ". 


was in every respect a child of the Welsh monastic system. CHAP. 
Other influences, Irish and Breton, had in all probability con- 
tributed to the ripening of his scholarship, 180 but it is undoubtedly 
to the credit of St. David's that, when King Alfred was en- 
deavouring to revive learning in Wessex by gathering foreign 
teachers at his court, it was able to furnish at least one man 
who could assist him in his design. Asser himself has told the 
tale, how the king summoned him from the further limits of 
the West to a conference which took place between them at 
Dean 1(J1 in Sussex, how Alfred appealed to him to leave Wales 
and join the circle of scholars at the court, how he hesitated to 
abandon the interests entrusted to him in his native country, 
and how he finally agreed to a division of his time between the 
old and the new responsibilities. A long illness which kept 
him a prisoner for many weary months at Caerwent 1(i2 postponed 
the fulfilment of the bargain, but it was ultimately approved, 
from motives of policy, by the monks of St. David's, and from 
this time forth Asser's connection with Alfred and with England 
was intimate, until he died bishop of Sherborne about 9io. 163 
Such a career leads one to look with interest into the work 
produced by Asser as an author. But the biography of Alfred 
is in many ways an unsatisfactory piece of writing. Its tone 
and spirit are admirable ; throughout there is genuine enthus- 
iasm for the patient, earnest hero who grappled with so many 
difficulties and ever set the loftiest ideals before him. But the 
arrangement is confused and shows no unity of purpose ; the 
style is of the inflated and rhetorical type which was affected 
by many Western writers at this period ; and nothing could be 
more abrupt than the conclusion. Asser stands on a distinctly 
higher level as an author than Nennius, but he is manifestly 

180 Stev. (xciii.) thinks the Prankish element in Asser's work may have come 
through Breton channels. 

181 Near Eastbourne (Stev. 312). 

18a So Stev. (313-4) understands the " Wintonia civitate " of the text, at 
once disposing of the serious difficulties raised by the ordinary rendering 
" Winchester ". Caerwent (which is " guentonia urbs " in Lib. Land. 220, 222) 
was an ancient ecclesiastical centre, owing its foundation to one Tathiu or 
Tatheus (Cambro-Br. SS. 255 64 ; Lib. Land. 222, 243; ibid. 270, mentions 
five " presbiteri tathiu "). It also lay on the old Roman road which would take 
Asser from Wessex to St. David's. 

188 A.S. Chr. MS. A. s.a. Harl. MS. 3859 gives the year as 908 (Cymr. ix 



CHAP, inferior as a literary artist to Gildas, and, while posterity will 
VIL ever be grateful to him for the picture of the good king he 
served, his work cannot be said to remove the impression that 
the life of the Welsh Church at this time ran a somewhat slug- 
gish and pedestrian course. 



(For the materials of this chapter I have had to depend mainly upon my 
own studies, save that invaluable help has been derived from the notes of Mr. 
Egerton Phillimore in the various Cymmrodorion volumes and elsewhere.) 


IT will now be convenient to undertake a general survey of the CHAP. 
political condition of Wales during the period 650 to 850, and VIII< 
this will of necessity be at the same time a topographical 
account of the country as it then was, since the story will not 
admit of being treated with reference to one central point, but 
must be separately told for each one of the many tribal areas 
into which Wales was at this time divided. 

Anglesey, 1 the. Mona of primitive times and the M6n of 
mediaeval and modern Welsh, has been throughout the historic 
period a political unit. It is true that it is divided into three 
cantrefs, but these bear upon them, in their names of Cemais, 
Aberffraw and Rhosyr, 2 the marks of their origin as areas 
mapped out by. the government with reference to the three 
principal courts of the chieftains of the island. The six 
commotes into which the three cantrefs are subdivided wear a 
more ancient aspect, but even they do not appear, with the 
possible exception of Tindaethwy, 3 to represent old tribal 
distinctions or anything but administrative convenience. Save 
for the great fen known as Malltraeth Marsh, which separates 
the commote of Menai from that of Malltraeth, the island has 
no important physical barriers ; it lies low and, in marked 
contrast to the opposite mainland, has no mountain ranges or 

1 For the derivation of the name see chap. vi. note 99. 
8 The documents which record the names of the cantrefs and commotes of 
Wales are discussed in the note appended to this chapter. 
8 See p. 41. 



CHAP, high table-lands. 4 Hence it acquired importance in very early 
' times as the great corn-growing district of North Wales ; its 
proud sons dubbed it " M6n, the mother of Wales," since the 
abundant crops it yielded were sufficient, they said, to maintain 
the whole of Wales for one year. 5 In Bede's time it was 
known for its fertility, 6 and the number of parish churches it 
contains is an incidental proof of its ability in the Middle Ages 
to support a large population. It thus became, as soon as the 
Goidelic elements in it had been thoroughly subdued to the 
Brythonic, the chief seat of political power in Gwynedd and the 
residence, in particular, of the line of kings which claimed to 
represent Cunedda Wledig and Maelgwn Gwynedd. Physical 
and political causes combined to prevent in Anglesey that 
division into minor chieftaincies which was so common a 
spectacle in other parts of Wales. 

After the fall of the great Cadwallon, the house of Cunedda 
was represented by his son Cadwaladr, who was king among 
the Britons in the days of Oswy of Northumbria. 7 None of his 
deeds have been recorded, yet he must have been a figure of 
some distinction, for the bards of later ages regarded his name 
as one to conjure with, and in days of national depression 
foretold his return, as was fabled of Arthur also, to lead the 
Cymry to victory. 8 He died in the great plague of 664 9 and 

4 According to App. Land Com. 259, only 188 acres of the surface of the 
county lie at a higher elevation than 500 feet above sea-level. 

8 Gir. Camb. vi. 127 (Itin. ii. 7), 177 (Descr. i. 6). Prydydd y Moch was 
familiar with the title; see Myv. Arch. I. 299 (211). 

6 H.E. ii. 9 (frugum prouentu atque ubertate). 

7 " Dum ipse (Osguid) regnabat, venit mortalitas hominum, Catgualart 
regnante apud Brittones post patrem suum, et in ea periit " (i.e., Cadwaladr) 
(Sax. Genealogies in Hist. Britt. c. 64). 

8 C/. the lines in the " Hoianau " (Blk. Bk. fo. 306 ; IV. Anc. Bks. ii. p. 26), 

A phan del Kadualadir y orescin mon 

dileaur Saeson o tirion prydein. 

(" And when Cadwaladr comes to seize Anglesey, the English will be driven from 
the pleasant isle of Britain.") 

9 The plague in the reign of Oswy which, according to the Saxon Genea- 
logies, carried off Cadwaladr, can hardly be any other than the famous pestilence 
of 664, for which cf. Bede, H.E. iii. 27. The chronicle in Harl. MS. 3859 
gives the year of the king's death as 682, but it is of inferior authority to the Sax. 
Gen. Geoff. Mon. introduced another element of confusion by identifying Cad- 
waladr with Caedualla of Wessex and making him die, accordingly, at Rome on 
2oth April, 689 (Hist. Reg. xii. 14, where the Berne MS. reads " Cheduallam 
iuvenem," and 18, where " mayarum ' is to be read for the "majurum" of 


it is likely, notwithstanding his martial reputation, that he CHAP, 
spent the close of his life as a monk, for the church of Eglwys 
Ael or Llangadwaladr in Anglesey claims him as its patron 
saint and founder, and churches were dedicated to him in other 
parts of Wales. 10 The situation of Llangadwaladr, some two 
miles from Aberffraw, suggests that this had already become 
the chief dwelling-place of the family, not to speak of the fact 
that Cadfan's tombstone is in the same church, carrying back 
the connection with the district a couple of generations earlier. 11 
Henceforth Aberffraw, a cluster of dwellings on the little lift 
which rises above the sand-flat at the mouth of the Ffraw, was 
the " principal seat " of Gwynedd, 12 and its possession was held 
to confer a dignity and precedence which no other title could 

The successors of Cadwaladr were men of no note, whose 
sway did not extend, it would seem, beyond the limits of 
Anglesey. The death of Rhodri Molwynog, son of Idwal, son 
of Cadwaladr, is recorded under the year 754, 13 and the family 
then passes out of sight until in the early part of the ninth 
century two sons of Rhodri, Hywel and Cynan, are found 
battling against each other for the lordship of M6n. In 816 
the death of Cynan, whose chief stronghold was in the commote 
of Tindaethwy, left the field clear for Hywel, who no doubt 
ruled over Anglesey until his death in 825. When Hwyel 
died, the male line of Maelgwn Gwynedd was at an end and 
its claims were transferred to another house by Ethyllt, the 
daughter of his brother Cynan. 14 

It may be mentioned that there were other royal courts in 
Anglesey than that of Aberffraw. In Talybolion, Cemais 15 
commanded the little harbour of Forth Wygyr, widely known 

10 Welsh SS. 299-301. In 1352 the vill of " Eglussell " was held " de sancto 
Cadewaladre rege " (Rec. Corn. 46). 

11 See p. 182. 

12 " Eisteddfa arbennig " is the phrase of the Dimetian Code (LL. i. 346). 

13 " Rotri rex brittonum moritur " (Harl. MS. 3859 in Cymr. ix. 161). Other 
sources add nothing of value save Rhodri's pedigree and distinguishing epithet, 
the latter not yet satisfactorily explained (Cymr. ix. 169-70 ; Bruts, 257 ; Cymr. 
viii. 87). 

14 The chief authority is Harl. MS. 3859 (chronicle and pedigrees). Ann. Ult. 
also record s.a. 815 (= 816) the death of " Conan mac Ruadhrach rex Britonum ". 
It has been very generally assumed that Merfyn Frych succeeded immediately on 
the death of Cynan Tindaethwy ; on this point see chap. vii. note 145. 

Rec. Cam. 63-5. 


CHAP, in the Middle Ages as the northernmost point of Wales, 16 while 
Twrcelyn had its royal manor at Penrhos Lligwy. 17 Llanfaes, 
not far from Beaumaris (which is a creation of Edward I.), was 
the court of Tindaethwy 18 and before Newborough commenced 
its career as an English-made borough, it had been, under the 
name of Rhosyr, the centre of the commote of Menai. 19 This 
was in accordance with the rule that each commote should 
have its own "llys" or royal vill, at which the lord of the 
country received the renders of the men of that commote, 
whether freemen or serfs. In addition, there were in the 
island two important ecclesiastical centres, the " clas " founded 
by Cybi at Caer Gybi, 20 under the shadow of the mountain 
the highest in Anglesey known to English sailors as the Holy 
Head, 21 and the similar foundation of Seiriol at Penmon, with 
its offshoot on the adjacent isle of Priestholm or Ynys Lannog. 22 
Penmon and Caergybi flank the island on its eastern and 
western sides respectively, and readers of Matthew Arnold will 
hardly need to be reminded of the tale told of the two founders, 
that they journeyed once a week to meet each other at the 
wells of Clorach, " in the bare midst of Anglesey," until the 
Western saint, ever facing the warm beams of the sun, became 
Cybi the Swarthy (Cybi Felyn), while his companion from the 
East, with the sunlight always falling upon his back, remained 
Seiriol the Fair (Seiriol Wyn). 23 

Facing Anglesey, in one long serrated line, are the heights 

16 Cymr. xi. 43. It is the " portus Yoiger " of Gir. Camb. vi. 165 (Descr. 
i. i) ; for other early references see Myv. Arch. I. 74 (62), 194 (143), 270 (193) ; 
Triad i. 5 = iii. 65. 

ll Rec. Cam. 70-2. 

18 There is an interesting survey of Llanfaes, as it was in 1294, *he Y ear 
before that in which the building of Beaumaris was commenced, in Trib. System, 
App. pp. 3, 4. 

19 Rec. Cam. 83-5. 20 See pp. 130, 218. 

21 The fourteenth-century English romance of "Gawain and the Green 
Knight " brings Gawain past " alle the iles of Anglesay " and " the Holy Hede " 
(w. 698-700, ed. Morris). 

22 See p. 216. That Penmon was a monastic church of the ancient type is 
made certain by the grant in 1237 to the prior and canons of Ynys Lannog of 
" totam abbadaeth (i.e., abbacy) de Penmon " (Man. Angl. iv. 582). 

23 The story first appears in the notes to Richard Lloyd's Beaumaris Bay 
(1800), p. 2, though it was known to Lewis Morris (Celt. Remains, p. 351). It 
should be explained that in his well-known sonnet (" East and West") Matthew 
Arnold misses the precise point of the two epithets and so tells the legend not 
quite convincingly. 


of the " stronghold of Gwynedd," 24 the region known to the CHAP. 
Welsh as Eryri, "the haunt of the eagles," 26 and to the V1IL 
English by the no less romantic name of Snowdon, " the hill of 
snows 'V 28 This was the mountainous rampart which, stretch- 
ing from the mouth of the Conway to the Rivals, at all times 
protected Anglesey and the intervening district of Arfon from 
serious attack on the landward side, and few sights are more 
impressive than the distant prospect of this mountain wall, 
rising in peak after peak along the horizon, as it may be seen 
from Aberffraw and many another point of vantage in Southern 
Anglesey. Nor was Eryri merely a barrier of crag and moor- 
land, a rocky, marshy wilderness. Hidden within its folds 
were mountain glens, such as Nant Peris in Arfon and Nant 
Ffrancon in Arllechwedd, where the herbage was of the finest 
and the woods sheltered deer and nurtured swine. Just as it 
was reckoned that Anglesey could feed with corn the entire 
population of Wales, so it was held that the pastures of Eryri 
could furnish grazing for all the sheep and cattle in the country. 27 
It was not only a citadel, but a citadel which, in the summer 
season, at any rate, could not easily be starved into submission. 
The region which lay opposite to Anglesey, from the 
summit of the Rivals to the river Cegin between Bangor and 
Llandegai, was appropriately known as Ar-fon, i.e., the land 
over against Mon. 28 The cantref thus named extended not 
only between the limits just specified, but also far into the 
heart of Eryri ; the vale of Nantlle and the pass of Llanberis 

24 " Kedernit gwyned " (Mob. 62, 63). Cf. Bruts, 292 : " mudaw hyt 
ymynyded eryri. Kanys kadarnaf lie adiogelaf y gael amdiffyn yndaw rac y llu 
oed hwnnw". 

25 The Hist. Britt. (c. 40) contains a reference to " montibus Hereri " (ac- 
cording to some MSS. " Heriri "), where was situated the " arx " of Ambrosius, 
i.e., Dinas Emrys, near Beddgelert. It is not so long since eagles ceased to 
haunt these mountains see Williams, Observations on the Snowdon Mountains 
(1802), pp. 2, 3. 

26 An early instance of the use of the name is to be found in A.S. Chr. s.a. 
1095, MS. E. (Snawdune). It is properly the equivalent of Eryri (Gir. Camb. vi. 
135 (Itin. ii, 9)) or Snowdonia, and was not used in the Middle Ages, as now, to 
denote merely the summit. The Welsh name of this, known to every Welshman 
to-day as " Y Wyddfa," was anciently " Y Wyddfa Fawr," or the Great Burial- 
place, since the bones of Rhita the Giant were supposed to be entombed in the 
cairn which crowned it (Celt. Folklore, pp. 474-9). 

27 Gir. Camb. vi. 135 (Itin. ii. 9), 170 (Descr. i. 5). 
M Gir. Camb. vi. 124 (Itin. ii. 6). 


CHAP, were within its borders. It is aptly described by the author of 
Breuddwyd Macsen Wledig as a land of which the seaboard 
ran side by side with the champaign, the woodlands with the 
mountain. 29 The cantref was conceived as divided into four 
strips or belts, the maritime, the agricultural, the forest and the 
highland belt, rising in terrace fashion above each other until 
the central mass of Snowdonia was reached. It was a rich and 
diversified country and nourished a race of great independence 
of spirit. The men of Arfon claimed it as their right to lead 
the van in the nestings of Gwynedd and therewith demanded 
many other privileges, such as the liberty to fish in the three chief 
rivers of the district (probably the Saint, the Gwyrfai and the 
Llyfni), the right to declare, against their neighbours of other 
cantrefs, the boundaries of Arfon, and the privilege of sleeping, 
when they visited the king's court, in the " neuadd " or common 
hall with the royal heir and the squires of the court. 30 Arfon 
was a true tribal district, in which tribal consciousness was 
keen and alert. At some period or other in its history, the 
ancient cantref was, like many another, divided into two com- 
motes. The river Gwyrfai, flowing from Llyn Cawellyn under 
the foot of Snowdon to the western end of the Menai Straits, 
furnished the dividing line, and from that day to this has 
separated the commote or hundred of Uch Gwyrfai (Above 
Gwyrfai) from that of Is Gwyrfai (Below Gwyrfai). The can- 
tref contained several notable civil and ecclesiastical sites. 
Oldest of all was Carnarvon, 31 the Segontium of the Romans, 
known to the Welsh as Caer Saint yn Arfon and often more 
briefly as " Y Gaer yn Arfon," which is the source of the modern 
name. Legend had much to say of the past glories of 
this place. It was the burial-place of Constantine the Great 
and the home of Elen of the Hosts, the British wife of the 
Emperor Maximus. Here Bran the Blessed was found by the 
bird which brought under its wing the tale of the woes endured 
by his sister Branwen in Ireland, and here Beuno the saint up- 

29 ic Q w j a t a oed kyhyt y maestir ae mor. kyhyt y mynyd ae choet " (Mob. 


80 " Breiniau Arfon " (The Privileges of Arfon), a tract found in two MSS. of 
the Yen. Code (LL. i. 104-6). 

31 For references to Carnarvon see Hist. Britt. c. 25 ; Mob. 34, 88 ; Cambro- 
Br. SS. 18. A charter in favour of Penmon priory was issued by Llywelyn ab 
lorwerth at " Kaerinarvon" on isth October, 1221 (Man. Angl. iv. 582). 


braided King Cadwallon for offering to him land which had CHAP, 
been unlawfully wrested from its infant proprietor. The fort 
on the hill of Llanbeblig for it was the building of the 
Edwardian castle which drew Carnarvon down to the margin 
of the strait was clearly the ancient centre of the cantref, the 
original home of its chieftains, though in later times it was 
eclipsed in importance by Dolbadarn. 82 The two chief sanctu- 
aries of Arfon were Bangor and Clynnog. Of the former, 
Bangor Fawr yn Arfon, the seat of the bishops of Gwynedd, 
some account has already been given ; 33 it is enough to say 
here that the whole of the north-eastern corner of the cantref, 
from the Cegin to the modern village of Portdinorwic, formed 
part of the possessions of the see, and that a solid barrier of 
Church land thus intervened between the men of Arfon and 
those of Arllechwedd. 34 The " clas " of Celynnog Fawr was 
little inferior in importance to that of Bangor ; it had lands in 
Lleyn and Anglesey as well as in the neighbourhood of the 
church, and St. Beuno, its founder and protector, was reckoned 
among the mightiest of saints. Not many years have elapsed 
since the whole countryside brought their children to Beuno's 
Well to be healed of their ailments, and paid an annual tribute 
to Beuno's Chest (Cyff Beuno) to ensure the prosperity of their 
flocks and their herds. 35 

To the east of Arfon lay Arllechwedd, for the most part a 
rugged, stony region, a land of declivities, as its name implies, 
and hence playing no important part in the early history of 
Wales. Its two strips of fertile, low-lying territory, the one 
bordering on the sea and the other on the river Conway, 
which was the eastern limit of the cantref, were known re- 

x Rec. Cam. (17-22) shows the tenants of Is Gwyrfai as joining in the main- 
tenance of the manor of Dolbadarn. 

88 See note appended to chap. vi. 84 Rec. Cam. 93-5, 231. 

85 The " clas " of " Beuno," with that of Bangor, was to protect the special 
rights of the men of Arfon (LL. i. 106). Clynnog is Celynnog, the hamlet of 
holly trees; see B. Sacs. s.a. 977 and 1151 (B.T. is wrong in both passages); 
Llyfr yr Ancr, 124 ; Buck. Gr. ap C., 36. Clynnog Fechan was close to 
Llangeinwen, Anglesey, and belonged to Clynnog Fawr (Arch. Camb. I. i. (1846), 
310-11 ; Rec. Cam. 257). There is a list of the possessions of the house in 
Rec. Cam. 257 (cf. Arch. Camb. I. iii. (1848), 253-5). For the antiquities and 
traditions of the place see B. Willis, Bangor, pp. 299-305 ; Penn. ii. 396-400 ; 
Arch. Camb. I. iii. (1848), 247-57; "Cyff Beuno," a Welsh account of the 
parish by Eben Fardd, schoolmaster and poet (Tremadoc, 1863). 


CHAP, spectively as Arllechwedd Uchaf (Upper) and Arllechwedd Isaf 
(Lower) and with these two commotes was associated that of 
Nant Conwy, 36 which lay west of the Conway from Dolgarrog 
to its source. The royal court of Arllechwedd was at Aber of 
the White Shells (Gwyn Gregin), 37 a favourite residence of the 
later rulers of Gwynedd ; Trefriw, perched on the hillside just 
above the highest reach of the tidal portion of the Conway, 
was the manor of Nant Conway when the district came into 
the hands of the English, but Dolwyddelan must have been at 
one time the chief stronghold of the lord of the commote. The 
Church had no great foothold in Arllechwedd until the found- 
ation of the Cistercian abbey of Aberconwy in 1 1 86 ; there 
was no important " clas " within its bounds, and probably none 
of its churches stood higher in popular repute than that of St. 
Tudclud at Pennant Machno (now condensed into Penmachno), 38 
which is shown by its ancient Christian tombstones to be a 
foundation of the fifth or sixth century. 39 

The great tongue of land thrown out by Eryri to the west 
has long been known by the name of Lleyn, which is said to 
signify the land of spearmen. 40 It is a remote, sea-locked 
region, lying off the track of the main currents of Welsh life, 
but its green straths and swelling knolls, only here and there 
broken into by mountain masses, are fertile and luxuriant, and 
the district has always maintained a considerable population. 
Three commotes went to make up the cantref of Lleyn. 
Dinllaen, so called from the " dinas " or cliff castle on the little 
peninsula of Porthdinlleyn, was the northernmost, stretching 
from the Rivals to Carn Fadryn ; its royal manor was at 
Nevin. 41 From Carn Fadryn to Aberdaron a second com- 

36 So Pen. MS. 163 (Evans, Rep. \. p. 952), which is right on this point against 
the other old lists. The five churches of the commote are assigned to the dean- 
ery of Arllechwedd in the Norwich Taxation (Arch. Camb. V. xi. (1894), 30) and 
the silence of the Statute of Rhuddlan as to Nant Conwy is only to be explained 
by the assumption that it was included in " candreda de Arthlegaph ". 

37 The full name is given in Pen. MS. 147 (Evans, Rep. i. p. 913) and Myv. 
Arch. II. 30 (419). 

38 Penanmagno in Rec. Carn. 9, Pennam'achno in the Norwich Taxation 
(Arch. Camb. V. xi. (1894), 30). Llyfr John Brooke (Evans, Rep. i. p. 913) and B. 
Willis, Bangor, p. 274, give the name of the saint correctly ; in Ecton's The- 
saurus (third edition, p. 495) it is printed as Tyddud, which misled Rees (Welsh 
SS. 332). 

39 Inscr. Chr. Nos. 135-7; Lap. W. 175-7. 

40 Celt. Folklore, i. p. 226. 41 Rec. Carn. 35. 


mote extended, which apparently took its name of " Cymyd- CHAP, 
maen " from a famous " Maen Melyn " or " Yellow Rock " 
forming part of the promontory which faces Bardsey. 42 The 
court of this commote was at Neigwl. 43 The third commote, 
which skirted the shores of Cardigan Bay from St. Tudwal's 
Isles to the river Erch, originally bore the name of Cunedda's 
son Afloeg, but Afloegion was in time corrupted into Gaflogion 
and Cafflogion. 44 Pwllheli was the ancient centre of this com- 
mote. 46 The church lands of Lleyn were extensive, for, in 
addition to the numerous vills which were the property of the 
see of Bangor and those which belonged to Clynnog Fawr, the 
cantref itself contained the important " clas " of Aberdaron, 
whose abbot was lord of a very considerable part of Cymyd- 
maen. 46 Pilgrims were constantly passing through to Bardsey, 
and the necessary provision for them helped to give the 
church a special title to the wide domains which it held in the 

There is scarcely anything to show who ruled in Arfon, 
Lleyn and Arllechwedd in the eighth century. But the mention 
in Harl. MS. 3859 of a King Caradog of Gwynedd who was 
slain by the English in 798 47 leads one to surmise that the 
pedigree of Hywel ap Caradog to be found in the same au- 
thority 48 is that of the royal line of his district. It goes back to 
Cynlas (no doubt the Cuneglasus of Gildas), cousin of Maelgwn 
Gwynedd and great-grandson of Cunedda. 49 As it is not carried 
beyond Hywel, who belonged to the early part of the ninth 
century, it may be conjectured that the three cantrefs were, 

43 " Maen Melyn Lleyn " is close to Braich y Pwll. It was famous in the 
fourteenth century ; see Dafydd Nanmor's reference in his ' cywydd " to the 
golden hair of Llio of Gogerddan 

Mae'r un lliw a'r maen yn Lly~n. 
(Tis of the same hue as the stone in Lleyn.) 

43 Rec. Cam. 38. A farm near Llandegwining still bears the name of Maer- 

44 See p. 117. 4B Under the name of Porthely (Rec. Cam. 25, 29, 31, 32). 
48 There are early references to the church and clergy of Aberdaron in Buck. 

Gr. ap C. (116; Myv. Arch. II. 596 (729)) and B.T. p. 122 (Bruts, 295; B. 
Saes. s.a. 1112 (= 1115)). For the "abadaeth" as a territorial area see Rec. 
Cam. 252 (composition of the year 1252 between the abbot of Bardsey and the 
secular canons of Aberdaron), and cf. the " abadaeth " of Penmon (Man. Angl. iv. 

47 " Caratauc rex guenedote apud saxones iugulatur " (Cymr. ix. 163). 

48 Cymr. ix. 172. 4 * See p. 133. 


CHAP, under Merfyn Frych, combined in one kingdom with the ancient 
VIII> realm of Anglesey. Henceforth Mon and Arfon are rarely 
divorced from each other. 

The rugged heights which surround the north-eastern corner 
of Cardigan Bay were, according to tradition, the portion of 
Cunedda's son Dunod (Donatus), and here, therefore, was the 
cantref of Dunoding. 50 It was, says the primitive narrator of 
the story of King Math, the best of all cantrefs for a young 
man to rule over, 51 which may be taken to mean that -as a 
rough and craggy region, it tried and disciplined the powers of 
the budding chieftain, not suffering him to fall into ignoble sloth 
and self-indulgence. It was cut in twain by the broad tidal 
estuary known as " Y Traeth Mawr " (The Great Sands), which 
was previous to its reclamation from the sea in 1 8 1 1 a most 
formidable barrier, and it thus fell at an early date into the 
commotes of Eifionydd and Ardudwy, names which soon 
eclipsed and drove out of current use the ancient one of 
Dunoding. Eifionydd, named after Dunod's son, Eifion, 52 was 
the northern commote ; it lay between the Erch and the Traeth, 
and, in historical times, the home of its lords was on the rock 
now crowned by the ruins of Criccieth Castle, though there is 
reason to think that the mound at Dolbenmaen marks the site 
of an earlier royal residence. 53 Ardudwy was a large but thinly 
populated area ; from the Festiniog valley in the north to the 
Mawddach estuary in the south it was chiefly moor and 
mountain, but a fertile strip along the coast was known as 
Dyffryn Ardudwy (The Plain of Ardudwy). Among the famous 
sites within it were Harlech, where Bran from the height of the 
Castle rock watched the coming of the ships which brought the 
king of Ireland and his train to beg the hand of his sister 
Branwen, 54 and Mur y Castell, where Llew held his court amid 

50 For the name see, in addition to the ordinary lists, Jesus Coll. MS. 20 
(Cymr. viii. 85), lolo MSS. 122, Mob. 73. On the analogy of Glywysing and 
Dogfeiling, the ending should be -ing and not-ig. Ffynnon Dunawd, near 
Eisteddfa in Eifionydd, probably commemorates a saint and not the son of 
Cunedda (Y Gestiana, gan Alltud Eifion : Tremadoc, 1892, p. 9). 

51 Mob. 73. 

52 " Ebiau(n) map Dunaut map Cunedda " (Cymr. ix. 178). For the form 
cf. Meirionydd, Elenydd and Maelienydd. 

55 Bye-Gones, viii. (1903-4), p. 180. 

54 Mob. 26 (hardlech yn ardudwy ynllys idaw). 


the broken walls of the dismantled Roman encampment. 65 It CHAP, 
was a land which bred hardy wielders of the lance, 59 a nurturer 
of warriors rather than of churchmen, for neither here nor in 
Eifionydd were there in early times any churches of the first 
rank and the ancient church holdings were not considerable. 
The local dynasty, tracing its origin to Dunod ap Cunedda, 
appears to have held its own until well on in the tenth century. 67 
The districts so far dealt with belonged to the ancient 
Gwynedd, sometimes called by way of distinction Gwynedd 
above Conway. 58 East of the river Conway came Gwynedd 
below Conway, which does not seem to have been entitled to 
the name originally, for the natural explanation of the name 
" Y Berfeddwlad " (The Middle Country) which it also bore is 
that it was the land which lay between Gwynedd and Powys. 59 
The four cantrefs of the Middle Country were Rhos, Rhufoniog, 
Dyffryn Clwyd and Tegeingl, 80 belonging for the most part to 
the great upland plateau of eastern North Wales, but cleft by the 
rich expanse of the Vale of Clwyd, a fertile, corn-growing tract 
of which each cantref had its share. Rhos was bounded by the 
Clwyd, the Elwy, the Conway and the sea ; along the east 
bank of the Conway it sent a long arm southwards as far as 
Capel Garmon. The Llandudno peninsula was included in it 
and formed the commote of Creuddyn, which is now in a 
different county from the main body of the cantref, but was 
anciently reckoned one of its members. One may, indeed, 
surmise from the name Eglwys Rhos that this was the original 
Rhos from which the cantref took its title and that the deriva- 
tion is to be sought in the Goidelic word for " promontory " 
rather than in the Brythonic for " moor ". 61 The two other 

65 Mob. 74 (lys idaw yn y lie a elwir mur y castell). The fort is described 
on p. 68. 

M " Sunt ... his in partibus lanceae longissimae " (Gir. Camb. vi. 123 
(Itin. ii. 5)). 

67 None of the persons named in the Dunoding pedigree in Cymr. ix. 177-8, 
appear elsewhere, so that dating is difficult, but the usual method of calculation 
(three generations to a century) will bring Cuhelyn, with whom the pedigree ends, 
to about 930. 

ss "Gwyned ewch Conwy " (B. Sacs. s.a. 1175). 69 Cymr. xi. 174. 

80 In a document in Rymer (4), i. 267, dated 3oth April, 1247, the four 
" cantredos" of " Pernechelad " are given as " Ros, Rowennok, Defrencluc et 
Anglefeld ". 

61 Both are derived from the same Celtic original (Urk. Spr. 312). 


CHAP, commotes were Uwch Dulas and Is Dulas, to the west and east 
respectively of the little river Dulas. Of the sites of the cantref, 
none ranked higher than Degannwy, the royal manor of 
Creuddyn, associated by tradition with the glories of the rule of 
Maelgwn, 62 known to have been a stronghold in the ninth 
century, 63 and only eclipsed as the principal fortress of the 
district when Edward I. built Conway on the opposite side of 
the river. Possession of the rock of Degannwy, which, like 
that pictured in " The Bard," 

Frowns o'er old Conway's foaming flood, 

was hotly contested in the long border strife between Welsh 
and English, and its vicissitudes furnish no bad index of the 
alternate rise and fall of the fortunes of the men of Gwynedd. 
The cantref also contained some important churches, such as 
Abergele, once the mother church of a wide district, but after- 
wards reduced to a state of dependence upon St. Asaph, 64 and 
Llandrillo or Dinerth, a church of similar standing. 65 

Rhufoniog was an inland cantref; the broad moorlands of 
Hiraethog were for the most part within its limits, which to 
the north and east were defined by the Elwy, the Clwyd and 
the Clywedog. The river Aled furnished the dividing line of 
the two commotes of Uwch Aled and Is Aled ; there was a 
third commote, called Ystrad or Cymeirch, which lay between 
the Lliwen and the Clywedog. Rhufoniog plays no important 
part in the early history of Wales ; though there must have been 
a fortress on the limestone rock of Denbigh, the Welsh Dinbych, 
long before the time of Edward I., it is only then that the place 
enters into the light of history, and no list has been preserved 
of the princes of this region or of the neighbouring cantref of 
Rhos during the period which preceded the absorption of the 
district in the realm of Gwynedd. 66 Much of it, in fact, was in 
episcopal hands ; the bishop of Bangor had Llanrhaeadr, 67 and 

62 P. 129. 6S P. 202. 

64 Thomas, St. Asaph, pp. 350-1. For the " princeps opergelei " who was 
head of the " clas " in 856 see Cymr. xi. 165. 

68 Thomas, St. Asaph, p. 548. It was an important church in 1137 (Buck. 
Gr. ap C. 128 (734)). 

66 Ann. C. MS. B. s.a. 816 mentions a " regnum Roweynauc " in a notice 
clearly derived from an ancient source. 

67 Rec. Cam. 112. 


no small proportion of the revenues of the bishop of St. Asaph CHAP, 
was drawn from the manors of the see in Henllan, Llansannan, 
Llannefydd and Llangerniew. 68 

Dyffryn Clwyd was a much richer cantref, though it did not 
cover the whole of the famous vale, but only the southern por- 
tion, from Bodffari to Derwen. The most important of the 
three commotes into which it was divided bore the name of 
Dogfeiling 6y and it is a reasonable assumption that this was once 
the designation of the whole cantref, marking it out as the realm 
of Dogfael, son of Cunedda Wledig. 70 Like Rhufoniog, it has 
no recorded early history, though there is in this cantref as in 
the other a site pretty clearly indicated by its name and its 
natural advantages as that of an early stronghold. Ruthin, 
which is by interpretation Rhudd-ddin, the red fort, occupying 
a ridge of red sandstone in the centre of the Vale, was the centre 
of the commote of Dogfeiling and became the seat of the lord's 
castle when at the end of the thirteenth century Dyffryn Clwyd 
was transformed into a marcher lordship. Ecclesiastically, the 
cantref, with the adjacent commote of Cymeirch, was associ- 
ated with Gwynedd above Conway, for, when evidence is forth- 
coming as to the boundaries of the Welsh dioceses, it is found 
to be part of the diocese of Bangor, though surrounded by that 
of St. Asaph. 71 The fact has also to be noted and here may 
lie the explanation of the anomaly that the bishop of Bangor 
was a principal owner of land in the cantref, being lord of 
almost the whole of the commote of Llannerch, which consisted 
of the modern parishes of Llanfair and Llanelidan. 72 

The fourth cantref included in the Middle Country was 
known to the Welsh as Tegeingl. 73 Under the later princes 
of Gwynedd, Tegeingl was a region which took in the whole 
of our Flintshire as far south as Connah's Quay and Cilcen, and 

88 Thomas, St. Asaph, p. 180. 

"Evans (Rep. \. p. 914, footnote 9) gives, from Cardiff MS. 14, the constituent 
parishes of the three commotes of Dyffryn Clwyd. 

Descr. Pemb. 201 (note by E. P.). 

71 This was the case at the time of the Norwich Taxation of 1254 (Arch. 
Camb. V. xi. (1894), 31). In 1859 Bangor gave up the deanery of Dyffryn Clwyd 
in exchange for that of Cyfeiliog. 

Rec. Cam. 113-5. 

73 The accent falls on the first syllable. For the suggested derivation from 
the tribal name " Deceangli," see Celt. Br. (2), pp. 81, 290; Arch. Camb. V. ix. 
(1892), 165 ; W. People, p. 94. 

VOL. I. 1 6 


CHAP, was divided into the three commotes of Rhuddlan, Coleshill and 
Prestatyn. But it has already been shown that this district 
was seized by the English of Mercia not later than the end of 
the eighth century, 74 and all the evidence goes to show that the 
bulk of it remained in English hands until it was reconquered 
by Owain Gwynedd in the twelfth. Even the names of the 
commotes show how late was the organisation of this country 
as a Welsh cantref, for they bear the names of the castles built 
by the Normans to secure their hold upon the district. At the 
period which is now under consideration, Tegeingl was a part 
of Mercia, save perhaps the portion around Llanelwy or St. 
Asaph, where the bishops of this quarter of Wales maintained 
a precarious and uneasy footing on the storm-swept frontier be- 
tween the two races. It is true that nothing is known with any 
certainty of the history of the see before the twelfth century, 
but the place was then regarded as a bishop's seat of old stand- 
ing and some ancient title no doubt underlay the extensive 
claims of the successors of St. Kentigern to territorial lordship 
in the northern part of the Vale of Clwyd. 


Central Wales may be regarded as a broad table-land, through 
which rivers great and small furrow their way in winding courses 
to the sea, but which has few clearly marked mountain ranges or 
stretches of fertile plain. The ancient kingdom of Powys 75 took 
in most of this region, extending in its widest limits from the 
neighbourhood of Mold to the river Wye, near Glasbury and 
Hay. 76 It included some productive districts, such as the lower 
valley of the Dee and the well-watered meadows of the upper 
Severn, so that its children were not altogether without warrant 
in hailing it as " Powys, the Eden of Wales ". 77 But most of it 

74 Pp. 201-2. 

75 The view of Zeuss (Gr. Celt. (2), pp. 1053-4), tnat the name is to be con- 
nected with the "poues = quies" of the Oxford glosses and to be interpreted as 
meaning " settlement," still holds the field (Celt. Br. (2), p. 218). 

76 In the " Dream of Rhonabwy " (Mab. 144) the limits of Powys are given as 
" oporford (i.e., Pulford, near Chester) hyt yg gwauan yg gwarthaf arwystli " (some 
point near Llangurig in South Montgomeryshire). But this was in the twelfth 
century, after the separation from the province of " Rhwng Gwy a Hafren ". 

77 "Powys paradwys Cymry " (Myv. Arch. I. 114 (92); IV. Anc. Bks. ii. 
P- 259). 


was pastoral upland, a country well fitted to be the nurse of a CHAP, 
race of hardy, independent warriors, lovers of tribal freedom, 
haters of the sluggish and toilsome life of the lowland tiller of 
the soil, and tenacious holders of ancient privileges. 78 Such 
were the men of Powys, inheritors of the old Brythonic tradi- 
tions, in whom incessant warfare with the Mercian English 
kept alive the ancient tribal characteristics. 

Tribal independence does not seem, however, in this case 
to have brought about the division of the country into separate 
chieftaincies, for, so far as the scanty records show, there was 
but one kingdom of Powys and this maintained its unity until 
it was absorbed in the kingdom of Gwynedd in the ninth cen- 
tury. The founder of the dynasty was one Cadell Ddyrnllug 
(the Blacklisted ?), a contemporary of St. Germanus. 79 He was, 
so the story ran, the swineherd of the tyrant King Benlli, and, 
when the gates of the royal stronghold were churlishly shut 
against the saint and his companions, offered the shelter of his 
cottage to the men of God and killed and dressed for their 
supper the calf of his solitary cow. The inevitable sequel fol- 
lows ; fire from heaven struck the citadel of Benlli (which was 
probably perched on Moel Fenlli 80 in the Clwydian range), so 
that it was instantly consumed and never again rebuilt, while 
Cadell and his nine sons became kings of Powys. Their pro- 
geny still ruled the country when the Historia Brtttonum, which 
is our authority for the legend, was composed. 81 Whatever 
may be the historical basis of the story, it is certain that the 
family which governed Powys during the period 600-850 traced 
their origin to Cadell. The Selyf ap Cynan who fell in the 
battle of Chester was his descendant in the seventh generation, 82 

78 See Cynddelw's poem entitled " Breiniau Gwyr Powys" in Myv. Arch. 
I. 257 (186). 

79 Hist. Britt. cc. 32-5 ; pedigrees in Harl. MS. 3859, as given in Cymr. ix. 
179, 181, 182, with Phillimore's notes. The old Welsh " Durnluc," for which 
see Cymr. vii. 119, was altered at an early date (see, for example, Jesus Coll. 
MS. 20 in Cymr. viii. 87) to " Deyrnllwg," and thus was evolved the mythical 
region of Teyrnllwg, placed by lolo MSS. 86, between the Dee and (apparently) 
the Cumbrian Derwent. 

80 The hill-fort on Moel Fenlli was occupied in the Roman period (p. 89 
above) and there is a Llanarmon, under the patronage of St. Germanus, not far 

81 " Et a semine illorum omnis regio Povisorum regitur usque in hodiernum 
diem " (Hist. Britt. c. 35). The passage is found in all the MSS. of the Historia. 

" P. 181. 



CHAP, and the EHse to whose memory was erected the monument 
popularly known as " Eliseg's Pillar" stood as his representative 
in the tenth. 83 Elise was the son of Gwylog and Sanan, daughter 
of Noe ab Arthur of Dyfed ; 84 he flourished in the middle of 
the eighth century, and, so far as can be ascertained from old 
readings of the now obliterated inscription on his monument, 
waged successful war against the English. His grandson Cadell 
died king of Powys in 808 85 and was succeeded by his son 
Cyngen. Under Cyngen the realm of Powys was reduced to 
sore straits, for in 822 it was overrun by the English, but this 
was but a temporary check, since Cyngen, who set up the stone 
cross to his great-grandfather's memory, was clearly in posses- 
sion of lal, and speaks, indeed, of the whole realm of Powys 
as though it were under his sway. After a long reign he closed 
his career with a pilgrimage to Rome, where he died in or about 
854, the last king of Powys of the line of Cadell. 86 

During the period which is now being dealt with much of 
what was reckoned Powys in later ages was in the hands of 
the English. The line of Offa's Dyke not only ran through 
Tegeingl, but also intersected the districts known in the Middle 
Ages as the commotes of Ystrad Alun and Yr Hob and of 
Maelor, the English Bromfield ; it left little to the Welsh save 
moor and mountain. It would appear that the ancient home 
of the kings of Powys in this part of their dominions was the 
commote of lal, a long strip of upland which took in the 
western part of the valley of the Alun and abutted upon the 
Dee where it works its devious way through the gorges of 
Llangollen. Here was the hill-fort, named after Bran, a famous 
figure of Celtic story, 87 which guarded the upper waters of the 

83 The inscription on "Eliseg's Pillar "is only known from a transcript by 
Edward Llwyd. See Penn. ii. 7-10; Arch. Camb. I. i. (1846), 17, 32; VI. ix. 
(1909), 43-8. At present hardly a word can be read (Lap. W. 200). Romilly 
Allen points out that in form the monument belongs to a well-known Mercian 
type (Arch. Camb. V. xvi. [1899], 19). According to Phillimore, Llwyd's Eliseg' 
is probably a misreading of Elisef, which represents Elisedd, the old form of 
Elise and Elis (Cymr. ix. 181). 

84 Cymr. ix. 175. 86 Harl. MS. 3859 s.a. Ibid. 

87 A doublet of the name Dinas Bran appears to be preserved in Dinbren or 
Dinbran, the name of the township in which the castle stands. See Goss. 
Guide, pp. 130-1. The township is now, with the rest of the parish of Llangollen, 
in the Nanheudwy division of the hundred of Chirk, but appears to have been 
originally in lal. 


" wizard stream " ; here was the pillar which told of the valour CHAP, 
of King Elise, and here was the ancient church, Llanarmon yn 
lal, which commemorated the services of St. Germanus to the 
founder of the fortunes of the house of Powys. 88 lal was the 
heart of Northern Powys and remained in the possession of the 
Welsh as long as they retained their independence. Higher 
up the Dee was the small commote of Glyndyfrdwy, not yet 
rendered famous by the exploits of its great chieftain, Owen 
Glyndwr, and still more to the west the valley broadened out 
into the plain of Edeyrnion, a commote which took its name 
from Edern, or Eternus, ap Cunedda. No record is preserved 
of any dynasty deriving its origin from Edern and it is probable 
that the district, with the two subsidiary commotes of Dinmael 
and Glyndyfrdwy, was at an early date brought under the sway 
of the rulers of Powys. It included one important church, that 
of Corwen or Corfaen, 89 where there is an ancient cross of the 
same type as Elise's Pillar. 90 

Westward of Edeyrnion lay the border cantref 91 of Penllyn, 
of which the centre was Llyn Tegid or Pimblemere, 92 now 
Bala Lake. Ardudwy and Meirionydd hemmed it in on the 
west and it was thus exposed to the attacks of the men of 
Gwynedd, to whom it ultimately fell in the days of Llywelyn 
ab lorwerth. Encircled by lofty mountains, which enclose as 
in a girdle the broad waters of the lake and their fringe of 
meadows, the land is one of legend rather than of history, 
playing but a small part in the conflicts of mediaeval Wales, but 
rich in romantic memories. It was the realm of Gronw the 
Radiant, who stole the love of the magic bride of Llew Llaw 

88 It was probably the Llanarmon to which Gruffydd ap Cynan made a 
death-bed gift of ten shillings (Buck. Gr. ap C. 128 (734)). There was a special 
cult of Germanus in Powys, for the five principal churches dedicated to him are 
within the old bounds of the province (Rees, Welsh SS. 131 ; add to the four 
mentioned Castle Caereinion). In this connection it is interesting to note that 
the saint was mentioned in the inscription on Elise's pillar. 

8 For the old form see Bruts, 324; B. Saes. s.a. 1164; Tax. Nick. 286; 
Arch. Camb. I. ii. (1847), 241. 

80 Arch. Camb. V. xvi. (1899), 19. 

91 So treated in all the lists, though styled a cymwd in the Statute of 
Rhuddlan (LL. ii. 908) and the Record of Carnarvon (261). 

M The name first appears in Gir. Camb. vi. 176 (Descr. i. 5), where the form 
is " Pemmelesmere ". Rejecting the impossible derivation from " pum plwyf," the 
five parishes of Penllyn, one may suggest that Penllyn itself furnishes the first 
element of the compound. 


CHAP. Gyffes and perished miserably by the hand of the injured hus- 
' band ; 93 Llywarch the Aged, warrior and poet, is said to have 
lived there in the sixth century and to have held his court on 
the mound near Llanfor Church which bears his name ; 94 even 
the great Arthur took his place among the legendary heroes of 
the lakeside, for was not Caer Gai the 'home of his foster-father 
Cynyr and th&young Cai ap Cynyr the companion of his boyish 
sports ? 95 Caer Gai may have been the royal residence at the 
head of the lake from which the district took its name of 
Penllyn (Lakehead), although in later times Y Bala, which 
signifies " The Outlet," 9G was the seat of the chief stronghold 
of the region. 

The mountain ridge which divides the valley of the Dee 
from that of the Ceiriog formed the backbone of the commote 
of Nanheudwy, 97 which included an important church, once the 
mother church of a wide area, 98 but apart from Llangollen 
was undistinguished. Another commote of the second rank 
was Cynllaith, watered by the stream from which it took its 
name and having as its centre the church of Llansilin. Sych- 
arth, which was perhaps the ancient " llys " of the commote, had 
not yet become famous as the home of Owen Glyndwr. 
Further south came the rich commote 99 of Mochnant, the 
" fair wooded country " of a mediaeval Welsh poet, 100 where the 
valley of the Tanat broadens out into a fertile plain beneath 
the very shadow of the Berwyn range. In the twelfth century 
political conflicts brought about a division of this region into 
the two commotes of Uch Rhaeadr and Is Rhaeadr (Above and 
Below Rhaeadr), and from that day to this the rushing stream 

93 Mab. 74-81. The mound at Pont Mwnwgl y Llyn anciently bore, it is 
said, the name of Gronw Befr (Penn. ii. 215). 

94 " Pabell Llywarch Hen," a stone enclosure, and " Castell Llywarch," a 
mound, have long been known at Llanfor (Camb. Reg. i. 192 ; Evans, Rep. ii. 
p. 453 ; Penn. ii. 209 ; Lewis, Top. Diet. s.v. Llanvawr). 

95 See Goss. Guide, p. 153 ; Celtic Folklore, p. 693 ; note 62 to chap. iii. 

98 Gibson, 662 ; Evans, Diet. s.v. 

97 This form may represent " Nannau Dwy," " the glens of Dee ". Cf. the 
" Nanneudui" of Ann. C. MS. B. s.a. 1132. 

98 Thomas, St. Asaph, p. 505-6. 

99 The original Mochnant was a commote (Mab. 62, kymwt ym powys a 
elwir . . . mochnant). 

100 Llywarch ap Llywelyn calls it "uochnant kein amgant coedawc" (Myv 
Arch. I. 305 (215)). 


which takes its name from the well-known waterfall in the CHAP. 
Berwyns has been a boundary line, first separating Powys VIII> 
Fadog from Powys Wenwynwyn and then the county of Den- 
bigh from that of Montgomery. But the region is naturally 
one ; it found its ecclesiastical centre in Llanrhaeadr Mochnant, 
the church of St. Doewan, which was still served by a college 
or community of clergy as late as the days of Edward I. 101 
Pennant, the church of the female saint Melangell, was also an 
ancient sanctuary, remarkable to this day for the girth of its 
venerable yews, but it lay hid in a fold of the mountains, re- 
mote from the centre of the commote, and can never have held 
the dominant position of Llanrhaeadr. 

The river Cain gave its name to the cantref of Mechain, 102 
which took in the whole of the valley of this stream and was 
bounded on the south by the river Efyrnwy, now known as the 
Vyrnwy. A great forest in its centre divided it into the com- 
motes of Mechain Uwch Coed and Mechain Is Coed (Above 
and Below the Wood). 103 It was one of the most desirable 
parts of Powys, with abundance of land suited for tillage and 
no lack, at the same time, of fish and game. " Llys Fechain," 
the court of the cantref, stood on the banks of the Cain to the 
east of Llanfyllin, and its site is probably indicated by the 
mound known as Tomen Gastell. 104 " Llan Fechain," the 
prince's chapel, dedicated to that special patron of Powys, St. 
Garmon, was not far off. But Llanfechain was far outshone 
in importance as a holy place by the church of Meifod, on the 
southern border of the cantref. 105 Founded by a scion of the 
royal house of Powys, Tysilio son of Brochwel of the Tusks 
(Ysgythrog), ruled by a "clas" which could offer bards and 
other wayfarers a regal hospitality, and wielding authority over 
the daughter churches of Pool, Guilsfield, Alberbury and 

101 Tax. Nick. 286a (ecclesia de Rauraeader) ; Thomas, St. Asaph, p. 520. 

102 For the first part of the name ( = field, plain) see Urk. Spr. 198, s.v. 
magos. It is seen also in Machynlleth, Mathafarn, Mathrafal and Mallwyd 
(Goss. Guide, cxix.). 

10 ' The townships of the two commotes of Mechain, save for the parish of 
Llanfyllin, are given in App. Land Com. 451. 

104 Llys Fechain is a township in the parish of Llanfechain (Thomas, St. 
Asaph, p. 755). 

106 For the church of Meifod see Thomas, St. Asaph, pp. 774-81 ; for 
Tysilio, Welsh SS. 277-79. Cynddelw's poem is printed in Myv. Arch. I. 243-7 


CHAP. Llanfair Caereinion, Meifod was the premier church of Powys, 
and, until the foundation of the Cistercian abbeys of Ystrad 
Marchell and Valle Crucis, was the chosen burial-place of its 
kings. 106 Its praises were sung by Cynddelw 

Stately is the holy place by candle shine, 

Gracious its men with their long drinking-horns of flashing blue. 107 

It was " a privileged monastery," a " sepulchre of kings," lift- 
ing its proud head above the flooded meadows of Efyrnwy. 108 
A number of smaller commotes lined the Severn from 
Melverley to Berriew. Deuddwr, the land of the " two waters," 
occupied the peninsula formed by the confluence of the Vyrnwy 
and the Severn ; Ystrad Marchell lay around the church of 
Guilsfield, of which the Welsh name is Cegidfa (Hemlock 
Field) ; Llannerch Hudol was a small tract to the south and 
west of Welshpool. The dwellers in these lands fronted the 
Mercian villages on the western slopes of the Long Mountain, 
Buttington, Leighton, Forden, and the rest, and, keepers as 
they were of the fords of the Severn, were chiefly intent upon 
the border warfare. Romance might tell of the halcyon days 
when Arthur and his attendant knights had spread their camp 
for a full mile along the river in the plain of Argyngroeg, 109 
but such ease and security were not known in the age of 
Offa. It is possible that Pool itself, called Welsh Pool to dis- 
tinguish it from Poole in Dorset, may owe its origin to an 
English settlement, 110 thrown across for the defence of a ford, 
though there was on the hillside hard by a Welsh ecclesiastical 
foundation, known to the natives as Trallwng Llywelyn, or 
Llywelyn's Bog. 111 A little further south was the strong post 

106 Madog ap Maredud (d. 1160) and Gruffydd Maelor (d. 1191) were buried 
at Meifod. 

107 " Berth y Hoc wrth lieu babir, 

Berth y chlas ae chyrn glas gloyuhir." 

" Pabir " is a variant of " pabwyr," the two being derived from different forms 
of "papyrus" (Mots Latins, p. 192). " Clas " and "Hoc" both suggest the 
monastic origin of the church ; the latter, from the Lat. " locus" (Mots Latins, 
p. 182 ; Gildas, ed. Williams, p. 262), is retained in the compound form " mynach- 

IDS Breiniauc loc," " guydva brenhined," and "balch y Hoc rac y llifeir- 
yeint " are the phrases referred to. 

109 Mab. 148. 

110 It would otherwise have hardly had an English name. 

ul Bruts, 290, 339. The name is explained by Edward Llwyd (Gibson, 
6 53); c f- also Cymr. xi. 39, where a number of similar forms are given. 


of Castell Coch, which became in later times, under the name CHAP, 
of Powis Castle, the principal fortress of Southern Powys. 
The ancient capital of this region was, however, in a remoter 
and safer quarter. Mathrafal, regarded in Welsh mediaeval 
literature as the " principal seat " of Powys and one of the 
three royal residences of Wales, 112 stood on the banks of a 
tributary of the Vyrnwy, a little above Meifod. In front was 
a fertile plain, yielding the necessary produce for the royal 
kitchen and stable ; behind stretched in wave after wave the 
rolling hills of the great commote of Caereinion. This was the 
heart of Southern Powys, where its princes might hope to find 
in their direst extremity a refuge from English or Norman 

In the cantref of Cydewain, which lay between the Rhiw 
and the Severn, was the church of Aberriw (now Berriew), 
where St. Beuno fled from before the face of the invading 
English, 113 and the fortified steep of Dolforwyn, fabled by some 
to have witnessed the drowning of the hapless Sabrina in those 
dim days of which Geoffrey of Monmouth has told the story 
with such surprising particularity. 114 A cantref which filled a 
far larger space in Welsh history was Arwystli, the land of the 
head waters of the Severn. It was originally, no doubt, a part 
of Powys, but an early connection sprang up between its 
chieftains and the realm of Gwynedd, so that it was frequently 
treated as belonging rather to Aberffraw than to Mathrafal, 
and, in particular, it became permanently attached to the 
bishopric of Bangor. 115 As in the case of Mechain, its two 
commotes were known as Uwch Coed and Is Coed, from a 
stretch of forest in the midst of the cantref; the respective 
manors or royal courts appear to have been at Talgarth, near 
Trefeglwys, and Caersws. There were also in the cantref two 
notable churches. Llandinam was the mother church of 

Trallwng is now pronounced Trallwm ; cf. " carlwm " for the older " carlwng " 
(Evans, Diet. s.v.). 

lia LL. ii. 50, 380, 584. It should be added that the position of Mathrafal 
is not as well attested as that of Aberffraw and Dinefwr, and the story of the 
transference of the capital thither from Shrewsbury in the eighth century is of 
modern growth (note 10 to chap. vii.). 

U3 Chap. vi. note 38. 

U4 Hist. Reg. ii. 5. Local tradition, and not Geoffrey, is responsible for 
fixing the site of the tragedy here (Penn. iii. 186). 

118 Trans. Cymr. 1899-1900, 157. 


CHAP. Llanidloes and Llanwnnog ; it had an abbot as late as the 
' twelfth century and was still held by a group of clergy at the 
time of the Edwardian conquest. 116 Llangurig was not so 
wealthy, but could also boast of a " clas," as well as of a patron 
saint, St. Curig, who was in high esteem as a worker of miracles 
in the surrounding district. 117 The bishop of Bangor had a 
manor at Llanwnnog, to which he no doubt resorted when his 
duties brought him to Arwystli. 118 Beyond Arwystli was 
Cyfeiliog, the westernmost of the commotes of Powys proper, 
which just touched the sea at the head of the tidal estuary of 
the Dovey. Its position on the confines of Gwynedd, Powys 
and Deheubarth made it the scene of many a conflict, and often 
did war rage around the stronghold of the chief at Tafolwern. 
The sister commote of Mawddwy was sheltered, on the other 
hand, by the beetling crags of Aran Fawddwy, and its history 
was uneventful ; in its inmost retreat was " y Llan ym 
Mawddwy," the spot hallowed by the bones of St. Tydecho, 
who had withstood the great Maelgwn Gwynedd himself, and, 
it was reverently believed, still kept watch and ward over the 
fair fields of the commote. 119 

The cantref of Meirionydd may be reckoned as in some 
measure a part of Powys, for its ancient name was Cantref 
Orddwy, or the cantref of the Ordovices. 120 But, from the 
time of the Meirion or Marianus (grandson, as was alleged, of 
Cunedda) who gave it its more familiar name, it appears to 
have been for many generations under the rule of its own 
dynasty, represented about 870 by a certain Cynan ap 
Brochwel. 121 One or two interesting names occur in the 
pedigree of this line. Gwrin Farfdrwch (of the Ragged Beard), 

110 Arch. Camb. V. xi. (1894), 29 (Arustly) ; III. vi. (1860), 331 (Dolfin abbas 
Llandinan) ; Tax. Nick. 2gib. Llandinaw is the old spelling; see Cymr. xi. 

117 The township of Llangurig (divided into Llan i fyny and Llan i waered) 
formed the manor of" Clas Arwystli " (App. Land Com. 451). See Gir. Camb. 
vi. 17-8 (Itin. i. i) for the virtues of St. Curig's crosier. 

118 Rec. Cam. 115. 

119 Camb. Reg. ii. 375-8. The poem there printed calls Tydecho 

" Crefyddwr cryf o Fawddwy 
Ceidwad ar eu holl wlad hwy." 
(" A great man of religion from Mawddwy and guardian of all their land.") 

120 Celt. Br. (2), pp. 302-3. 

121 Pedigree XVIII. in Harl. MS. 3859 (Cymr. ix. 178). 


who perhaps gave his name to the neighbouring church of CHAP. 
Llanwrin, furnished Geoffrey of Monmouth m with one of the 
figures of his ever-shifting panorama, that of the just and cle- 
ment king who found the ancestors of the Irish cruising aim- 
lessly about in the seas to the north of Britain and gave them 
the island in which they afterwards became so mighty 
a people. Another of the line bore the name of Idris, 
still commemorated in that of the loftiest of the many heights 
of Meirionydd ; for Cader Idris was the chair of no Druid or 
astronomer, but takes its name from a brave descendant of 
Meirion, who fell in battle on the banks of the Severn about 
63O. 123 The realm which obeyed this forgotten dynasty ex- 
tended from the estuary of the Dovey to that of the Mawd- 
dach and further included the valley of the Wnion. It was a 
land of rocky confusion, where crag rose above crag ; " shaggy 
and fearsome," says Giraldus, beyond any other region of 
Wales. 124 Shepherds perched on two opposing peaks might 
exchange abuse in the morning, but the day, he avers, would 
be far advanced ere they could meet in the valley bottom to 
settle their differences. Yet Meirionydd had its stretches of 
fertile soil. In the heart of the cantref, where the Disynni 
spreads itself out ere it is lost in Cardigan Bay, was " Tywyn 
Meirionydd," the Sandy Plain of Merioneth, where St. Cadfan 
founded in the sixth century a notable church. 125 Towyn was 
not inferior in importance to any save the two cathedral 
churches of North Wales ; it was the mother church of the 
whole of the commote of Ystum Anner, 126 it had an abbot in 
the twelfth century, 127 and, again recalling Llandinam, was 
held by a number of clergy in the days of Edward I. 128 Like 

122 His*. Reg. iii. n, 12. 

123 " Strages sabrine et iugulatio iudris " (Harl. MS. 3859, s.a. 632 Cytnr. 
ix. 158). Cf. the notices of Tighernach " cath ludruis rig Bretan (the battle of 
Idris, king of the Britons) qui in eo cecidit " (Rev. Celt. xvii. (1896), p. 182) and 
of the Annals of Ulster, s.a. 632 ( = 633) " bellum Iudris regis britonum ". The 
date is in harmony with the position of Idris in the pedigree. Cf. Goss. Guide, 
79 ; it should, however, be added that in the Celtic annals " jugulatio " often 
refers, as probably here, to ordinary killing in battle, and does not of necessity 
imply throat-cutting. 

124 Gir. Camb. vi. 122-3 (Itin. ii. 5). 

125 There is a full account of Cadfan in Welsh SS. 213-5. For the so- 
called Cadfan stone see note 135 to chap. vii. 

128 B. Willis, Bangor, p. 276. 187 Note 57 to chap. vii. 

128 Tax. Nick. 2916. 


CHAP. Meifod, the place had its poet, one Llywelyn Fardd, 129 who 
' sang in the twelfth century of the 

Lofty fane of Cadfan by the margin of the blue sea, 
and prayed for its welfare 

Prosper may its turf and its corn and its seed, 

nor forgot its abbot, the liberal rewarder of poetic merit, for at 
Towyn there was 

Peace and mead borne in vessels 

And easy converse exchanged with a bard. 

Less is known of the civil sites of the cantref. It was divided 
by the Disynni into the two commotes of Talybont and Ystum 
Anner, two names which may indicate two ancient seats of the 
lords of Meirionydd, for there is a mound at Talybont (near 
Llanegryn), and Ystum Anner can be located at Llanfihangel 
y Pennant, where stands the fortified rock of Castell y Bere. 
But history speaks of castles also at Cynfael, near Towyn, and 
Cymer, near Dolgelly, and it is not easy to say which strong- 
hold was the temporal, as Towyn was the spiritual, centre of 
the district in the days of its independence. 

The southern portion of the ancient Fowys was known in the 
Middle Ages as " Rhwng Gwy a Hafren," the land betwixt Wye 
and Severn. 130 It was a district which formed no part of the 
patrimony of the rulers of Powys descended from Bleddyn ap 
Cynfyn (d. 1075), DUt many circumstances point to its having 
once been included in the province. Tradition alleged that 
the southern limit of Powys was Rhyd Helyg on the Wye, near 
Hay ; 131 in 1 176 Bishop Adam of St. Asaph sought to estab- 
lish a claim to the whole of Rhwng Gwy a Hafren, no doubt 
on the plea that his diocese covered the whole of Powys, 132 

129 Myv. Arch. I. 360-2 (248-50). If one sets aside as the work of a later 
hand the poem to Owain ap Gruffydd ap Gwenwynwyn, the rest of the poems 
attributed in the Myvyrian Archaiology to Llywelyn Fardd fall naturally into 
the latter half of the twelfth century. The common date, 1230-1280, is a clumsy 
attempt to reconcile hopeless anachronisms. 

130 The form occurs, not only in the lists of cantrefs and commotes, but also 
(in Latin) in Gir. Camb. i. 30, 35 (De Rebus, i. 5, 6) ; vi. 19 (Itin. i. i). 

131 lolo MSS. 31 ; Camb. Qu. Mag. iii. 403. 

132 The claim to Kerry was avowedly only a first step or trial of strength. 
See Gir. Camb. i. 32-3 (totam . . . terram usque Vagam (Wye) . . . occupare 


while the early history of our Radnorshire, so far as it can be CHAP, 
unravelled, seems to make it an Ordovician country. 188 

Setting aside the two small com motes of Keri m and Cymwd 
Deuddwr, 135 of which the former is now in Montgomeryshire, 188 
Rhwng Gwy a Hafren comprised the commote of Gwerthrynion 
and the cantrefs of Buellt, Elfael and Maelienydd. Buellt, 
being so much of our Brecknockshire as lies north of the Epynt 
range, was not entitled geographically to be reckoned among 
the lands between the Wye and the Severn, but the Historia 
Brittonum shows that about 800 it was joined with Gwerthry- 
nion under the rule of Ffernfael, 137 and thus it was no doubt 
included in the term as the definition of a political area. The 
mountains of the north of Buellt are separated from those of 
the south by the broad valley of the Irfon, and here, or close 
by, were the chief sites of the cantref, the courts of the four 
commotes of Treflys, Penbuellt, Dinan and Is Irfon, 138 the 
ancient church of Llanafan Fawr, where the tomb of Afan 
Buellt was shown, chief of the saints of this region, 139 and St. 
David's church of Maes Mynis, which dominated Is Irfon and 
was the mother church of Llanfair ym Muellt (St. Mary's in 
Buellt), the modern Builth. 140 Gwerthrynion was on the 
opposite side of the Wye, occupying the region between that 
river and the leithon. The name is obviously formed from 

138 On the whole question see the discussion in Owen, Pemb. i. 203 (Philli- 

1M "Keri" is mentioned in Mab. 62. The modern village bearing the 
name is properly Llanfihangel yng Ngheri (Pen. MS. 147 in Evans, Rep. i. 

P- 915)- 

138 Omitted from the lists. It is represented by the parish of Llansant- 
ffraid " Cwm Toyddwr " see Pen. MS. 147 (Evans, Rep. i. p. 915, K. dayddwr) ; 
Reg. Conway, p. 9 (Comottewthur) ; Owen's Pemb. i. 203. The "two waters" 
are the Wye and the Elan ; from the latter is probably derived another name of 
the commote, viz., Elenyddor Elenid (Mab. 62; Gir. Camb. i. 117; vi. 119, 138, 
171, 173 ; W. People, p. 45 ; Owen's Pemb. i. 203). 

138 In all the lists of cantrefs and commotes Ceri is associated with Gwerthry- 
nion or Maelienydd, not with Cydewain or Arwystli ; this is also implied in its 
position in the diocese of St. David's and was common ground between the dis- 
putants in 1176 (Gir. Camb. i. 35 (De Rebus, i. 6)). 

187 Hist. Britt, cc. 48, 49; note 144 to chap. vii. above. 

138 Treflys is close to Llangammarch, Llys Dinan, near Newbridge on Wye. 
The court of Is Irfon was probably at Llanfair, where Builth Castle was after- 
wards raised ; Penbuellt would seem to have been in the southern part of Llan- 
gammarch parish (Carlisle). 

Welsh SS. 208-9. i ibid. 53. 


CHAP. Gwrtheyrn, better known in its older guise of Vortigern, and 
VIII> the legend of the country was that the notorious bearer of that 
name had fled hither to escape the fulminations of St. Germanus, 
and that after his death his descendants had been settled in the 
commote as its rulers. 141 Some alleged that the tyrant had 
actually come to his shameful end in "the castle of Gwer- 
thrynion on the river Wye," by which is probably meant 
Rhayader, but there were several other claimants for this 
distinction. 142 It is certain, however, that the early chieftains 
of Gwerthrynion traced their origin to Gwrtheyrn Gwrthenau, 
and it should further be noted that here, as in other quarters of 
Powys, there was a church under the patronage of Germanus. 
St. Harmon's, where there is known to have been a " clas," 143 
was probably pointed out by tradition as the spot where the 
saint, with all the clergy of Britain, prayed on the bare rock 
for forty days and forty nights for the conversion of the 
reprobate king. 144 

South Radnorshire was occupied by the cantref of Elfael, 
separated from Buellt and Brycheiniog by the swift-running 
stream of the Wye. Save along the banks of this river the 
cantref was a region of grassy highlands and did not rank as 
one of the coveted tracts of Wales. The broad-backed chain 
of hills to the south of Aberedw served to divide it into the 
commotes of Uwch Mynydd and Is Mynydd (Above and Below 
the Mountain), which had in Norman times their respective 
castles of Colwyn and Painscastle ; what strongholds had in 
earlier days been held by the native chieftains in these two 
commotes there is little evidence to show. Of the ancient 
churches of Elfael the most notable was undoubtedly Glascwm. 
It was one of the greater Dewi churches, founded, as was 
believed, by the saint himself, 145 and here was kept a most 
precious legacy of his, the portable bell called " Bangu " which 
was endowed with miraculous powers. 146 In the White Bard 

141 Hist. Britt. cc. 47-9. 

142 Triad i. 91 = ii. 6. For the other stories see Diet. Nat. Biog. s.v. 

143 ''Clas Garmon " is the name of one of the two townships of St. 
Harmon's (Carlisle). 

144 Hist. Britt. c. 47. 

145 Life by Rhygyfarch in Cambro-Br. SS. 123. 

146 Gir. Camb. vi. 18 (Descr. i. i). Cf. Cambro-Br. SS. 136 (nola). 


of Brycheiniog's catalogue of the churches raised in honour of CHAP. 
St. David it is the 

Church by the green hillside, 
Towering among the thickets, a sanctuary that faileth not. 147 

Its " clas " was a daughter community of the great house of 
Mynyw, and Glascwm thus came to be regarded as a manor of 
the bishop of St. David's, whose rights in the place have not 
yet become wholly extinguished. 148 

In its physical features the cantref of Maelienydd, 149 ex- 
tending from the Teme to Radnor Forest and the neighbour- 
hood of Llandrindod, was not dissimilar. The leithon was its 
principal river, winding its way south through miles of gorse- 
clad moor and upland pasture. From an early period the 
cantref was divided into three commotes, 150 which took their 
names from the principal royal residences within them. 
Rhiwlallt lay along the Lug, where stood, near Llangynllo, 
the spot, known to the English as Weston, which originally 
bore the name. 151 The " swydd " or " shire " of Buddugre 
was the northern portion of the cantref, the particular Buddugre 
or " Hill of Victory " which formed the centre of this commote 
being on the banks of the leithon. 152 The " swydd " of Diniei- 
thon was the southern limb of Maelienydd, and no doubt the 
" din," or fortress, on the leithon from which it took its name 
was that of Cefn Llys, where the river winds around and almost 
isolates a steep hill which must have been in primitive as in 
Norman days a natural stronghold of great importance. Of 
the saints of Maelienydd none could compare in importance 

147 A glascwm ae eglwys gyr glas uynyt 

Gwyteluod aruchel nawd ny achwyt " (Myv, Arch. I. 271 (194)). 
The Red Book text has " chwyd," with the same meaning. 

148 Black Book of St. David's, 291, 331 ; Carlisle, Top. Diet.; App. Land 
Com. 445. 

149 No doubt derived from a proper name which would in old Welsh be 
Mailgen. There is no ancient authority for the form " Moelynaidd " favoured 
by Jonathan Williams (Hist. Radn. passim). The deanery of Maelienydd in- 
cluded Keri and Gwerthrynion, and some of the lists appear to make the cantref 
no less extensive ; in the oldest, however, that in Dom. viii., it has three com- 
motes only. 

150 For a list (not very carefully put together) of the parishes and townships 
of these commotes see Radnorsh. (2), 142 (Nos. 3, 4, 8). 

181 L. G. Cothi, 239, 243, 245, 246, 330. 

188 It is the " Bedd Ugre " of the maps. For the true form see Bruts, 409, 
Cymr. ix. 328 ; for the meaning and other instances, Evans, Diet. s.v. 


CHAP, with Cynllo, the sphere of whose influence, in this and the 


neighbouring district of Gwerthrynion, included no less than 
thirteen modern parishes. 183 His principal church was at 
Llanbister, to which accrued the largest income of any church 
in the archdeaconry of Brecon. 154 


The original " dextralis pars Britanniae " or " Deheubarth 
Kymry " embraced the whole of South Wales, in which sense 
the term is used by Asser 155 and by those who drafted the 
charters in the Liber Landavensis. 1 But in later parlance 157 
the name Deheubarth came to be restricted to the realm, which 
included most of the South Welsh area, formed by the accretion 
of Ceredigion, Ystrad Tywi and Brycheiniog around the ancient 
kingdom of Dyfed, and in this sense it was exclusive of Gwent 
and Morgannwg. In the present section an attempt will be 
made to trace the earlier history of the constituent elements of 
this later Deheubarth. 

Ceredigion, the territory which Ceredig ap Cunedda is said 
to have carved for himself out of the country of the Demetae, 
appears to have had throughout its history the same borders 
as the modern county of Cardigan, 158 which thus hands on the 
tradition of the territorial unity of this district. Its bounds 
are given by a man of Ceredigion in the eleventh century 159 
as follows to the east, lofty mountains ; to the west, the ocean ; 
north and south, broad rivers (i.e., the Dovey and the Teifi) 
and this is still a good general description. The eastern 
boundary, from Glandovey to Lampeter, crosses and recrosses 
the watershed of Plynlimmon in somewhat devious fashion, but 
it is still true to say that Ceredigion is a land between the 
mountains and the sea, with all its rivers, including the two 

163 1 refer to Llanbister, Llangynllo, Nantmel and their chapels as given by 
Rees (Welsh SS. 351). 

154 Tax. Nich. 2746 (Lanbyst, 30 : 13 : 4). 

155 C. 80. Cf. note in Stev. 233-4. Asser more than once has " Britannia " 
in the sense of Wales (cc. 7, 79 (bis)). 

166 Life. Land. 161, 162, 163, 169, 192, 212, 223, 230, 237. 

157 In B.T., for instance. 

158 There is some evidence of an extension of Ceredigion at one period so 
as to include the cantref of Cemais (Owen, Pemb. i. 222-3), but this can only 
have been a temporary success of its chieftains. 

159 leuan ap Sulien of Llanbadarn about 1090 (H. and St. i. 664-5). 


which make its northern and its southern border, running CHAP. 
westward. Its eastern barrier" of lonely moorland, though 
lacking the beauty of outline of the Snowdonian range, served 
a similar defensive purpose, and the stranger who set greedy 
eyes on Ceredigion, whether Brython or Norman or Englishman, 
usually made his attack from the north or from the south. 

For some 400 years, if the evidence of the royal pedigree 18 
is to be trusted, this country was ruled by the descendants of 
Ceredig. Some of the names recorded are undoubtedly his- 
torical. Seisyll, whoiwas king about 730, embarked on a career 
of conquest and added to Ceredigion the three cantrefs of Ystrad 
Tywi, the whole dominion being henceforth known from the 
name of its founder as Seisyllwg. 161 The death of his son 
Arthen is chronicled under the year So/. 162 The last of the 
line was Gwgon ap Meurig, a great grandson of Arthen, who 
is recorded to have been drowned in or about 871 ; 163 his sister 
Angharad was married to Rhodri the Great, king of North 
Wales, 164 and it was thus, no doubt, that Seisyllwg became 
incorporated with Gwynedd and Powys in one realm, which 
was shortly afterwards divided among the sons of Rhodri. 

Ceredigion is always spoken of as a land of four cantrefs. 165 
But the names of the cantrefs were, with the exception of that 
of the northernmost, not generally current ; the well-known 
divisions of the country were the ten commotes into which the 
four cantrefs were divided. Penweddig, 166 or, as it was some- 
times called, " Y Cantref Gwarthaf," m " the uppermost " cantref, 
formed an exception ; the Ystwyth, stretching across the whole 
of Ceredigion, supplied it with a well-marked southern border, 
and it was thus a clearly defined and well-known area. It 
contained three commotes, of which Geneu'r Glyn lay north of 
the Clarach, Perfedd (so called as the midmost of the three) 

180 Cymr. ix. 180, No. xxvi. 

181 This is inferred from Mab. 25, where " Seisyllwch " is explained to be 
Ceredigion and Ystrad Tywi combined. Cf. Basil Jones, Vestiges of the Gael 
in Gwynedd (Lond. and Tenby, 1851), pp. 61-2 ; Cymr. xi. 56. Seisyllwg is also 
mentioned in LL. ii. 50 and Triad iii. 14. 

IBS Arthgen rex cereticiaun moritur " (Harl. MS. 3859 in Cymr. ix. 163). 
IBS < Quoccaun mersus est rex cetericiaun " (ibid. 166). 184 See p. 325. 

lM Mab. 25, 59 ; Str. Flor. Appendix, xxi. 

188 The dd is implied in the " penwefic " of Blk. Bk. fo. 330 and the 
" PenwefAig " of Gir. Camb. vi. 175 (Descr. i. 5). 

187 Bruts, 410. Cf. " hyt yggwarthaf keredigyawn " (Mab. 62). 
VOL. I. 17 


CHAP, between the Clarach and the Rheidol, and Creuddyn between 
VIIL the Rheidol and the Ystwyth. 168 Forts and castles of all ages 
were dotted over this border land between North and South 
Wales, among the most notable being the hill-fortress of 
Pendinas, anciently known as Dinas Maelor. 169 The neigh- 
bouring harbour of Aberystwyth had not yet been formed, and 
Ystwyth and Rheidol fell into the sea a mile or so apart. 
Penweddig had ecclesiastical no less than military importance, 
for there was in it a " clas " of the first rank, that of Llanbadarn 
Fawr, founded by Padarn or Paternus of Brittany, which was 
once the seat of a bishop and still retained its abbot in the 
twelfth century. 170 Llanbadarn was the mother church of the 
whole of Penweddig, 171 and the community held as church land 
that part of the commote of Perfedd which lay next the sea. 172 
A slender Celtic cross over 8 feet high still stands in the 
churchyard 173 as a memorial of the pre-Norman glories of the 
place, which are also faintly suggested by the chronicler who 
tells us that it was pillaged by the Danes at the end of the 
tenth century. 174 

Of the remaining commotes of Ceredigion, the northernmost 
was Mefenydd, which lay along the southern bank of the 
Ystwyth, from the sea to Cymwd Deuddwr. To the south- 
west came the little commote of Anhuniog, so called, no doubt, 
from an Annun or Antonius, 175 long forgotten, who ruled over 
it ; it reached as far as the Aeron. The south-eastern angle 
of Ceredigion belonged to the commote of Pennardd, an ex- 
tensive region which took in Tregaron, Llanddewi Brefi and 

168 Lewis Morris gives the bounds of the crown manor of Creuddyn in a docu- 
ment printed in Meyrick, Card, (i), 555-67. The (old) borough of Aberystwyth 
lay outside this manor, but it was probably taken out of the commote of Creuddyn, 
the northern and southern boundaries representing the old channels of the 
Rheidol and Ystwyth respectively. See the plan in the Report of the Boundary 
Commissioners of 1831. Morris also gives (ibid. 568) the bounds of the crown 
manor of Perfedd ; the commote of the name included in addition Llanbadarn 
and its manor of Vainor. 

169 Brython (1860}, Hi. 331; Evans, Rep. i. p. 724. It was also known in 
Welsh legend as Rhiw Faelor (Triad ii. n). 

170 See pp. 159, 208. 

171 It is the only church in the cantref mentioned in Tax. Nick. 

172 The church claimed all the land between the Clarach and the Rheidol by 
donation of Maelgwn Gwynedd (Cambro-Br. SS. 192). 

173 Arch. Camb. V. xiv. (1897), 152-3. 174 Ann. C. s.a. 988. 

178 Haminiog and Harminiog are modern corruptions. For Annun = An- 
tonius see Cymr. x. 86 ; Mots Latins, pp. 132-3. 


the site of the future abbey of Strata Florida. In these three CHAP, 
commotes, sometimes grouped with Penweddig under the name 
of Uch Aeron (Above Aeron), the most famous of the early 
sites was Llanddewi Brefi, known through the length and 
breadth of Wales as the scene of the miracle wrought by St. 
David when the hill of Brefi rose beneath his feet and formed 
a pulpit from which he addressed the synod and the attendant 
crowds there gathered together. 178 The " clas " which was sub- 
sequently formed there was naturally a daughter-house of St. 
David's, and Llanddewi, with the moorlands which rise above 
it to the east, became an episcopal manor and forest, 177 

From Caron 

To the banks of the Towy, that fair and beauteous stream ; 
From the Black Pool, where once was wrathful encounter, 
To the Twrch, divider of lands and rocks. 178 

Is Aeron, or Ceredigion below Aeron, consisted of four 
commotes. Of these Caerwedros was a coast region, extending 
to Llangrannog ; its castle stood, at any rate after the Norman 
Conquest, a short distance from Llwyn Dafydd. 179 Mabwnion 
spread its length from the neighbourhood of Aberaeron to that 
of Lampeter ; part of the rich vale of Aeron, praised by the 
White Bard for its meadows of clover and the wealth of acorns 
in its woods, 180 fell within the compass of this commote. The 
two remaining commotes lay along the northern bank of the 
Teifi, one of the greater streams of Wales, in which the beaver 
still built in the days of Giraldus Cambrensis, though no longer 
to be found in any other river of Southern Britain. 181 It was 
a river of salmon, also, in which the coracle men took, as they 
do to this day, great catches of fish in the swirling eddies below 

178 See p. 157. 

177 Elk. Bk. of St. David's, 197-203; App. Land Com. 444 ; Meyrick, Card. 
(2), 250. The act of 1536 added " Landway Ureny," until then a marcher lord- 
ship, to Cardiganshire. 

" O garawn gan yawn gan ehoec 

Hyd ar dywi auon uirein a thec 

Or llyndu lied vu Hid gyhydrec 

Hyd ar twrch teruyn tir a charrec." 

Gwynfardd Brycheiniog in Myv. Arch. I. 272 (195). 

179 Meyrick, Card. (2), 233. 

" A henuyniw dec o du glennyt aeron 
Hyfaes y meillyon hyfes goedyt." 

Myv. Arch. I. 271 (194). 

181 Gir. Camb. vi. 114 (Itin. ii. 3), 173 (Descr. i. 5). 



CHAP, the rocks of Cilgerran. 182 Iscoed and Gwinionydd, with their 
' southward-sloping valleys and their meadows fringing the Teifi, 
were the two most desirable commotes of the land of Ceredig, 
and here was Rhuddlan Teifi, the Red Bank on the Teifi, where 
Pryderi, the legendary king of South Wales, entertained to his 
loss and dire misfortune the cunning Northern wizard, Gwydion 
ap Don, and his companions. 183 In later times the royal 
residence of Gwinionydd was at Pen Coed y Foel, near Llan- 
dysil, 184 and the bishop of St. David's had also an important 
manor in this region of which the ancient church of Dewi at 
Bangor was the centre. 185 Iscoed, too, had its sites of legendary 
and historical interest, though Cardigan town and castle had as 
yet no existence. At Crug Mawr was the conical hill, still a 
conspicuous object in the landscape around Cardigan, which 
legend associated with the cure of melancholy ; he who bowed 
thrice before it would never again suffer weariness of soul, 
"though he might wander solitary in the uttermost parts of 
the earth". 186 Penbryn, on the sea coast, was another im- 
portant early centre ; here was the standing stone of Corba- 
lengi, 187 and the discovery in the same neighbourhood of Late 
Celtic bronze objects resembling spoons 188 shows that it was 
a scene of active life in the shadowy period following the 
Roman Conquest: 

Crossing the Teifi from this part of Ceredigion, the traveller 
of olden time found himself in the kingdom of Dyfed, which 
occupied the south-western peninsula of Wales. This was one 
of the last regions to feel the force of the Brythonic conquest, 
and more than one indication is afforded of the continuity here 
between the older life and the new. The country kept the name 
of its pre-Roman settlers ; it was ruled by the same dynasty 
from the fifth century to the tenth, and further by a house which 
had no connection with Cunedda Wledig. No conquests by 
the kings of Dyfed are recorded ; they represented the ancient 
and declining race ; theirs was a waning star, and it was their 
part gradually to yield to the encroachments of the Brythons 

182 Gir. Camb. ibid. For the coracles see pp. 201-2 (Descr. i. 17). 

183 M ab. 61. Pentre Rhuddlan, near High Mead, indicates the site. 

184 Meyrick, Card. (2), 192. 186 /A. Bk. of St. David's, 211-35. 

186 This is the last of the " mirabilia Britanniae " (Hist. Britt. c. 74). 

187 See p. 119. 196 Arck. Camb. IV. i. (1870), 205-7. 


until their whole realm fell at last into the lap of a scion of the CHAP, 
house of Gwynedd. 

At one time the Demetian country had formed a consider- 
able part of the area of South Wales, and men spoke loosely 
of the whole land to the north of the Bristol Channel as 
Demetia. 189 But first had come the loss of Ceredigion ; then, 
in the early part of the eighth century, the broad lands between 
the Teifi and the Tawe had been reft from Dyfed by the 
prowess of Seisyll ; 19 until at the time of the Danish invasions 
the name had come to be restricted to the narrower region 
which is the Dyfed of mediaeval literature. It was now bounded 
by the sea, the Teifi, the Tywi, and a line connecting these 
two rivers which may be taken as running due north from 
Carmarthen. The 'modern county of Pembroke was thus 
entirely included in Dyfed, but to it has to be added a sub- 
stantial part of Carmarthenshire if an idea, however rough, is to 
be formed of the extent of the ancient kingdom. Its seven 
cantrefs 191 were of very diverse physical aspect, from the barren, 
rocky moors of Pebidiog to the sandy shores of Penfro, warmed 
by the southern sun, and the rich alluvial meadows of Cantref 
Gwarthaf. But, taken as a whole, it was a goodly realm, as 
may without difficulty be gathered from the circumstance that 
in the eleventh century it fell not only speedily, but irredeem- 
ably, into Norman hands. 192 

The kings of Dyfed drew their origin, according to one 
account, from a national hero of the name of Dimet, 193 ac- 
cording to another, from a branch of the royal line of the 
Deisi in Ireland, 194 but in either case from no Brythonic source. 
At the beginning of the sixth century the dynasty was repre- 

189 The various applications of the name Dyfed are fully discussed by Mr. 
Phillimore in Cymr. xi. 56-7 ; Owen, Pemb. i. 45-6, 199, 224, 257-8. 

190 Hist. Britt. c. 47 mentions an " arcem Guorthigirni quae est in regione 
Demetorum iuxta flumen Teibi ". The reference is, no doubt, to the fortified 
" Craig Gwrtheyrn " which stands near Llandysul, but on the south side of the 
river and therefore in the later Cantref Mawr. 

191 Mab. i, 25, 44, 57, 59; Gir. Camb. vi. 93 (I tin. i. 12) ; 166 (Descr. i. 2). 
1M The Welshman and the Norman united in its praise; " nyt oes seith 

cantref well noc wy," says Pryderi in the story of Manawyddan (Mab. 44) : 
" terrarum omnium Kambriae totius . . . tarn pulcherrima est quam potissima," 
are the words of Giraldus (vi. 93). 

1M See the pedigree (No. ii.) of the house in Harl. MS. 3859 (Cymr. ix. 

194 Cymr. xiv. 112. 


CHAP, sented by Aircol or Agricola of the Long Hand, 195 whose name 
indicates that he inherited some Roman traditions, and who is 
further alleged to have been a good friend of the monastic 
movement of his day. St. Teilo was a subject of his, born at 
or near Pen Alun (now Penally) in the cantref of Penfro, and 
some of the lands in this neighbourhood, claimed in a later age 
by the see of Llandaff as heir to the saint, are maintained to 
have been given by the liberal hand of Aircol. His son and 
successor was the Vortiporius of Gildas, who bore the Roman 
title of protector and was buried at Castell Dwyran Church in 
the centre of his kingdom. 196 He, it need scarcely be said, 
was no patron of the monks, though it is equally certain that 
he professed the Christian religion. The crown of Dyfed 
passed on to men of lesser note ; early in the eighth century 
it was worn by Rhain ap Cadwgan, in whose reign, no doubt, 
befell the loss of Ystrad Tywi, for the diminished realm thus 
left to the Western dynasty is occasionally called Rheinwg, by 
contrast with the Seisyllwg of the conqueror. 197 Rhain's 
grandson Maredudd died in 796, 198 on the eve of the Danish 
invasions, and it may be conjectured that the disorder which 
followed, nowhere more signally than in Dyfed, pushed the 
ancient dynasty into the background, for, after the death of 
Maredudd's grandson Tryffin in 8i4, 199 all trace of it is lost, 
until in the age of Asser one Hyfaidd ap Bledri is found ruling 
over the district, with a hereditary claim based on the descent 
of his mother from Maredudd. 

The northernmost cantref of Dyfed was Cemais, which 
stretched along the coast from the Teifi to the Gwaun and 
southward across the Preselly range as far as the moors which 
fringe its southern slopes. It thus included the only mountain- 

196 Lib. Land. 118, 125-7, 12 9> r 3- His court of " Lis castell" is tradition- 
ally located at Lydstep, near Tenby (Arch. Camb. III. xiii. (1867), 366; Lit. Eng. 
pp. 81, 188). In the Blk. Bk. his grave is vaguely said to be in Dyfed (fo. 350; 
IV. Anc. Bks. ii. p. 35). Cf. Owen, Pemb. i. 223, for a reference to Aircol in the 
Book ofTaliesin, quite misunderstood in IV. Anc. Bks. i. p. 448, and Evans, Diet, 
s.v. Aercol. For the Latin origin of the name see Celt. Br. (2), p. 256, and 
Mots Latins, p. 131. 

196 Pp. 115, 132. He was " boni regis nequam fili(us)," says Gildas (c. 31). 

197 See note on Rheinwg, Esyllwg and Fferyllwg appended to this chapter. 
198 "Morgetiudrex demetorum (moritur)," Had. MS. 3859 (Cymr. ix. 162-3). 
199 (i Trifun films regin moritur" (ibid.). His uncle Owain, through whose 

daughter the crown descended to Hyfaidd, died in 811 (ibid.). 


ous tract to be found in Dyfed, a region dotted over with CHAP, 
cromlechs, cairns and barrows, fraught with ancient memories, 
the scene of many a romantic story. It was at Preselly the 
barons of the legendary King Pwyll met to expostulate with 
him upon the absence of an heir to the crown ; it was there 
that the giant boar called " Y Twrch Trwyth " killed no less 
than eight of Arthur's knights. 200 The saint of the district 
was Brynach Wyddel (the Irishman), whose principal church 
was at Nanhyfer (now Nevern), 201 lying half-concealed in a fold 
of the leafy valley of the Nyfer. This was for centuries the 
most important ecclesiastical site in Cemais ; ogam inscriptions 
mark it out as a Christian burying-place of the early period ; 20 
an elaborately carved Celtic cross over 1 2 feet high bears testi- 
mony to its standing in the age of the Danish incursions ; 203 
the massive church is a witness to its mediaeval importance. 
Cemais was divided by the river Nyfer into the two commotes 
of Uch Nyfer and Is Nyfer, but there is no evidence to enable 
one to fix the ancient royal seats of these divisions. To the 
west of the Gwaun came the cantref of Pebidiog, which opposed 
a rocky front to the sea at St. David's Head ; it resembled 
Anglesey, Giraldus observes, 204 in its bare and wind-swept 
aspect, but had not the kindly soil which made amends for the 
uncomely mien of the northern island. Poor as the country 
was, it was the home of a notable community, for here, in the 
little valley of Hodnant, within a few miles of the westernmost 
cape of Dyfed, was the church of Mynyw or Menevia, the 
mother-monastery of all the churches of Dewi. It is unneces- 
sary to repeat here what has been said in earlier chapters of the 
fame of " Ty Ddewi " (David's House), of its origin in the sixth 
century, of the austerity of its " clas," and of the tradition of 
learning there maintained. 205 Towards the end of the eighth 

ao Afa6. 18, 138. 

301 Nanhyfer is for Nant Nyfer, " the valley of the Nyfer," and appears in 
Harl. MS. 3859 (if a little slip of the scribe, who wrote " n/mer," be corrected) 
as " nant nimer " (Cymr. ix. 165). 

aoa The " Vitaliani " stone came originally from Nevern, and, after serving as 
a gatepost on Cwm Gloyn farm, is now there once more (Gibson, 638 ; Arch. 
Camb. V. xiii. (1896), 291). Another ogam inscription was discovered in the 
church in 1904. 

303 Gibson, 638-9; Inscr. Chr. No. 103; Lap. W. 100; Arch. Camb. V. xvi. 
(1899), 26. 

804 Wks. vi. 127 (Itin. ii. 7). oSee pp . i 54 -6, 227. 


CHAP, century the chronicle now most nearly represented by Harl. 
MS. 3859 began to be regularly kept at St. David's, and from 
this time forth there is a tolerably complete record of bishops 
of the see, beginning with Sadyrnfyw the Liberal, who died in 
8 3 1. 206 No stone remains of the pre-Norman church and 
buildings, save a few fragments of tombstones bearing Celtic 
ornament, and no account has survived of the revenues of the 
"clas" at this period, though it may be assumed that they 
included, not only rents from distant manors, but also a very 
considerable annual render from estates in Pebidiog itself. It 
is, indeed, possible that the cantref as a whole had no other 
lord in early times save the successor of St. David. 207 

In the three southern cantrefs of Dyfed the traces of the 
ancient Welsh life have been almost entirely obliterated by the 
Norman Conquest and the Flemish and English settlements. 
The old place-names have disappeared, save in a few isolated 
cases ; the old legends have been forgotten only the carved 
and inscribed stones, monuments like the great Celtic cross at 
Carew, remain to tell, and this in the barest outline, the story 
of a civilisation which has been swept away as thoroughly as 
that which the invading English found in South-eastern Britain. 
Especially has this been the case in Rhos, the cantref which, 
lying between the West Cleddau, Milford Haven, and St. Bride's 
Bay, perhaps took its name, as did its namesake in the Berfedd- 
wlad, from the promontory which it pushed out into the sea. 
Of the pre-Norman history of Rhos all that can be said is that 
it had a notable saint in Ismail or Ysfael, nephew of St. Teilo, 
whose church at St. Ishmael's or " Llanisan yn Rhos " was one 
of the seven bishop-houses of Dyfed. 208 Deugleddyf, a small 
cantref which lay, as its name indicates, between the eastern 
and the western Cleddau, furnishes little more to record ; the 
church of Llanhuadein (now Lawhaden) was its chief ecclesi- 
astical site and this had from an early period been a dependent 
church of St. David's, so that the place became an important 

ace Saturbiu hail miniu moritur " (Cymr. ix. 164). The name appears as 
" Sadurnueu " in the catalogue transcribed by Giraldus (not, it would seem, 
without some confusion of the chronological order) from the records of the see 
(Wks. vi. 102-3 (Itin. ii. i)). 

207 LL, i. 558 says Mynyw was to render nothing to the lord of Dyfed. 

208 LL. i. 558 ; ii. 790, 869; Welsh SS. 252 ; Owen, Pemb. i. 307. 


episcopal manor. 309 Penfro, the " land's end " of Dyfed, 210 a CHAP, 
long, low peninsula skirting the southern edge of Milford 
Haven, is in better case ; the early history of this cantref has 
not been so completely forgotten. It had in Rhoscrowther, 
formerly Llandegeman, one of the bishop-houses of Dyfed ; 211 
even better known were Penalun (now Penally), one of the 
leading churches which honoured St. Teilo, and the ancient 
monastery of Ynys Byr, or Caldy Island, with its ogam inscrip- 
tion linking the settlement with the beginnings of monastic 
enthusiasm. Other famous sites, of which the associations 
were secular, were Arberth (our Narberth) in the north, the 
chief seat of Pwyll and his son Pryderi and the scene of most 
of their surprising adventures, Llonion, now absorbed in the 
bustling activities of Pembroke Dock, which was famous 
throughout Wales for its crops of barley, 212 and Dinbych y Pys- 
god (of the Fish), which was thus distinguished from its name- . 
sake in Rhufoniog, and was a stronghold of the lords of Dyfed 
ages before the English learned to call it Tenby. 213 Though 
Penfro was among the first of the seven cantrefs to fall under 
foreign rule and to lose its Welsh features and traditions, its past 
had been too brilliant to be altogether forgotten or to be obliter- 
ated by English "ton" and Norman "castle". 

Eastern Dyfed was almost entirely included in the great 
Cantref Gwarthaf (the Uppermost Cantref), which extended 
from Narberth to Carmarthen and was made up of no less 
than eight commotes. Efelffre lay east of Narberth, Peuliniog 
the land of Peulin or Paulinus to the north, between the Taf 
and the eastern Cleddau ; 2U Laugharne and Henllan Amgoed 

*<Blk. Bk. of St. David's, 137-67. 

310 Pembroke, like Builth, Brecknock, Cardigan and Kidwelly, is properly 
the name of a district, usurped in course of time by its principal castle. The 
cantref of Penfro is probably represented by the deanery of Pembroke or Castle- 
martin, for which see Owen, Pemb. i. 304-9. 

211 LL. i. 558 ; Fenton (2), 218-9. 

*"" Amhynny y diharhebir o heid llonyon" (Pen. MS. 12 = Heng. MS. 
202, in Cymr. vii. 132). The Red Bk. has " llouyon " (Mab. 307), and in the 
first series of triads (No. 30) the name has become " llonwen ". There is a 
reference to " llys Llonion " in " Cywrysedd Gwynedd a Deheubarth " (Myv. 
Arch. I. 73 (62)) and to its maer in LL. ii. 879, line 4 (amitei and llonion should 
be separated) ; cf. also Lib. Land. 124, 255 ; LL. ii. 306 ; Lit. Eng. p. 91. 

413 The maer of " Dyubyt " (read " Dynbyc ") appears in LL. ii. 879 (cf. 
306). Gir. Camb. (Wks. iii. 353) speaks of the abundance of fish in the harbour. 

214 Owen, Pemb. i. 388-9. 


CHAP, indicate the position of Talacharn and Amgoed ; Ystlwyf or 
Ysterlwyf centred in Llanddowror ; 215 Penrhyn, as its name 
shows, occupied the peninsula between the Taf and the 
Tywi ; 216 Derllys came north of this, 217 and, last of all, Elfed, 
stretching along the western bank of the Gwili, was the border 
commote on the side of Ystrad Tywi. In this wide region 
there were many places which had importance in early times, 
but it suffices to mention Llan Deulyddog, one of the seven 
bishop-houses, built within the shelter of the ruined walls of the 
Roman Maridunum, 218 Meidrum, one of the greater Dewi 
churches, the " high-roofed fane, within the churchyard walls of 
which a host might meet," of the White Bard of Brycheiniog, 219 
and Y Ty Gwyn ar Daf (The White House on the Taf), a 
royal residence of the princes of Dyfed which was afterwards 
made for ever famous by Hywel Dda. The seventh cantref of 
Dyfed was Emlyn, lying opposite to Ceredigion on the south 
bank of the Teifi. It was divided by the river Cuch into the 
commotes of Uwch Cuch and Is Cuch, which have ultimately 
fallen into different counties, so that the Cuch forms the 
boundary here between Pembrokeshire and Carmarthenshire. 
The most notable ecclesiastical site of Emlyn was Cenarth on 
the Teifi, the church of St. Llawddog, where there was a re- 
markable salmon-leap, said to be the work of the holy man 
himself. 220 Whether any earlier fortress preceded the Norman 
castle of Cilgerran there is no evidence to show. 

The three cantrefs annexed by Seisyll to the realm of 
Ceredigion were known collectively as Ystrad Tywi, or Strath 
Towy, which was a convenient geographical description, but 
brings out unmistakably the fact that the district had no 
ancient or traditional bond of union. A number of commotes 
to the north and west of the Tywi were grouped together 
under the name of Y Cantref Mawr (The Great Cantref) ; others 

215 Owen, Pemb. i. pp. 206, 213. 216 Ibid. p. 206. 

217 The place now known as Derllys is, curiously, not within the modern 
hundred of Derllys at all, but in the parish of Merthyr and, therefore, in the 
hundred of Elfed. 

218 " Lann toulidauc ig cair mirdin " (Lib. Land. 62 ; cf. 124, 254). 

219 < A dewi bieu bangeibyr yssyt 

Meidrym le ae mynwent y luossyt" (Myv. Arch. I. 271 (194)). 
For the episcopal manor there see Blk. Bk. of St. David's, 234-41. 

220 Gir. Camb. vi. 114 (Itin. ii. 3). 


between the Tywi and Brycheiniog formed Y Cantref Bychan CHAP. 
(The Little Cantref), while the third cantref (Cantref Eginog) VIIL 
was made up of the three commotes which lay between the 
Tywi and the Tawe. The arrangement was artificial, being the 
result of a reconstruction which followed upon the ruin of the 
older and wider Dyfed. The Great Cantref was a broad upland 
region, bounded by the Tywi, the Teifi and the Gwili ; there 
was good land along the margin of these rivers, but most of 
the soil was covered with a dense scrub, which afforded excel- 
lent cover for fugitives or for irregular troops, and thus made 
Y Cantref Mawr a stronghold of Welsh independence in the 
troublous times which were experienced by South Wales after 
the coming of the Normans. 221 In area the cantref rivalled Y 
Cantref Gwarthaf in Dyfed, containing no less than seven com- 
motes, of which the situation may be roughly indicated as 
follows. Mabelfyw and Mabudryd, names which explain them- 
selves as of dynastic origin, were the commotes which lay 
along the Teifi, the former including Pencarreg, the later 
Pencader. Widigada, extending from the Gwili to the Cothi, 
marched with Dyfed and was near neighbour to Carmarthen. 
Catheiniog, the land of St. Cathen ap Cawrdaf, 222 came next 
in order in the eastward direction ; the church of Llangathen 
was within its limits. Maenor Deilo, or the Manor of St. 
Teilo, occupied the rich tract along the banks of the Tywi 
from Dinefwr to Llanwrda: 223 north of this was Mallaen, which 
ended in the moorlands of southern Ceredigion. The seventh 
commote was Caeo, which took in the upper valley of the 
Cothi and thus formed the heart of the cantref, the inmost 
retreat in a land where nature ever gave the sons of the soil 
her kindly, overshadowing protection. 

In the Cantref Mawr was situated, in the commote of 
Maenor Deilo, the " principal seat" of the realm of Deheubarth, 

321 " Kantaredum magnum, . . . copiosa silvarum condensitate australis 
Kambriae civibus tutissimum in necessitate refugium " (Gir. Camb. vi. 80 (Itin. 
i. 10)). 

a Welsh SS. 280. 

2as Part of this district, in the parishes of Llandeilo and Llansadwrn, lies 
in the modern hundred of Perfedd, which represents the two northern commotes 
of the Cantref Bychan. But the arrangement seems to be too complicated to be 
really ancient, not to speak of the express statement of Gir. Camb. (Wks. vi. 172 
(Descr. i. 5)) that the Tywi was the dividing line between the Great and the Little 


CHAP, the royal residence, the possession of which carried with it 
' the primacy among the rulers of South Wales. This was 
Dinefwr, 224 an ancient stronghold which stood on the summit 
of a little hill overhanging the north bank of the Tywi. Its 
early history, like that of Aberffraw and Mathrafal, is unre- 
corded, but there was, no doubt, good reason for the position of 
honour it held in mediaeval tradition. Not far off was the im- 
portant church of Llandeilo Fawr, a principal " clas " of that 
saint, which contested with Llandaff and Penally the distinction 
of being the resting-place of his body. 225 As a Teilo church it 
was claimed in the twelfth century by the see of Llandaff, 226 but 
the hold of the bishop of St. David's over the place, however 
acquired, was substantial and the name Tir Esgob (Bishop's 
land) still preserves the memory of the episcopal manor of 
Llandeilo. 227 St. David's had other possessions in the cantref, 
notably at Abergwili, which after the Reformation became 
the principal residence of the bishop. 

The Cantref Bychan only deserved its name by comparison 
with the Cantref Mawr, for it lined the southern bank of the 
Tywi for a distance of thirty miles, from Ystrad Ffin (The 
Border Strath) to a point not much above Carmarthen. It 
was, on the other hand, of no great breadth, forming a long 
strip of territory which was divided at two points so as to make 
three commotes. Of these Hirfryn was the northernmost, 
lying between the Tywi and Cantref Buellt ; Y Cymwd Perfedd, 
or the Middle Commote, came next, extending to the Black 
Mountains ; Is Cennen, so called because it was, for the most 
part, on the seaward side of the river Cennen, was the western 
limb, including within its bounds the upper waters of the 
Llwchwr. In the Cantref Bychan were Llanarthneu, a " clas " 
of St. David which, according to the saint's legend, had the 
custody of the mystic gifts bestowed before his birth upon his 
father, 228 Myddfai, a royal manor, where in later times there 

224 First mentioned in Lib. Land. 78 (" gueithtineuur," the "work" of 
Tinevwr). Cf. Gir. Camb. vi. 80-1 (Itin. i. 10), 172 (Descr. i. 5) ; LL. i. 346, 
ii. 831 ; Letters, Hen. III. i. 176 ; Cymr. xi. 45. 

225 A peaceful solution was provided, we are assured, by the miraculous ap- 
pearance of three bodies (Lib. Land. 116-7). 

226 Lib. Land. 56, 62, 77-8, 133, 254. 227 Blk. Bk. of St. David's, 262-75. 
228 Cambro-Br. SS. 117. For the grounds of the identification see chap. v. 

note 165. 


dwelt a remarkable race of hereditary physicians, 229 and that CHAP, 
eagle's nest of the lords of Is Cennen, perched on the top of a VIII- 
tall limestone bluff, the castle of Carreg Cennen. 280 . 

According to the lists, the third cantref of Ystrad Tywi 
was known as Cantref Eginog. But this was a name which 
gained no popular currency ; men knew the district only by the 
names of its three commotes, Cydweli, Carnwyllion and Gwyr. 
The occurrence of the form " Gwyr and Cydweli " 231 suggests 
that originally there were but two, and that Carnwyllion was 
carved out of Cydweli, which continued to be the name of the 
western half of this land between the Tywi and the Llwchwr. 
Thus the " land of Cadwal " 232 bordered on the former of these 
two rivers, dipping in gentle curves to its broad and sandy 
estuary, while the " land of Carnwyll " 233 lay along the other, 
rising in mountainous masses which confronted the eastern sun. 
The third commote of this cantref, which retains its ancient name 
in the slightly altered form of Gower, was the largest in all Wales ; 
it included not only the cliff-bound peninsula which stretches 
westward from Swansea, but also the wilder country to the 
north, between the Llwchwr and the Tawe, as far as the foot 
of the Black Mountains. Within it were the ancient fanes of 
Llangyfelach, one of the most important of the Dewi churches, 
which is still an episcopal manor, 234 Ystum Llwynarth, now 
known as Oystermouth, where Illtud lived in holy seclusion 
and the credulous in later ages pointed out an 'altar which hung 
in mid air, with no visible support, 235 and Llandeilo Ferwallt, 
which as a "clas" of Teilo was claimed by the bishops of 

219 For " Meddygon Myddfai " see the work published under that name by 
the Welsh MSS. Society in 1861 (Llandovery). 

230 The castle is first mentioned in 1248 (B.T. s.a.), but it would seem to be 
an ancient stronghold from the fact that from it is derived the name of the town- 
ship of Trecastell in which it stands. 

3i Guir (et) Cetgueli " (Hist. Britt. c. 14). Gwyr and Cydweli are frequently 
mentioned in Lib. Land., but there is only one reference to Carnwyllion (p. 247), 
and this in a thirteenth-century addition to the original MS. 

889 Owen, Pemb. i. 200. 

233 According to Cambro-Br. SS. 22, corrected as to the spelling in Cymr. 
xiii. 77, " Cornouguill " gave his name to " Cornoguatlaun ". 

234 Cambro-Br. SS. 123, 136; Blk. Bk. of St. David's, 284-9; App. Land 
Com. 445. The place is styled a " monasterium " in both the passages from 
Rhygyf arch's life. 

235 Hist. Britt. c. 71, where the form is "Loyngarth"; Cambro-Br. SS. 
177 ; Arch. Camb. IV. xi. (1880), 155. 


CHAP. Llandaff and thus acquired its modern name of Bishopston. 236 
VIII. ^ke | lomes O f t. ne ancient lords of Gower are not so easy to 
indicate ; at Castell Llwchwr (or Loughor) some stronghold may 
have formed a link between the Roman fort and the Norman 
castle, but there is ho good evidence of the fact. Swansea was 
in all probability but a stretch of sand, the haunt of sea-gull 
and plover. 

Between the broad-backed "Epynt range, the towering mass 
of the Black Mountains of Talgarth and the graceful peaks of 
the Brecknock Beacons lie the fertile valleys of the Usk and 
the Llynfi, here contracting in narrow woodland passes and 
there broadening out in luxuriant meadows. This was the 
land of Brycheiniog, which extended from Buellt and Eliael on 
the north to Gwent and Morgannwg on the south, and thus 
included the whole of our Brecknockshire save the first named 
of these four districts. From the earliest time of which Welsh 
tradition speaks, Brycheiniog was an independent realm ; its 
foundation was attributed, not to any scion of Cunedda or re- 
presentative of the Brythonic race, but to one Brychan, who 
was, on the father's side, at least, of Irish, i.e., Goidelic, origin 
and from whom the country acquired the name it has ever 
since borne. 237 Brychan is one of the most shadowy figures of 
Welsh legend ; twelve sons and twenty-four daughters, most 
of whom are said to have adopted the religious life, are assigned 
to him in the ancient lists, and the story tellers of Brecknock 
had much to tell of him and his deeds of prowess, of the hot 
temper which threw bold critics into jail and flung whatever 
was nearest to hand at the bearer of evil tidings, of his marvel- 
lous ride to meet the hosts of Deheubarth 238 with the fetter 

338 Lib. Land. Index, s.v. L. Mergualt. 

237 The earliest references to Brychan are in Gir. Camb. vi. 31-2 (Itin. i. 2) 
and De Nugis, 77-9, both especially valuable because of the familiarity of Giral- 
dus and Walter with Brecknock and its people. The tract " De Situ Brychen- 
iauc" in Cottonian MS. Vespasian A. xiv. (ff. 106-116), written about 1200 and 
printed in Cymr. xix. 24-7, contains the earliest account of Brychan's origin 
and descendants, and was probably copied from a still older source (Cymr. 
vii. 105-6; Owen, Catalogue, p. 22). For other accounts see Cymr. viii. 83-4 
(Jesus Coll. MS. 20), Breconsh. (2), Appendix No. vi. (Cognacio Brychan, 
from Cott. MS. Dom. i.), Myv. Arch. II. 29 (418-9), lolo MSS. in, 118-21, 
140, Welsh SS. 136-60. I follow the figures of Jesus Coll. MS. 20, supported, 
as far as the daughters are concerned, by Giraldus. 

238 The MS. of Map (Bodl. 851) has " regem . . . de Heulard " (De Nugis, 
77), which I take to be a copyist's error for "regem Deheubard". 


hanging to his horse's foot, and of the crushing defeat he CHAP, 
inflicted on the invaders, whose mutilated bodies furnished VIIL 
material for three great cairns heaped up on the field of 
slaughter. It was to Brychan, through his son Rhain Drem- 
rudd (the Red-eyed), that the later kings of Brycheiniog traced 
their descent, and, notwithstanding the large element of fable 
in the usual accounts of him, he may perhaps be regarded as 
the real founder of the dynasty. The pedigree has, however, 
come down in a corrupt form,* 39 and it is not possible to fix 
with any certainty the place in it of the two or three historical 
kings of the country. All that can be said is that in the 
seventh century one Awst (Augustus) seems to have been the 
ruling prince of Brycheiniog, 240 that somewhat later Tewdwr 
ap Rhain and Elwystl ab Awst divided .the sovereignty between 
them, until Elwystl was treacherously murdered by Tewdwr, 241 
and that in the age of Alfred the realm was possessed by 
Elise ap Tewdwr. 242 Brycheiniog retained its independence 
until the tenth century was well advanced. 

The region was divided into three cantrefs, 243 and the names 
of two of these, viz., Cantref Selyf (the cantref of Solomon) and 
Cantref Tewdos (the cantref of Theodosius), suggest that this 
took place in connection with the division of the territory at 
some period in its history between three coinheritors. 244 But 
the early conquest of Brycheiniog by the Normans had, as in 
Rhos and Penfro, the effect of obscuring, though it did not 
quite obliterate, the ancient local divisions, so that little can be 
said with confidence of the cantrefs of this district or of the 
commotes into which they were divided. 245 Previous to the 

439 It was for some reason or other not included in the collection in Harl. 
MS. 3859. In Jesus Coll. MS. 20, No. viii. (Cymr. viii. 85), there appears to be 
some confusion with the line of Dyfed. 

840 Lib. Land. 146, 154. 

841 Ibid. 167. There are no data to enable one to fix the age of Gwrfan. 
348 Asser, c. 80. 

343 "Tribus cantaredis distincta conseritur" (Gir. Camb. vi. 28 (Itin. i. 2)). 

244 An annotator of the list of cantrefs and commotes in Cott. MS. Domitian 
i. fo. 1246, says that this happened on the division of Brycheiniog between the 
three sons, Selyf, Tewdos and Einon by name, of Einon ap Gruffydd ab Elise. 
Einon had the cantref of Talgarth. 

943 A comparison of the lists will show that Dom. viii. and the Red Book 
give radically different accounts of the composition of the three cantrefs ; the 
former is followed by Pen. MS. 163 and (substantially) by Pen. MS. 50, but the 
account in the Red Book is to be preferred for several reasons in particular, it 


CHAP. Norman invasion, the chief royal residence of the country was 
vin * Talgarth, where Brychan himself is said to have held kingly 
state ; 246 commanding the vale of Llynfi, it guarded the ap- 
proach which was most likely to be chosen by an invading 
English host. One of the commotes of Brycheiniog bore the 
name of Llywel, so that it is not unreasonable to suppose that 
Trecastell, near Llywel Church, was another of the ancient strong- 
holds of the kingdom. A third was Ystrad Yw, which stood 
in the little valley of the Rhiangoll and gave its name to a dis- 
trict which formed the south-eastern corner of Brycheiniog. 247 
Of notable churches there was no lack in this quarter of Wales. 
The principal saint of the region was Cynog, son of Brychan, 
who suffered martyrdom ; Merthyr Cynog (the " martirium " of 
Cynog) was the place of his burial and held a position of cor- 
responding importance. 248 Another member of the same family, 
Cynidr, a grandson of Brychan, was also of high repute as a 
patron in Brycheiniog ; to him belonged Y Clas ar Wy (the 
"clas" on the Wye), which the English termed Glasbury. 249 
Llangors, one of the few churches which preserved the memory 
of Paulinus, was another ancient ecclesiastical centre ; it is ex- 
pressly termed a monastery, i.e., a " clas," and was held by the 
bishops of Llandaff, who paid occasional visits to the reedy 
banks of Llyn Syfaddon. 250 Nor was the sister see of St. David's 
unrepresented, for as the White Bard, himself a native of Bry- 
cheiniog, reminds us 

is not tainted with the error which divides Ystrad Yw into three cymwds, viz., 
Ystrad Yw, Crug Hywel and Eglwys lail. The latter two represent in reality 
Ystrad Yw Isaf and Ystrad Yw Uchaf, Eglwys lail being an old name of Llangy- 

246 Cambro-Br. SS. 23. Talgarth is, perhaps, the Garth Matrun of p. 272. 

Arch. Camb. VI. iii. (1903), 82-4. 

248 Welsh SS. 138-9. The torque of St. Cynog was famous in the twelfth 
century (Gir. Camb. vi. 25 (Itin. i. 2), 112 (ii. 2)) and the church of Merthyr was 
a hundred years later the best endowed in the deanery of Brecon (Tax. Nick. 


249 The connection of Cynidr (for whom see Welsh SS. 148-9) with Glasbury 
is clear, notwithstanding that his place as patron of the parish has long been 
taken by St. Peter. See Cambro-Br. SS. 274 (sancti Kenider de Glesbyri), Cart. 
Glouc. i. 314-6, iii. 5 (ecclesiam sancti Kenedri) and Arch. Camb. IV. xiv. (1883), 
227 (Kenedereschirch). There is still a Ffynnon Gynid(r) in the parish. Other 
churches originally dedicated to the same saint are Kenderchurch in Archenfield, 
which is the " lann cinitir lann i cruc " of Lib. Land. 277, Llangynidr and 

350 Lib. Land. 146, 2-38, 255. 


Garth Bryngi is Dewi's honourable hill, CHAP. 

And Trallwng Cynfyn above the meadows ; VIII. 

Llanfaes the lofty no breath of war shall touch it, 
No host shall disturb the churchmen of Llywel. 281 

It may not be amiss to recall the fact that these posses- 
sions of St. David's brought here in the twelfth century, to re- 
side at Llandduw as Archdeacon of Brecon, a scholar of Penfro 
who did much to preserve for future ages the traditions of his 
adopted country. Giraldus will not admit the claim of any region 
in Wales to rival his beloved Dyfed, but he is nevertheless 
hearty in his commendation of the sheltered vales, the teeming 
rivers and the well-stocked pastures of Brycheiniog. 262 


The well-sunned plains which, from the mouth of the Tawe 
to that of the Wye, skirt the northern shore of the Bristol 
Channel enjoy a mild and genial climate and have from the 
earliest times been the seat of important settlements. Roman 
civilisation gained a firm foothold in the district, as may be seen 
from its remains at Cardiff, Caerleon and Caerwent. Monastic 
centres of the first rank were established here, at Llanilltud, 
Llancarfan and Llandaff, during the age of early Christian en- 
thusiasm. Politically, too, the region stood apart from the rest 
of South Wales, in virtue, it may be, of the strength of the old 
Silurian traditions, and it maintained, through many vicissitudes, 
its . independence under its own princes until the eve of the 
Norman Conquest. It had its own bishop, seated at Llandaff, 
and never acknowledged the supremacy of David, whose sway 
was so mighty in the rest of Deheubarth. 

Until the middle of the seventh century the political history 
of the district is obscure. 253 Tradition spoke of a King Glywys, 
who ruled over the greater part of it, namely, that portion be- 
tween the Tawe and the Usk afterwards known as Glywysing, 
and whose sons, including Gwynllyw, the father of St. Cadog, 
divided their father's realm between them.' 254 But the dynasty 

K1 Myv. Arch. I. 271 (194). Wks. vi. 33, 36 (I tin. i. 2). 

188 Without accepting every document contained in the Liber Landavensis 
as authentic, one may use the evidence supplied by the compilation in drawing 
the broad outlines of the history of the period, and this I have endeavoured to do. 

454 Cambro-Br. SS. 22, 145. The antiquity of the form Glywysing is shown 
by its appearance in Hist. Britt. c. 41 (Gleguissing in the best texts), the chronicle 
VOL. I. 1 8 


CHAP was short-lived ; in a generation or two its place was taken by 
VIII. another, represented about>63O by one Meurig ap Tewdrig, who 
held not only Glywysing but also the region between the Usk 
and the Wye known from the ancient tribal centre of Venta or 
Caerwent as Gwent. The legend ran that Meurig's father Tew- 
drig had been mortally wounded in conflict with the English 
at the ford of Tintern on the Wye, 255 and it is most probable 
that this river now formed the boundary between the two races 
for a considerable distance from its mouth. In due course, 
Meurig was succeeded by his grandson, Morgan ab Athrwys, 286 
known as Morgan Mwynfawr or the Benefactor ; this prince 
was a contemporary of Rhain of Dyfed and Seisyll of Ceredigion 
(circa 730), so that it was pretty certainly from him the realm 
gained the name, in later ages so familiar, of Mofgannwg. 257 
Morgan was succeeded by his son Ithel ; in the next generation 
a division of the realm seems to have taken place between the 
sons of Ithel. Ffernfael ab Ithel, who died in 77 5, 258 was king 
of Gwent, where also his descendants bore rule, 259 until the line 
ended with his grandson, Ithel ab Athrwys, in 848. 260 The 
other sons, Rhys, Rhodri and Meurig, seem to have been kings 
of Glywysing, but the course of events west of the Usk in the 

in Harl. MS. 3859, s.a. 864 (Cymr. ix. 165), and Asser, c. 80 (Gleguising). It 
did not include Gwent, and, on the other hand, while it included Gwynllwg 
(Cambro-Br. SS. 95), it was not, as is alleged in lolo MSS. 18, a mere alias ot 
it. In Lib. Land, the term seems often to be used loosely as an equivalent of 
Morgannwg (see pp. 137, 156, where Gwent is included), but this may well be 
due to the ignorance of the compiler. 

285 Lib. Land. 141-2. 

266 Athrwys does not seem to have ruled himself, unless he was under king 
in Gwent (Lib. Land. 165-6). 

257 This view differs from that of Mr. Phillimore (Owen, Pemb. i. 208), who 
ascribes the origin of the name to Morgan the Aged (d. 974). It certainly does 
not occur in any good authority of older date ; on the other hand, the " seven can- 
trefs of Morgannwg " were not as a whole under the rule of the later Morgan, and 
it seems but natural to suppose the name came into existence at the same time as 
Rheinwg and Seisyllwg. Since the elder Morgan's grandson Ffernfael died in 
775, I do not think he can well be the " Morcant" of Harl. MS. 3859 s.a. 665 ; 
he belongs rather, with Rhain and Seisyll, to the beginning of the eighth century. 

258 Harl. MS. 3859 s.a. (Cymr. ix. 162). 

259 This may be inferred from the grants ascribed to Ffernfael and his sons 
in Lib. Land. 

280 ludhail rex guent a uiris broceniauc (Brycheiniog) occisus est " (Harl. 
MS. 3859, s.a. in Cymr. ix. 165). His pedigree is given in the same MS. (Cymr. 
ix. 181-2) as far back as Tewdrig and, stopping short with him, implies that he 
left no descendants. 


middle of the ninth century is involved in much obscurity, for, CHAP, 
when the political arrangements of the district of Morgannwg VIIIt 
are once more clearly revealed, about 870, Gwent is under the 
sway of a great-grandson of Ithel, one Meurig ab Arthfael ap 
Rhys, 2 " 1 while Glywysing has as its ruler one Hywel ap Rhys, 
of quite uncertain pedigree. 262 

Like Dyfed, Morgannwg was reputed a land of seven 
cantrefs. 2 " 3 Six of these were generally known ; they were 
Gorfynydd, Penychen, Y Cantref Breiniol ("The Privileged 
Cantref "), Gwynllwg, Gwent Iscoed and Gwent Uchcoed (" Be- 
low " and " Above the Wood "). As to the seventh there was 
less agreement ; the likeliest view is that it lay in our Hereford- 
shire, where the two regions of Erging (Archenfield) and Ewias 
remained thoroughly Welsh up to the time of the Norman Con- 
quest. The first five of these cantrefs bordered on the sea ; 
each had its tract of fertile land along the coast and behind this 
a wide extent of mountain or forest ; the " bro " or champaign 
country was thickly peopled, while the " blaenau " or mountain 
glens, now among the busiest seats of industry in the Empire, 
were left to the browsing cattle and the hunter's quarry. From 
the mouth of the Tawe to that of the Thaw (anciently the Nadd- 
awan) stretched the cantref of Gorfynydd ; 264 here were the 
ancient churches of Margam and Llanilltud Fawr, the former 
of unknown history, 265 the latter a wealthy foundation, main- 

at)1 For Meurig's pedigree see Harl. MS. 3859 in Cymr. ix. 182. His posi- 
tion is indicated in Lib. Land. 200, 226. 

888 The pedigree of Hywel ap Rhys is not to be found in Harl. MS. 3859, 
probably because Owain ap Hywel Dda was unwilling to recognise the rights of 
the family. In Jesus Coll. MS. 20 (Cymr. viii. 85, No. ix.) he is connected with 
Ithel ap Morgan, but the pedigree is a generation or two too long and its details 
are not attested by other authorities. 

16S Mob. 59. Cf. Cambro-Br. SS. 145 : " septem pagos rexit Gulat mor- 
gantie," though this is a wrong use of Gwlad Forgan. 

184 So termed in the Red Book of Hergest (Bruts, 412) and therefore to be 
explained, it may be, as the land beyond the mountain (from the point of view of 
Gower). Other forms found are Gorenydd (so practically in Cymr. ix. 331), Gor- 
wennydd (Triad iii. 14) and, most unsatisfactory of all, Gro Nedd. For the bound- 
ary between Gorfynydd and Penychen see Cambro-Br. SS. 53. The commotes 
of the two cantrefs, as usually given, appear to me to be subsequent in date to the 
Norman Conquest. They are the " members " of the lordship of Glamorgan, and 
do not include its main body, now known as the Vale or " Bro". Cf. Owen, 
Pcmb. i. 427. 

968 Margan (the m does not occur in any ancient authority) was originally the 
name of a district, probably a commote of Gorfynydd; see Geoff. Mon. ii. 15 : 

18 * 


CHAP, taining with dignity the traditions of the days of Illtud, and 
' both remarkable for the many examples they had to show of 
elaborate carving in the Celtic fashion. The wheel crosses of 
Margam and Llanilltud, adorned with intricate plait-work, form 
a group of monuments of great interest and bear witness to the 
existence in Gorfynydd in the ninth and tenth centuries of a 
school of carvers in stone of considerable technical facility. 260 
The inscriptions which many of them bear have not cast much 
light upon the history of the period ; probably, however, one 
may recognise the Hywel ap Rhys of Asser and the Liber Landa- 
vensis in the person who speaks in the following epitaph from 
Llanilltud : " In the name of God the Father and the Holy 
Spirit Houelt set up this cross for the soul of his father Res ". 26T 
Llanilltud may well have been a royal burying-place, for its 
abbot was one of the three great ecclesiastics of the diocese of 
Llandaff and the revenues of its " clas " were drawn from many 
a manor of Morgannwg. 268 Across the Thaw was the cantref 
of Penychen, 269 extending as far as the Taff ; this also contained 
two ecclesiastical centres of the first rank, the one the seat of 
the bishop of Morgannwg, the other at Nant Carfan (corruptly, 
Llancarfan), 270 the principal " clas " of St. Cadog and a match 
in affluence and historic dignity to the notifar distant Llanilltud 
Fawr. The abbot of Nant Carfan ruled over a community of 
thirty-six canons, who included a priest, a master or teacher, a 
sexton, and three custodians of sacred relics ; broad lands around 

" in pago Kambriae qui, post interfectionem Margani, eius nomine, videlicet Mar- 
gan, hucusque a pagensibus appellatus est ". Merthyr Mawr was within it (Lib. 
Land. 224). 

266 Lap. W. 8-15, 25-30; Margam Abb. chap. x. ; Arch. Camb. V. xvi. (1899), 
136-68 ; Allen, Celtic Art, p. 186. 

267 Rhys does not accept this identification (Arch. Camb. V. xvi. (1899), 
155), but this is because he is concerned to show that the stone may be of as 
early a date as the seventh century. As against this date see Allen, Celtic Art, 
p. 179. 

268 See Lib. Land, passim for "abbas sancti Ilduti" ("abbas Lannildut," 
145). The "abbas Carbani vallis" (or "Sancti Catoci") and the "abbas Doc- 
guinni " (of Llandough near Penarth Margam Abb. 3) appear no less frequently. 
There is no direct evidence as to the possessions of Llanilltud, but the statement 
of Cambro-Br. SS. 168 (habentes . . . singuli suam villam) rests, no doubt, upon 

269 The district appears (as Penn Ohen) in Wrmonoc's life of Paul Aurelian 
(Rev. Celt. v. p. 418), and was therefore a well-known area in the ninth 

270 See chap. vii. note 52. 


the settlement provided the means of its support 271 There CHAP, 
was nothing essentially different in the organisation of the VIII< 
cathedral church of Llandaff. The place of the abbot was 
taken, indeed, by the bishop, but in other respects the analogy 
was close. A " priest of Teilo " represented the later dean ; 
twenty-four canons (if not a greater number) formed the "clas" 
or " household " of the saint, 272 and if the territorial claims of 
Llandaff, as put forward by enterprising bishops, were more 
ambitious than those of Llancarfan, this was a circumstance due 
to the wider distribution of Teilo churches than of those which 
claimed the protection of Cadog. 

The Taff and the Rhymni, flowing south from the mountains 
of Brycheiniog in parallel valleys but a few miles apart, were 
the western and eastern boundaries of Cantref Breiniol. Wherein 
the privilege of this cantref stood is not stated by tradition ; it 
is possible, indeed, that the name is of later date than the 
Norman Conquest of Glamorgan, for there are some indications 
that the ancient title was Senghenydd, a term limited in later 
times to that part of the cantref which lay north of the ridge 
of Cefn On. 273 The district contained no important church or 
ecclesiastical manor and may, therefore, have always been, as 
it certainly was in later ages, the chief seat of civil power in 
Morgannwg. But there is no good evidence that the site of 
Cardiff Castle was occupied by any royal residence in the 
interval between the ruin of the Roman fort and the choice of 

271 Cambro-Br. SS. 82-96. Both the " presbiter " or "sacerdos" and the 
" magister " or " doctor " of St. Cadog find a place in Lib. Land. (258, 268, 272, 

273, 274). 

97S The " sacerdos teiliav " and the " familia teliaui " are mentioned in the 
marginalia of the Book of St. Chad (Lib. Land, xliii. xlvi.) ; for the former see 
also Lib. Land. 247, 258, 264, 273. Bishop Urban gives the number of canons 
(Lib. Land. 88). 

273 The " Tref Eliau in Seghenid " of Lib. Land. 255-7 has been supposed 
to be Roath (ibid. 382) and was certainly on the coast. Thus the term at this 
date included Cibwyr or Kibor, as well as the Senghenydd of later times, for the 
limits of which see Owen, Pemb. i. 258, and Arch. Camb. IV. viii. (1877), 264-9. 
Before its division into the commotes of Uwch Caeach and Is Caeach, Senghenydd 
was itself a commote, which is what Giraldus means when he calls it the fourth 
part of a cantref (vi. 170 (Descr. i. 4) ; cf. 34 (I tin. i. 2) " kemmoti, id est, 
quartae partis cantaredi "). It may be added that the way in which he seems in 
this passage to leave Cibwyr out of account in his analysis of the constitution 
of the diocese of Llandaff suggests that the authority he followed used Seng- 
henydd in the older and wider sense. The " Sein Henyd " of B.T. (Bruts, 353, 
355. 359. 360, 363) is a place in Gower. 


CHAP, the spot by Robert Fitz Hamon as the head of his newly won 

VIII - * 

lordship. 274 The remaining cantref of the old realm of King 
Glywys lay between the Rhymni and the Usk ; it took from 
his son Gwynllyw the name of Gwynllwg which, in the form 
Wentloog, has survived to the present day. 275 On a height 
near the mouth of the Usk, looking out over the marshy flats 
of the Severn estuary, was the church of Gwynllyw (now St. 
Woollo's), served by a " clas," ruled by an officer, who, though 
he once may have been an abbot, was in more recent times a 
dean, and consecrated in the affections of the cantref by many 
a tale of miraculous help rendered to the men of Gwynllwg 
in their days of sore need and tribulation. 276 Another church of 
high standing in the cantref was Bassaleg, a " basilica " 277 of 
which the ancient traditions were submerged by the devouring 
tide of the Norman Conquest, but which is known to have been 
the mother church of most of the land between the Rhymni 
and the Ebbw. 278 

The two cantrefs of the kingdom of Gwent occupied the 
region enclosed by the Usk, the Wye, the Monnow and the sea. 
A great forest, of which a large portion still remains under the 
name of Wentwood, divided the low-lying tract along the Bristol 
Channel from the northern uplands and thus parted the realm 
into Gwent Iscoed and Gwent Uchcoed. 279 Of these two 
divisions the seaboard one, though much the smaller, was the 
more important. It was famed for its fertility; the renown 
of the wheat and the bees of Maes Gwenith, on the banks of 
the Troggy, passed for a proverb throughout the whole of 

274 So Mr. J. S. Corbett in the Transactions of the Cardiff Naturalists 1 
Society, vol. xxxiii. (1900-1), pp. 26-7. Mounds of the type of that on which the 
keep is erected are no longer regarded as pre-Norman. 

275 See chap. v. note 168. 

276 Cambro-Br. SS. 145-57 (Vita S. Gundleii). The " decanus ecclesie " 
is mentioned on p. 156. 

277 The derivation is suggested by Prof. Hugh Williams (Giidas, p. 29). 
Notwithstanding the occurrence in mediaeval Welsh literature of the form 
Maesaleg, this cannot be the Campus Elleti of Hist. Britt. c. 41, which is 
rather to be looked for in the neighbourhood of the Palus Elleti of Lib. Land. 
148, i.e., near the river Thaw in S. Glamorgan. 

278 In addition to the chapels of Henllys and Risca assigned to it by Rees 
(Welsh SS. 342), Machen, Bedwas, Mynydd Islwyn, and Coed Cernyw were 
regarded as chapels of Bassaleg about noo (Cartae Glam. i. 2). 

279 " Coit guent" is mentioned in Lib. Land. 262 (bounds of " Hennriu in 
Lebinid "). 


Wales. 880 Here the princes of Gwent held court, 281 and here CHAP. 
was Caerwent, from which the region took its name, once the 
Silurian capital if we may dignify it with such a name but 
since the days of St. Tathan the ecclesiastical and not the civil 
centre. - s - The saint was reputed to have been a famous 
teacher, the head of a " studium " or college of the same mon- 
astic type as that of Illtud ; he founded a " clas " of the first 
rank, of which an abbot had the direction. All the men of 
Gwent revered him as the father of their land, its guardian and 
the avenger of its wrongs. In Upper Gwent there were no 
sites of like importance. It was a thriving land, dotted over 
with churches, but its traditions were matters of local interest, 
which had not caught the fancy of Wales at large. The bishop 
of Llandaflf, it should be added, drew no small part of his in- 
come from the prosperous plains of Gwent ; important manors 
at Llangadwaladr (now Bishton),>Merthyr Tewdrigor Mathern, 
Llaneuddogwy or Llandogo, and in the valley of the Trothy 
sent their produce to maintain the state of the great monastery 
on the Taff. 283 

Two outlying members of the older Morgannwg remain to be 
noticed, namely, Ewias and Erging. 284 The former lay between 
Brycheiniog and the valley of the Dore a land of long and 
narrow mountain glens, of which the streams run southward 
side by side to the Monnow. Most of it is now included in 
the county of Hereford, which has also absorbed the richer 

280 Triad i. 30 = ii. 56 = iii. 101. The place is a little north of Llanfair 
Discoed (Owen, Pemb. i. 236). 

281 The royal court of " Lisarcors " was somewhere in lower Gwent 
(Cambro-Br. SS. 156), and so too the " palacium " to which King Caradog ab 
Ynyr moved when he resigned Caerwent to St. Tathan, for it was between that 
city and the Severn, perhaps at Caldicot (Cambro-Br. SS. 259). 

282 The life of St. Tathan, a saint's day homily (" cuius hodiernam festivi- 
tatem celebramus," p. 263) composed by a Norman writer (observe the use of 
" indigene " on p. 264), is to be found in Cambro-Br. SS. 255-64. Caerwent 
Church is now dedicated to St. Stephen, but evidence is not lacking to show that 
Tathan was the original patron see Ann. Theokesb. s.a. 1235 (? 9*>) ', lolo 
MSS. 114, 132, 151. Hence the "presbiteri tathiu " of Lib. Land. 270 are 
clergy of Caerwent ; the " abbas guentonie urbis " appears on p. 222, and on pp. 
243, 245 a "lector urbis guenti ". 

888 Lib. Land. 180-3, 141-3, 156, etc. ; Tax. Nick. 280 ; App. Land Com. 

884 The translator of " Brut y Brenhinoedd " perversely renders the 
" Wissei " and " Gewissei " of Geoff. Mon. as " Ergig ac Euas," the latter, it 
may be remarked, a late Welsh form (Bruts, 109, 127, 252). 


CHAP, district of Erging, known to the English by the name of Arch- 
' enfield. Erging was bounded by the Wye, the Worm and the 
Monnow ; though so close to the gates of Hereford, it was a 
stronghold of Welsh customs and ideas as late as the end of the 
twelfth century. The Welsh saints were honoured throughout 
the district, and among them St. David had a great church at 
Much Dewchurch, 285 and Dyfrig, who was (if we may believe 
his legend) by birth and residence a man of Erging, a group of 
churches which commanded the allegiance of the dwellers 
along the winding banks of the Wye. 286 


Four lists of the cantrefs and commotes of Wales are to be found in MS., 
representing the work of four editors or compilers. The oldest is probably that 
contained in Cottonian MS. Domitian viii. ff. 119-206 (printed, not very ac- 
curately, in Leland's Itinerary, ed. 1769, v. 16-20), for, though the writing is said 
to be of the fifteenth century, the forms of the names imply an original of the 
twelfth or thirteenth (Cymr. xi. 168). Next comes the list in the Red Book of 
Hergest, cols. 377-80, written about 1400 and printed, first in the Myvyrian 
Archaiology, II. 606-12 (737-40), where it is printed onjthe lower half of the 
page, and more recently by Rhys and Evans in Bruts, 407-12. A third list 
occurs in Hengwrt MS. 34 = Peniarth MS. 50 (Y Cwta Cyfarwydd), pp. 133-8, 
written about 1450 and printed in Cymr. ix. 326-31. The fourth is in Hengwrt 
MS. 352 = Peniarth MS. 163, pp. 57-60, and was transcribed by Gruffydd 
Hiraethog in the year 1543 ; it will be found in full in Evans, Rep. i. p. 952-54. 
The upper list in the Myvyrian Archaiology (II. 606-13 (735-7)) is substantially 
that of Gruffydd Hiraethog. 

Not one of these lists can be implicitly trusted, though they go far to correct 
each other's errors. The Cottonian list is defective in the section Ceredigion and 
throughout is atrociously spelt, but in other respects it is fairly accurate. Its 
order is Y Berfeddwlad, Powys (including Arwystli), Gwynedd (including Pen- 
llyn), Rhwng Gwy a Hafren, Deheubarth, Morgannwg. The chief mistakes are 
the misplacement of Nant Conwy, the transposition of Uwch and Is Rhaeadr 
(this runs through all the lists save that of Gruffydd Hiraethog), the omission of 
Buellt and serious confusion in Gwynllwg and Gwent. The Red Book list fol- 
lows the order Y Berfeddwlad, Gwynedd (including Penllyn, Cyfeiliog, and 
other border districts), Powys (including Arwystli), Rhwng Gwy a Hafren, 
Deheubarth, Morgannwg. In the North Wales portion there are many errors, 
but the Dyfed and Ystrad Tywi sections are almost flawless. The older Peniarth 
list follows the same order as Dom. viii. but places Arwystli at the end of Gwy- 
nedd ; its chief defect is wrong bracketing, which extends to nearly every section. 
Morgannwg, where it was written, naturally shows the fewest blunders. Gruff- 
ydd Hiraethog's order is his own, viz., Gwynedd (including Arwystli and Penllyn), 
Powys, Rhwng Gwy a Hafren, Deheubarth, with Morgannwg and Gwent sand- 
wiched between Brycheiniog and Dyfed. This list professes to be based on a 

285 Rees, Welsh SS. 53. 286 Chap. v. note 114. 


survey (y messvrwyd ac i rranwyd ac i rivwyd) of all Wales made by Llywelyn ap CHAP. 
Gruffydd (ob. 1282), who, however, was never in a position, despite the extent of VIII. 
his rule between 1267 and 1277, to make any survey of the kind. It is full of in- 
accuracies, such as those which mark the treatment of Cantref Bychan and Can- 
tref Mawr, and names many commotes, such as Penal, Hafren and Trefdraeth, 
which had no real existence. Its unsupported evidence is of the very slightest 


Really ancient evidence on this head is most difficult to obtain, but there was 
so much continuity in the matter of local divisions, notwithstanding political 
changes, that late authorities may often be used with advantage. In many parts 
of Wales the rural deaneries corresponded closely to the civil areas, and assistance 
may therefore be derived from Tax. Nick. 272-94. The survey printed in Rec. 
Cam. 1-89 is arranged under commotes, so that for the counties of Anglesey and 
Carnarvon our information is pretty full. Modern hundreds and manors often 
preserve the ancient boundaries ; for a list of the former and their constituent 
elements, see App. Land Com. 362-76, and for particulars as to crown, episcopal 
and private manors, ibid. 437-75. Hengwrt MS. 99 = Peniarth MS. 147, 
written about 1566, contains (pp. 5-22) a list of parishes (printed in Evans, Rep. 
\. pp. 911-20), grouped to a large extent under the old territorial names, but the 
scheme is only partially carried out and not always quite correctly. Special 
sources of information for particular districts are indicated in the footnotes to this 


No ancient map of the cantrefs and commotes of Wales is known to me. 
That of William Owen (Pughe), published in 1788 in the third edition of War- 
rington's History of Wales, is very largely guess-work of a clumsy kind, and it 
is to be regretted that the authors of The Welsh People should have given it a 
place in their book. 

NOTE TO CHAPTER VIII. Rheinwg, Esyllwg and Fferyllwg. 

That names of districts might be formed in Welsh by the addition of -wg 
to personal names is clear from the well-established cases of Morgannwg, 
Gwynllwg (for Gwynllyw wg) and Seisyllwg. An instance is to be found in 
Hist. Britt. c. 70, where mention is made of a region styled "Cinlipiuc," of un- 
known situation, but certainly named after some Cunalipi (Arch. Camb. IV. 
xiii. (1882), 163-4) or Cynllib. The cases of Rheinwg, Esyllwg and Fferyllwg, 
nevertheless, present in one way or another no small difficulty. The clearest in- 
dication of the position of Rheinwg is to be found in the life of St. Padarn in 
Vesp. A. xiv., in which a tripartite division of South Wales between Padarn, 
Teilo and Dewi is said to have been made ; Padarn took Seisyllwg, Teilo Mor- 
gannwg " regnum autem Rein hec predicta iura ab episcopatu Sancti David 
accepit " (Cambro-Br. SS. 196-7). St. David's domain can have been none other 
than Dyfed, and as the pedigrees show that Rhain (in Old Welsh, Regin) ap 
Cadwgan, Seisyll ap Clydog, and Morgan ab Athrwys ruled over Dyfed, Cere- 
digion and Glywysing respectively about the beginning of the eighth century, the 
three names appear to fit easily, on this explanation, into their places. They 
are also found in conjunction in Pen. MS. 32 (MS. D. of the Welsh Laws), where 
they appear to be intended to explain what was meant by Deheubarth (LL. it. 
50). Rheinwg is, therefore, taken to be Dyfed by Basil Jones (Vestiges of the 
Gael, pp. 61-2) and Phillimore (Cymr. xi. 141). Some other passages which might 
be cited do not so easily lend themselves to this conclusion ; for instance, the 


CHAP. "Vastatio Rienuch ab Offa " of Ann. Camb. MS. C. s.a. 795 and the allusions in 
VIII. Cambro-Br. SS.jj, 79, to attacks on Glamorgan by kings of " Reinuc ". Aneurin 
Owen was probably led by these references to suppose a connection with Rhain 
Dremrudd, son of Brychan, and hence his gloss to Reinwg in LL. ii. 50, note b, 
" a district in Brecknockshire ". For this view there is something to be said ; the 
notion that " Ereinwg " was the Welsh name for Southern Herefordshire has, on 
the other hand, nothing to support it. It was first put forward by Humphrey 
Llwyd (Comment. (2), 94) and popularised by Camden (Britannia, 550) ; in all 
likelihood it owes its origin to the Offa passage quoted above. 

Esyllwg, there can be little doubt, is an antiquaries' form, having no genuine 
root in history. Welshmen did not give to their territories the names of wo- 
men ; moreover, the only Esyllt who appears in Welsh records is the famous 
Iseult of romance. Those who used the name claimed, in fact, a different origin 
for it, which, in the light of modern philology, has only to be stated to be 
promptly dismissed ; they regarded it as the Welsh equivalent of Siluria ! Its 
real source, as can easily be shown, was a misunderstanding of Seisyllwg. This 
name having become obsolete and its application forgotten, the passage in the 
Laws (already mentioned) in which it occurs became corrupt (cf. LL. ii. 50, 
584, Comment. (2), 169, and lolo MSS. 74). Humphrey Llwyd found the form 
" Syllwc " in some MS. and forthwith leapt to the conclusion that the region 
meant was that of the Silures (Comment. (2), 102). The view gained acceptance 
and " Esyllwg " (the reading of some copies) found its way as a supposed ancient 
name for Morgannwg into the third series of Triads (Nos. 14, 16, 37) and the 
notices printed in the lolo MSS. (86). Camb. Reg. ii. 8 contains a tremendous 
list of alleged variants of Esyllwg (land) and Esyllwyr (people) ; like other lists in 
the same article, it is the coinage of the ingenious and original contributor. 

Fferyllwg is another form open to the gravest suspicion. Fferyll or Fferyllt 
is the Welsh mediaeval name of the poet Vergil, and, owing to the bard's reputa- 
tion in the Middle Ages as a necromancer, became a common noun, denoting an 
alchemist (whence the modern " fferylliaeth," chemistry) or worker in metal 
(Mots Latins, pp. 167-8 ; W. Ph. (2), 205). But no character in Welsh history 
bears the name, and there is no early instance of the use of Fferyllwg to denote 
" Rhwng Gwy a Hafren " (lolo MSS. 86). I believe the origin of the form is to 
be found in the old name of Hereford which appears as Fernleg (Camden, 
Britannia, 553), Ferleg (Comment. (2), 94) and Fferleia (Radnorsh. (2), 108). 
This became Ferlex (Camb. Reg. i. 57 ; Breconsh. (2), p. 36) ; Fferregs (Bre- 
consh. (2), p. 38), and, when written by Welshmen, Fferyllwg (Gw. Brut. s.a. 838). 



(As indicated in the text, this chapter is primarily based upon the evidence 
afforded by the Welsh Laws, though much valuable matter for purposes of 
illustration is to be gleaned elsewhere. For an account of the MSS., editions 
and history of the Law of Hywel see chap, x., appended note. The following 
are the chief modern works which deal with the subject of the chapter ; Das Alte 
Wales, von F. Walter (Bonn, 1859) ; Hubert Lewis's Ancient Laws of Wales 
(1889) ; Seebohm's Tribal System in Wales (1895) ; Rhys and^rynmor Jones's 
Welsh People (chap, vi.) ; VinogradofTs Growth of the Manor, book i. (1905); 
Wade- Evans's Welsh Medieval Law (1909). In the notes references to the laws 
are only given in support of statements not easily verified ; the note appended 
to the chapter explains why no use has been made of the Triads of Dyfnwal 


FOUR leading institutions supplied the framework of the civil CHAP. 
organisation of Wales in the early mediaeval period. They IX - 
were the cenedl, the tref, the cantref and the brenin, or, to use 
terms of more general application, the kindred, the hamlet, the 
tribe and the chief. The first was the basis of society, for the 
Welsh were, and long continued to be, in that stage in which 
the tie of kinship is paramount, overshadowing all other rela- 
tions. " They are above all things," wrote a keen observer in 
the twelfth century, "devoted to their clan, and will fiercely 
avenge any injury or dishonour suffered by those of their 
blood." l The second was the economic unit, the area of co- 
operation for the production of food, both by tillage of the soil 
and otherwise. The third was the political and judicial unit, 
the district within which men acted together for peace and 
war, for the trial of causes, both criminal and civil, and for 
the maintenance of the chieftain and his court. In the fourth 
appears the monarchical element, binding together the com- 
munity under one authority a costly burden from the economic 

1 Gir. Camb. vi. 200 (Descr. i. 17). 


CHAP, point of view, but able to offer in return, not only guarantees for 
the preservation of order within the state, but also what was 
no less prized satisfaction to the spirit of tribal pride and 
security from the inroads of detested rivals. 

Our knowledge of these institutions in their detailed work- 
ing comes chiefly from the documents known as the Welsh 
Laws, which are to be found in various widely differing editions 
of the twelfth and succeeding centuries, but are ultimately to 
be traced to a code prepared, as tradition avers, under the 
direction of Hywel the Good in the early part of the tenth. 
What precisely was done by Hywel and his advisers will be 
discussed in a later chapter ; here it is sufficient to say that, 
while the code is our chief source of information for these 
institutions, it did not bring them into existence. The cenedl, 
closely bound up as it is with the blood- feud, is incontestably 
of ancient origin, linking the Welsh of the Middle Ages with 
their prehistoric ancestors. The tref appears, under that name, 
in the ninth century records entered in the Book of St. Chad, 
and is already an organisation for the regular supply of food. 2 
Gildas bears ample witness to the fact that in the sixth century 
his countrymen were ruled by kings, whose power rested on 
the possession of military force, and who tried, imprisoned and 
punished criminals as part of the daily business of kingship. 3 
As to the cantref, it may be in name of more recent upspring- 
ing than the other three, but the tribe or " gwlad " it repre- 
sented undoubtedly had its roots far back in the history of the 
Welsh people. 

The cenedl was the kindred or clan, extending far beyond 
the household or family, but not to be confounded, on the 
other hand, with the larger community formed by the people 
or tribe. It corresponded to the Latin "gens," the Greek 
761/05, the Teutonic " sib," and the Irish " fine ". It was the 
body of kinsmen descended from a common known ancestor 
who, recognising their relationship, acted in concert in all 
family matters, such as giving in marriage, acknowledging 
sons, and, above all, waging the family feuds and ending them 

2 " Treb guidauc . . . hie est census eius" (Lib. Land. xlv.). 

3 See especially c. 27("reges habet Britannia . . . iudices . . . belliger- 
antes . . . fures insectantes ... in sede arbitraturi sedentes . . . vinctos in 
carceribus habentes "). In c. 109 there is a reference to " catastam poenalem ". 


by the payment and receipt of compensation. The unity of CHAP. 
this body was maintained by the agnatic principle ; the kin- 
ship, that is to say, which bound it together was reckoned ex- 
clusively through males, so that a man could only belong to 
one cenedl, which did not include his wife, his mother, or any 
maternal relative. In this respect the Welsh appear to have 
followed a rule which was widely current among the Aryan 
races. 4 Unlike the "gens," but like the "fine," the cenedl was 
further limited by being confined to kindred within a certain 
degree of relationship. The fifth cousin (in Gwynedd the 
sixth) was, at least for matters of the first importance, the out- 
side man ; " beyond this degree," the lawyers alleged, " there 
can be no computation of kinship". 5 Within this limit, on the 
other hand, every man knew his pedigree accurately ; to quote 
Giraldus once more, " the most ordinary folk among this people 
keep careful count of the family pedigree ". c 

The cenedl was so far organised as to have regular officers, 
with privileges substantial enough to make it worth while to 
pay the king a fee for admission to them. 7 Of these officers 
the chief was the pencenedl, the " caput " or " magister gentis," 
whose duty is described in general terms as " to act with his 
kinsman 'in every need that might beset him ". 8 He was the 
guardian, the champion against oppression, the corrector, too, 
if there should be occasion, 9 of the men and women of his clan. 
The office was not hereditary, in the sense of descending 
directly from father to son, 10 and so differed from the kingship 
of the tribe ; rather it was elective, the heads of households 
of the cenedl choosing one of their number, when a vacancy 
occurred, to hold it for life. 11 Among a people who had no 

4 Schrader, Prehistoric Antiquities of the Aryan Peoples, tr. F. B. Jevons 
(London, 1890), pp. 397-9. 

6 LL. i. 226 (Ven. MS. B.). The difference in practice between Gwynedd 
and the rest of Wales explains the discrepancies which caused perplexity to 
Hubert Lewis (Ancient Laws, pp. 83-4) and Seebohm (Trib. System, pp. 78-81). 

8 Gir. Camb. ut supra. 

7 Dim. II. xxiii. 54, 55 = Lat. A. II. xvii. 5,3 = Lat. B. II. xxxix. 3, i. 

8 Ven. II. xix. 3. For the expression "ymyrru ag un yn rhaid " see Mab. 


9 Dim. II. viii. 20. 

10 Gw. II. xl. 10, where it is added : " oes uodawc yw pen kynedlaeth ". 

11 Details are wanting as to the method of election, but several points are 
clear. The office was held for life (see previous note) ; it did not pass automati- 
cally (as suggested in LL. ii. 516 " hynav o wr cyvallwy yn y genedl "), for the 


CHAP, nobility, as distinguished from the free tribesmen,) the pencenedl 
IXt naturally ranked high in the social scale, and, judged by the 
worth of his life and honour in the tribal tariff, was only in- 
ferior in dignity to the immediate kin of the crown. 

It was one of the specific duties of the pencenedl to stand at 
the head of his clan when it was necessary for the body to take 
action in respect of the receiving or denying of a son. The 
institution of marriage was, as one might expect in a Christian 
land, recognised and in honour among the Welsh, and in- 
volved important legal rights and obligations. But it was not 
a rule of Welsh law, at any rate in Gwynedd, that only sons 
born in wedlock should be deemed legitimate, should be 
members of their father's family and inherit his possessions. 12 
The evidence of the legal texts, both direct and indirect, leaves 
no doubt that foreign censors had in the main good grounds 
for asserting that in Wales no distinction was made between 
lawful and natural children. 13 Not only had the teaching of 
the Church been without effect in this particular, but, what was 
still more remarkable in a community so tenacious of ancient 
customs, there had been a departure, too, from the Aryan ideal, 
which required in the freeborn son pure blood on the mother's 
no less than on the father's side. 14 In Wales it is clear that, 
when once the fact of paternity had been established by the 
proper legal procedure the oath of affiliation not rebutted by 
legal denial the son passed at once into his father's cenedl 
and could not henceforth be shut out from any of its privileges. 
It was the father who ordinarily denied or received, but, as 
the law allowed affiliation many years after the birth of the child, 

law finds it necessary to provide for a possible vacancy (Ven. II. xxxi. 19 = Dim. 
II. viii. 30 = Gw. II. xxxix. 41 = Lat. A. II. viii. 19 = Lat. B. II. xliv. 6) ; it was 
held by an " uchelwr " or head of household (Ven. II. xviii. 8). 

12 Ven. II. xvi. 2 is quite explicit as to the opposition on this point between 
" Cyfraith Hywel " and " Cyfraith Eglvvys ". Some of the other texts draw a 
distinction between lawful and unlawful sons (e.g., Dim. II. viii. 27), but the 
rules as to affiliation everywhere observed leave it hardly doubtful what the real 
practice was. 

is > Paternam hereditatem filii inter se, tam naturales quam legitimi, herili 
portione dividere contendunt" (Gir. Camb. vi. 225. (Descr. ii. 9)). Cf. H. and 
St. i. 514 (gravamen No. 18 of church of St. Asaph against Llywelyn ap 
Gruffydd) and the provision of the Statute of Rhuddlan (LL. ii. 925) : " quod 
bastardi de cetero non habeant hereditates et eciam quod non habeant propartes 
cum legitimis nee sine legitimis ". 

14 Schrader, op. cit. pp. 391-2. 


some provision was necessary in the event of his being dead. CHAP. 
Then it was that the pencenedl, with seven men of the clan, 
took the place of his deceased kinsman and either " swore the 
lad away from the cenedl " 15 or formally received him into it. 
In Gwynedd there was a ceremony of reception, no doubt of 
ancient origin. The pencenedl took the hands of the youth 
within his own and kissed him, " for a kiss is the token of kin- 
ship ". He then placed the right hand of the youth in that of 
the eldest of the assembled kinsmen, who also kissed him in 
token of the new relationship. Thus the boy passed from 
hand to hand until the last of the group was reached. 16 

The admission of a son to a cenedl was not only a matter 
of importance for the near kin, whose rights of inheritance 
were affected ; it touched every member of the clan as adding 
another name to the long list of those who might claim his 
help in the matter of the payment of galanas. The vendetta, 
or blood-feud, was still an institution of undoubted vigour and 
vitality in mediaeval Wales. So keen was the clan feeling of 
solidarity, so strong the bonds which united the cenedl, that no 
man could injure or insult, still less pursue to the death, the 
humblest of its members without drawing down upon himself 
the unanimous hostility and vengeance of the whole kin. Time 
could not extinguish the mortal quarrel ; " they are ready," says 
Giraldus, "to avenge, not only new and recent injuries, but 
also ancient and bygone ones, as though but lately received ". 17 
At an early stage, however, the idea of compensation emerged 
as the corrective of the evils of a perpetual state of private war ; 
the feud, if there was no other way of ending it, might be 
bought off. To this problem, then, the law addressed itself; in 
a case of "galanas" 18 or "enmity" arising out of a violent 
death and how the killing cameabout, whether by accident or 

15 " Gwadu mab o genedyl " (Ven. II. xxxi. 7 ; Dim. II. viii. 30). The oath, 
as Hubert Lewis says (Ancient Laws, pp. 15-16), had strict reference to the fact 
and was not a mere refusal to receive. 

18 Ven. II. xxxi. 25. 

17 Gir. Camb. vi. 200 (Descr. i. 17). There was a proverb which ran 
" Hir y byddchwerw hen alanas " (Evans, Proverbs, p. 560). 

18 From gal, gelyn, a foe. Galanas came to mean not only the effect of the 
feud, the wergild, or blood money, but also its cause, the deed of slaughter. 
Incidental evidence of the antiquity of the term is furnished by its survival, in 
the form " galnes," " galnys " ( = satisfaction for slaughter), among the Strath- 
clyde Britons (Acts of the Parliament of Scotland, vol. i. (1845), pp. 299-300). 


CHAP, in self-defence or in cold blood, did not in the least affect the 


situation 19 under what conditions might the honour of the 
one clan be appeased and the safety of the other secured. In 
the first place, provision was made for a formal declaration ot 
war ; the charge of manslaughter must be publicly made and 
an interval left for the legal reply. This might take the form 
of a denial, supported by the oaths of a body of compurgators, 
or, failing this, there was an undertaking to pay the " galanas," 
wergild, or murder fine, the amount of which was fixed, in a 
scale duly graduated according to rank, for every member 
of the community. It is unnecessary to enter here into the 
elaborate arrangements prescribed by the laws for the collection 
and division of the galanas fund ; for the present purpose it is 
enough to say that every member of the cenedl of the 
" llofrudd " (red hand) had to contribute a share fixed in 
amount by the degree of his relationship, and every member 
of the cenedl of the slain man received a share calculated upon 
the same principle. The matter was further complicated by 
the inclusion for the purposes of galanas of kin on the female 
side, a practice which was not strictly consistent with the idea 
of the cenedl as a purely agnatic body, but which finds more 
than one parallel elsewhere. 20 At last, after many conferences 
and much spinning of legal subtleties, the last penny was paid 
and " everlasting concord " established between the two kins. 
Should a single penny, however, be wanting of the galanas 
money, all that had been done was of no avail ; the law " set 
vengeance free," since " a part is not the whole," and, though 
all the rest of the money was retained, the " llofrudd " might, 
for want of this one penny, be slain with impunity. 21 So 
narrowly did the ancient spirit of tribal honour watch its rights, 
so grudgingly did it yield to the state for it was the king 
who, for a substantial consideration, took the leading part in 
the exaction of galanas the power of setting limits to its in- 
born liberty of self-defence. 

19 LL. ii. 42, 44 ( ii, 12). The rule Dim. II. i. 35 = Lat. A. II. ii. 33 
= Lat. B. II. i. 14 shows that killing in battle might give rise to galanas ; the 
West Saxon law under Ine ( 34) was in almost identical terms. 

20 It was also the English custom. Cf. Vinogradoff, Growth of the Manor, 
pp. 9-1 1. 

21 Yen. III. i. 18 (MS. B.). The phrase, " cyfraith a ryddha ddial," occurs 
in Dim. II. viii. 14. 


Considerable independence was enjoyed by the adult CHAP, 
members of the cenedl, and there are few traces, in the case of IX ' 
men, of that absolute paternal power which gave Roman fathers 
entire control of the destinies of their sons. In the matter of 
holding land, the patriarchal principle, which conceded no rights 
to the son while the father was alive, still held good among the 
higher classes, but otherwise the attainment of majority brought 
with it emancipation. At the age of fourteen 22 the young 
tribesman was withdrawn from the control of his father, who 
might no longer correct him, 23 and he became in the eye of 
the law a fully responsible person. His independence was 
secured by the provision that he was now to come into posses- 
sion of his rightful share of the family goods ; 24 he might then 
choose his career, might enter the royal service in some 
capacity, might tempt fortune in some distant land, or, if of 
home-keeping disposition, might settle in a house of his own on 
some corner of the family land and join in the cultivation of 
the patrimony under his father's direction. 25 Such was the 
position of the " bonheddig cynwynol," the " gentleman born," 
the scion of a free stock ; the son of an aillt or villein reached 
manhood and responsibility at the same age, but his freedom 
to rove was limited by the laws of villein tenure. 

The independent householder was naturally also the married 
man, and the laws facilitated early marriage by providing that 
at the age of twelve a girl should come into possession of her 
goods, the share of the family property which she was entitled 
to carry with her to a husband. 26 She still remained, however, 
under the government of her father and other relatives until 
marriage was effected, for the essence of a regular marriage 
from the point of view of Welsh law was the formal bestowal 
of the bride by her kindred. 27 Whatever may have been usual 
in the way of invoking a blessing from the Church, a ceremony 
which, it is to be remembered, the canon law did not treat as 

18 Yen. II. xxviii. ; Dim. III. ii. 8; Lat. B. II. xliv. 3, 4. 

23 Yen. II. xxviii. 8. 

84 " Medu y da " (ibid.). " Da," i.e., goods, chattels, must always be care- 
fully distinguished from land, the ownership of which passed under entirely 
different conditions. 

36 The rules for the division of land on an uchelwr's death show that his 
sons were already settled upon it in various detached homesteads. 

"Yen. II. xxx. "See especially Yen. II. i. 75. 

VOL. I. 19 


CHAP, indispensable, the Welsh conception of a lawful marriage was 
one which recognised the ancient rights of the cenedl over its 
daughters, in which the father uttered the fateful words, " Maiden, 
I have given thee to a husband," 28 and the pencenedl received his 
fee from the bridegroom in recognition of the rights of the 
kindred as a whole. 29 It is likely enough that in primitive 
times the whole clan was solemnly consulted ; in the story of 
Cilhwch and Olwen, the giant Yspaddaden Pencawr refuses to 
give the hand of his daughter until the matter has been laid 
before her four great-grandfathers and her four great-grand- 
mothers, all of whom are alive ; 30 in later practice, no doubt, 
the business was arranged more simply. It appears to have 
been more usual to wed within the clan than without it ; 
" marry in the kin and fight the feud with the stranger " 31 was 
an old tribal saying, and Giraldus testifies how widely diffused 
in Wales among all classes was the custom, notwithstanding 
that it was i reprobated by the Church, of marrying near kindred. 32 
The wedded wife, though no special privileges were enjoyed 
by her children on the ground of their legitimacy, was herself 
protected in many ways by the law. Her husband had not, as 
in ancient Rome, the power of life and death over her ; in 
everything which concerned galanas she was still a member of 
her native cenedl and liable to be avenged by her father and her 
brothers. 33 Nor might she be beaten, save for serious offences 
specified in the codes. 34 While she had a limited control over 
all the joint property of the> household, including the portion 
she had herself brought to the common stock, she had also 
certain personal possessions which were exclusively hers and of 
which nothing could deprive her. 35 Most important of all, the 

28 Dim. II. viii. 73 = Gw. II. xxxix. 35 = Lat. A. II. xx. 41 = Lat. B. II. 
xxiii. 37. 

29 Ven. II. xix. i. It was but 24d. (a usual figure in the case of official fees), 
and cannot, therefore, be treated as a relic of the bride-price. As no other 
payment was made by the bridegroom to the relatives of the bride, it must be 
supposed that all traces had disappeared of the fact that originally the husband 
bought his wife from her kindred. 

30 Mab. 119. 

si Dyweddi o wngc, galanas o bell " (Evans, Proverbs, p. 550). " Dyweddi " 
= marriage in old Welsh, the preliminary ceremony of betrothal being unknown. 
32 Gir. Camb. vi. 213 (Descr. ii. 6). 33 Ven. III. i. 38 (MS. B.). 

34 Ven. II. i. 39 ; Dim. II. xviii. 6. 

35 These included the " cowyll " or " hood," which is the Morgengabe of 
the Germans, the gift of the bridegroom to the bride on the morning after the 


law gave her considerable protection against arbitrary divorce. CHAP. 
The Church had not succeeded in making the marriage tie IX> 
indissoluble ; the ancient rule that a man might put away his 
wife, if so minded, was still valid, and in certain cases, for 
instance, when a king's wife bore him no heir, 36 was deemed to 
be very salutary. But the law took care that it should not be 
frivolously put into operation by providing for a substantial 
payment, known as "agweddi," to the divorced woman, and 
allowed even this only during the first seven years of marriage ; 37 
after the lapse of this period, the husband who had grown weary 
of his spouse and desired a younger partner had to resign to 
the first wife the half of all his possessions, exactly as if the 
separation had been brought about by his death. 38 

The study of the cenedl reveals to us the oldest elements 
in Welsh society, those which had resisted the influence of 
Roman law and government, of Christian ethical teaching, and 
of royal authority as exercised by Welsh chiefs. It carries us 
back into the Celtic foreworld, and discloses a system not at 
all unlike, in spite of variations of detail, that which prevailed 
among the Irish. It was the continuance of this system far 
on into the Middle Ages, when feudalism and the canon law 
had elsewhere wrought such mighty changes, which gave Welsh 
life its piquant interest, its individual tone and colour. 


The structure of Welsh society as an aggregation of kins 
having been examined, it next falls to consider its relation to 
the land which it occupied and from which it drew its susten- 
ance. Largely pastoral in its activities, concerned in the rearing 
of horses, cattle, sheep and swine, it was also to some ex- 
wedding, and anything which the wife might have subsequently received from her 
husband in atonement for his offences against her. 

36 In the story of King Pwyll of Dyfed, it is related how the men of the land 
began to clamour at the end of the third year against an alliance which had 
yielded no royal heir (Mab. 17-18). 

37 From the mention of " resipiscendi poena statuta" one may infer that it 
was this system which led Giraldus to talk of marriages on trial as usual among 
the Welsh (vi. 213-4 (Descr. ii. 6)). 

38 A wife irregularly married, without the concurrence of the kindred, was 
not entitled to the full " agweddi," though she might not be put away without 
some compensation. At the end of seven years she attained the status of a 
wedded wife. 



CHAP, tent agricultural, tilling the soil for crops of oats, barley and, 
more rarely, of wheat. The woodlands nurtured bees, which 
furnished the mediaeval substitute for sugar; firewood and 
building material were supplied from the same quarter. Hunt- 
ing, hawking and fishing were open to the free tribesman as 
means of adding to the family larder. Clothing was made at 
home by the inmates of the household, and only articles of 
special value were imported. Thus each community was to 
a very great extent self-supporting and economically complete ; 
the calling of the smith, which was in great honour because it 
furnished the warrior with his weapons, was the only industrial 
one which appears to have been definitely specialised. 39 At 
the same time there was within the community itself some 
degree of organisation and division of labour, the lower classes 
serving the higher, and this leads to the consideration of the 
various grades into which Welsh society was anciently divided. 
As among other primitive peoples, there was at the bottom 
of the social ladder a slave class, supplying most of the manual 
labour needed and possessing hardly any rights. The " caeth," 
whether male or female, was the absolute property of his owner ; 
he belonged to no cenedl, and, if he were killed, no galanas 
arose ; the slayer had merely to recompense his master for the 
loss. 40 If a slave struck a free man, he was liable to the loss 
of his right hand ; 41 if he sought to escape from his thraldom, 
he might be recaptured anywhere, and the captor received 
a reward. 42 The class was probably at one time largely re- 
cruited by war, but in a later age it was no doubt chiefly main- 
tained by that active slave trade which was kept going by the 
Danes of Ireland, notably at the ports of Bristol and Chester. 43 
It was by means of slaves that much of the field-work of a 
Welsh homestead was performed ; cutting and clearing wood, 
digging, and the like were thus accomplished, and so, too, the 
menial housework, grinding corn in the family quern, baking 

39 Smithcraft ranked with bardism and holy orders as one of three free 
vocations which a villein might not follow (Ven. I. xliii. n = Dim. II. viii. 7 = 
Lat. A. II. viii. 10 = Lat. B. I. xix. 2). 

40 Dim. III. iii. 8 ; Gw. II. xl. 23. 

41 Ven. III. i. 34; Dim. II. xvii. 44; Gw. II. v. 32; Lat. A. II. xv. 18 ; Lat. 
B. II. xxii. 17. 

42 Dim. II. xvii. 49 ; Lat. A. II. xvi. i, 2 ; Lat. B. II. xxii. 23, 24. 

43 Cong. Eng. (2), pp. 443-5. 


and washing. 44 The lot of a slave was more or less irksome CHAP. 


and unpalatable according as the owner was of lower or higher 
rank ; the sewing-maid who served the queen was happily 
placed, 45 while life was not worth living, if a current proverb 
is to be believed, for the thrall who was fated to do service to 
a villein. 46 . 

A second class which stood outside the tribal system was 
that of " alltudion " or aliens. The alltud, or " other-country- 
man," 47 was naturally in very many cases the foreigner from 
England, Ireland or the Continent, living in Wales as an exile, 
a hostage, or an adventurer, but foreign speech and alien birth 
did not, in the sight of the law, form the essence of the alltud 
status. What really made the alltud was the want of attach- 
ment to the soil ; he was the man who had no claim upon any 
land in the district, whether as villein or free tribesman, the fifth 
wheel in the coach, for whom there was no place in the system 
of the country. 48 Nevertheless, if he chose to settle in the 
district, he received full legal protection, and his descendants 
might in time be recognised as proprietors. He was required 
to place himself under the protection of some owner of land, 
who gave him the foothold on the soil which he required ; 
during three generations the bargain was revocable and the 
alltud tenant might leave, on condition of halving his chattels 
with his lord ; the fourth man such, at least, was the custom 
of Gwynedd 49 became a landowner, and at the same time a 
villein bound to perpetual service. 

All who held land or were entitled in due course to succeed 
to it were reckoned proprietors, inheritors or Cymry. 60 Each 
was a member of some cenedl, with a definite " braint " or legal 
status ; high and low, they were parts of one system. Never- 
theless, a broad and well-defined barrier separated the villein 

44 Ven. III. i. 33; Dim. II. xvii. 38, 47; Gw. II. vii. 17. For lavatrix = 
ancilla, see Lat. A. II. xviii. 12; Lat. B. II. xlii. 12, Iv. i, Ivi. i. 

48 Dim. II. viii. 49 mentions " nottwyd gwenigyawl y vrenhines," which is 
to be translated " the needle of the queen's domestic ". 

4 " Da angheu ar eidywc taeawc " (Evans, Proverbs, p. 546). 

47 From a primitive " allo-touto-s," perhaps seen in the " matribus ollototis 
sive transmarinis " of a Roman inscription from Binchester (Urk. Spr. 22). 

48 See especially Ven. II. xiv. i, 2, and xv. 8. 49 Ven. II. xvi. 20-6. 

80 It is important to observe that the term " Cymro " was not a badge of 
freedom, but included the aillt as well as the bonheddig. See note appended to 
chap. v. 


CHAP, holder of land from the free-born member of the tribe. The 
aillt or taeog, for he was known indifferently by either name, 
as well as by the borrowed one of " bilain," 51 was distinguished 
in many ways as belonging to an inferior social grade. He was 
subject to a lord, who might be the king, or one of the free 
landowners. He had no place in the free court of the district 
in which he lived. Hunting and hawking were for him forbidden 
amusements, and therefore he received no more compensation 
for any dog of his which might be killed than was paid for a 
farmyard cur, or for any bird he lost than the value of a hen. 52 
He might not leave the tillage of the soil for any liberal occupa- 
tion ; smithcraft, the* bardic order, the Church were all closed 
against him. 53 He might not, indeed, quit the community in 
which he was bred and born, the corner of the social system 
which the law provided for him, and if he took to flight and 
deserted his lawful station, he paid forfeit with all his posses- 
sions, which were immediately seized by his lord. 54 All the 
conditions of his life show that he was regarded as by pre- 
eminence the cultivator of the soil, the man whose industry as 
farmer and stockbreeder was necessary to the life of the com- 

Little light is thrown by the laws upon the way in which 
the aillt of an uchelwr or free landholder served his master and 
was associated for work with his fellows. 55 The code of Hywel 
was put together, primarily, in the interests of the crown, and 
accordingly in the documents derived from it matters in which 
the king had no concern are only incidentally treated. As to 
the king's aillts, there is, on the other hand, some information, 
and this reveals the villeins who provided for the needs of 
the court as grouped together in communities known as trefs. 

81 Aillt is the usual term in Ven., taeog and bilain in the other codes. Aillt 
is connected by Whitley Stokes (Urk. Spr. 21, as modified by 327) with the 
Irish " alt, ailt" = house, and taeog is no doubt from " ty," but the meaning is 
" a slave having a house " and not, as the " verna " of Davies, Diet, suggests,' 
" a slave born in the house ". " Alltud " and " aillt " have been much confused 
by late writers, but in the authoritative legal texts they are always carefully 

52 Ven. III. xiv. 17 ; Dim. II. xiv. 18 ; Gw. II. xxi. 12 ; Ven. III. xv. 10. 

53 See note 39 above. 

54 Dim. I. xxix. 6 ; II. xii. 5 ; Gw. I. xxxv. u ; xxxvi. 12. 

55 Ven. II. xvi. 20-6 deal with uchelwrs' alltuds and aillts, and suggest that 
they were under much the same rules as those of the king. 


The tref was originally the "house" or "dwelling-place"; 59 a CHAP, 
secondary meaning which it acquired was "hamlet, village," 
and this is what seems to be denoted in the case of these 
villeins the group of villein homesteads clustered together for 
the purpose of common cultivation of the surrounding land. 67 
Each taeogdref or villein hamlet was responsible as a whole for 
what was due from it to the king, and, while each man had his 
own house, farm stock and farm buildings, it is clear that for 
the important business of ploughing there was co-operation. 
The heavy wooden plough was drawn by a team of eight 
oxen, and to furnish these the tref clubbed together ; each 
taeog became a member of what was known as a " cyfar " or 
plough fellowship, and it was a rule that in no taeogdref might 
the annual ploughing begin until every man had found his 
cyfar. 58 Each day's ploughing produced an " erw " or " cyfair," 
a strip of ploughed land of which the dimensions varied in 
different districts, but which was usually about ten times as long 
as it was broad. 59 When twelve erws had been ploughed, they 
were assigned in accordance with a standing rule among the 
contributors to the ploughing, including not only the owners of 
the oxen and the plough, but also the ploughman and the ox- 
tender, the latter being not a " driver " but a " caller " (geilwad), 
for he walked backward in front of the team and relied even 
more upon the power of his voice than upon his ox-goad. 80 
It is likely that each hamlet had its ploughman and its caller, 
who made ploughcraft their special business. Whether the 
land ploughed by a villein plough-team had been previously 

68 Urk. Spr. 137, 334. The word occurs in the sense of "house" in the 
ninth-century glosses of Martianus Capella (Arch. Camb. IV. iv. (1873) 4) ; this 
is also the meaning which explains " adref," "athref," and "gartref" and very 
many of the trefs of local nomenclature. 

87 " Trefgordd " is also used to denote the villein hamlet; Dim. II. ii. 12, 
compared with Gw. II. ii. 12, shows that it was used interchangeably with tref. 

88 Gw. II. xix. 2. Ploughing began on gth February (Ven. III. vii. 4). 

88 " Cyfar" = co-tillage, and " cyfeir " (now " cyfer," an acre) = a day's 
tilth, are carefully distinguished in the texts (for " cyfeir " see Ven. III. vii. 4 ; 
xxii. 236, 237; xxiv. 13), though not in Aneurin Owen's translation. MSS. D. B. 
C. J. K. of the Ven. code clearly identify the "cyfeir " and the " erw" in III. 
xxiv. 3 (" erw y gwydd a honno a elwir cyfeir y casnad "). A. N. Palmer has 
made a special study of the various customary acres of Wales ; see especially 
Arch. Camb. V. xiii. (1896), 1-19. 

60 " Stimulatore praeambulo sed retrogradu " (Gir. Camb. vi. 201 (Descr. 
i. 17)). " Geilwad" = driver is found in Job xxxix. 7. 


CHAP, allotted as arable land to the villeins individually or was 
IX ' common for this purpose to the whole of the tref is not made 
clear by the laws, 61 but it is certain that the erws lay in open 
fields, unprotected by ditches or hedges, and that it was cus- 
tomary to plough virgin soil from time to time in order to get 
the benefit of unexhausted land. 

The tref of villeins, which as a cluster of homesteads is 
often styled "trefgordd," acted together for other ends than 
that of tillage. The cattle of the tref grazed together in the 
wide pastures which surrounded it under the eye of the village 
herdsman or " bugail," 62 whose dog walked out at the head of 
the herd in the morning and followed its rear at night, a valiant 
protector against the wolves of the forest, not a penny inferior 
in legal worth to the best ox in the herd. 63 There was also a 
village smithy, a village kiln for the drying of corn, and a bath- 
hut, in which water was heated for special ablutions. 64 

Perhaps the most remarkable fact about the " taeogdref " or 
villein hamlet is that its folk held their land, not by right of 
kinship, but as members of the community. Land of this 
description, the Venedotian code explains, is not to be divided 
(like family land) between brothers, but the king's officers are 
to apportion it equally between all the men of the tref, whence 
it is known as " tir cyfrif " (reckoned land). 65 It is further ex- 
plained that no land of this kind can revert to the lord for 
want of heirs, since, as another passage puts the case, the son's 
share of his father's erw is, in such a tref, no greater than that 

81 Ven. II. xvi. 22, referring to men who have become aillts, says they have 
their homesteads and land attached to these and " eu tir namyn hynny yn dir 
swch a chwlltyr (i.e., arable land) rhyngddynt". On the other hand, III. xxiv. 
17 suggests that the partners in the "cyfar" had land of their own which they 
could require the team to plough when it was their turn to receive an " erw ". 

82 " Bugail," which now means shepherd, is properly neat-herd or cattle-tender 
(Urk. Spr. 178). For " bugail trefgordd," tending the cattle of different owners, 
see Ven. II. iv. 8 ; Dim. II. viii. 81 ; Gw. II. xxviii. 24. Lat. A. calls him " pastor 
communis uille " (II. v. 8) ; so Lat. B. II. xvi. 8. 

63 Ven. III. xiv. 19 ; Dim. II. xiv. 18 ; Gw. II. xxi. 14 ; Lat. A. II. xxiv. 30 ; 
Lat. B. I. xiii. 26. 

64 Ven. III. iii. 20 ; Dim. II. viii. 36 ; Gw. II. xxxix. 16 ; Lat. A. II. iii. 14 ; 
Lat. B. II. vi. ii. 

65 Ven. II. xii. 6 ; xviii. 7. It may be said, with reference to Vinogradoff, 
Growth of the Manor, p. 93, that in the laws, as in the extents, "tir cyfrif" is 
clearly always villein land. Save in the Triads of Dyfnwal Moelmud (for which 
see note appended to this chapter), there is nothing to suggest that a system of 
co-tillage existed in mediaeval Wales among the free tribesmen. 


of the furthest man in the tref. 88 There was no general shift- CHAP. 


ing of the homesteads, for the law provided that each man 
should remain in undisturbed possession of his " tyddyn," with 
the home croft of four erws, 67 and further laid down the rule 
that a taeog should be succeeded in this holding by his young- 
est son. 08 But as to the general arable land, on whatever 
principle it was tilled, whether as private or as communal 
property, it is clear that it was not subject to any rights of 
inheritance, but was regarded as a common fund for the main- 
tenance of the whole tref. Every taeog's son in the village, if 
he had attained the age of fourteen and was not, as the young- 
est son, his father's destined successor, was entitled, on each dis- 
tribution, to a man's share of the land which was being divided. 

Thus the village community is to be found on Welsh soil, 
though only among the unfree cultivators. It is natural to re- 
gard this class, holding a servile position and having few privileges, 
as the descendants of a conquered race, 89 and the system of ten- 
ure under which they worked is obviously ancient, telling of the 
long settlement on the land of the cultivating community. In 
view of what has been said in the early chapters of this history, 
one may therefore without undue boldness recognise in these 
aillts and taeogs the remnant of the Iberian people, the oldest 
tillers of the soil in Wales, reduced to servitude by wave after 
wave of Celtic conquest, by the might of the ancestors of the 
free tribesmen, whose institutions are now to be examined. 

Besides the true tref, or village community of servile tenants, 
there was another tref, termed the free tref, which seems to 
have acquired its name through being, like the other, a definite 
area within which there was joint responsibility for the render 
due to the king. Every commote was divided into trefs, 
some of which were bond and some free ; the free trefs, no 
less than the bond, paid each its annual contribution towards 
the maintenance of the court. 70 But the free tref was not a 

88 Ven. II. xxi. 2 ; LL. ii. 56, 64. 

87 For the home croft of four or eight erws, which went with every home- 
stead or " tyddyn," see Ven. II. xii. i ; Dim. II. xxiii. i ; Gw. II. xxxi. i. 

88 LL. ii. 64. 

89 Lewis, Ancient Laws of Wales, p. 47 ; W. People, p. 215 ; Vinogradoff, 
Growth of the Manor, p. 25. 

70 The Venedotian code makes the " maenol," containing four trefs, the unit 
which paid the king the free render of i (i. 188). But all the particulars it gives 


CHAP, hamlet or body of villagers ; there is clear evidence that the 
households of the better class in Wales were not grouped to- 
gether in villages, but were scattered here and there over the 
face of the country. Observers were struck by the difference 
in this respect between England and Wales ; " they do not 
foregather in cities, villages or castles," says Giraldus, 71 "but 
inhabit the woods like anchorites," and Archbishop Peckham a 
century later noticed the same love of solitude. 72 It may be 
pointed out that the map of Wales bears witness to this day 
to the divergence between English and Welsh custom in the 
matter of the distribution of the rural population. 73 Most of 
the Welsh parish churches stand almost alone, the houses, in- 
stead of forming a group as in England, with the church as its 
centre, being pretty evenly dispersed over the whole parish. 
The free tref was constituted by marking off a number of these 
scattered holdings and associating them in responsibility for 
the payment of a fixed portion of the free render of the com- 
mote. Such a tref might very well be occupied by a body 
of kinsmen, since kinsfolk would naturally settle together; it 
would be separated from other trefs by well-marked barriers. 74 
But it would not be, like the taeogdref, a society of joint tillers 
of the soil, with interests closely intertwined, but merely a group 
of private owners, each pursuing his own way and holding his 
land separately. 

In the free tref the landowner was the man variously known 
as the uchelwr (high man), gwrda (goodman) and breyr. 75 For 

under this head have an air of unreality ; from the phrase " y dref uchaf (i.e., 
next to Arfon) o arllechwed " used in Mab. 63 of Creuwyrion (for which see Rec. 
Cam. 12-13), it would appear tnat m Gwynedd, as elsewhere, the term "tref" 
was used to denote the fiscal unit which became the " villa " and township of 
later times. Vinogradoff (148) points out how among the English artificial and 
composite " tuns " were formed for administrative and judicial purposes on the 
analogy of the real village settlements. 

71 " Non urbe, non vico, non castris cohabitant ; sed quasi solitarii silvis 
inhaerent" (Gir. Camb. vi. 200 (Descr. i. 17)). 

72 " II ne habitent pas en semble, mes menit chescun loinz de autre " (H. and 
St. i. 570). 

73 The distinction is one between Celtic and Teutonic practice, as pointed out 
by Meitzen. See the contrasted maps in Social England (1901), i. pp. 164, 165. 

74 For the meer between two trefs see Ven. II. xxi. i ; xxv. i ; Gw. II. xxxii. i, 5. 

75 Breyr is a purely South- Welsh term, while uchelwr occurs chiefly in Ven., 
often in the form "mab uchelwr," in which, as in " mab aillt," "mab sant," 
the prefix merely indicates the gender. Gwrda is commoner in general literature 
than in the laws. 


the attainment of this position it was necessary not only to be CHAP, 
of free birth, a member of a free cenedl, but also to have been 
fully emancipated, to have been freed by death from the con- 
trol of father and paternal grandfather. This, it has been 
stated above, was not necessary for the holding of personal pro- 
perty or " da " ; the " boneddig cynwynol " might be a house- 
holder and owner of cattle, but he had no rights over the family 
land while his father was alive. 7 " Thus the uchelwrs were not 
only a wealthy, but also an experienced class, who naturally 
took a leading position in all the affairs of the country ; they 
were the nearest approach to a nobility to be found in Welsh 
society 77 and are sometimes dignified by that title. 78 They 
had alltuds and aillts who did them service ; they hunted the 
deer in their forests and took the fish which were caught in 
their weirs. Each lived in a " neuadd " or hall, which, although 
rudely built of timber and wattle, like all the buildings of the 
Welsh, was a much more substantial affair than the mere " ty " 
of the taeog. 79 It was the privilege of the uchelwr to possess 
a harp 80 and he was a recognised patron of travelling bards. 81 
He was the man of leisure and dignified ease, the organ of the 
public opinion of his commote, the adviser, the critic and the 
defender of the crown. 

Nevertheless, the uchelwr was not the absolute owner of 
the land under his control. He had but a life interest in it, and 
could not give, sell or bequeath any part of it, to the detriment 
of the family whose patrimony it was. 82 It passed on from 
generation to generation by fixed laws of inheritance, which 
might not be set aside. The first principle observed was equal 
division on the death of an uchelwr among all sons. Daughters 
were excluded, because it was not for them to carry on the 
family traditions, but by marriage to secure heirs for other 

78 Ven. II. xxviii. 9. 

77 The triad defining the three sorts of men as a king, a breyr, and a villein 
(LL. i. 350) excludes a nobility in the ordinary sense. 

78 See LL. ii. 1083 (index, s.v. nobilis) and Gir. Camb. vi. 166 (Descr. i. 2 : 
" nobiles qui Kambrice Hucheilwer quasi superiores viri vocantur "). 

79 Ven. III. xxi. 2, 3. 80 Ven. I. xliii. 2 Dim. II. viii. 10. 

81 Dim. I. xviii. i ; Gw. I. xix. 2 ; Lat. A. I. xxii. 2 ; Lat. B. I. xxi. 10. I 
take the reference to the taeog to be contemptuous and a reflection on the taste 
of the bard who could choose such a patron. 

88 Ven. II. xv. 8; LL. 11.270. 


CHAP, proprietors. 83 All sons, on the other hand, took equal shares, 
since the purpose of the law was to maintain the tribe, not to 
keep the estate intact. Such favour as was shown fell to the 
lot, not of the eldest, but of the youngest son, for it was he 
who, among uchelwrs and villeins alike, succeeded to his father's 
homestead, 84 the others being expected to remain in the 
homes they had made for themselves on the family land during 
their father's lifetime. But the law went further even than 
this in its benevolent determination to provide a fit main- 
tenance for every uchelwr and ordained that, after the equal 
division between brothers, there might be on occasion a similar 
division between first cousins and even one between second 
cousins, so as to prevent the rise of marked inequalities 
between uchelwrs who were nearly akin to each other. Thus 
arose a subdivision of the cenedl the group of men descended 
in the male line from a common great-grandfather ; they could 
inherit from each other in default of issue and formed the body 
which it was necessary to consult before any part of the land 
could, for however good a reason, be disposed of to an out- 
sider. 85 


The survey of mediaeval Wales which was undertaken in 
the eighth chapter showed us a country divided into regions 
called cantrefs, each of which was again divided into two, three 
or more commotes. Both the cantref and the commote are 
repeatedly mentioned in the laws, but upon examination it be- 
comes evident that they are not two separate institutions, the 
one subordinate in its working to the other, like the English 
shire and hundred ; though the geographical facts at first sight 
suggest this, they must not be so interpreted. The leading 
feature of the commote is its court, for the trial of disputes 
among the free tribesmen, and this is sometimes called the 
court of the commote or the cantref. But nowhere is there 

83 Ven. II. xv. i ; Dim. II. xxiii. 7 is less rigid, recognising a daughter's 
claim in the absence of any male heir. This relaxation is said in the Statute of 
Rhuddlan (LL. ii. 925) to be " contra consuetudinem Wallensem antea usitatam," 
but it may be that the stricter rule was confined to Gwynedd. 

84 The obvious exception in the case of the successor of a chief is duly noted 
in LL. ii. 578, 686. 

86 LL. ii. 270. 


any shadow of a suggestion that a case might be taken from CHAP, 
the commote to the cantref, as from a lower to a higher court, 
or that there were two distinct sets of officers for the two areas. 
In the laws it is the commote which appears as the living and 
active body, the references to the cantref being for the most 
part perfunctory, with a smack of antiquity about them. 88 It 
is further to be noticed that the commote does not appear in 
the earliest Welsh records, 87 while the names of the Welsh 
commotes wear in the main a much more modern air than 
those of the cantrefs. The conclusion is, therefore, not to be 
resisted that the commote or " neighbourhood " is of compara- 
tively recent origin ; when the cantrefs had become, through 
the growth of population and the development of more settled 
habits, too large to be convenient areas of tribal co-operation, 
they were divided by agreement into smaller districts, each of 
which had henceforth its own independent organisation. 88 
There is good reason to think that this step had not been 
taken in South Wales before the appearance of the Normans, 
and that the commote dates from the end of the eleventh 
century. 89 

The machinery of the commote, therefore, is in origin that 
of the cantref, and, during the period which is now under con- 
sideration, it is with the cantref only that one has to deal. A 
further question naturally arises, whether the cantref itself is a 
primitive institution and may not have been the fruit of an 
early attempt at constitutional reform. This is certainly sug- 
gested by the derivation of the name ; the area formed by the 

88 Ven. knows only of the court of the cymwd, but Dim. II. viii. no and 1 16 
have " brawdwr cymwd neu gantref" and " llys cymwd neu gantref" is a 
common phrase in LL. ii. The only passage in which an office is associated 
with the cantref is Gw. II. xl. 9, where the king's footholder or " troediog " is 
called " eisteddiad cantref, " a phrase which is no doubt to be regarded as an 
archaism. The cantref continued, of course, to be of importance as a geographi- 
cal area after losing its political importance ; see Dim. II. viii. 14 ; xxiii. 47-9 ; 
Gw. II. xxxix. 5. 

87 The term is not to be found in Lib. Land. It occurs twice in Mab. (31, 62). 

88 Gr. Celt (2). pp. 192, 207 ; Urk. Spr. 87. 

89 The usual list of the commotes of Morgannwg, besides including such 
obviously late forms as Tir yr larll and Tir yr Hwndrwd (" Tirhundred" in the 
Cottonian list), does not appear to account for Bro Morgannwg at all. In Dyfed 
again, Penfro, which was originally the name of a cantref, could not have 
become the name of a commote also until " Castell Penfro " had come to be 
known by the shorter title. 


CHAP, grouping of a "hundred houses " or " hamlets " (can tref) would 
IX ' seem to be of unmistakably artificial origin. In point of fact, 
there is good evidence that the cantref is the historical suc- 
cessor of the " gwlad " or " tud," i.e., country or tribe, the body 
of free tribesmen who, either as ancient settlers or as Brythonic 
immigrants, held sway as an independent community within 
bounds which clearly marked them off from their neighbours. 90 
At some period which cannot now be precisely ascertained, the 
larger gwlads, such as Mon, Ceredigion, Brycheiniog, Ystrad 
Tywi, were divided into cantrefs, the smaller, such as Meiri- 
onydd, Dyffryn Clwyd, Buellt, became cantrefs themselves, 91 
and thus the cantref everywhere took the place of the tribe as 
the means of enforcing justice and as the link between people 
and crown. 

The cantref court was an assembly of the uchelwrs of the 
cantref, who as heads of households represented all the freemen 
of the tribe. There was no fixed place of meeting ; indeed, it 
was customary in cases of dispute as to land to meet on the 
land which was the subject of the lawsuit 92 The king pre- 
sided, or, in his absence, his official representative ; it was his 
" llys " or " gorsedd," the " high seat " of his jurisdiction. 93 In 
South Wales, the assembly formed, nevertheless, a ' considerable 
check upon his power, for it might, by its " dedfryd gwlad," or 
"judgment of country," declare him to have acted oppressively 
and obtain the reversal of the deed which had incurred its 
censure. 94 Giraldus bears witness to the extent to which in 

90 " Gulat " is used for cantref (or cymwd) in the " Privilegium Teliaui " in 
Lib. Land. 120 (line 10), and the " gladoet" of Ven. II. ii. 4 must be cantrefs. 
Gwlad and cantref are treated as synonymous in Dim. II. viii. 14 ; Gw. II. xxxix. 
5 ; the " henwryeyth gwelat " of Lat. A. II. ix. 15 are the " heneuydyon cantref" 
of Lat. B. II. xxiv. 45. "Tud," the Irish tuath (Urk. Spr. 131), is of rarer 
occurrence ; cf. however, the "castell teirtut" of Lib. Land. 134, said to be the 
meeting-point of Cantref Bychan, Cantref Selyf and Buellt. 

81 It is quite possible that at the time of the later division some of the smaller 
centres became commotes, as not requiring to be divided, and were then grouped 
together in artificial cantrefs, which never had a separate existence. This 
appears to have been the case in Powys, where many of the cantref names have 
an unreal ring and are not known otherwise than from the lists. 

92 Ven. II. xi. g (a hynny ar y tir), 10 (dyfod ar y tir). 

93 For "gorsedd" = court see Ven. II. xi. 51; Dim. II. viii. 15 ; LL. ii. 8 
( 24). In Mob. 8, 32, 166, and in place names, it has the meaning of " mound," 
" tumulus," possibly because it was a royal habit to hold session on elevations of 
the kind. 

94 Dim. III. i. 17 ; cf. II. viii. 139. 


the south the royal dignity and supremacy were prejudiced by CHAP. 
tliis aristocratic independence; the realm of Deheubarth, he 
says, though much larger than either Gwynedd or Powys, is 
not so desirable a heritage, by reason of the number and 
insubordinate temper of its uchelwrs. 95 It is hardly fanciful to 
find in this difference between the political atmosphere of the 
north and that of the south the chief explanation of the fact 
that the princes who worked most successfully for Welsh unity 
were in the main of northern origin. At the stage of political 
development which the Welsh people had now reached, aristo- 
cratic freedom meant tribal isolation and weakness, while royal 
power in capable hands made for national union and strength. 
And in South Wales, not only did the uchelwrs keep a tight 
hold of the reins in matters of government, they also retained 
to the full their ancient judicial powers. In Gwynedd and 
Powys, though the king's gwrdas still sat on his right and on 
his left as the assessors of his court, their functions were purely 
ornamental ; in every commote there was an official or pro- 
fessional judge, who took the business of judgment out of 
their hands. 96 There was no such officer in the south ; " in 
Deheubarth," says the Dimetian code, " there is in every court 
a great number of judges, as aforetime before the days of 
Hywel the Good, to wit, every owner of land by privilege of 
that land, though it be not land of office ". 97 

The tribal court dealt, it would appear, with all such 
questions requiring judicial settlement as arose among the free 
tribesmen. Of these the most important was the determina- 
tion of disputes as to who was the lawful owner of land, a 
matter settled, as has been said, on the land itself, in full court, 
after elaborate pleadings and counter-pleadings. Besides the 
judges and the parties, there were in attendance the clerk or 
recorder, the " rhingyll " or usher and summoner of the court, 
and, in many cases, two professional pleaders, whose duty it 
was to conduct the parties through the mazes and pitfalls 
spread for the unwary by primitive rules of legal procedure. 98 

96 Gir. Camb. vi. 166 (Descr. i. 2). 

98 Ven. II. xi. 10; Dim. I. xxxi. i ; II. viii. no. * 7 Dim. I. xxxi. 2. 

98 The employment of a " tafodiog " or pleader was optional in the south 
(Dim. II. ix. 5) ; in Ven. II. xi. 10, the presence of a " cynghaws " on each side 
is assumed. That the advocate was a professional lawyer appears from LL. ii. 
98 ( 7), 146 ( 33). For the " gwallawgeir " or "faulty word " see Dim. II. viii. 78. 


CHAP. Silence was proclaimed in the field, "and no one might help 
IX- plaintiff or defendant by any suggestion save his " canllaw " or 
"hand-rest," the friend specially assigned for this purpose. 
Witnesses gave evidence as to recent facts ; the important 
question 'of ancestry was settled by the testimony of neighbour- 
ing landowners, which might not be< gainsaid. The decision 
rested with the judges, who in South Wales chose one of their 
number to preside and announce the verdict." Such in a few 
words was the nature of the proceedings in an action of "kin 
and stock " ; other suits are not so fully described, but it may 
be assumed that they followed in great measure the lines of 
this, the most important. There was a process for obtaining 
temporary possession of land which was being claimed in the 
court, a process open only to the son of a former occupant and 
known by the remarkable name of " dadanhudd " or " uncover- 
ing ". It is explained, and no doubt rightly, as the assertion by 
the son of his right to uncover the fire on the family hearth in 
succession to his father. 100 In the Welsh house, it is 'known 
from other sources, the fire, burning on an open hearth in the 
centre of the floor, was at bedtime covered up with peats or 
logs, so as to keep it gently smouldering until the morning, 
when it was again uncovered and set blazing with the aid of 
fresh fuel. 101 If the son claimed the right to uncover his dead 
father's hearth, it can only have been as the priest of a long- 
forgotten ritual, as the lawful head of the house ministering to 
the ancestral hearth-spirits, and thus the action of " dadanhudd " 
preserves in its name the one known trace among the Welsh of 
that ancestor-worship which was so widely practised among 
Aryan peoples, and has been held to be of greater significance 
for their religious history than the nature-worship rendered so 
familiar to us by the classical writers. 102 

"Dim. II. viii. 114. 

100 LL.ii. 140-2 (26, 27) ; Wotton, 565; Evans, Diet. s.v. Moses Williams 
aptly cites a couplet : 

" Anhuddwyd aelwyd ; Duw a welo 
Dodi un heddyw a'i dadanhuddo ". 

101 Gir. Camb. vi. 184 (Descr. i. 10) says " igne sicut die, sic et nocte tola, 
ad pedes (concubantium) accenso". Cf. Ven. II. i. 31 ; Lat. A. II. xix. n ; Lat. 
B. II. xxiii. 33. 

102 The subject in general is discussed in Coulanges, La Cite Antique and 
W. E. Hearn, The Aryan Household (London, 1870), its bearing on " dadanhudd " 
in Trib. System, pp. 81-3. There is no authority for the form " dadanhudd " and 
little warrant for the translation " reuncovering ". 


Another matter which came under the notice of the court CHAP. 


of the cantref was the determination of boundaries. 102 It is to 
be observed, however, that more was involved in this than the 
ascertaining of the facts. Facts were, in an early Welsh court, 
often quite subordinate to status, and the mere will of the 
landowner of " higher privilege " was enough to give effect to 
his desire to extend his boundaries at the expense of a less 
privileged neighbour. Thus the Church meered against the 
king, the king against his subjects, the uchelwr against the 
aillt, the aillt against the alltud. Fable reflects the spirit of the 
law when it tells how the young prince Geraint of Cornwall 
traversed the borders of his realm with a company of guides, 
the best men of the land, and " the furthest mark which was 
shown him, this he ever seized on as his own ". 104 The law for- 
bade the carrying of the custom to ridiculous and violent ex- 
tremes, but the love of encroachment remained deeply rooted 
in the national disposition ; " the digging up of boundary 
ditches, the removal of landmarks, the outstripping of bounds, 
the occupation and extension of lands by hook or by crook," 
says Giraldus Cambrensis, " are a passion with this people be- 
yond any other race." 106 

There was much that was primitive in the criminal law ad- 
ministered in the cantref court, while at the same time modern 
conceptions of crime as an offence against the well-being of the 
whole community and of penalty as a matter for the state and 
not for the injured individual had made considerable headway. 
The law of galanas, discussed in the first section of this chapter, 
retained most of the ancient leaven, being little more than an 
attempt to regulate and keep within bounds the primitive right 
of revenge. Where, however, there was no question of loss of 
life, offences were for the most part dealt with, not on the basis 
of tribal custom, but on the assumption that they touched the 
honour and prestige of the king, for whose broken peace 
due atonement must be made. Fighting, violent seizure of 
another's goods, attacks on the honour of women, involved 
the payment of a "dinvy " or special fine of twelve cows ; a 
" camlwrw " or fine of three cows was levied for a large number 

108 For the law of meering see LL. i. 196, 536, 542, 762, 764, 774; ii. 76, 
90, 148-50. 

104 Mob. 268. Gir. Cam. vi. 211 (Descr. ii. 4). 

VOL. I. 2O 


CHAP, of minor offences, such as contempt of court, neglect of official 
x> duty, unauthorised meddling with another's property. Penal- 
ties were imposed not only upon principal offenders, in case of 
murder, theft or house-burning, but also upon those who in one 
way or another were sharers in their guilty enterprises ; this 
inclusion within the scope of the law of the " affeithiau " 106 or 
"affections" of a crime was no doubt an innovation due to 
some early reformer, for that it was of recent origin is shown 
by the fact that no part of the fine exacted from an accomplice 
to murder went to the relatives of the murdered man. 107 
Theft, as was usual in primitive communities, was punished 
with great severity ; it was taken for granted that the thief 
caught with the stolen property in his possession would fall a 
just victim to his captor's rage, 108 and the law only interposed 
to secure a fair trial of the issue whether there had been theft 
or not ; upon condemnation the flagrant thief was ignomini- 
ously hanged. 109 The suspected thief who did not succeed in 
legally establishing his innocence was not executed, but escaped 
on payment of a heavy fine. As was inevitable, the adaman- 
tine rigour of the law of theft led to very careful definition of 
the crime. It was distinguished from violent seizure, which 
was esteemed comparatively venial, from taking in ignorance, 
and even from taking without permission, provided that there 
was in this last case no attempt at concealment or denial. 110 
Finally, a remarkable provision in some of the codes exempts 
from the doom of theft the starving man who, after begging 
for three days, and receiving nothing, helps himself to the food 
which he needs to keep him alive. 111 That every man had the 
right to live was a principle of the law/and the sentiment of the 
country demanded that every person of substance should keep 
open house, not only for ordinary travellers, but also for the 

108 A loan-word from the Lat. affectus, which had in law the sense of 
" animus, consilium". 

107 So Ven. III. i. n, where the question is argued on grounds of principle 
against the " rey " whose view is doubtless embodied in Dim. II. iii. 13. 

108 Cf, Maine, Ancient Law, tenth ed. (1885), pp. 378-9. 

109 The captured thief, apparently a mouse, but really a wizard in that shape , 
is destined for the gallows in the story of Manawyddan (Mab. 54-5). 

110 " Lledrad," " trais," " anoddeu " and " anghyfarch " are denned in Ven. 
III. ii. 52-5 ; Dim. III. vi. 23-6. There was no penalty for "anoddeu " and only 
a camlwrw for " anghyfarch ". 

111 Dim. II. viii. 94 = Lat. B. II. x. i. 


destitute and the friendless. Giraldus did but state the posi- CHAP. 


tion in round, set terms when he said : " Beggars are unknown 
among this people, for all houses are open to every one alike. 
Liberality, especially in the form of hospitable entertainment, 
is deemed by them to be the chief of virtues." 112 

It is scarcely necessary to say that criminal procedure ex- 
hibited little of that careful sifting of evidence which is char- 
acteristic of a modern law court. The business of the cantref 
authorities was to get the parties face to face and then to com- 
mit the issue to the arbitrament of the powers above, invoked 
on either side by solemn oaths. Every charge was made upon 
oath and, broadly speaking, had to be met in the same way, 
testimony as to the facts being usually not admitted. But the 
single oath of the accused was rarely sufficient to rebut the 
charge ; the presumption was too strong against him, and he 
was required to clear himself by means of a " rhaith," or jury, 
who supported him in his protestation that he was innocent. 
In South Wales the "rhaith gwlad" or "jury of the country" 
upon which an accused person had to rely was apparently 
chosen from among the members of the court 113 and numbered 
no less than fifty men. What had originally been an appeal 
to heaven on the part of the man's kinsmen had become some- 
thing very like an appeal to the public opinion of the cantref. 
The religious character of the proceedings was, however, never 
entirely forgotten ; oaths were sworn upon relics of the saints, 
and, as special penalties were believed to be in reserve for the 
men who by perjury dishonoured these visible symbols of a 
power not yet defunct, there was great anxiety to secure and 
tender to the opposing party the relic which had the greatest 
fame in the district as the avenger of any false oath that might 
be sworn upon it. 114 Gospels were of less account in this re- 
spect than such precious objects as the torque of St. Cynog 
or the bell of St. David kept at Glascwm, for the saints of 
Wales, like those of Ireland, were held to be pitiless in venge- 
ance when their ire was kindled by an indignity. 116 

The system of civil and criminal law which has been lightly 

u Gir. Camb. vi. 182 (Descr. i. 10). us Dim. II. viii. 135, 136. 

U4 For the seeking of relics see Ven. II. xxxi. 6; Dim. III. vi. 19; Gw. II. 
xxxvii. i. 

m Gir. Camb. vi. 26, 27 (Itin. i. 2), 130 (ii. 7). 

2O * 


CHAP, outlined in this section and limits of space will not allow the 
IX- treatment of the subject on a more extended scale was not 
confined in its working to the cantref court. Its rules were 
observed in the supreme court, in which the " ynad llys " or 
royal judge decided the disputes of the circle surrounding the 
king, in the courts held by the " maer " and the " canghellor " 
for the villein hamlets of the cantref, 116 and in those, altogether 
exempt from royal jurisdiction, in which bishop and abbot dis- 
pensed justice to their tenants. Yet it may well be believed that 
the court of the cantref, representing the ancient tribal freedom, 
was its true source and that other courts borrowed from this, 
the oldest, their ideas as to law and legal procedure. 


From the earliest period at which it is possible to study 
the organisation of the Welsh tribes, they were under the rule of 
chiefs or kings, whose power was substantial and unquestioned, 
backed as it was by the possession of troops. Of Ireland it 
has been said that it is " doubtful whether the public force at 
the command of any ruler or rulers was ever systematically 
exerted through the mechanism of Courts' of Justice ". m The 
evidence for Wales is in quite the opposite direction ; as has 
already been pointed out, the kings of Gildas were energetic 
in the administration of justice, 118 and in the codes the figure 
of the chief or " arglwydd " is everywhere the dominant one, 
securing the observance of the law and profiting heavily by its 
penalties. This much had been done for the Celts of Britain 
by the spectacle of the majesty of Roman justice ; they never 
abandoned the conception of the magistrate as one who " bear- 
eth not the sword in vain ". The law of Hywel was compiled 
at the instance of a chief; it began with and discussed on the 
amplest scale the rites and customs of the court ; it protected 
the royal interests at every turn. No satisfactory view of the 
social and political institutions of the early Welsh is, therefore, 
to be obtained without taking into fullest account the position, 
the privileges and the power of the king. 

116 The court of the "maerdref" appears clearly in Ven. I. xxxiv. 8; xxxvi. 
10; II. xx. 3, and it may be inferred from II. xviii. i that there were similar 
courts, under the " maer " and the " canghellor," in the other unfree hamlets. 

117 Maine, Early History of Institutions, third ed. 1880, p. 43. 

118 P. 128. 



The head of the tribe was no longer known by the old Celtic CHAP, 
title of rex or " rhi " ; 119 he was the " brenin," the " high " or 
" noble " one, I2 the " arglwydd," the " lord " or " master," and less 
commonly, the " tywysog," the " captain " or " leader ". m He 
might be the lord of a single cantref, playing with all ceremony 
and dignity the part of a monarch on that narrow stage, or 
master of the fifteen cantrefs of Gwynedd, with a principal seat 
at Aberffraw. No distinction is made by the codes in respect 
of title or ordinary legal status between the great and the little 
chief; all are " brenhinoedd " and "arglwyddi" alike, and the 
supreme rulers of Deheubarth and of Gwynedd are only dis- 
tinguished from their fellows by the larger amount due to them 
as compensation for dishonour. 122 The number of minor chiefs 
is. a standing characteristic of Welsh political history down to 
the age of the loss of national independence ; while the success- 
ful ambition of the bolder princes, bent on conquest and 
aggrandisement, was a force which ever tended to reduce it, 
there was another which no less persistently worked in the 
direction of increase. This was the habit of regarding the 
kingdom as an ordinary estate, liable on the death of the owner 
to division among all his sons. 

No rule is given in the codes on the important question 
of the succession to the crown. The one point which they 
make clear is that the matter was settled, not on the occurrence 
of a vacancy, but in anticipation, during the lifetime of the 
reigning chief. The " edling," 123 or next successor, held a 

119 Urk. Spr. 230. " Rhi " appears in poetry only, and the feminine " rhiain " 
usually denotes " dame, lady " (cf. however, the " rieingylch " of Gw. II. xxxv. 2). 

180 Urk. Spr. 171 ; Celt. Br. (2), p. 282. 

m Urk. Spr. 269. Though " tovisaci " appears in an early Welsh inscrip- 
tion (Inscr. Chr. No. 159; W. Ph. (2), p. 372; Arch. Camb. V. xv. (1898), 373-7), 
it is a title unknown to the laws, and appears to have come into ordinary use as a 
translation of the Lat. princeps. " Teyrn " (from " tegernos," Urk. Spr. 126) is 
occasionally found (LL. i. 342, 660, 678). 

122 For the use of "brenin" and " arglwydd " interchangeably see especi- 
ally Ven. II. vi. 28 ; the " guastraut arglyides " of Lat. C. I. i. 24 is the " gwast- 
rod y frenhines " (queen's groom or squire) of the other codes. Dim. I. ii. 5 is 
decisive as to the use of " brenin " for chiefs of the second rank. 

183 A loan-word from the English " aetheling " = one of noble or royal 
birth. The native term " gwrthrych " or " gwrthrychiad " is occasionally found ; 
see Ven. I. v. i (MSS. B. D.) ; Dim. I. v. i ; Gw. I. xiii. 2 ; Lat. A. I. v. (heading) ; 
Mab. 105. " Gwrthrych " is literally " what is looked at, an object " ; the edling 
probably got the name from his place in the royal hall, where he sat facing the 

3 1 o HIS TOR Y OF W 'ALES. 

CHAP, prominent and honoured position in the court of every king ; 
he had no land of his own, but was maintained by the royal 
bounty, being the chief personage in that motley company of 
officials, troopers, menials, youths and vagrants which made up 
the king's recognised following or " gosgordd ". He was of 
the king's near kin, is described, indeed, as his son, his brother 
or his nephew, but it is nowhere said in the ancient authorities m 
that he was of necessity the eldest son, and might only be a 
brother, if sons were wanting. It is, therefore, possible that 
the edling, like the Irish tanist, at one time obtained his position 
by election and did not step into it by mere right of birth. 
On the other hand, it should be remembered that, owing to 
the existence of the office of " penteulu," or captain of the 
guard, one usually important reason for making the monarchy 
elective, viz., the military needs of the tribe, did not here apply. 
And, if regard be had, not to the legal authorities, but to the 
historical facts, it will appear at once that what was customary 
was the succession of the eldest son to the principal part of 
the royal inheritance, with the assignment to younger sons of 
certain cantrefs of less importance. The completeness of the 
cantref in itself, as a political and economic area, made this a 
fatally easy policy, and the tendency to division was further 
encouraged, as Giraldus points out, 125 by the institution of 
fosterage. In accordance with ancient custom, 126 royal infants 
were not brought up at the court, but at a very tender age 
were placed under the care of foster-parents, to whom they 
became united by lifelong ties of affection. "Foster-father" 
and " foster-brother " were terms of respect and endearment, 
while kings' sons who were brethren by blood grew up strangers 
to each other, knowing each other as brothers only by repute. 
Moreover, as the foster-parents were uchelwrs of wealth and 
influence, each one a power in his own cantref, every claimant 
for a share in the division of the realm had behind him the 
weight of the support of some locality, desirous of honour for 
its favourite prince. 

124 LL. ii. 304 (xxxix. 2) is from Peniarth MS. 36^, which is not earlier than 
the end of the fifteenth century (Evans, Rep. i. p. 370). 

125 Gir. Camb. vi. 211-2 (Descr. ii. 4). 

126 The ancestors of St. Samson are described as " altrices regum " ; see 
chap. v. note 107. 


The privileges which fell to the lot of the successful candi- CHAP. 


date for the dignity of " brenin " were many and various. In 
the figurative language of the law, he had eight " pack-horses," 
or agencies which, without effort or trouble on his part, brought 
wealth and laid it at his feet. 127 One was the sea, for any 
wreckage which its waves 'might bring ashore was accounted 
royal spoil. 128 Another was the waste, the land claimed by no 
tref and subject to no kind of occupation ; this he might grant 
to alltuds for settlement in return for a fee. 129 Another was 
the thief who, not having been caught with his theft in hand, 
was allowed to pay a price for his neck ; that price was forfeit 
to the king. Another was the offender whose crime was 
expiated by the payment of a " dirwy," or the lesser offender 
who made his peace with a " camlwrw " of three kine ; " dirwy " 
and " camlwrw " alike went to swell the wealth of the chieftain. 
Every man and every woman not subject to a lesser lordship 
paid to the crown a fee of specified amount, the woman on 
marriage, the man on succeeding to land or other property; 
the man's "ebediw" (often treated as a death duty) and the 
woman's " amobyr " were usually of the same amount, and it was 
a rule of law that neither was due from the same person more 
than once. 130 Most sweeping claim of all, the " brenin " was, 
in respect of movable property, treated as the universal heir 
in default of children. While the dead man's land might be 
claimed by any relative within the degree of second cousin, 
none but a son or a daughter could touch his cattle and house- 
hold belongings, and if he had left no child, the " rhingyll " of 
the cantref forthwith made his appearance and seized the whole 
as the "marwdy" (house of death) of the lord. In South 
Wales, the influence of the Church had secured recognition of 
the right of a dying man to make bequests out of his property 

127 Yen. I. xliii. 12 = Dim. II. xi. 2 = Lat. A. II. xii. = Lat. B. I. xx. 

128 Yen. II. xvi. 6 ; Dim. II. xxiii. 37, 38. 

129 The waste is always the king's (in Latin " desertum regis ") and there 
is no trace of any authority exercised over it by the community or tribe. See 
Yen. II. xvi. 21; xviii. 7; Dim. II. xii. 9; Gw. I. xxxv. i; Lat. A. I. xxvii. 2; 
Lat. B. II. xxiv. 19 ; xl. 21. 

130 yen. II. i. 55. The "ebediw" is the heriot of Anglo-Saxon law, as 
may be seen from the case of the " penteulu " (Yen. I. vii. 23) and from the fact 
that the portion of the ebediw of a king's villein anciently payable to the lord 
was sixty pence, i.e., the value of an ox (Lat. A. II. xxii. 8). For the equality 
of " amobyr " and " ebediw " see Lat. B. II. xxxix. 5 ; LL. ii. 574. 


CHAP, on his death-bed, and thus the lord only got a " marwdy " in 

IX ' case of sudden death. In Gwynedd it was not open to a man to 

bequeath anything he had, and even the goods of a bishop, 

save his books and official vestments, passed on his death into 

the royal hands. 181 

These manifold sources of income, being casual and 
uncertain in their nature, were not to be relied upon for the 
daily maintenance of the chief and his retinue. This was pro- 
vided by a system of food renders, which were brought as 
tribute by free and bond subjects alike for consumption in the 
royal court of the cantref ; the king had also, it should be 
added, his own demesne land, tilled for him by the aillts of 
the maerdref or hamlet of the court, and considerable use was 
made of the right of " cylch " or free quarters. There may 
have been a time when the king himself went on progress 
among his people, who received him in hospitable wise in 
their own homes ; 132 in the laws, however, there is no mention 
of a royal " cylch " of this description, and only the members of 
the court appear as going on circuit. The queen, the captain 
of the guard with his men at arms, itinerant bards from 
another country, the royal horses, dogs, and hawks with their 
custodians, were at various seasons sent round the cantrefs, 
to quarter themselves chiefly upon the villeins. But the 
" gwest " 133 or entertainment of the king, if it ever took this 
form, had in historical times assumed a different aspect. The 
cantref was divided into a number of trefs, of which some 
were bond and some free ; each tref was made jointly respon- 
sible for the render twice a year to the royal residence in the 
cantref of a specified quantity of food and provender. The 
tribute of the free tref was known by the honourable name of 
" gwestfa," or entertainment due ; it was a " gwledd," or 
feast, 134 provided by a " cwynosog," or supperer, 135 who prob- 

131 The Venedotian rule appears clearly from Ven. II. i. 13 ; for its special 
application to bishops (but not abbots, whose goods passed to their convents), 
see LL. ii. 10 and the St. Asaph gravamina of 1276 (H. and St. i. 512). The 
South Welsh custom (of which there are traces in the north : cf. Ven. II. xxviii. 9, 
MS. D, and H. and St. i. 513, No. 7) is mentioned in Dim. II. viii. 62. " Cym- 
ynnu " is the Lat. commendo. 

132 This is suggested by the Mab. phrases "cylchu Dyfed" (46) and 
" cylchu ei wlad" (59) used of Pryderi and Math respectively. 

133 Used for " cylch " in Gw. I. xv. 13. 

134 Ven. I. ix. 25. 135 Gw. II. xxxix. 45 ; xl. 8. 


ably attended to represent his tref on the evening when it CHAP, 
was consumed. 18 " Its chief constituents were flour, beef, mead 
or other liquor of the kind, and oats for the horses ; 137 in later 
times it might, instead of being paid in its original form, be 
commuted for the sum of one pound, and it is of interest to 
note that in 1352 the render of the free vill of Gloddaeth in 
Creuddyn to the Prince of Wales still stood at this amount. 188 
The villein trefs, notwithstanding the extent to which they 
were burdened by the system of progresses, were not free from 
the liability to provide food ; their contributions were known 
as " food gifts " and consisted chiefly of meat (mutton in 
summer, pork in winter), cheese, butter and loaves of bread. 139 
Thus the whole cantref worked together for the mainten- 
ance of the royal court which formed its centre. This was 
the " llys " or " castell," at one time also called the " manor " 
or " maenol ". 14 Here were the king's summer pastures, 
ranged over by his great herds of cattle, the produce of fine 
and forfeit and border foray. Here was the "board land," 
tilled for the service of the court by the only Welsh villeins 
who, like those of mediaeval England, gave their labour as the 
rent of their holdings. Here was the " maerdref," the hamlet 
in which they lived together at the castle gates, ruled by the 
" maer biswail," the " dung bailiff," whose epithet contemptu- 
ously distinguished him from the great "maer" or bailiff of the 
cantref. And here was the group of buildings, enclosed within 
a strong wall, which constituted the royal palace of the Welsh 

138 Cambro-Br. SS. 156 mentions a dean of the " das " of S. Woollo's, 
Newport, who "visitavit curiam Lisarcors apud inferiorem Guentoniam, con- 
vivio regali functus ; sic consuetudo erat tune temporis per patriam ". 

137 The pork and bacon of Ven. II. xxvi. i do not find a place in the free 
renders of South Wales. 

138 Rec. Cam. i. 

139 The Book of St. Chad furnishes two early instances of the renders of 
servile trefs ; see Lib. Land. xlv. and chap. vii. note 105. The transference of 
the lordship from the tribal chief to the Church in these cases, typical of scores 
of others, did not, of course, affect the obligation of the tenants. 

140 "Maenor" (in Gwynedd, "maeno/") is a word of undoubted native 
origin, occurring in the marginalia of the Book of St. Chad (Lib. Land, xlvii.). 
It has nothing to do with the English manor, but is connected with " maen " 
(stone) ; probably it was first applied to the stone-girt residence of the chief, so 
as to distinguish it from the ordinary tref. Thence were derived the later mean- 
ings (i) group of villein trefs (Lat. A. II. xiii. 9) ; (ii) a division of the cymwd in 
Gwynedd, scarcely to be distinguished from the ordinary tref (Ven. II. xvii.). 
For fuller discussion see Cymr. xi. 32-4, 57-8 ; W. People, pp. 218-9. 


CHAP, chief. The testimony of Giraldus is hardly needed to convince 
us that the Welsh of his day had no lofty stone-built towers or 
stately halls ; 141 for the language of the codes and the poverty 
in remains of such sites as Aber and Aberffraw show indubit- 
ably that all the buildings of the " llys " were timber structures, 
more substantial, no doubt, than those of the ordinary free 
household, yet essentially temporary and of little value. 
Security was not neglected ; the castle wall had a gate, 
narrowly watched by a porter, whose house was just behind 
it ; the place was strong enough to be used for the custody of 
prisoners ; U2 and all night long, from the evening horn-blast 
until the opening of the gate in the morning, the watchman, 
a freeman of the country who had no daylight responsibilities, 
kept a close vigil and guarded the castle from nocturnal attack. 
But the buildings so protected were of the simplest kind. 143 
The chief was the " neuadd " or hall, an oblong structure rest- 
ing on six wooden uprights, of which two were placed at the one 
end, with the door between them, and two at the other ; the 
central couple, having between them the open hearth, divided 
the hall into an upper section, or " cyntedd," where the king 
and the greater officials sat at their meat, and a lower section, 
or "is cyntedd," assigned to the less distinguished members of 
the royal train. Next in importance was the chamber, or 
" ystafell " ; this was the king's private apartment, where he 
passed the night, 144 and it was also the queen's ordinary day- 
room, since it was not in accordance with ancient custom for 
women to join in the festivities of the neuadd. The outhouses 
or subordinate structures included a kitchen and larder, a 
mead brewery, a kiln or drying-house, a stable, a barn and a 
dog-house. All these the villeins of the cantref were to put up 
and keep in repair, a further proof, if one were needed, that they 
were of light construction and called for no exercise of the 
builder's art. 

The king was served by a number of officers, whose duties 

141 Gir. Camb. vi. 200-1 (Descr. i. 17). 

142 Yen. I. xxxvi. 6; Dim. I. xxii. 2; Gw. I. xxxix. 7; Lat. A. I. xviii. 7; 
Lat. B. I. xxi. 16. 

143 In Mab. 46 mention is made of "teiyllys" at Arberth and in parti- 
cular of the " neuadd," " ystafell " (so Wht. Bk.), " hundy," " meddgell " and 
" cegin ". 

144 i Ystafell y brenin yr hon y bo yn cysgu ynddi " (Yen. I. xi. 4). 


and privileges were carefully defined by the laws. Each had CHAP, 
his appointed seat in the neuadd when the court was gathered 
together for the evening's festivity, each his proper lodging for 
the night in or about the castell, each his horse from the royal 
stable and his clothing from the royal store. The most re- 
sponsible was the judge of the court, who decided all the dis- 
putes of the court and was, moreover, the king's perpetual 
counsellor, always with him, and required to be always sober. 145 
He was admitted to his office at a solemn ceremony held in 
the chapel of the court, when he swore to do justice ; he then 
received the symbols of his dignity, a throw-board from the 
king and a gold ring from the queen. 148 The judge of the 
court was the official examiner of all candidates for minor 
judicial posts in Gwynedd ; 147 the great gate of the castle was 
thrown open to receive him, and he slept at night either in the 
king's chamber or with his head on the cushion on which the 
king had sat in the neuadd on the previous day. Thus was 
concrete expression given to the idea, which redounds not a 
little to the credit of a warlike people, that "judgeship is the 
greatest of all temporal things ". u8 Other officers of the first 
rank were the " offeiriad teulu " or court priest, who was the 
king's secretary and the incumbent of the royal chapel, the 
" distain " or steward, 149 who was in supreme control of the 
castle, its furniture and the store of food and drink within it, 
and the " gwas ystafell " or chamberlain, who looked after the 
ystafell, guarded the royal treasure, and was the king's special 
messenger. A chief falconer, a chief huntsman, and a chief 
groom, each with a troop of underlings, had custody of the 
hawks, the dogs, and the horses which ministered to the out- 
door recreations of the court. For the minstrelsy of the 
neuadd the " bardd teulu," or court poet, was responsible, 
though the pencerdd, or chaired bard, the head of the bardic 
fraternity in his country, might occasionally be present and in 
that case would take the lead. There were also attached to 
the court a physician, a mead-brewer, a cook, a doorkeeper, a 
candle-bearer, a smith, a woodman (who replenished the hearth) , 

146 Ven. I. xliii. i ; Dim. II. viii. 19, no. 148 Dim. I. xiv. 20, 21. 

147 Yen. III. preface. 148 Gw. I. xiii. 29. 

149 The "distain "is the English " discthegn " (Cod. Dipl. No. 715 cf. 
Kemble, Saxons in England, ii. p. 109). 


CHAP, and a cupbearer. The footholder held the king's feet in his 
lap during the evening's revelry and guarded him from mis- 
chance during the riotous hours of carousal ; the silentiary, who 
through the day was the distain's agent and deputy, struck the 
pillar opposite the king, and from time to time demanded 
silence ; the rhingyll, the usher of the commote court, had also 
duties in the neuadd, for it was his business to stand in front of 
the hearth and protect the timber-built hall from the destruction 
by fire in which the recklessness of the festive company might 
but too easily involve it. 

But beyond a doubt the most important element in the con- 
stitution of the court was the " teulu ". In modern Welsh this 
word signifies " family," and there has been in consequence a 
very general failure to understand the nature of the old institu- 
tion so termed, efforts being made to interpret it in terms of 
kinship. 150 As a matter of fact, members of a royal " teulu " 
were completely divorced from family life, for the word has in 
this connection its original meaning of " house-host," 151 and the 
" gwr ar deulu " was a guardsman or trooper, belonging to the 
royal bodyguard or standing company of household troops. 152 
The teulu was a body of horsemen, fed, clothed and mounted 
by the king and in constant attendance upon him. 153 It was 
commanded by the " pen teulu," or captain of the guard, who 
was always a near relative of the reigning chief and in his ab- 
sence took his place in the hall. 154 The company might 
number as many as one hundred and twenty, 155 and it is easy 
to see that the possession of this little force, constantly under 
arms, was a source of great strength to the lord of a cantref in 
his dealings with his subjects. All the more was this the case 
in that the loyalty of the teulu was no mere affair of contract ; 

l50 Camb. Reg. i. 205. 

151 Gr. Celt. (2), p. 1068 ; Urk. Spr. 321. " Gosgordd " had anciently the 
same sense; see Gr. Celt. (2), p. 1067 (Familia, goscor pi teilu). 
162 Ven. I. vii. 13 makes this definition certain. 

153 Ven. I. vii. 16, 22 ; LL. ii. 68 ( 69). 

154 In Mab. 144 Madog ap Maredudd of Powys offers his brother lorwerth 
the office of " penteulu ". Their father Maredudd appears to have been " pen- 
teulu " to his nephew Owain ap Cadwgan in 1113 (Bruts, 291). None but very 
late authorities use the term to denote an ordinary pater familias. 

158 This was the number of each of the three faithful and of the three faith- 
less warbands (Mab. 305) ; Gruffydd ap Llywelyn had a warband of 140 in 1047 
(Trans. Cymr. 1899-1900, 132-3, 168-9). 


the bond which united its members to their lord was one of CHAP. 


sentiment and honour, and it was held to be their duty to cling 
to him in the hour of his sorest need, to die in his defence and, 
if need should be, even in his stead. " Three faithless warbands 
of the Isle of Britain," runs an ancient triad, " the warband of 
Gronw the Radiant of Penllyn, who refused to receive instead 
of their lord the poisoned shaft aimed by Llew Llaw Gyffes ; 
the warband of Gwrgi and Peredur, who deserted their lords 
at Caer Greu, and they under bond to fight Eda of the Great 
Knee on the morrow, in which fight they were both slain ; and 
the warband of Alan Fergant, who stealthily left their lord on 
the road to Camlan the number of each of these warbands 
was one hundred and twenty men." 168 In the teulu, in short, 
is to be found the Welsh representative of the English warband 
or company of gesiths ; one may go still further back and say 
it embodies the spirit of the "comitatus," as described by 
Tacitus in his account of the Teutonic foreworld. In Wales, 
as in the primeval forests of Germany, it was a lifelong disgrace 
for the warrior-client to return alive from the battle-field on 
which the master lay dead ; to defend the life and honour of 
one's lord, to make all one's own achievements merely tributary 
to his renown, was the holiest of obligations. 167 

It is abundantly clear that the teulu was not merely used 
by the king to suppress domestic disorder and to repel ex- 
ternal enemies. A plundering expedition into a " gorwlad " or 
neighbouring kingdom is treated by the laws as an ordinary 
incident in the routine of its duties, and minute directions are 
given for the division of the spoil, which, it is assumed, will be 
chiefly in the form of cattle. 158 Once a year, indeed, for a 
period of six weeks, the king might lead, not the teulu merely, 
but all the men of his realm upon a warlike incursion into some 
distant land ; even the aillts sent from each villein tref a man 
with an axe and a pack-horse to make rough quarters for the 
host 159 Yet life was not quite so hazardous and insecure as 

188 Triad i. 35 = ii. 42 (Mab. 305) = iii. 81. The story of Gronw will be 
found in Mab. 80 ; "Eda Glinfawr (or gawr) " is Eata, father of Archbishop Egbert 
(Hist. Britt. c. 61), whose association with Gwrgi and Peredur (ob. circa 580) is 
an anachronism. 

157 Tacitus, Germania, cap. xiv. 

158 Ven. I. vii. 14, 18 ; xi. n ; xiv. 7; Dim. I. xvii. n ; x/iii. a. 

189 Ven. I. xliii. 15 ; II. xix. 7, n ; Dim. II. xi. 5, 6 ; Gw. II. xxxv. 2, 7 ; Lat. 
A. I. xxvi. i ; Lat. B. I. xviii. 12, 17. 


CHAP, these plans for systematic border warfare would suggest. 
The lands of the church enjoyed unbroken peace, and, as they 
were extensive and often lay along the borders of cantrefs, the 
points at which a conflict might take place were not so numer- 
ous as might at first be supposed. Moreover, it was not the 
fashion to fight at all seasons of the year, for instance, in the 
middle of winter, when the weather was cold and cattle were out 
of condition. 160 Such limitations as these explain the popular 
saying quoted by Giraldus, that the eagle knows the place where 
it may find its booty, but not the time, while the raven, that 
other grim satellite of war and carnage, can tell the time, but 
knows not the place. 161 

NOTE TO CHAPTER IX. The Triads of Dyfnwal Moelmud. 

In the royal pedigrees in Harleian MS. 3859 a " Dumngual moilmut " ap- 
pears as a grandson of Coel Hen and grandfather of " Morcant bulc," who may 
be the " Morcant " of Hist. Britt. c. 63. This reference makes it likely that there 
was a historical person who bore the name, a Northern British prince flourishing 
about A.D. 500. The next mention of Dyfnwal Moelmud is in the Historia 
Regum of Geoffrey of Monmouth, who makes him the son of Cloten, king of 
Cornwall, and ruler over all Britain about 450 B.C. Dyfnwal is represented as 
the author of the " leges quae Molmutinae dicebantur . . . quae usque ad hoc 
tempus inter Anglos celebrantur " ; they dealt with such matters as the right of 
sanctuary in temples and the peace of the great roads, and were translated from 
the original British into Latin by Gildas and by King Alfred from Latin into 
English (Hist. Reg. ii. 17 ; iii. 5)- It is surely a mistake to find in these par- 
ticulars any echo of an old tradition current among the Welsh as to Dyfnwal, 
which handed down his name as the primeval British legislator ; the " Molmu- 
tine " laws of Geoffrey are in repute, not in Wales, but in England, and are ex- 
tant in the English tongue. A much more natural explanation presents itself ; 
Geoffrey, who undoubtedly used the old Welsh pedigrees as a quarry for proper 
names, found there his " Dunwallo Molmutius " and was subsequently led, by 
some haphazard similarity of names, to ascribe to him certain old English laws 
of the same general character as those embodied in " Leges Henrici Primi " and 
similar compilations. Such a course would have been quite consistent with his 
literary methods, as witness his assignment of the Mercian law (" Merchenelage " 
in the Berne MS.), also said to have been translated into English by Alfred, to 
an early British queen of the name of Marcia (iii. 13). 

In the Codes there is no mention of Dyfnwal, save in one passage (Ven. II. 
xvii.) of the Venedotian Code, where he appears as the author of the institu- 
tions, and, in particular, of the measures of length and of area in force in the 
island until the age of Hywel Dda. It is pretty certain that this Code was com- 
piled about the beginning of the thirteenth century, in an age when the narrative 
of Geoffrey had become the common literary property of Western Europe, and 
when one finds Dyfnwal described in it as "mab iarll Cernyw," the conclusion 

160 The teulu left the king and went on progress after Christmas (Ven. I. vii. 

161 Gir. Camb. vi. 136 (Itin. ii. 9). 


is inevitable that the lawyer's ideas on this subject were derived, not from native CHAP, 
tradition, but from the pages of romance. Genuine Welsh tradition knows, in IX. 
fact, nothing of this shadowy figure but his name (cf. Mab. 109) ; he is not 
mentioned in the older triads and first makes his appearance in the untrustworthy 
third series (Nos. 4, u, 36, 57, 58). 

In 1807 the editors of the Myvyrinn Archaiology printed in their third 
volume (pp. 283-318) " Triodd Dyvnwal Moelmud, a elwir Triodd y Cludau a 
Thriodd y Cargludau," to the number of thirty-four, followed by a much more 
bulky collection of " Triodd Gwladoldeb a Chywladoldeb," numbering 248. The 
MS. was a transcript made in 1685 by Thomas ab Ifan of Tre Bryn, near Coy- 
church in Glamorgan, "o hen lyfrau Syr Edward Mawnsel o Vargam" (1634- 
1706). Its originals are not now known to exist, and can in no case have been 
much older than the sixteenth century. In language and spirit the triads are 
thoroughly modern, and everything goes to show that they are the work of some 
Glamorganshire antiquary who at the close of the Middle Ages adopted this 
vehicle for the expression of his political aspirations, delineating the Wales 
that he desired to see under the guise of a description of a golden age in the 
prehistoric past. The writer had obviously a considerable knowledge of the old 
Welsh system, but in his day it was largely obsolete and accordingly he often 
goes astray in a manner which would not have been possible for a genuine Welsh 
lawyer. A conspicuous instance is the way in which he confounds the " alltud " 
and the " aillt," e.g., in No. 66, "aillt neu estron a wladycho yng Nghymru ". 
The statements made as to certain persons having a " spear allowance " for 
their support (Nos. 166, 199, 200, 239, 240) receive no warrant from any other 
authority and seem to rest on a misunderstanding of the " ceiniog baladr " or 
" spear penny " which could be claimed in certain cases by those who were mak- 
ing up the amount of a " galanas " fine. A similar mistake is probably at the 
root of the theory advanced that every Cymro was entitled to five free erws (Nos. 
61, 65, 68, 70, etc.) ; No. 83 adds " cyfar gobaith " or " co-tillage of the waste," 
but neither the erws nor the co-tillage find any place in the genuine old docu- 
ments and it is probable that the former, ample as is the space they occupy in 
the compiler's ideal reconstruction, have been evolved out of the four erws of the 
home croft which went with each homestead when a proprietor's land was 
divided on his death. 

It will thus be seen that the Triads of Dyfnwal Moelmud ( = book xiii. in 
LL. ii.) are not only valueless as records of the age before Hywel Dda, but do 
not even furnish good evidence as to ordinary questions of Welsh law and custom 
in a later age. One cannot but feel that to use them, in however guarded a 
fashion, is to introduce an element of unreality into the discussion of the history 
of Welsh institutions, and in the foregoing pages they have been left entirely out 
of account. 



(Asset and the various editions and versions (B.T. and B. Saes.) of the 
Annales Cambriae have supplied most of the material for this chapter. Mr. 
Stevenson's edition of the Vita Mlfredi, Green's Conquest of England, and 
Todd's War of the Gaedhil with the Gaill, among modern books I have used 
with great profit.) 


CHAP. WITH the opening of the ninth century <a new element entered 
x> into the history of Western Europe, which profoundly 
affected the course of events in every country between the 
Baltic and the Mediterranean. The Western Seas, undisturbed 
since the Low German marauders had made their home in 
Britain by the shadow of war or piracy, and fearlessly used as 
a peaceful highway by monastic pilgrims without number, sud- 
denly swarmed with fierce and intrepid buccaneers, whose 
vessels sailed from Scandinavian creeks to i the uttermost parts 
of Europe, and whose piratical impulses were not tempered by 
the least tincture of respect for the Christian religion. The 
advent of the Northman everywhere marked a new era, checked 
the growth of the nascent civilisation which was slowly rising 
out of the ruins of the old, gave the reins of power to prowess 
and physical force, and created the feudal system. No region 
was more thoroughly shaken by the Norse inroads than the 
British Isles ; in few countries did they leave a more lasting 
impression, and it becomes important, therefore, to consider in 
what manner they affected Wales. 

The fact soon becomes evident that, much as the country 
was exposed to the attacks of the Northmen, and long as it 
continued to be in danger from them, their total effect was 
comparatively slight. Further research may yet establish points 



of contact between the Welsh and the old Norse language l CHAP. 
;uiil literature, and even between the institutions of the two 
races, but the salient fact remains that nowhere on Welsh soil 
was there any permanent Scandinavian settlement. History 
records no sustained attempt at anything of the kind, and, if it 
be replied that the contemporary annals are meagre, it may be 
added that the evidence of Welsh place names does but confirm 
the negative conclusion drawn from the silence of the chroniclers. 
There is, it need hardly be said, abundant evidence of the 
familiarity of the sea-rovers with the features of the Welsh 
coast ; headland, islet and harbour still bear the names which 
were given them by the bold navigators who cruised around 
them on their adventurous quests, and the Norse names, through 
the medium of English, are now more widely known than the 
original Celtic ones. Thus the ancient Mon became " Anglesey," 2 
the " island of the strait " ; " Gwales," far out in the Western 
Sea, became " Grasholm," the " grassy islet "; 3 "Abergwaun" 
became " Fishguard," the " place of fish ". 4 Holms, like Priest- 
holm and Flatholm, fiords like Milford and Haverford, wicks 
like Goodwick and Oxwich, eys like Ramsey, Caldy and 
Swansea tell their tale of Scandinavian flitting to and fro among 
the rocks and inlets of the Welsh seaboard. 5 But all traces of 
the Northmen disappear as one leaves high-water mark and 
strikes inland. Even in Pembrokeshire, where the undoubted 
evidences of Teutonic settlement on a large scale have beenicon- 
nected by some writers with the viking movement, it is clear on 
examination of the place names in point that they are much 

1 Prof. Kuno Meyer informs me that there are but two loan-words of un- 
doubted Norse origin in Welsh, viz., " iarll " and " tarian". 

3 See note 99 to chap. vi. Angle, on Milford Haven, also takes its name 
from the strait (old Norse ongull) at the mouth of the Haven (Owen, Pemb. i. 

3 Mab. 40, 41 ; Owen, Pemb. i. 112 ; Lit. Eng. p. 72. The thirteenth- 
century MS. of the De Excidio of Gildas known as Camb. Ff. I. 27 has a marginal 
note stating that the work was written in " Guales insula marina " (ed. Mommsen, 
1 1 ), so that this lonely islet, like many another of its kind, was traditionally as- 
sociated with the early anchorites of Wales. 

4 Owen, Pemb. i. 225 (note by Sir John Rhys). 

6 A list (which, however, requires some revision) is given of the Welsh 
names of this class in Taylor, Words and Places, pp. 117-8 (1888 ed.). The early 
forms of Swansea, viz., Sweinesie (Gir. Camb. vi. 73, reading of MSS. R. B.), 
Sweynesia (ibid. 172), Sweinesham (Ann. Marg. s.a. 1212), Sweynese (Ann. C. 
p. 109), and " Sweynesse " (Tax. Nick. 272), show that it is unconnected with 
" swan " or " sea " and is really the " ey " of some piratical Sweyn. 
VOL. I. 21 


CHAP, more likely to be of Anglo-Norman than of Scandinavian origin. 6 
The prevailing element is the English " -ton " ; no instance is 
to be found of the Northern " -by ". 7 Williamston and Johnston 
and Jeffreyston are intermingled with Gumfridston and Her- 
brandston and Haroldston in such a manner as to forbid the 
assumption that we have to do with a heathen settlement of the 
viking age, and the historian is relieved from the necessity of 
finding a place for an encroachment upon the kingdom of Dyfed 
of which no hint is to be iound in the ancient chronicle of St. 
David's or in the laws of Hywel the Good. 

But, if there was no colonisation, such as took place in East 
Anglia, in the Hebrides, around Dublin and York, there were 
marauding expeditions without number ; in this respect Wales 
had no immunity. The attacks probably began as soon as the 
viking boats had learnt to range over the seas which divide 
Ireland from Great Britain, that is to say, in the closing years 
of the eighth century. If, as seems most probable, the Isle of 
Man was visited in 798, 8 Anglesey cannot have had a long re- 
spite and the monastic communities of Ynys Seiriol, Caer Gybi, 
and Bangor must have early borne the brunt of an attack which 
was specially directed against the defenceless sanctuaries of the 
Celtic coast. The " gentiles," or heathen folk, as they are termed 
in the Welsh annals, are first definitely recorded as having 
relations with Wales in an annal of the year 850, when they are 
said to have slain a certain Cyngen ; 9 three years later Anglesey 
was ravaged by the "black gentiles". 10 Isolated notices like 
these merely afford a glimpse of what must have been going on 

6 Taylor (loc. cit.) and Laws (Lit. Eng. pp. 70-3) argue in favour of a Norse 
settlement, but Rhys and Jones (W. People, p. 27) will allow nothing substantial, 
and the Rev. J. Sephton, in a letter from which I am kindly permitted to quote, 
says that he "can see nothing very distinctively Norse" in the -ton names of 
Pembrokeshire and therefore is " compelled to regard the wicking settlement as 
doubtful ". 

7 Tenby is not an instance, since it is merely a corruption of the Welsh 
Dinbych, well attested as the ancient name of the place. See chap. viii. note 213. 

8 Todd, War ofG. and G. Introd. xxxv. 

9 Harl. MS. 3859 (Cymr. ix. 165) has " cinnen," for which " cincen " should 
no doubt be read, in harmony with Ann. C. MS. B., B.T. and B. Saes. 

10 Harl. MS. 3859 (ibid.). The " gentiles nigri " are the Dubhgaill or "dark 
foreigners" of the Irish Chronicles, who appear upon the scene in 851 and con- 
test with the Finngall or " white foreigners " the possession of Ireland. The two 
sets of invaders are believed to have been Danes and Norsemen respectively 
(Todd, War ofG. and G. Introd. Ixii). 


in Wales throughout the ninth century ; they suggest, without CHAP. 
describing, an era of general devastation and insecurity, the 
detailed history of which can never be written. No doubt there 
were parts of the country which offered little temptation to the 
invader; the rocks of Eryri and Meirionydd, the uplands of 
Ceredigion and Ystrad Tywi gave scanty scope for plunder or 
profitable settlement, and the viking ships seem rarely to have 
visited the broad expanse of Cardigan Bay. 11 But the corn 
lands of Anglesey, so open to attack by sea, the pleasant creeks 
and anchorages of Dyfed, and the fertile regions to which the 
Severn estuary gave easy access were beyond a doubt exposed 
to continual inroads, a few of which have been recorded, while 
scar and holm and wick tell vaguely of the rest. 

In Wales, as in other parts of Western Europe, the attacks 
of the Northmen shook society to its foundations and in par- 
ticular did fatal injury to the work of culture carried on at the 
religious centres, but, as the crisisi produced in Wessex a 
deliverer in the person of Alfred, so among the Welsh it brought 
to the front a new dynasty, which henceforth sways the destinies 
alike of the North and of the South until the extinction of native 


Upon the death of Hywel ap Rhodri Molwynog in 825, the 
direct male line of Maelgwn Gwynedd appears to have come to 
an end. A stranger possessed himself of the throne of Gwynedd 
and the royal seat of Aberffraw. Merfyn Frych (the Freckled) 
was descended from Llywarch Hen ; his father, Gwriad, had 
married a daughter of Cynan ap Rhodri, so that he was not 
altogether without a hereditary claim to the crown, but it was 
a claim which would probably have been of little account had 
it not been backed by personal force and distinction. 12 Merfyn 
came, according to bardic tradition, " from the land of Manaw," 13 

11 No place name of Scandinavian origin is to be found between Bardsey 
and Fishguard, with the doubtful exception of Harlech (W. People, p. 27). 

18 Merfyn is the " Mermini regis " of Hist. Britt. c. 16 ; see note 145 to 
chap. vii. above. His paternal ancestry is given in Buck. Gr. ap C. 30 (721) and 
Jesus Coll. MS. 20 (Cymr. viii. 87) ; for his mother Etthil or Ethyllt see Harl. 
MS. 3859 (Cymr. ix. 169) ; Celt. Folklore, p. 480. 

18 " Meruin vrych o dir manaw " (" Synchronisms of Merlin" in IV. Anc. 
Bks. ii. p. 222; Myv. Arch. I. 141 (no)). 

21 * 


CHAP, and may thus be supposed to have appeared on the scene to 
put an end to the confusion which ensued on the death of 
Hywel, though whether his starting-point was the Isle of Man 
or that other Manaw on the banks of the Forth where it is more 
natural to look for a scion <of Llywarch Hen, must remain an 
open question. 14 He established himself firmly in Gwynedd 
and allied himself to the royal house of Powys by marrying 
Nest, daughter of Cadell ap Brochwel. 15 For nineteen years 
he maintained his power against all rivals and against the 
Danish irruptions, and on his death in 844 he was able to hand 
it on to his son Rhodri, surnamed the Great. 16 

No one will contest the right of Rhodri to a title which he 
earned, not only by strenuous and gallant resistance to the 
northern marauders, but even more by his success in uniting 
the greater part of Wales, so long divided into petty states, in 
a single realm. The kingdom he founded, though it did not 
retain its unity for any length of time, afforded future ages an 
instance of what could be achieved in this direction and set 
before ambitious princes a goal towards which their efforts 
might be directed. How deeply his countrymen were impressed 
by his achievements may be seen from the hold which his 
dynasty acquired upon Wales ; to be of the blood of Rhodri 
Mawr was henceforth the first qualification for rule, alike in 
Gwynedd and in Deheubarth. The story of Rhodri's rise to 
supreme power has not been preserved, for us by any chronicler, 
but the two principal steps may be dated within a year or two 
and partially explained. The first was the acquisition of 
Powys ; it was in 855 that Cyngen, the last of the ancient 
dynasty of that region, died a pilgrim at Rome, whither he had 

14 Skene (IV. Anc. Bks. i. p. 94) favoured Manaw Gododin, as did the present 
writer in the Diet. Nat. Biog. (1894). But the discovery in 1896 in the Isle of 
Man of an inscription of about the ninth century which runs "Crux Guriat " 
(The Cross of Gwriad) undoubtedly strengthens the case for the insular origin of 
Merfyn (Zeit. Celt. Ph. i. 48-53). 

15 Jesus Coll. MS. 20 (Cymr. viii. 87). This account, harmonising as it 
does with Harl. MS. 3859, is to be preferred to the common one, which reverses 
matters, making Nest the mother of Merfyn and Ethyllt or Esyllt his wife. 

18 " Annuscccc. mermin moritur. gueith cetill " (Harl. MS. 3859 in Cymr. ix. 
165). The attempt of Gw. Brut (s.a. 838) to connect these two notices (" Gwaith 
Cyfeiliawc, lie bu ymladd tra thost rhwng y Cymry a Berthwryd (i.e., Burhred) 
brenin Mers, ac yno y lladdwyd Merfyn Frych ") is a good instance of the way 
in which the compiler of this version of the Brut supplied the lack of material by 
pure invention. 


been driven by old age and misfortune ; " if he left sons, which CHAP, 
seems unlikely, they were forthwith ousted by Rhodri, who 
through his mother Nest was the old king's nephew. 18 Defeat 
had perhaps abated somewhat the high spirit of the men of 
Powys and prepared them to accept a deliverer from the fast- 
nesses of Snowdon ; for of late it had gone hardly with them in 
the perennial conflict with Mercia. Mercian greatness was, in- 
deed, at an end, but with the rise of a new power in the South a 
new danger had arisen ; the Mercians, no longer standing in their 
own strength, had begun to invoke the aid of their West Saxon 
overlords, and in 830 Egbert, in 853 yEthelwulf had led armies 
against the Welsh whose victories were no doubt chiefly gained 
at the expense of the border realm of Powys. 19 The second 
acquisition of importance made by Rhodri was that of Seisyllwg, 
the state formed rather more than a century earlier by the 
union of Ceredigion and Ystrad Tywi. This addition to his 
realm must have been made soon after 872, when Gwgon, the 
last of the kings of Ceredigion, met his death by drowning ; 20 
Rhodri's marriage to Angharad, the dead king's sister, while it 
gave him no sort of legal claim to the province, made it easy 
for him to intervene and invested his sons with rights there 
which would be more generally recognised. 21 At his death 
Rhodri held in his grasp the whole of North Wales and such 
portion of the South as was not included in the kingdoms of 
Dyfed, Brycheiniog, Gwent and Glywysing. 

Despite these successes, Rhodri was at no period of his 
long reign free from the menace of Danish invasion. Refer- 

17 " Cinnen (read Cincen)rex pouis in roma obiit " (Harl. MS. 3859 in Cymr. 
ix. 165 for the pedigree see ibid. 181). The year implied is 854, but there is 
reason to think that from about 850 this chronicle is one year behind the true 
reckoning. See especially the entry under 866, instead of 867, of the fall of York 
and the assignment to 878 of the death of Aedh Finnliath, an event known from 
the precise chronological data in Chron. Scot, to have occurred in 879. Chron. 
Scot, is believed to be correct from 805 to 904 (Introd. xlvi). 

18 Three sons of Cyngen are mentioned in Harl. MS. 3859, pedigree No. xxxi. 
(Cymr. ix. 182), but they do not appear in the chronicle as kings of Powys and 
probably did not survive their father. 

19 A.S. Chr., 828 (for the true date see Plummer, ii. 73) ; ibid. 853 (cf. Asser, 
c. 7). It will be observed that Green (Conq. Eng. (2), p. 80) has hastily credited 
/Ethelwulf with a conquest of Anglesey which was really a Danish achievement 
(see note ro above). 

20 Harl. MS. 3859, s.a. 871 (Cymr. ix. 166). 

21 Jesus Coll. MS. 20 in Cymr. viii. 87 (No. xxi.). 


CHAP, ence has already been made to the ravaging of Anglesey in 
X ' 853 ( or 854) by the Dubhgaill or " black " Danes who had newly 
appeared in Irish waters and attacked the earlier settlers of 
Norse origin. In 856 Rhodri avenged himself by killing their 
leader Horm. 22 Nevertheless, he is still found fighting the 
" black gentiles " at the close of his life ; the " gwaith dyw Sul " 
or "Sunday's fight" fought in Anglesey in 877 23 must have 
been an encounter with a heathen foe, and its issue is shown by 
the statement in the Irish Chronicles that Rhodri, king of the 
Welsh, in this year sought safety in Ireland from the attacks 
of the " black gentiles ", 24 In the following year he was back 
in Wales, to fall a victim, along with his son (or brother) 
Gwriad, to English enmity. 25 The manner of his death is un- 
known, but that the loss was fiercely resented may be gathered 
from the fact that a rout of the English some years later was 
triumphantly hailed as " God's vengeance for the slaughter of 
Rhodri ". 

Six sons were left by Rhodri to carry on his line, and, after 
a fashion which, to the injury of the country, widely prevailed 
in mediaeval Wales, his broad realm, so laboriously built up, 
was divided between them. Little is known of the distribution 
of the various provinces, indeed, only three of the sons, 
Anarawd, Cadell and Merfyn, are mentioned by name in con- 
temporary records. 26 But it is clear that Anarawd, as the 
eldest, took possession of Anglesey and the adjacent parts of 
Gwynedd, and most probable that Cadell received as his share 
a substantial domain in South Wales, where his descendants 
ruled for many generations. What portion fell to Merfyn can 
only be conjectured, for he founded no house and nothing is 
recorded of him in authentic sources save the bare fact of his 
death. 27 This is nevertheless known of the sons of Rhodri, on 

Ann. Ult. s.a. 855; Chron. Scot. s.a. 856. 

23 Gueith diu sul in mon " (Had. MS. 3859 in Cymr. ix. 166). 

24 Ann. Ult. s.a. 876 ; Chron. Scot. s.a. 877. 

25 " Rotri et films eius guriat (MS. B., B.T. and B. Saes. have " frater " and 
" brawd ") a saxonibus iugulatur " (Harl. MS. 3859). Ann. Ult. s.a. 877 (=878) 
also says that Rhodri was slain by the English. 

26 Jesus Coll. MS. 20 (Cymr. viii. 87, No. xx.) gives the names of the other 
three as " Aidan, Meuruc, Morgant". 

27 Anarawd is described by an editor of Nennius who wrote in 912 as "regis 
Moniae, id est mon, qui regit modo regnum Wenedotiae regionis, id est Guernet " 
(Hist. Britt. ed. Mommsen, p. 146). That he was the eldest son maybe inferred 


the testimony of their contemporary, Asser of St. David's, 28 CHAP. 
that they were a vigorous brood, working strenuously together 
for the overthrow of the remaining dynasties of the south. In 
Dyfed, Hyfaidd ap Bledri, himself the terror of the wealthy 
" clas " of St. David's, dreaded the violence of the new lords of 
Seisyllwg ; in Brycheiniog, Elise ap Tewdwr, of the line of old 
Brychan, also feared for his crown ; and, though for the moment 
Hywel ap Rhys of Gly vvysing 29 and Brochwel and Ffernfael ap 
Meurig of Gwent were chiefly perturbed by the activity of the 
Mercians, they too had much to apprehend from any revolution 
which might establish the house of Rhodri in Brycheiniog. Thus 
arose the political situation which is described by Asser as hav- 
ing existed for a good many years at the date of his composi- 
tion of the life of Alfred in 893. 

That great king's famous victory over the Danes in 878 
had given him a commanding position in Southern Britain. 
Not only did he gain undisputed authority over the whole of 
Wessex, but, on the death about this time of Ceolwulf, the last 
king of Mercia, he assumed the control of as much of the 
ancient province as remained in English hands, while entrusting 
its actual rule to an ealdorman named jEthelred, with the 
hand of his daughter yEthelflaed. It was but natural that the 
minor Welsh kings should seek from Alfred the protection 
which his known love of justice would dispose him to give, and 
thus it came about that Hyfaidd, Elise, Hywel, Brochwel and 
Ffernfael all placed themselves under his patronage and became 

from the prominence assigned to him by Asser and from the fact that he held 
his father's paternal inheritance. Late authorities furnish full details, which 
grow more precise as time goes on, as to the division of Wales between the sons 
of Rhodri Mawr ; see Gir. Camb. vi. 166-7 (Dtscr. i. 2, 3); Powel, 29; Gw. 
Brut. s.a. 873 ; lolo MSS. 30-1. Even the earliest of these accounts is full of 
errors ; the statement of Giraldus that Cadell survived his brothers and thus 
obtained for himself and his descendants the monarchy of all Wales is in flat 
opposition to the testimony of the chronicles and was no doubt concocted to sup- 
port certain South Welsh claims ; he has further complicated matters by trans- 
posing Merfyn and Anarawd and making the former's daughter Afandreg a son. 

28 In c. 80 of the life of Alfred (ed. Wise, pp. 49-50 ; Mon. Hist. Br. 488 ; 
Stev. 66-7) there is a most valuable account of Welsh politics in the period 880- 
93, and the ease with which the author moves in a field in which a later forger 
would have infallibly shown his ignorance is a weighty argument in favour of the 
authenticity of the work (Stev. Ixxv.). 

29 I think it very unlikely that this is the Hywel whose death at Rome is 
recorded in Harl. MS. 3859, s.a. 885, since Hywel ap Rhys is associated in Lib. 
Land. 236, with Bishop Cyfeiliog of LlandarT, who belongs to the tenth century. 


CHAP, the vassals of a monarch who could succour them by land and 


by sea. Asser explains, in his apology for his desertion of his 
native St. David's, how the situation thus created enabled him 
to serve his beloved monastery even better at Alfred's court 
than he could have done at home ; the friend and companion 
of Hyfaidd's overlord was in a position to set limits to Hyfaidd's 
tyranny in a way impossible for the simple bishop of St. David's. 
It was now the turn of Anarawd and his brothers to find 
their power fettered and their triumphant progress brought to 
a stand. With Mercia they had contended not unequally ; a 
raid upon Eryri conducted by ^Ethelred in 88 1 had been 
arrested by Anarawd at the mouth of the Conway, and the 
victory of Cymryd the day of divine vengeance for Rhodri 
had been won with great slaughter of the foe. 30 In order to 
secure himself against further attacks from Mercia, Anarawd 
had then entered into an alliance with the Danish king of York, 
whose realm, embracing as it did the ancient Deira, extended 
to the Mersey and possibly took in also the peninsula of Wir- 
ral. But 'the Danes proved indifferent allies, and gradually 
Anarawd came to the conclusion that it was his interest also, 
no less than that of the minor chieftains of the country, to 
make his peace with the strong ruler of Wessex. He found 
Alfred in no wise loth to respond to his advances ; paying 
him a ceremonious visit at his court the first of the kind on 
record paid by a Welsh to an English king he was received 
as befitted his rank and treated with marked generosity. It 
was a part of Alfred's statesmanship to lead the other Chris- 
tian princes of the island to regard him as their natural lord 
and protector against heathen attack, and thus it was that 
Wales came formally under the supremacy of Wessex ; Gwy- 
nedd under Anarawd was recognised as standing in the same 
relation to the West-Saxon king as did Mercia under ^Ethelred, 
and the basis was laid of the homage which in later ages was 
regularly demanded from all Welsh princes by the English crown. 

30 Had. MS. 3859 (s.a. 880 ; it is still a year in arrear cf. the annal 887 
with the Irish notices of the death of King Cerbhall of Dublin) has the simple 
entry: " Gueit conguoy digal rotri adeo " (the battle of the Conway: Rhodri 
avenged by God). In Wynne, 37-9, there is a detailed narrative, possibly derived 
from records of the see of Bangor (cf. B. Willis, Bangor, p. 184), which, while it 
contains some legendary features, appears nevertheless to embody a genuine 


At the moment when Asser was placing these events on CHAP, 
record, English overlord and Welsh king were alike preoccupied 
with a new danger in the renewal of the Scandinavian attacks 
on an extensive scale. For some years the peril had seemed 
to be over, and Asser's pages show no sense of its urgency. 
But the pirate bands had merely transferred their operations to 
the Continent, and, upon undergoing a crushing defeat at Louvain 
in 891, had laid their plans for fresh incursions into England, 
where they appeared at the end of 892. 31 The army led by 
Haesten included not only warriors from over sea, but also 
Danes of Deira and East Anglia, who looked upon the hostings 
as agreeable breaks in the monotony of farm life ; it swept 
Southern Britain during the four years 892-6 from sea to sea, 
ravaging without mercy in summer and passing the winter in 
some fastness specially chosen and prepared for the purpose. 
In the chronicle of its movements Wales and the Welsh border 
occupy a prominent place ; some impulse, possibly the desire to 
get into touch with the Danish settlements in Ireland, constantly 
drew it west. In 893 it is found at Buttington, close to the 
confluence of the Severn and the Wye, 32 where it was beset by 
a joint army of English and Welsh ; Mercians and West Saxons 
on the one hand, men of Gwent and Glywysing on the other, 33 
reduced it to sore straits, and, after great slaughter, drove what 
remained of it back to Essex. Before the close of the year it 
had repaired its broken ranks, had, become again formidable, 
and, after a long march which was not suspended by night or 
by day, had taken possession for the winter of the ruined walls of 
Chester. The Mercians followed hard upon its heels and cleared 
the country round of all supplies of food ; hence it was forced 
early in 894 to replenish its stores by a raid on North Wales. 
It may be conjectured that Anarawd received some English 

31 A.S. Chron. MSS. A. B. C. D. s.a. 893-7 tell the story in detail. 
Plummer shows (ii. 108, no) that the dating must be set back a year in the case 
of each of these annals. 

32 See Plummer, ii. 109-10. Buttington by Chepstow tallies much better 
with the data of the Chronicle especially with the composition of the English 
army and its leaders than does Buttington by Welshpool. 

33 So I interpret the " sum dasl thaes Norg Weal cynnes " of A.S. Chr. The 
idea of Green (Conq. Eng. (2), pp. 172, 173, 183) that other Welsh, under the 
leadership of Anarawd, were in alliance with Haesten is plainly inconsistent with 
the evidence, and seems to have arisen out of a baseless impression that Anarawd 
did not submit to Alfred until 897. 


CHAP, help to repel this invasion ; he had at any rate English troops 
x> at his command when in the following year he plundered 
Ceredigion and Ystrad Tywi, a blow directed most probably at 
his brother Cadell. 34 During most of 894 and 895 the Danes 
were busy in the neighbourhood of London ; at the end of the 
latter year they were once more in the Severn valley, encamped 
for the winter at Quatbridge, which cannot have been far from 
the modern Bridgenorth. This became the starting-point for 
the last great raid, in the spring of 896, which devastated not 
only the adjacent parts of Mercia, but also the Welsh districts of 
Brycheiniog, Gwent and Gwynllwg. 35 In the summer of this 
year the great confederacy was dissolved ; the men of Deira 
and East Anglia returned to their homes, while the wandering 
pirates turned their attention once more to the banks of the 

The closing years of Alfred's reign were comparatively 
peaceful, but his death, in or about 90 1, 36 was followed by 
another period of struggle and unrest. The Welsh now found 
themselves between two fires, for, while the danger of invasion 
from the Scandinavian settlements in Ireland and the North 
was no less a matter of apprehension than before, attention had 
to be paid to the doings of Mercia also, since Edward the Elder 
was too much occupied with schemes of his own to control its 
aggressive spirit as had been done by his peace-loving father. 
In 902 the Celtic element won a temporary triumph over the 
Scandinavian in Ireland ; Dublin was cleared of its heathen 
folk 37 and very many of them, under the leadership of one 
Ingimund, made their way to Anglesey, intending, no doubt, to 
found a new settlement in the island. 38 They were stoutly 
resisted by the inhabitants and forced to look elsewhere for 

34 "Anaraut cum anglis uenit uastare cereticiaun et strat tui"(Harl. MS. 
3859, s.a. 894 (= 895)). 

35 " (N)ordmani uenerunt et uastauerunt loycr et bricheniauc et guent et 
guinnliguiauc " (ibid. s.a. 895 (=896)). 

36 Harl. MS. 3859 records the death of "Albrit rex giuoys" (i.e., of the 
Gewissae) s.a. 900, and thus tends to support the ordinary dating in 901. 

37 Ann. Ult. s.a. 901 ; Chron. Scot. s.a. 902. 

38 Harl. MS. 3859 has s.a. 902 (= 903) '? Igmunt in insula mon uenit. et 
tenuit maes osmeliavn " (a place not yet identified). A more detailed account of 
the adventures of " Hingamund " (= Ingimundr) is to be found in a fragment of 
an Irish Chronicle printed by J. O'Donovan in 1860 for the Irish Archaeological 
and Celtic Society (pp. 225-37). Though the MS. is a late transcript, it seems 
to embody some genuinely historical material. 


a foothold, which they ultimately found, if an Irish account is CHAP, 
to be trusted, in the neighbourhood of Chester. This was a 
concession due, it would appear, to the famous " lady of the 
Mercians," yEthelflaed, daughter of Alfred and wife of the 
Mercian ealdorman, who during her husband's lifetime and for 
seven years after his death led and governed the Mercian folk 
with such intrepid energy that one accepts without surprise the 
Welsh and Irish designation of her as a queen. 39 She had 
already repaired the fallen ramparts of the ancient stronghold 
on the Dee, rightly estimating its great importance as the key 
to North- Western Mercia, and realising that in Danish hands it 
would be a perpetual menace to her power. 40 She maintained 
her position there, the Irish chronicler tells us, against fierce 
onslaughts on the part of Ingimund and his men, and then 
proceeded to buttress the Mercian state by building other forts 
along the borders specially open to attack. Scarcely anything 
is known of yEthelflaed's relations with the Welsh ; the one 
event recorded in this connection can have been but an incident. 
In the early summer of 916 she invaded the realm of Bry- 
cheiniog, stormed a royal stronghold near Llyn Safaddon, and 
captured the queen and a number of followers of the court. 41 
The reigning chief was probably Tewdwr ab Elise, the son of 
the contemporary of Asser, 42 and, whatever may have been his 
offence against his powerful neighbour, his attitude affords no 
clue to the solution of the much more important question as 
to the relations between ^Ethelflaed and the house of Rhodri. 
But a little light is thrown upon this matter by the disposition 
of the great lady's fortresses ; though the sites of all have not 
been identified, it seems likely that some were built on the 
eastern frontier of Mercia, for instance, at Chirbury, 43 and in 

39 "^Elfled regina obiit" (Harl. MS. 3859, s.a. 917); " Eithilfleith, famos- 
issima regina Saxonum, moritur" (Ann. Ult. s.a. 917). 
"M.S. Ckr. MSS. B. C. s.a. 907. 

41 Ibid. s.a. 916. Camden saw that " Brecenan mere " must be Llangorse 
Lake (Britannia, p. 562), but the exact position of the " llys " is not easily de- 

42 In Lib. Land. 237-9, " teudur filius elised," king of Brycheiniog, has dealings 
with Bishop Llibio of Llandaff, who succeeded Cyfeiliog in 927 and died in 929. 

43 " Cyric byrig " (A.S. Chr. MSS. B. C. s.a. 915) appears to be Chirbury, 
which had an ancient church and was head of a hundred before the Norman 
Conquest ("Cireberie" in Domesd. i. 2536 i), rather than Chirk, which takes its 
name from the river Ceiriog (old Welsh Ceriauc) and had no importance in 
Anglo-Saxon times. 


CHAP, this case she and the leaders of the Welsh can scarcely have 

been upon entirely friendly and easy terms. It is most probable 

that her death in 9i8, 44 in the midst of her career of conquest, 
came as a relief to men who knew not whither she might next 
turn her victorious arms. 

Edward the Elder at once seized upon the opportunity 
which was now afforded him of becoming direct ruler of the 
whole of Mercia and was thus brought into immediate relations 
with the princes of Wales. His policy was that of his father, 
one of friendship and protection in return for submission and 
homage. He had already given proof of kindly sentiments 
towards the Welsh people. In 915 a viking host had sailed 
from Brittany into the estuary of the Severn, and, landing on 
the southern coast of Wales, had spread ruin over Gwent and 
Glywysing ; their daring onslaught carried them as far as Erging, 
and here, not many miles from Here ford, .they captured Bishop 
Cyfeiliog of Llandaff, and, rejoicing in their good fortune, led 
him a prisoner to their ships. 45 Edward at once came to the 
relief of the hapless prelate, and, on payment of a ransom of 
forty pounds, obtained his release from the clutches of his 
heathen captors, nor was it until this transaction was complete 
that vigorous measures were taken for the expulsion of the 
Danes, who ultimately withdrew by way of Dyfed to their 
kinsmen in Ireland. Thus it was only to be expected that on 
yEthelflaed's death the Welsh princes should readily acknow- 
ledge the sway of the West-Saxon king ; " Hywel and Clydog 
and Idwal," according to the official chronicle kept at Win- 
chester, "with all the North Welsh (i.e., Welsh) race, sought 
him as their lord ", 46 A new generation of chiefs had arisen 
since the days of Asser ; of the sons of Rhodri, Merfyn had 
died in 904, Cadell in 909 (or 910), Anarawd in 9i6. 47 Idwal 

44 Harl. MS. 3859 gives the year as 917 and would seem to be still a year 
behind; see the entry under 907 of the death of " Guorchiguil, " placed by B. 
Saes. in the same year as the death of King Cormac of Munster, i.e., 908. 

4 M.S. Chr. MSS. A. B. C. D. The date is no doubt rightly given by B. C. 
D. as 915. Lib. Land. (231-7) makes " Cimeiliiauc " a contemporary of Brochwel 
ap Meurig of Gwent and Hywel ap Rhys of Glywysing and gives 927 as the year 
of his death. There is fairly good evidence that the early bishops of Llandaff 
exercised authority in Erging. 

46 A. S. Chr. MS. A. s.a. 922, the year being actually 918. 

47 Harl. MS. 3859 s.a. gog, 915 ; Chron. Scot, (after 904 a year in arrear) 
s.a. 908, 915 ; B. Saes. s.a. 901, B.T. p. 18 (for Merfyn). 


the Bald, son of Anarawd, now ruled over Gwynedd, Hywel CHAP, 
and Clydog, the sons of Cadell, in the south. 


One prince only, among the many who bore rule in Wales 
in the Middle Ages, was honoured by posterity with the title 
of " Good " a circumstance which in itself imparts a peculiar 
interest to the reign of Hywel Dda. 48 Yet history has hardly 
anything to tell of the personal traits of one who gained for 
himself so secure a place in the affections of his fellow-country- 
men ; the facts of his life, as handed down by the ancient 
sources, are extremely meagre, and, though they reveal in 
some measure the greatness and distinction of the man, leave 
much unexplained, so that even in the briefest outline of his 
career the historian must not seldom invoke the aid of con- 

He was the son of Cadell ap Rhodri, and when he and his 
brother Clydog a younger brother, it may be inferred from 
the order of the names offered submission to Edward the 
Elder in 918, it is safe to conclude that they were rulers of 
Seisyllwg, which they had divided between them in accordance 
with Welsh custom. But it is most probable that Hywel had 
also by this time come into possession of Dyfed. No king of 
that region appears after the death in 904 of Llywarch ap 
Hyfaidd, who was doubtless the last of the old line, 49 and, as 
Hywel is known to have married his daughter Elen, 50 it seems 
clear that, either then or shortly afterwards, the realm came 
into the hands of Cadell's eldest son as the result of the marriage 
alliance. Clydog did not long survive the submission of 918 ; 
two years later he died, 61 when it may be presumed that Hywel 
obtained the whole of Seisyllwg ; united to Dyfed, it formed 
the kingdom of Deheubarth, a compact area which covered 
the whole of the South-west of Wales from the Dovey to the 
Tawe. The foundation of this kingdom, which was that of 

48 The title does not appear in Harl. MS. 3859 or any other contemporary 
source, but obtained an early currency from the prefaces to the editions of the 

49 " Loumarch filius hiemid moritur " (Harl. MS. 3859 s.a. 903). 

50 Pedigrees Nos. i. and ii. (Owain ap Hywel) in Harl. MS. 3859. 
" " Clitauc rex occisus est " (Harl. MS. 3859, s.a. gig). 


CHAP. Rhys ap Tevvdwr and, save for Norman acquisitions, of Rhys 
ap Gruffydd, is the first notable event in the history of Hywel. 
The next was his visit to Rome in 928, an undertaking 
which finds no true parallel in the life of any other Welsh 
prince. 52 Two earlier instances are no doubt recorded of pil- 
grimage on the part of Welsh chieftains to the Holy City, 53 but 
of both Cyngen of Powys and the unknown Hywel of 886 it 
is said that they died at Rome, so that it is clear that the journey 
was a penitential effort at the close of a busy and not too 
scrupulous reign, intended to smooth the pathway to a better 
world. Hywel's visit does not at all suggest the repentance of 
a dying man ; he returned to Wales to wield the sceptre with 
vigour and marked intelligence for twenty years, and must have 
been in this year of pilgrimage a man in the full prime of his 
career. What took him so far out of the common-place round 
of princely life in Wales in his day must have been, one cannot 
doubt, that breadth of sympathy, that enlightened interest in 
the life of other nations than his own which come to light in 


other parts of his history. There is good reason to think 
that he made Alfred, in whose reign he was doubtless born, his 
model and exemplar ; Alfred had twice visited Rome as a boy 
and had maintained the connection by frequent gifts to the 
holy see, and it was thus to be expected that, when Hywel 
found himself free to undertake this journey, he should adven- 
ture upon it. Rome itself was in this age scarcely worthy of 
the veneration lavished upon it by the Church of the West ; 
whether the pope to whom Hywel made his obeisance was 
Leo VI. (928-9) or Stephen VII. (929-31), he was probably in 
either case a creature of the notorious Marozia. 54 But, though 
there may have been little in the atmosphere of the city to 
sustain the fires of a high spiritual enthusiasm, the historic 
memories of the spot must have cast their spell over the 
wanderer from distant Dyfed, and each league of the long and 

52 "Higuelrex perrexit ad romam " (Harl. MS. 3859, s.a. 928). The date 
may be correct, as the chronicle seems hereabouts to right itself and soon after- 
wards to get a year in advance of the true dating. The battle of Brunanburh, the 
death of Athelstan, that of Olaf Godfreyson (Abloyc rex), the devastation of 
Strathclyde and the murder of Edmund are all post-dated a year. 

53 See note 44 to chap. vii. 

84 Gregorovius, History of the City of Rome in the Middle Ages, Eng. trans. 
Hi. pp. 282-3. 


perilous journey must have left its impress upon the mind of CHAP, 
one so observant and so ready to learn. 66 

The remarkable feature of the years which followed Hywel's 
return is his close association with the English court. Edward 
the Elder had died in 924 or 925 at Farndon on the Dee, hav- 
ing been summoned to this region by a revolt of the city of 
Chester, in which the neighbouring Welsh were concerned. 50 
He was succeeded by Athelstan, who showed the utmost energy 
and resolution in maintaining and extending the power which 
of late had fallen to the king of Wessex. It would appear 
that about 926 or 927 he summoned the leading Welsh princes 
to Hereford, imposed a tribute upon them of gold, silver, cattle, 
hunting dogs and hawks, and fixed the Wye as the boundary 
between the two races in that part of the country. 57 Hywel 
of West Wales and Owain of Gwent (son of Hywel ap Rhys) 
are expressly mentioned in a MS. of the English Chronicle as 
having submitted to Athelstan ; 58 of Idwal Foel there was a 
story current that the English king for a time deprived him of 
his realm, 59 so that his position at this juncture is not quite cer- 

55 According to an early editor of Ven. (LL. i. 216), whose statements were 
copied into certain South Welsh MSS. (LL. i. 342, MSS. S. and Z.), Hywel when 
at Rome sought the approval of the pope for his laws and obtained it. Not only 
is it highly improbable that papal authority was ever given to laws which so fre- 
quently run counter to those of the Church, but there are also chronological ob- 
jections. The compilation of the code belongs to Hywel's later years, not to the 
early period of his reign, for he was not in a position to undertake the task until 942. 

68 Wm. Malm. G.R. i. 144 (i. 210). The authority used by William for this 
part of his account of Athelstan's reign was an ancient metrical life, written most 
probably during the king's lifetime (Stev. Asser, 184, n. 4). The mention of 
Chester shows that Farndon by Holt is intended, not Faringdon (suggested 
by Plummer, ii. 380-1). " Fearndun " was literally translated into Welsh as 
" Rhedyn-fre" in the name " Siat Rhadynfre" found for St. Chad in Myv. Arch. 
II. 52 (429). 

87 Wm. Malm. G.R. i. 148 (i. 214). 

98 A. S. Chr. MS. D. s.a. 926. The names are introduced parenthetically 
(" Huwal West Wala cyning . . . Uwen Wenta cyning") into an account of 
Athelstan's dealings with the Danes of Deira on the death of their king Sitric in 
927, and it is not really intended to state that the Welsh and other kings men- 
tioned met at " Ea mot " on the i2th of July (Plummer, ii. 136). Plummer (ii. 
addenda viii.) thinks that " West Wala " must be Cornwall, and that the reference 
is not to Hywel the Good, but to an otherwise unknown Cornish ruler. This is 
not, of course, impossible, but it is surely a simpler supposition that West Wales 
is used in this passage in an unusual sense. " Yuein," son of Hywel ap Rhys, 
appears in a notice in Lib. Land. 236 of the time of Bishop Cyfeiliog. 

59 Wm. Malm. G.R. i. 142 (i. 206). This story does not come, it would 
seem, from the metrical life, and has the suspicious feature of making Idwal king 
of all the Welsh. 


CHAP. tain. It will be seen that Athelstan was bent upon turning 
into a real subordination the formal submission demanded from 
the Welsh by Alfred and Edward, and what is notable is the 
manner in which Hywel accepted the situation and did his best 
to turn it to account. Valuable evidence is furnished on this 
head by the English land-charters. 60 These documents are 
usually attested by a large number of witnesses, members of 
the assembly of " witan " or " wise men " which approved the 
grant, and, as Athelstan adopted the policy of summoning his 
" subreguli" or under-kings to occasional meetings of this body, 
the names of Welsh princes are often found among those of 
the attesting counsellors, taking precedence of those of earls 
and thegns, and sometimes even of those of bishops. Some of 
these charters rest under considerable suspicion, since their text 
has only been preserved in cartularies or charter-books compiled 
at a much later date by monastic houses which were very jeal- 
ous of their rights and not unwilling to forge documentary 
evidence in support of them in case of need. Nevertheless, 
though it is not possible to accept the testimony of every charter 
for the facts which it professes to record, some are undoubtedly 
genuine, and even in the forged charters the names of the wit- 
nesses are probably drawn from authentic documents. It is, 
therefore, a significant fact that Hywel is of all the Welsh princes 
the most prominent in this connection ; from '92 8 to 949 his 
name is appended to every charter which has Welsh signatures 
and is among them placed first ; in three cases he is the only 
under-king who joins in the grant. He is often supported by 
Idwal Foel and Morgan ab Owain of Morgannwg, and once by 
Tewdwr ab Elisedd of Brycheiniog, but no prince seems to have 
entered so heartily into Athelstan's design of linking Wales with 
England by this system of attendance at the English court. All 
that is known of Hywel points him out as a warm admirer, not 
only of Alfred, but also of English civilisation ; he led no ex- 
pedition across the border, but instead secured to Athelstan the 
faithful allegiance of his brother chiefs, even in that year of re- 
bellion, 937, when the league against Wessex included the 
Scots, the Danes and the Strathclyde Britons, and only the 
Southern Britons held aloof. English influence is manifest in 
the law of Hywel and betrays itself even in the naming of his 

60 See note A appended to this chapter. 


sons, for Edwin ap Hywel Dda bore an English name, 81 which CHAP, 
was possibly given him out of compliment to the young son of 
Edward the Elder who perished in 933. 

Hywel was now about to reach the zenith of his career and 
to become, as he is termed in the codes, " by the grace of God 
ruler of all Wales ". 62 This was a position he certainly did not 
hold during the life of Idwal Foel ab Anarawd, for the evidence 
makes it clear that Idwal and Hywel ruled North and South 
Wales respectively under the overlordship of Athelstan. 63 But 
in 942 Idwal, never quite easy under the English supremacy, 
seems to have broken into revolt against the power of the new 
king Edmund ; he and his brother Elisedd met the Saxon 
in battle and were both slain. 64 In the natural course the 
sovereignty of Gwynedd should have passed to the sons of 
Idwal, lago and Idwal, the latter often called leuaf (or 
" Junior ") to distinguish him from his father. Hywel, how- 
ever, now appeared upon the scene, and, driving out the young 
heirs, 65 made himself master of Gwynedd and, most probably, 
of Powys also. It is quite impossible to say whether there 
were good grounds, recognised as valid by Welsh public opinion, 
for this act of aggression ; one may nevertheless conjecture that 
the revolution was favourably viewed by the English, and no 
attempt to reverse it was made while Hywel lived. Whether 

81 He died in 954 (Ann. C. ; B.T. s.a. 952). The form " Guin," found in 
Ann. C. MS. B., is due to a misunderstanding of " Etgu'm ". 

M " O rat Duw . . . brenhin Kymry oil " (Dim. i. 338). Cf. Lat. B. ii. 814 : 
" Dei gratia atque providentia rex Howel . . . totius Wallie principatu prae- 
sidebat pacifice ". All the texts of the laws emphasise in their prefaces the ex- 
tent of his rule. 

68 See the charters mentioned in the note at the end of this chapter and note 
58 above. 

84 " ludgual et filius eius elized a saxonibus occiduntur " (Harl. MS. 3859, 
s.a. 943). The year was that which followed the death of Olaf Godfreyson, 
shown by Irish records to have happened in 941. The MS. followed by Ann. 
C. MS. B., and B.T. makes Idwal son of Rhodri, but the ordinary account, 
contained in Jesus Coll. MS. 20, No. xxvi. (Cymr. vii. 88), is to be preferred as 
better suiting the chronological data. As to Elisedd, " filius " would seem to be 
an early error of transcription for " frater," corrected in the older texts of B.T. 
(but not in B. Saes.), which have " urawt " (Bruts, 261 ; B.T. p. 20); see 
Dwnn, ii. to, 16; Harl. MS. 3859, s.a. 946 ( = 945) " Cincenn filius elized 
ueneno periit " ; Carlisle, s.v. Hawarden (" Cynan ap Elis ap Anarawd"). 

88 " lago et leuaf, quos Howel e regno expulerat " (Ann. C. MS. C. p. 18, 
n. 20). It was at this period of his life, no doubt, that Hywel issued the coin, 
the only one known to have been struck for a Welsh prince, which is described 
by Mr. Carlyon-Britton in Trans. Cymr. 1905-6, 1-13. 
VOL. I. 22 


CHAP, justified or not, it united Wales almost as completely as any 
recorded movement in its history. Morgannwg and Gwent 
still retained their independence under the sons of Owain, 
Cadwgan and Morgan the Aged, 66 but this was a part of Wales 
which never, save for a few years under Gruffydd ap Llywelyn, 
entered into any wide-embracing Welsh realm. 

The events of 942 prepared the way for the crowning and 
distinctive achievement of Hywel's life, namely, the reduction 
of the varying royal and tribal usages of Wales to a uniform 
and consistent system, accepted by the whole country and 
permanently embodied in a written code. No direct evidence 
exists that this work was actually done by Hywel save the 
statements made in the various copies of the code itself, 
and none of these copies are of older date than the end of the 
twelfth century. But the unanimity with which they ascribe 
this great legal reform to the " good " king, and the un- 
challenged assumption throughout Welsh literature that the 
work was his, without suggestion of a rival for the honour, 
constitute as strong a proof as tradition can well supply; if 
Hywel was not the author of the code which bears his name, 
how came he to be singled out from among his fellows and 
invested by later generations of Welshmen with this unique 
distinction ? It may well be believed that the task was the 
principal object he set before himself in his career ; the union 
of Dyfed and Seisyllwg under his rule would bring him in early 
life face to face with the inconveniences of conflicting tribal 
custom, the example of Alfred, always ihis guiding star and 
embodiment of the princely ideal, would invite him to the 
beneficent path of legislative reform, and his foreign experience 
would teach him how much the Welsh had still to learn ere 
social institutions among them could rest on a firm foundation. 

66 Morgan Hn succeeded his father Owain ap Hywel (the " Uwen Wenta 
cyning" of ^4.S. Chr. MS. D. s.a. 926) about 930 and long survived Hywel, 
dying at a patriarchal age in 974. His brother Gruffydd, who was king of 
Gower in 928 (Lib. Land. 239-40), was killed in 935 (or 934) by the men of 
Ceredigion (Ann. C. MSS. B. C. ; B.T. ; B. Saes.). The third brother, 
Cadwgan, appears in Lib. Land. 224-5 as " Catguocaun rex filius Ouein " and 
lord of the region of Margan about 940; his death in battle against the English 
is recorded about 950 (Harl. MS. 3859; Ann. C. MS. B. ; B.T.). "Caducon" 
may be the correct reading, instead of " Cadmon," in Cart. Sax. iii. 39, in which 
case Cadwgan, with his brother Morgan and Hywel Dda, was at the court of 
Edred in 949. 


He was obliged to wait for many years for the realisation of CHAP. 
his hopes, but at last the whole of Wales lay at his feet and the 
hour was ripe for his enterprise. 

According to the unanimous testimony of the prefaces to 
the various versions of the code, 67 Hywel's first step was to 
summon a representative assembly, in which each cantref 88 
was represented by six men, to his hunting lodge of Y Ty 
Gwyn ar Daf (The White House on the Tafif), now marked 
by the village of Whitland in Carmarthenshire. 69 No other 
national gathering of the kind is recorded in Welsh history 
until the fifteenth century, but the occasion, it should be 
remembered, called for very exceptional measures. The laws 
of the Welsh were not the creation of any legislator, owing 
their origin to the royal fiat and capable of being altered by the 
same authority. They were the ancient customs of the race, 
handed on in each tribe as a precious heirloom from generation 
to generation, preserved by the tradition of the elders, and 
having the sanction of immemorial antiquity. In order to 
annul or modify them, nothing short of the assent of a national 
conclave was. sufficient, and an editor of the Venedotian code 
lays down the principle that no part of the law of Hywel itself 
can be abrogated save by leave of a body as large as that 
which met at Hywel's behest. 70 

The prefaces state that the conference was held in Lent 71 
and occupied the whole six weeks of that season, so that the 
deliberations must have been long and the revision of the laws 
entered into with great thoroughness. They give discrepant 
accounts of the way in which the final text of the revision was 
arrived at, the Dimetian and the Gwentian prefaces ascribing 

87 LL. i. 2, 214-6 (Yen.), 338-42 (Dim.), 620-2 (Gw.) ; ii. 749 (Lat. A.), 814 
(Lat. B.), 893 (Lat. C.). 

88 Only MSS. B. D. of Yen. have "cantref " (cf. however, C. D. K. in pref. 
to bk. iii. p. 214), the other MSS. and codes reading " cymwd ". But the sub- 
stitution of the cymwd, the actual, for the cantref, the archaic area in MSS. of 
the twelfth and thirteenth centuries is more easy to explain than would be the 
reverse process. Reasons have already been given (p. 301) for thinking that 
the cymwd had not come into existence in the age of Hywel Dda. 

89 For Whitland see chap. v. note 129. 

70 Yen. i. 216 (MSS. C. D. K.). Other passages which illustrate the doctrine 
of the necessity of national assent for all new laws are to be found in Dim. i. 594 
(heading of III. ii.) ; LL. ii. 394 ( 43) ; De Nugis, ii. 22 (antequam in contrarium 
decreta ducent/>6/tVa, nihil novum proferamus). 

71 Gw. and Lat. B. omit this detail. 

22 * 


CHAP, the actual compilation of the new code to a commission of 
twelve laymen and one cleric, while the other MSS. seem to 
assign the task to the whole assembly. 72 It is most probable 
that some small body of men practised in the law was charged 
at the end of the conference or during its sessions with the 
business of embodying its decisions in writing, but one can 
hardly go further and believe with the author of the Dimetian 
code that the commission did everything and that the Lenten 
assembly was solely occupied with fasting and prayer. Many 
hostile interests had to be reconciled, many discordant usages 
brought into harmony and the general will enforced against 
local peculiarities all of it work, not for a drafting committee, 
but for a genuine popular assembly. It was, no doubt, ere the 
assembly broke up that the formal proclamation of the new 
laws was carried out and a solemn malediction laid upon all 
evil-disposed men who should infringe the dooms thus set forth 
by national assent. 

Many lawyers of repute must have lent their aid to Hywel 
in his great undertaking, but nothing is known of any of them 
save Blegywryd ab Einon. In the various copies of the 
Dimetian Code, Blegywryd is made to play a leading part in 
the proceedings at Whitland ; he is the " master " or " scholar " 
who acts as clerk to the twelve lay commissioners, and later 
texts dub him Archdeacon of Llandaff and send him to Rome 
with the deputation which, it is alleged, went thither to obtain 
for the new code the benison of the holy see. 73 It is clear that 
from this source no trustworthy information about him is to be 
gleaned, but there are other references to him of a less doubt- 
ful complexion. Some rudely formed Latin hexameters found 
in two or three late copies of the laws would seem to have 
been originally written at the end of a copy of the code trans- 
cribed by Blegywryd for the use of the court of Hywel and 
its judge, Gwrnerth Lwyd of Is Cennen. 74 In these he is 

72 MSS. S. Z. name the commissioners, but the list is merely one of famous 
lawyers, without regard to chronology. Morgeneu and Cyfnerth were the com- 
pilers of the Gwentian Code (Gw. i. 622, MSS. U. X.), Goronwy ap Moreiddig 
was a lawyer one of whose dicta is quoted in Dim. II. xviii. 37, and both he and 
Gwair ap Rhufon were authors of editions of the laws said to have been used by 
lorwerth ap Madog in the compilation of Ven. (i. 218). 

73 MSS. S. Z. For the supposed visit to Rome on this special business see 
note 55 above. 

74 Pref. toLL. i. xxxiv. (Bodl. MS.) and MSS. S. Z. (i. 342). Sir J. Rhys, in 


described as teacher of law to the household of Hyvvel, and it CHAP. 


is implied that his knowledge of the law and his power of 
exposition were of an exceptional kind. In agreement with 
this the Liber Landavensis speaks of "that most famous 
man, Blegywryd son of Einon," as intervening in 955 in a 
quarrel between the see of Llandaff and the king of Gwent 
and forbidding a breach of the law of sanctuary. 76 Another 
reference in the same authority suggests that his home lay in 
Gwent and, in direct opposition to the statements in the codes, 
sets him down, with his brother Rhydderch, as a layman. 78 
It would seem likely, on the whole, that Blegywryd was not a 
churchman (Welsh lawyers, in point of fact, were not drawn 
from the ranks of the clergy), 77 and that while he gave his 
services, so highly prized in his generation, to the king of 
Deheubarth in the great enterprise of his reign, he returned 
after Hywel's death to his native hamlet in the plains of 
Gwent. 78 

No copy of the original code of Hywel has survived in any 
form, 79 for not only are all extant MSS. of the laws of later 
date that 1150, but they represent improved and enlarged 
editions of the law book of Whitland, compiled from time to 
time by distinguished lawyers for the use of particular districts 
or communities. Three of these recensions, written in Welsh, 
the original language of the code, have, by reason, no doubt, 

Cymr. xviii. 115-7, discusses these verses and supplies a corrected text from the 
Oxford MS. They are clearly ancient, having been torn from their true place 
as a colophon, and are the source of many of the statements made in late MSS., 
such as that of Dim. i. 340 as to the writing of three law books and the assertion 
of MSS. S. Z. (i. 342) that Blegywryd was doctor of civil and canon law ! The 
" iudex cotidianus " is the " ynad Hys," who dispenses justice from day to day 
in the " llys peunyddiol " of Dim. i. 344. 

75 Lib. Land. 219. The mention of "houel britannicum regem " is, of 
course, an anachronism, but it is possible that the name read originally Owain 
ap Hywel. 

n lbid. 222. The " Bledcuurit " who appears among the clergy on p. 230 can 
hardly be identified with the lawyer of Hywel's day, for the grant he witnesses 
was made by the Hywel ap Rhys of Asser. 

77 See especially Dim. II. viii. 124 for the disability which excluded a cleric 
from the office of judge in a secular court. 

78 That Blegywryd was of this region is suggested by his appearance in 
955 at the church of SS. larmen and Febric, probably St. Arvan's (Lib. Land. 
219, 405), and by his witnessing with his brother (ibid. 222) the grant of " Cair 
Nonou," which was not far from Caerwent and Llanfihangel (by Roggiet ?). 

79 For a discussion of the history and relationship of the chief MSS. of the 
laws see note B. appended to this chapter. 


CHAP, of their excellence of form and substance, long ago ousted all 
x * others, so that all the existing MSS. in Welsh are copies of the 
Venedotian, the Dimetian, or the Gwentian Code, or have been 
compiled out of these three sources. The Venedotian Code has 
the advantage of the other two in respect of the age of its 
oldest MS., for the Black Book of Chirk dates from about 1200, 
while no extant copy of either of the other two can be ascribed 
to an earlier date than the fall of Llywelyn ap Gruffydd. Never- 
theless, there is good reason to suppose that it is in substance 
comparatively recent, having been compiled about or not long 
before 1200 by one lorwerth ap Madog as a special law book 
for North-west Wales. The Gwentian Code (to give it the 
title by which it has of late become generally known, though 
no evidence for its connection with Gwent or with Morgannwg 
has been adduced) appears to be also a compilation, to be 
ascribed, perhaps, to Morgeneu and his son Cyfnerth, who are 
mentioned as the authors in two or three of the MSS. The 
Dimetian Code seems to have best preserved, both in substance 
and in arrangement, the original Law of Hywel, but it is ob- 
viously a greatly amplified edition of that law and contains a 
reference to an enactment made by the Lord Rhys ap Gruffydd, 
who died in 1 1 97. Of greater antiquity, probably, than any 
of the Welsh texts are the Latin versions, of which three are 
printed in the Record edition of the laws. Their divergence 
is too marked to allow it to be supposed that they represent an 
original Latin draft by one of Hywel's clerks, from which the 
earliest Welsh editions were translated ; they were clearly 
themselves translated at different times from different Welsh 
originals, in order to enable foreign ecclesiastics to understand 
and administer the national customs. But they appear to be 
based on older Welsh texts than any which have come down 
to modern times, and the matter which is common to them 
may be taken as the nearest approximation we are ever likely 
to attain to the code which was promulgated by Hywel. 

Hywel died in 949 or 95O, 80 having done something for 
Welsh unity by his career of conquest, but far more by his work 
as a legislator. The realm he founded died with him, but the 

80 Harl. MS. 3859 has " Higuel rex brittonum (a title only previously be- 
stowed in this MS. upon Rhodri Molwynog) obiit " under the year corresponding 
to 950. 


code he gave to Wales was the beginning of Welsh jurisprud- CHAP, 
ence ; it was the solid foundation upon which the lawyers of 
later ages built at their leisure in the practice of a noble and 
peaceful profession. Though local customs still continued and 
were embodied in local documents, the conception of one law, 
valid for the whole of Wales, took its rise from the measures of 
Hywel and was developed by the activity of the legal com- 
mentators in this domain the Welsh people early arrived at 
national self-consciousness. 


Hywel the Good was not able to hand over to his descend- 
ants the authority he had acquired in North Wales, and accord- 
ingly the realms of Gwynedd and Deheubarth once more owned 
different rulers and waged war upon each other. The ninety 
years which followed the death of the legislator are filled with 
barren strife, with provincial feuds and family quarrels, seeming 
to lead to no result ; in spite of a multitude of claims, the man 
had not yet arisen who could gather the whole nation around 
his banner and breathe life and force into the national aspira- 
tions. In the background, meanwhile, was the double menace 
of the sea-rovers and the English, which the Welsh might only 
forget at their peril ; ever and anon they were roughly re- 
minded, by piratical raid or border foray, that their ancient 
enemies were at their gates, in readiness for the hour of weak- 
ness or disunion. 

It will perhaps be best to take the dynastic history of 
Gwynedd, Deheubarth and Morgannwg separately and there- 
after to discuss briefly the relations of these states with the 
English and with the "gentiles" of the opposite coast of 
Ireland. 81 On the death of Hywel, the men of Gwynedd at 

81 The last entry in Harl. MS. 3859 belongs to the year 954, and henceforth the 
historian has to rely upon the two later editions of the Menevian annals, known as 
MSS. B. and C. of Ann. C. B.T. and B. Sags, often supply valuable additions. 
The notices are meagre at this point and it is often difficult to distinguish the 
various princes named Idwal, Rhodri, etc. As to chronology, it should be 
observed that B. Saes. antedates by two years the killing of Congalach, king of 
Ireland, in 956 (Ann. Ult. and Chron. Scot. s.a. 955 ; War of G. and G. xcvii. 
44) and the visit of Edgar to Chester in 973 (for the year see Plummer ii. 160), 
but only by one year the death of Edgar in 975 and the expedition of Godfrey 
Haroldson (see note 120 below) in 987. One adds, therefore, two years to its 
dates until 972 ( 974) and one year afterwards. 


CHAP, once threw off the southern yoke and marched to meet the 
sons of the dead king under the leadership of their own princes, 
lago or Jacob and Idwal or leuaf, the sons of Idwal the Bald. 
The battle was fought at Nant Carno, in the region of Arwystli, 
on the borders of North and South Wales, and was a victory 
for the sons of Idwal, who were thereby secured in the posses- 
sion of Gwynedd, and, it may be, of Powys also. Peace and 
good neighbourhood, nevertheless, were not at once established 
between the two houses. In 952 lago and Idwal led their 
men as far south as Dyfed on an errand of fire and slaughter ; 
the sons of Hywel in 954 retaliated by a march into the Conway 
valley, where their progress was checked not far from Llanrwst 
and a defeat inflicted upon them which emboldened the men of 
Gwynedd to pursue them into Ceredigion. After these reverses, 
the southern folk forbore for a time to harass the house of 
Idwal and devoted their energies to other enterprises. Left to 
themselves, the rulers of Gwynedd spent their strength in civil 
war ; in 969 leuaf was taken prisoner by his brother lago and 
henceforth plays no part in the affairs of the kingdom, though 
he seems to have languished in captivity until 988. It was 
lago's turn next to feel the edge of misfortune ; after a tempor- 
ary defeat in 974, from which he seems to have recovered, he 
was taken captive in 979 by the son of his dispossessed brother, 
and Hywel ab leuaf thereupon became king of Gwynedd. 
Hywel retained this position for six years, and on his death in 
985 was succeeded by his brother Cadwallon. The line of 
Idwal the Bald then lost for some years its royal rights in 
Gwynedd, for in 986 Maredudd ab Owain of Deheubarth 
invaded the realm, slew the new king and annexed the northern 
to the southern state. 

Of the three sons of Hywel the Good who fought with the 
sons of Idwal, Rhodri died in 953 and Edwin in 954, 82 leaving 
Owain undisputed ruler of Deheubarth, a dignity which he held 
until his death in 988. The dynastic position of Owain was a 
strong one ; his brothers had left no heirs, and thus he was the 
sole representative of Cadell ap Rhodri, as well as of the native 
line of Dyfed, which had ended with his mother Elen. It is 
not surprising, therefore, that he should have taken special pains 

82 See note 61 above. 


to get together a permanent record of his claims. Both the CHAP, 
chronicle and the genealogies contained in Harleian MS. 3859 
were unmistakably compiled in the reign of Owain ap Hywel 
Dda, though the actual copy in which they are preserved is of 
later date. 88 The chronicle is based upon the annals kept 
throughout this period at St. David's ; its purpose is to record 
particulars of the persons mentioned in the genealogies, and it 
was originally intended to bring it as far as the year 977 an 
intention not quite fulfilled. The pedigrees immediately follow ; 
the first is that of Owain, tracing his descent from Cunedda and 
Maelgwn and Rhodri the Great ; the second is that of his mother, 
going back to Voteporix of Dyfed and that prince's father, 
Agricola of the Long Hand. A number of other pedigrees are 
then inserted, giving the lineage of the ancient princes of Powys, 
Strathclyde, Morgannwg, Ceredigion, Meirionydd, and other dis- 
tricts, but it is remarkable that in not a single instance are 
these lists brought down to Owain's own day ; 84 there is no 
mention of Idwal and his sons or of Morgan of Morgannwg 
it must be supposed that Owain desired to represent himself as 
the one lawful heir of the ancient dynasties, alike in North and 
South Wales. Nevertheless, as was noted above, he took no 
active steps after the battle of Llanrwst to enforce his preten- 
sions to the throne of Gwynedd ; he and his son Einon, who 
now appears upon the stage, turned their arms instead upon 
Morgannwg : in 960 Owain crossed the Tawe and ravaged the 
border cantref of Gorfynydd ; 85 in 970 and again in 977 Einon 
laid waste the plains of Gower, of which the chiefs of Morgannwg 
had perhaps got temporary possession. It was on an enterprise 
of this kind that Einon met his death ; in 984 the " uchelwyr " 
of Gwent fell upon him and slew him, in vindication, no doubt, 
of provincial liberties which he was seeking to destroy. Owain 
was now growing too old for the labours which were imposed 
by tribal ideas upon a Welsh prince, and his place was taken by 
his son Maredudd, who proved his mettle at the outset of his 
career by his conquest of Gwynedd from Cadwallon ab leuaf. 

88 See note on this MS. appended to chap. v. 

84 No. xvii. (line of Dunoding) is the only one which is brought down to the 
tenth century, ending about 930. 

85 B. Sues. s.a. 958 has " goryuyd" (for " goruynyd "), of which I take the 
" Goher " of Ann. C. MS. C. and the " Gorwyd " of B.T., MS. C. to be corrupt 


CHAP. The thirteen years of the rule of Maredudd (986-999 88 ) 
form something of a contrast to the time of confusion just 
described, in that this prince maintained a hold over both North 
and South Wales and opposed a bold front both to the English 
and to the Norse buccaneers. He is recorded to have led a 
raid into Maes Hyfaidd, or the plain of Radnor, where he no 
doubt sacked the Mercian villages of the neighbourhood ; 87 in 
his dealings with the sea-rovers, too, he showed an alert and 
resourceful spirit, redeeming their captives by the payment of a 
large ransom. For the times he was a man of mark, not un- 
deserving of the title bestowed upon him by the Bruts of " most 
famous king of the Britons ", 88 Yet his reign was a troublous 
one, disturbed by foreign attacks and by movements in favour 
of his nephew, Edwin ab Einon, and the sons of Meurig ab 
Idwal, who sought to win back Gwynedd for the old line. On 
his death his work was entirely undone ; Gwynedd was regained 
by a scion of Idwal the Bald in the person of Cynan ap Hywel 
ab leuaf, who ruled for six years (999-IOO5), 89 while a veil falls 
over the history of Deheubarth which suggests the beginning 
of a period of anarchy unexampled even in that turbulent 

What is most noteworthy in the period now reached is the 
success of men who were out of the direct line of succession 
from Rhodri the Great in seizing royal authority in Gwynedd 
and Deheubarth. Such a successful pretender was Aeddan ap 

86 The chronology of this reign is tolerably certain, the true date being found 
by adding one year to the date of B. Sues. Observe the following correspond- 
ences : 

B. Saes. 986 " marwolaeth ar yr ysgrybyl " = the murrain of Ann. Ult. 
986 (= 987), A.S. Chr. MSS. C. D. E. F. 986, and Fl. Wig. 987. 

B. Saes. 988 D. of" glumayn vab abloyc" = the killing of Gluniarainn, 
son of Olaf Cuaran, in Ann. Ult. 988 (= 989) and Chron. Scot. 987 (= 989). 

B. Saes. 995 Burning of " Arthmatha" = destruction of Armagh in Ann. 
Ult. 995 (= 996), Chron. Scot. 994 (= 996) and Tighernach (Rev. Celt. xvii. 
P- 350). 

87 " Maes Hyfaidd" means, of course, the vill, and not, as Woodward sup- 
posed (i. 203), the county of Radnor. See Trans. Cymr. 1899-1900, p. 125. 

88 " Clotuorussaf vrenhin y brytanyeit" (Bruts, 264). So B. Saes. s.a. 

89 About the year 1000 there appears to be confusion in the dating of B.Saes. 
The " Ivor porthalarchi " of 1001 is Ivarof Port Lairge or Waterford, who died in 
1000 (Ann. Ult. s.a. 999 ; Chron. Scot. 998). With the annal 1005 we return, 
however, to the system of adding two years, for the first year " decemnouenalis 
cicli " was, of course, 1007 (Giry, Manuel de Diplomatique, p.193). 


Blegywryd, who, after a reign of uncertain length, was killed, CHAP, 
with his four sons, by Llywelyn ap Seisyll in 1018. Another 
was Rhydderch ab lestyn, king of South Wales from 1023 to 
1033, and founder of a house which, though it failed to retain 
its hold upon the crown of Deheubarth, nevertheless played a 
prominent part in Welsh history during the eleventh and twelfth 
centuries. Of wider fame than either was the Llywelyn men- 
tioned above. Nothing is known of the origin of his father 
Seisyll ; he had, however, if the pedigrees are to be trusted, some 
royal blood in his veins through his mother Prawst, daughter 
of Elise ab Anarawd 90 and he further strengthened his position 
by marrying Angharad, daughter of King Maredudd. 91 His 
own energy and force of character did the rest : by his overthrow 
of Aeddan and his defeat at Abergwili (1022) of the Irish pre- 
tender Rhain, who claimed to be a son of Maredudd, he attained 
a commanding position in Wales, which, despite his brief 
enjoyment of it (he died in IO23), 92 was long remembered by 
his fellow-countrymen 93 and not only stimulated the ambition 
of his son Gruffydd but gave him a great initial advantage in 
the struggle for supreme power. 

During these vicissitudes the region of Morgannwg remained 
under its own princes and for the most part escaped the revolu- 
tions which distracted the rest of Wales. Glywysing, to which 
the name of Morgannwg was now frequently restricted, and 
Gwent still formed separate realms, but the kings of Gwent of 
Asser's day had, apparently, left no descendants, and both 
kingdoms fell into the hands of the posterity of Hywel ap Rhys 
of Glywysing. Arthfael ap Hywel was king of Gwent in the 
time of Bishop Cyfeiliog of Llandaff, who died in 927 ; e * he was 

90 Dwnn ii. 10, 16 (Prawst is the right spelling Owen, Pemb. i. 294). See 
note 64 above and Trans. Cymr. 1899-1900, p. 125, to which it may be added that 
Hawarden was probably in English hands and not at all likely to have been the 
home of a Welsh chief. 

91 Bruts, 296-7; Jesus Coll. MS. 20 in Cymr. viii. 88. 

M The date, given by B. Saes. as 1021, is fixed by the Irish evidence, for the 
death of " Leobelin," king of the Britons, is entered in Chron. Scot. s.a. 1021 
= 1023), Ann. Ult. s.a. 1023, and in Tighernach (Rev. Celt. xvii. p. 363). 

98 B.T. (p. 36) and B. Saes. (s.a. 1020) tell a wonderful tale of the pro- 
sperity enjoyed by the Welsh in Llywelyn'r time, with which should be compared 
what is said in Lib. Land. 253 of his ' ccessor, Rhydderch ab lestyn, and in War 
ofG. and G. pp. 136-40, of Brian 'l>oru. 

94 Lib. Land. 236 (h* u gel rex filius ris . . . filiorum suorum yuein et 
arthuail), 237. 


CHAP, succeeded by his son Cadell, who died in 942. 95 A new dynasty 
then came into possession, represented in 955 by Noe ap 
Gwriad 96 and carried on by his son Arthfael, whose murder 
of his brother Elise and solemn atonement by the gift of land 
to Bishop Gwgon (d. 982) are recorded in the Liber Landa- 
vensis.^ The sons of Elise, named Rhodri and Gruffydd, 
next ruled over Gwent, 98 but about 1020 gave place to one 
Edwin ap Gwriad, 99 of unknown origin, who was the last 
independent king of this region, holding it until he was dis- 
possessed, blinded and imprisoned by Meurig ab Hywel of 
Morgannwg. 100 Meanwhile, the house of Hywel ap Rhys had 
retained a firm hold of the country west of the Usk. The three 
sons of Owain ap Hywel had each some authority in this region, 
but the death of Gruffydd in 934 (or 935) and of Cadwgan in 
949 (or 950) finally left the whole land under the sway of 
Morgan the Aged, 101 whose long reign did not close until 974. 
The next in succession was Owain ap Morgan, whose sons Rhys 
and Hywel carried on the line into the eleventh century. 102 

The close connection between the Welsh princes and the 
court of Wessex which marked the reign of Hywel the Good 
continued for some years after his death. It would appear 
that Edred and the advisers of Edwy were anxious to main- 
tain the policy of Athelstan and to encourage the attendance 
of the Welsh at meetings of West Saxon magnates. Accord- 
ingly there is evidence of the presence in 955 of three leading 
chiefs at the court of Edred, viz., Morgan the Aged, Owen ap 
Hywel Dda and lago ab Idwal. But, when Edgar came to the 
throne, there was a reversal of this line of action, and henceforth 
no Welsh attestations are to be found in the charters of English 
kings. What occasioned the change it is not easy to discover ; 
possibly it was held that the attempt at conciliation had been 
a failure, and with no Hywel to advocate peace and friendship 
with the English, it was in fact little likely to succeed. Of one 

98 Lib. Land. 223, 224 ; Ann. C., B.T., B. Saes. (s.a. 941). Harl. MS. 3859 
has (s.a. 943) " Catt. Ilius artmail ueneno moritur ". For the true date see note 
64 above. , , . 

96 Lib. Land. 218. 97 Ibid. 244-6. 

98 Ibid. 251, 252. 99 Ibid. 249-51. 

100 Ibid. 255-7. 1(U ^ ee note 6( >. 

102 Lib. Land. 246, 252, 246-7. Hywel ab Ov-:rin died in 1043 (Ann. C., 
B.T. and B. Saes.). 


thing'one may' feel certain, namely,'; that the return to the old CHAP, 
relations of mutual suspicion betokened no real weakening of 
the English hold upon Wales. 103 Edgar, as is well known, was 
an active and strenuous monarch, and his possession of a fleet 
would make him specially formidable to the Welsh. The 
story that he imposed upon leuaf an annual tribute of three 
hundred wolves may be mere legend, 104 but the account of the 
submission to him at Chester in 973 of all the chief kings of 
these islands, including those of the Welsh, is not to be disposed 
of so readily. 105 Both the Welsh and the English contempor- 
ary chronicles speak of a great gathering of vessels at the "city 
of the legion," presided over by Edgar himself; the English 
narrative is further supported by independent evidence when 
it states in addition that six kings there met the overlord of 
Britain and swore to be faithful to him and to act with him 
by sea and by land. That the six (or the eight, if the latter 
account be accepted) included Welsh princes is highly probable, 
even though it may not be possible to identify more Welsh 
names than that of lago ab Idwal in the list handed down by 
tradition. It is, of course, a good deal more difficult to accept 
the picturesque detail that Edgar sat at the helm while the 
eight kings rowed him in his barge from the castle to St. John's 
Church and back again surely a romantic embellishment of 
the plain, unvarnished fact of the submission. Edgar the 
Peaceable, the statesman and wise administrator, was not the 
man to hazard the substance of his power by a theatrical dis- 

10S Green (Conq. Eng. (2), pp. 323-4) seems to misinterpret the situation. 

104 \Vm. Malm. G.R. i. 177 (i. 251). " ludval(us) " is Idwal, the true name 
of leuaf. 

105 For the naval gathering see Ann. C. MS. C., Bruts, 262 (where " edwart " 
is a slip), and B. Saes. s.a. 971. The two Welsh chronicles wrongly translate 
" urbe legionum " as " kaer llion ar wysc," instead of Caerlleon ar Ddyfrdwy ; 
in the Gw. Brut (p. 691), the inconsistency of this with the English account is 
solved by bringing Edgar to Caerleon in 967 and to Chester in 968 ! For the 
submission see A.S. Chr. MSS. D. E. F. s.a. 972 (=973 see note above); 
Stevenson has shown (Eng. Hist. Rev. xiii. pp. 505-7) that there is good evidence, 
dating from about 996, of the trustworthy character not only of this entry, but 
also of the fuller account in Fl. Wig. i. 142-3, so far as it names eight kings, and 
among them, those of the Cumbrians and the Scots. Florence's " Jacob " is, 
no doubt, lago ab Idwal ; " Huwal " and " Juchil " offer some difficulty, for Hywel 
ab leuaf did not obtain power until 979, while leuaf (or Idwal) lost it in 969. 
Wm. Malm. G.R. i. 165 (i. 236) probably borrowed his account from Fl. Wig. 
(Asser, ed. Stev. Ixii). 


CHAP, play of it, to the humiliation and mortification of his royal 
vassals. His hold upon the coasts of Wales was patent to all 
and needed no such advertisement. 106 

The English king did not long survive his great triumph 
at Chester, and after his death in 975 the forces of dissolution 
began to make themselves felt in the West Saxon realm, so 
that, with England parcelled out among great ealdormen, the 
crown ceased to be much concerned with the doings of the 
princes of Wales. As in the days before Alfred, it was the 
ruler of Mercia who now kept an eye upon the movements of 
the Welsh and directed punitive expeditions against them. 
Already, in 956, ^Ifhere had been invested with the Mercian 
earldom, 107 and in 967 had ravaged the lands of lago and 
leuaf; 108 in 983, at the end of his long tenure of office, he is 
found acting with Hywel ab leuaf of Gwynedd in an 'attack 
upon Einon ab Owain of Deheubarth, which the latter repelled 
with much slaughter. 109 After the banishment of ./Elfhere's 
son ^Elfric in 985, there was for many years a vacancy in the 
earldom of Mercia, and it is not easy to say who the ^Ethelsige 
was of the expedition of 992, when Edwin ab Einon obtained 
the help of an English force to harry the dominions of his uncle 
Maredudd. 110 But in 1007 Mercia received an earl once more 
in the person of the notorious Eadric Streona ; hence it is 
Eadric who in 1012 leads the English attack upon Mynyw, an 
attack which may well have been made by sea, with the aid of 
some of the Danish ships taken this year by King Ethelred 
into his service. 111 The Eilaf who in 1022 repeated this raid 
upon the ancient seat of St. David 112 was not Earl of Mercia 

106 According to the records of the cathedral church of Bangor, Edgar 
founded the neighbouring church of St. Mary, or Llanfair Garth Brannan (B. 
Willis, Bangor, p. 183 cf. 72). 

107 Crawford Charters, 84. 

108 B.T. alone introduces ^Ifhere (MS. C. Alfre) into this annal, and it 
may be the case that its compiler has inadvertently doubled the notice of 983. 

109 Ann. C. (Alfre), B.T. (Aluryt), B. Saes. s.a. 982 (Alfred). 

110 Ann. C. MS. B. supplies the form " Edelisi " from which ^Ethelsige is 
inferred (e.g., by Freeman, Norm. Conq. i. (3) 285). 

111 Ann. C. MS. B. has " Edris," MS. C. " Edrich," pointing clearly to 
Eadric Streona (Norm. Conq. i. (3), p. 351). His companion " Ubis " has not 
been traced. 

112 "Eilaph venit in Britanniam (i.e., Wales) et vastavit Dyvet et Mene- 
viam " (Ann. C. MS. C.). This was Eglaf or Eilifr, a Dane in the service of Cnut 
(Norm. Conq. i. (3), p. 447). 


(a position held^under Cnut by Leofwine), but he clearly filled CHAP. 
some post of authority on the border, for the life of St. Cadog 
calls him "an English sheriff" and tells how he invaded Mor- 
ganmvg with a mixed force of Englishmen and Danes and so 
alarmed the clergy of Llancarfan that they carried off the 
shrine of their saint to the mountain retreat of Mamheilad. 113 
Moreover, if there is no record of any Welsh invasion conducted 
by Earl Leofwine, it was his son Edwin who led the English on 
the fatal day of Rhyd y Groes. 114 

Throughout the whole of this period the hapless inhabitants 
of Wales had to bear the brunt of the piratical onslaughts of 
the sea-rovers, which were more frequent and more difficult to 
foresee and ward off than invasions from over the English 
border. It was unusual during the half-century between 950 
and 1000 for more than five years to pass without a Danish 
attack upon some quarter of Wales important enough to be 
recorded in some chronicle of the time. 115 Anglesey, Lleyn, 
Dyfed and the shores of the Severn especially suffered from this 
scourge, but no part of the coast was wholly secure. As in 
the ninth century, the raiders were chiefly attracted by the 
plunder of the monasteries ; the sack of Aberffraw, the royal 
seat of Gwynedd, in 968 stands alone, for the other places said 
to have been raided by the foreigners were all the sites of 
important churches. Holyhead was despoiled in 961, Towyn 
in 963, Penmon in 971, Clynnog in 978, 116 Mynyw (St. David's) 
in 982, 988 and 999, and in 988 a whole series of sanctuaries, 
including Llanbadarn Fawr, Llandudoch (St. Dogmael's), Llan- 
illtud and Llancarfan. It need scarcely be said that much of 
the plunder took the form of saleable captives, for the Danes 
were traders as well as pirates and the slave trade was one 
of the most flourishing branches of their commerce. 117 Two 
thousand captives are said to have been carried off by Godfrey 
son of Harold from Anglesey in 987, and in 989 Maredudd 

113 Cambro-Br. SS. 77. 114 See p. 360. 

118 Ann. C. records the attacks of 971, 982, 987, 988, 999 and 1002, B.T. and 
B. Sues, in addition those of 961, 963, 968, 972, 978, 980 and 993. A.S. Chr. 
MSS. C. D. E. s.a. 997 mention a raid on the coast of Morgannwg (" on 
NorSwalum") not noticed in the Welsh annals. 

116 B. Sues. (s.a. 977) has here the right reading " y diffeithwyt lleyn a 
chelynnauc vaur". As Ann. C. MS. C. says that Hywel won his victory of the 
following year with " gentile" aid, it is probable that " saeson" in this annal is 
a mistake of the original of B.T. and B. Saes. 

117 Cong. Eng. (2), p. 118. 


CHAP, redeemed very many of his subjects from thraldom at a cost of 
one penny a-head. The centre of Danish power in the western 
seas was the city of Dublin, where Anlaf Cuaran ruled from 
about 945 until his abdication in 980 and was succeeded by 
his sons Gluniarainn (i.e., Iron-knee) and Sitric of the Silken 
Beard. 118 But, though the "sons of Abloec" (Anlaf) in 961 
ravaged Holyhead and Lleyn, this house is less prominent in 
the tale of ruin and slaughter than that of Limerick, where a 
Danish dynasty had been established in the ninth century 
which also became master of the Hebrides and the Isle of 
Man. 119 Magnus or Maccus son of Harold, of this line, in 971 
made a descent upon Penmon, while his brother Godfrey, who 
succeeded him about 977, appears on four occasions as the 
leader of a flotilla bound for Wales in pursuit of booty. In 
972 he ravaged Anglesey ; in 980 he helped Cystennin ab lago 
in an attack upon the same island which was directed against 
Hywel ab leuaf ; in 982 he invaded Dyfed ; in 987 he and his 
Danish host, in a third irruption into Anglesey, won a victory 
over the Welsh, the fame of which for a thousand of the 
enemy were left dead on the field and two thousand carried 
into captivity penetrated to Ireland and was thought worthy to 
be preserved in the annals of that country. 120 The Welsh of 
the age of Maredudd ab Owain had no more persistent or 
pitiless foe than Godfrey Haroldson. 

At the very end of the tenth century the Danish peril was 
still as instant and real as at any time during the previous two 
hundred years. In 999 St. David's was pillaged and its Bishop 
Morgeneu was slain. The evil day was long remembered, not 
only by reason of its tragic history, but also because the 
bishop's death was deemed to have been of the nature of a 
judgment. He had been the first of all the long line of suc- 
cessors of Dewi to break the custom of the see and to eat meat ; 
no one, therefore, was surprised to hear that on the night of his 
death a bishop in Ireland had encountered his ghost, wailing 
and showing his wounds, with the pitiful cry " I ate flesh and 
am become carrion ". 121 

118 War ofG. and G. pp. 280-7, 288. lis lbid. pp. 271-2. 

120 The "Cath Manand" of Ann. Ult. s.a. 986 (= 987), won by "mac 
Aralt" and the Danes, with the slaughter of a thousand men, must be the same 
as that recorded in Ann. C., B.T. and B. Saes. 

121 Gir. Camb. vi. 104 (Itin. ii. i). 





rt o 


-> o 



1 J 'c 





* '5 




a < <8 ii* 


Authority for the 

S 3-g 

g *<-> 

Cott. Charter viii. iC 
inal charter in Br. 
The same as No. i. 

MS. Vitellius A. xiii 





> * * 

*; < c 
rt u 

o Ji 6 

5| a 

9J I-" M U t& 


f "|' ? I s -<i 

rt O^-So" 

l|rf- S c".l 
S^oq .8. -2 



S oo o-< S 





tf C!<0 



s 1 

5 | 




1 it 

i i* i 

. . W) 


- 1 

CQ rt 

? ? 

"rt 15 
.50 ^ 




. o 

1 1 11 1 

o 1 86 . 

jg rt o 


3 5 

co 3 n jJ 

a- 2 S| 



uT ^ * ^ 




"3 13 

s * 

u _- g 



' - 

O rt 'rt 

Illl |||1 


^K ffi 



~* O 3 

r|| 11^ 




. . 






*rt *O 


w w 







0) V 



2 2 



"i 3 

1 1 1 1 1 


1 I 

j is 



o 2 

z fa 

n c 
5 .S 


M (^ 

M m 


^^. t g 



^3 OO w 

M . M . (V 


4- .. Tf- . 

^.1^.1^.^ o^m vo 

PO ^ PO bkO f* 1 

) CJ 

'O ^ 


en o f^) f^o *^ ^ *^ *o 

-4 O^ 

O^ O O^ 3 C 

1 1) 

O* O* W 

O^ 4> ^ O^ ON O^ O^ O* 

< > 

z < 


. Q 



O oo 

M H 

P. *8 

M N 




o> n 

M N 

s 2"& a? 


OQ > 

> s 

a > 



= > 

j^ is i ; : 



O O 

u-> o 



10 M 

O VO M N CO tft 


M M 
M M 




CO CO * * * * 

W K rS 

N M 



I 5. 

N C4 CO CO f^ 

It ft 




* o* 

r*^ oo 



o 1 o 1 

O M N CO O\ t** 
M N OO 00 O fO 
l > ^ t^30 00 O\ O^ 





t> oo 

O> O M N CO 't- 




NOTE B. TO CHAPTER X. MSS. and Editions of the Welsh 


CHAP. The history of the various texts of the law of Hywel in Welsh and Latin is far 

X. too large a subject to be discussed within the limits of a note, but an attempt may 
perhaps be made to indicate the principal features of the problem and to furnish 
some evidence on behalf of the statements made in the body of this work. 

No progress can be made with the study of the subject until it is recognised 
that the Welsh MSS. fall into three distinct groups, representing three recensions 
of the original law of Hywel. The Venedotian, the Dimetian, and the Gwentian 
Codes, to use the names commonly applied, must be separately dealt with in any 
edition of the laws which is to be of service to the historical student. Ignorance 
of this fundamental fact very largely destroyed the value of the editio princeps of 
the laws, prepared by Dr. William Wotton (1666-1726), with the assistance of 
Moses Williams (1686-1742), and issued after Dr. Wotton's death by William 
Clarke, his son-in-law, under the title, Cyfreithjeu Hywel Dda ac Eraill seu 
Leges Wallicae, etc. (London, 1730). Moses Williams was a good Welsh 
scholar, but by selecting as the main text to be printed Titus D. II., a MS. of the 
Venedotian Code (= Aneurin Owen's B.), and representing all departures from it 
in the form of various readings, he introduced a confusion upon which learning 
spent itself in vain. Leges Wallicae preserves for us some readings not else- 
where to be found in print, notably from the lost Wynnstay MS. which Wotton 
styles LI., and the translation and glossary were valuable pioneer work. But 
fruitful study of the laws only became possible on the appearance of the edition 
undertaken by Aneurin Owen (1792-1851) for the Record Commission, which 
was published in 1841. In this the three codes are separately printed, each 
with the various readings of the MSS. of its class ; the supplemental matter 
found in many MSS. forms a second volume ; each of the Latin MSS. is printed 
in extenso. Another advantage possessed by this edition over that of 1730 is 
that adequate use is made in it of the great Hengwrt (now the Peniarth) collection 
of MSS., which is rich in copies of the laws, but was little used by Wotton and 
Williams. Exception may be taken to the editor's choice of MSS. in some cases 
(for the strange neglect of Peniarth MS. 30 (= Hengwrt MS. 12), see Evans, 
Rep. i. pp. 361, 367), but in the main the Record edition is of such a character as to 
place the student under very great obligations to it, and, short of printing each 
MS. as a separate text, it is difficult to see how, in the main, it could have been 
more usefully arranged. 

The oldest extant Welsh MS. is a Venedotian text, viz., Owen's A. (Pen. 
MS. 29 = Heng. 26), known as the " Black Book of Chirk". Old as it is, its 
ascription by Owen (i. Pref. xxvi.) to " the early part of the twelfth century " is a 
mistake ; the date suggested by Dr. Gwenogvryn Evans (Rep. i. p. 359) is about 
1200, and reasons will be given below for thinking an earlier one to be unlikely 
on other than palseographical grounds. Other important Venedotian MSS. are 
B. (Cott. MS. Titus D. II.), which dates from the end of the thirteenth century 
and records the customs of Gwynedd below Conway (e.g., as to the amount of the 
amobyr payable by a king's aillt ; cf. LL. i. 94, note 38, with the sum said in 
Rec. Cam. to be due under this head from the villein trefs of Creuddyn) ; C. 
(Caligula A. III.), written between 1225 and 1250 (Evans, Rep. i. Pref. to Pt. 2, p. 
viii) ; Pen. MS. 30, which is of the thirteenth century, and G. (Pen. MS. 35), 
which is of the last quarter of that century (Rep. i. p. 367). In spite of the age of 
the oldest MS., the Venedotian Code appears to embody a late recension of the 
law of Hywel. The passage in C. (LL. i. 218 " Ar llevyr hvn ") in which this is 
expressly asserted and the name of the compiler given as lorwerth ap Madog 
would, no doubt, carry little weight if unsupported by other evidence (see Evans, 


Rep. i. Pt. 2, Introd. viii), but the name of lorwerth occurs in other passages of CHAP. 
undoubted authority (i. 104, 292 ; ii. 20) as the editor of the code, dealing freely X. 
with older material, and the whole structure of the work, departing as it docs 
widely from all other arrangements of the laws (note, for instance, its treatment of 
the protections, sarhads and lodgings of the king's officers in bk. i), supports the 
view that it is an independent compilation. Moreover, a reference to the Historic* 
Rtgnm of Geoffrey of Monmouth (i. 184 : " Dywynwal moel mud . . . mab . . . 
yarll kernyw " cf. Hist. Reg. ii. 17) and another to the order of Knights Hospi- 
tallers (i. 170 : " Teyr gorsetua . . . abat ac escop ac hyspyty "), both occurring 
in the original text, as evidenced by E. (a transcript supplying the lacuna: of A.), 
show that this compilation cannot have been made before the middle of the twelfth 
century. Indeed, if C. is right in giving the full name of lorwerth as lorwerth ap 
Madog " vap Rahawt " (i. 292), the work can hardly have been put together before 
the first years of the thirteenth (to which A. is ascribed), for his brother, Einion 
ap Madog " ab Rahawd" (in the case of so rare a name as Rhahawd, a mere 
coincidence is not to be thought of), was the contemporary of Gruffydd, the 
eldest son of Llywelyn ab lorwerth (Myv. Arch. I. 391 (266)). It may well be 
the case that the code was compiled at the bidding of Llywelyn, who desired to 
emphasise the supremacy of Gwynedd by the issue of the laws in a distinctively 
Venedotian form. 

The oldest extant MSS. of the Dimetian Code belong to the end of the 
thirteenth and the beginning of the fourteenth century, that is, to the period 
immediately following the loss of independence. O. (Pen. MS. 36A) is deemed by 
Dr. Evans (Rep. i. p. 369) to be the oldest of all, preceding by some years L. (Titus 
D. IX.), which was adopted by Owen as the main text of his edition of the code and 
is assigned to the neighbourhood of the year 1300. This version of the laws 
embodies legislation by Rhys ap Gruffydd of South Wales (II. xxxi. 9) and its 
compilation must therefore be ascribed to the period 1150-1250. In its structure it 
conforms pretty closely to what one may suppose the original code of Hywel to 
have been, but the material has been amplified and supplemented. The special 
references to South-West Wales (I. ii. 6 ; II. xxiv. ; II. xxxi. 9) quite justify the 
title which has been bestowed upon it of the Dimetian Code. 

The so-called " Gwentian" Code is found in a group of MSS. of which the 
oldest are U. (Pen. MS. 37) and V. (Harl. MS. 4353), assigned by Dr. Evans to 
the end of the thirteenth century. It may be inferred from U. that the compilers 
of the code were one Morgeneu and his son Cyfnerth, who wrought the old law of 
Hywel into a version resembling the Dimetian Code, but shorter and more con- 
cise. According to i. 218, lorwerth ap Madog used, among other sources, " the 
book of Cyfnerth ap Morgeneu," and this is confirmed by the reference in Ven. I. 
xvi. 8 (end), introduced by " rey adeueyt " (some maintain), to a rule which is 
peculiar to the Gwentian Code (see Gw. I. xvi. 21). The code is therefore older 
than 1200, but cannot, notwithstanding what is said in the late MS. S. (i. 340) of 
" Morgeneu ynat " and his son " Kyfnerth," be pushed back as far as the age of 
Hywel himself, for "yn oes Hywel Dda" was already, when it was drawn up, 
a distant past (I. xxxv. 19). As to its local connections, there is nothing to 
connect it with Gwent, where indeed Welsh rule came to an end before noo; 
I. ii. 3 points rather to Deheubarth. 

Three Latin versions of the laws are printed in the second volume of Aneurin 
Owen's edition. The first (Lat. A.) is taken from Pen. MS. 28, which is assigned 
by Dr. Evans to 1175-1200, and is, therefore, older than any Welsh text. If one 
may judge from the inclusion in it of the section (II. xviii.) on the "Bishop- 
houses " of Dyfed (with the unauthorised addition to the list of Llawhaden), it is a 
Dimetian MS. and perhaps belonged to a bishop of St. David's. The second 
Lat. B.) is from Br. Mus. MS. Vespasian E. XI. of the beginning of the four 


CHAP, teenth century; it is also distinctly Dimetian (cf. II. xlii. lv. Ivi.), but contains 
X. material drawn from various sources. The third (Lat. C.) is also from a British 
Museum MS., Harl. 1796, and is ascribed to the thirteenth century. Though in 
no way connected with the Venedotian Code, it is a Venedotian MS., as shown 
by the reference to the supremacy of the king of Aberffraw (I. v.), while the dis- 
tinction drawn (p. 906) between the cattle of " Mon " and those from " vltra 
Menei " is evidence of an Anglesey origin. 

It has been maintained that the law of Hywel was originally written in Latin 
and that the Welsh codes are therefore translations, in so far as they draw upon 
a common stock. In the absence of direct evidence, it is difficult to say what 
happened at Ty Gwyn ar Daf, but, so far as the extant Latin texts are concerned, 
they may safely be ranked as adaptations for a special purpose from Welsh ori- 
ginals. Not only single words, but whole clauses and sentences are left un- 
translated (see Lat. A. II. ii. 22 ; xi. 9 ; Lat. B. I. xxvi. 2 ; II. xx. 10; Lat. C. xiii. 
9) ; technical terms are differently rendered in the three texts thus, for penteulu 
we have " penteylu " in the first, " princeps militie " in the second, and " pater- 
familias" in the third; while such various renderings as the following are incon- 
sistent with the notion that there was current in Wales a primitive Latin text as 
old as the time of Hywel : 

Lat. A. II. i. i. Animalium que usui hominum sunt necessaria. 

Lat. B. I. ix. 29. Animalium que necessaria sunt ad opus hominum. 

Lat. C. xiv. 7. Animalium quae necessaria sunt ad usus humanos. 
The versions would appear to have been made for the benefit of ecclesiastical 
landowners and judges who did not know Welsh. Lat. B. II. lv. I, for instance, 
explains ("solent enim Wallenses ") the Welsh system of counting " galanas " 
scores ; Lat. C. i. i adds to the usual statement that the laws of the court have 
the first place the pious qualification, " secundum seculum ". Lat. B. II. xlix. 
(De Variis Iniuriis) embodies a number of ecclesiastical canons, ancient but hav- 
ing no reference to the forms of Welsh law (H. and St. i. 127). 

It thus becomes clear that no MS. in Welsh or Latin preserves for us the 
original code of Hywel. The Latin, no less than the Welsh, MSS. speak of the 
time of the great legislator as a bygone age (Lat. A. II. xxiv. 32 ; xlii. 2 ; Lat. 
B. I. xiii. 31 ; II. xi. 4). The nearest approach to evidence of what was con- 
tained in the first law-book is the consensus of all codes and versions, and there is, 
in point of fact, so much in common between them as to make this criterion not 





A history of Wales. .LB