The Welsh wars of Edward I : a contribution to mediaeval military history, based on original documents

Preface

If it be true that the English learnt on the battlefields of Scotland the experience which was destined to make them the victors at Crecy and Poitiers, it is equally true that the preliminary experience had been learnt in Wales. It was precisely because I wanted to trace back the evolution of the typically English combination of knights with archers beyond the battle of Falkirk, which is usually considered to have been the first great victory of the longbow, that I began to study the documents of the Welsh wars. The period, rather more than eighteen years, is of peculiar interest. The rolls are tolerably exhaustive ; we have minute details concerning the men in the king's pay in most of his armies, besides evidence which enables us to conjecture the strength of others. Feudal and paid troops are found fighting side by side. We can form a good picture of Edward as he steadily fought Llewelyn and David for fifteen months without pause, grimly determined to succeed, though he had to keep the field through the winter and to bring up mercenaries from Gascony after a serious defeat. We see him able to devote his whole attention to the task, for there was no agitation of earls and barons hampering him in his military needs by insisting that he should confirm the charters, whereas the first and foremost of the reasons of his failure in Scotland, in spite of^ the victory at Falkirk, was the baronial agitation in England. We have an insight into the various subsidiary tasks of

as

VI

PREFACE the war, details of the woodmen who cut the paths through which the soldiers marched, of the expenses of castle-building, and so forth. But especially important is the knowledge, which documents alone can give, of the earliest efforts in four wars of that infantry, a raw and untried infantry, which, under Edward's organization, learnt the art of war so that their grandsons under his grandson were able to startle Europe by their good shooting on the terraces of Crecy. It is clear that the earliest foot, partly spearmen, partly bow-armed, were drawn alike from the Welsh 'friendlies' of marcher lordships and from the border counties between Lancashire and Gloucestershire, while the men of nearly all the other English counties had yet to be trained in the Scottish wars. Edward's employment of his late enemies in Wales to fight for him in Flanders and Scotland is suggestive.

The documents are chiefly the pay-rolls among the Exchequer Accounts preserved in the Record Office. Very often the pipe rolls give us facts which are not to be found elsewhere, and the patent and close rolls are helpful ; it is more than likely that I have missed some material among the mass of writs, for the great length of these rolls is bewildering when there is no printed calendar at hand. The marshal's rolls for 1277 and 1282 are printed in the Parliamentary Writs. The most useful of the chroniclers I have found to be Hemingburgh and Trivet (English Historical Society), and Wykes of Osney {Annates Monastici in the Rolls Series), while Mr. Martin's Preface to the Letters of Archbishop Peckham (Rolls Series) has led me to work out several points.

Although my chief object has been to depict an English army in the field at the most critical point in

PREFACE vii

mediaeval military history, namely at the very beginning of a systematic organization of infantry, I have thought it best to treat the story of the Welsh struggles consecutively, sketching roughly the course of the Anglo- Norman aggressions, the creation of the marcher lordships, the Welsh reaction in the days of civil war in England, and Edward's reasons, indeed his justification, in attacking Llewelyn. It is clear that there was no bond of union between the Welsh of the north and the centre and south, and independence was represented by the princely house of Gwynedd alone. The strength of the one strong section of the Welsh race was the mountain fortress of the range of Snowdon, where guerilla tactics were possible. The two centuries of the successful resistance of Gwynedd up to Edward's reign are quite intelligible, and we have an object-lesson in Wales of the geographical problem in military undertakings. It is easy to sneer at the powerful tyrant who crushes a brave little race struggling for liberty, but military students must admit that resources and good organization alone can prevail where geography helps the numerically weak defenders. The larger and stronger race often has to incur the accusation of tyranny, even if provocation is undoubted and the need of union is great.

I submit that the Welsh wars have an influence on English constitutional history. I have tried to show the weakness of the feudal system and the genesis of the paid soldier, who was a native Englishman or a Welsh friendly, and no longer a foreign mercenary as under the Angevin kings, an occasional importation of Gascons notwithstanding. The custom of the march explains much of the obstinacy of some of the barons in their defiance of the crown, and grievances which

Vlll

PREFACE seem to have originated in Wales come to the front when the new enemy is Wallace. The money question is also important, because Edward could afford to fight to the end in Wales, but the expense incurred in Scotland was very considerable.

I have nowhere found a satisfactory detailed account of these campaigns, for the chroniclers write so vaguely that they are most misleading unless compared with each other and with the documents. It is only after close research that I venture to offer this contribution to mediaeval military history.

Lastly, I have to apologize, if one ought to apologize for anything done deliberately, for repeating in many cases the exact words which have been used in a previous chapter, for I have often, in reading some book, felt annoyed at having to turn back for information.

J. E. M.

Contents

PAGES

Preface v-viii

Chapter I

The * Normanization' of the march, i ; counties palatine, 3; the ' Laws of Breteuil,' 4 ; the Normans in North Wales, 6 ; in South Wales, 10; the house of Clare, 12; Giraldus Cambrensis, 15 ; invasion of Ireland, 17 ; results of civil war in England, 19 ; state of the marches in the thirteenth century, 20 ; Llewelyn ap Grufifudd, 23 ; his alliance with Simon de Montfort, 24 ; retrospect of military developments since Hastings, 25 ; battle of Lewes, 30 ; Evesham and onwards, 32.

Chapter II

Knight- service, 35 ; scutage in 1279, 36; details of great fiefs, 39 ; how scutage was assessed, 42 ; cavalry raised by feudal system, 45 ; principle of a quota, 46 ; question of mounts and equipment, 49; the constitution of a retinue, 55; the retinues of earls, 57 ; of lords marchers, 61 ; of other barons, 63 ; lack of military experience, 66 ; introduction of a system of pay, 68 ; military legislation, 72 ; feudal and paid service, 74 ; maximum of available cavalry, 80 ; horses and armour, 82 ; the household, 84 ; tenure by serjeanty, 87 ; crossbowmen and crossbows, 88 ; other infantry, 92 ; Edward's improvements in organization, 95 ; the longbow, 99 ; tactics, 104 ; the ships of the Cinque Ports, ic6 ; a note on scutage, 108.

X CONTENTS

Chapter III Pages

Llewelyn and Gloucester, no; the treaty of 1267, in; Llewelyn during Edward's absence, in; fear of revival of Montfortian party, 114 ; preparations for war, 115 ; state of the march, 116; Earl of Warwick at Chester, 118; Mortimer and Lincoln at Montgomery, 120 ; Pain de Chaworth in South Wales, and treaty with Rhys, 123 ; results of preliminary campaigns, 125 ; the muster of the feudal host, 126 ; the fleet, 128 ; the plan of campaign, 129; Flint, 130; reinforcements, and end of feudal forty days, 131 ; Grey's horse, 133 ; Rhuddlan, Diganwy, and Anglesey, 134 ; Llewelyn submits, 135 ; Edmund at Aberystwith, 136; review of details of infantry and workmen, 138 ; the budget of the war, 140 ; the treaty with Llewelyn, 142 ; English ideas in Wales, 142; Flint and the Cantreds, 144; Rhuddlan, 145 ; Llanbadarn and South Wales, 146; the central march, 147 ; Builth, 148.

Chapter IV

Features of the second war, 149 ; list of Welsh grievances, 150; David seizes Hawarden, 153; raises South Wales, 154; Edward's writs from Devizes, 155 ; the first forces concentrated, 155 ; new writs, question of paid and feudal service, 156 ; muster in June at Chester, 159; advance on Hope by Grey, 161 ; by Edward on Flint and Rhuddlan, 162 ; Gloucester's defeat at Llandeilo, 166 ; appointment of William de Valence, 167 ; feudal force in South Wales, 168 ; advance into Cardiganshire, 169; Llewelyn in Carmarthenshire, 170; Earl of Hereford in Brecknock, 170; Mortimer on the Severn, 171 ; fleet and feudal quotas at Rhuddlan, 173; infantry in main army, 174; Tany in Anglesey, 176; Edward occupies Denbigh and the Clwyd, 178; defeat of Tany near Bangor in November, 179; Llewelyn on the Wye in December, 181 ; battle of Orewin bridge, 183; reoccupation of Llanbadarn, 185 ; finances, 185 ; aid from Gascony, 186; numbers of the Gascons, 188; renewed advance in January, and fall of Dolwyddelan, 190; occupation of Conway in March, 191; David in Merioneth, 192; siege of Bere in April, 192 ; David captured in June, 194 ; the budget of the war, 196.

CONTENTS xi

Chapter V Pages

The Peace Settlement and Rhys's Rising . . . 198-219

Castles and disposition of troops in 1283 and 1284, ig8 ; the annexation, 199 ; Edward's royal progress, the northern counties, 200 ; Tibotot and the southern counties, 200 ; Gififard and the Earl of Hereford, 201 ; the Earl of Gloucester, 202 ; details of the household, 203 ; rising of Rhys, general state of South Wales, and his grievances, 204 ; concentration of the English and Welsh friendlies, 207 ; Tibotot's corps, 207 ; the Earl of Cornwall's, 208; Grey's and I'Estrange's, 209; cavalry, 210; the Earl of Gloucester in Brecknock, 211 ; siege of Dryslwyn, 212; army disbanded, 213; Plukenet in garrison at Dryslwyn, 214; fresh outbreak, 214; relief of Dynevor, 215; siege of Emlyn, 216; pursuit of Rhys, 217 ; settlement of Carmarthenshire, 217 ; new appointments, 218; cost of this war, 219.

Chapter VI

Edward returns from Gascony, 220 ; position of the Earl of Gloucester, 221 ; position of the Earl of Hereford, 222 ; his feud with Giffard, 222 ; feud between the two earls, 224 ; Edward forbids private war, 225 ; Gloucester's defiance and raids, 225 ; Hereford's suit against Gloucester, 227 ; marcher lords refuse to swear in court, 229 ; facts proved against Gloucester in his absence, 231 ; the king takes the case himself before a full court, 231 ; Hereford's transgressions in the interval, and sentence, 232 ; Gloucester's excuses for previous absence, and defence concerning the three raids, 233 ; sentence against Gloucester on the old charge, 235 ; nobles come forward to offer security, 235 ; final sentence at Westminster, imprisonment, and confiscation, 236 ; other cases of march disobedience, 237 ; Edward taxes the lords marchers, 239.

Chapter VII

The Last Rising; Madoc, Maelgwn, and Morgan; 1294 and 1295 ......... 240-270

Expedition to Gascony of 1294, 24c ; enrolment of Welsh foot, 241 ; causes of Welsh rising, 242 ; unexpected outbreak, 244 ; Edward's steps to concentrate troops, 244 ; letters of protection for the war, 245 ; campaign in South Wales, 248 ; case of the Earl of Norfolk, 250; revolt in Glamorgan, 251; Brecknock, xii CONTENTS

PAGES

Builth, Bere, 252 ; Carnarvon and Denbigh, 253 ; Edward's advance, 254 ; disaster at Bangor, and Edward blockaded in Conway, 255 ; victory of Earl of Warwick over the Welsh, and relief of Conway, 256 ; Edward pauses before advancing afresh, 258 ; relief of western castles by sea, 259 ; English army reconstituted, 260 ; Grey's paid cavalry, 262 ; king's progress to Anglesey, thence along west coast, 263 ; thence into South Wales, 264 ; return northwards and thence to England, 265 ; revolt finally quelled by Havering, 265 ; investigation into grievances of Welsh, 266 ; castle-building at Carnarvon and Beaumaris, 267.

Chapter VIII

Evidence of armies in Flanders and Scotland, 271 ; cavalry of the campaign of 1296, 273; the constitutional troubles of 1297, 274 ; cavalry in Flanders, 277 ; infantry in Flanders, 280 ; question of the Confirniatio Cartarum in Edward's absence, 280 ; effect of Wallace's victory on English constitutional history, 283; forces at Newcastle and Berwick early in 1298, 284; new army enrolled for summer of 1298, 286; the foot nearly all Welsh, 287 ; the paid cavalry, 288 ; the feudal cavalry, 290 ; approximate total, 292 ; Hemingburgh's account of Falkirk, 292 ; Edward's difficulties in spite of victory, 293 ; no campaign in 1299, 297 ; the Caerlaverock campaign, 299; Edward and his barons, 303 ; last war in Scotland, 304.

APPENDIX :

I. Note on the Death of Llewelyn . . 306

11. Madoc's War 308

III, Castles in Wales . . . . . . 310

IV. Cavalry at Falkirk 313

PEDIGREES :

I. Royal Family and connexion with Baronage. II. Clare — Hastings — Valence — Bigod.

III. BoHUN — Braose — Hastings — Mortimer.

IV. Lacy — Lancaster — Chaworth.

V. FiTZALAN WaRENNE.

VI. Braose — Clifford.

VII. Geneville — Lacy — Verdun.

INDEX 315

THE WELSH WARS OF EDWARD I

Chapter I

The conquest of Wales was first taken in hand by the restless adventure-loving Normans who had come over with William the Conqueror. It is quite true that Saxon Harold had already made an impression, and the final annexation by Edward may very fairly be regarded as the assertion of English supremacy. But, during the two centuries which separated Harold and Edward, the prominent figure on the march of Wales was the Norman adventurer. He had fought at Hastings as an adventurer looking forward to a grant of English land as his reward, and he had been disappointed ; for William conferred manors, but took good care that the baron should not become independent of the crown of England. On the borders of Wales, however, things were different. There was a hardy and active race to be conquered, and regular methods of war were out of place. The men of the mountains fought in the guerilla fashion. They had natural fortresses to which to retreat, and it could never be expected that, being beaten in one battle, they would submit at once and for ever. Therefore William and his sons could authorize, or else tacitly allow, what in England they were determined to prevent. If their Norman lords loved adventure and fighting, they might settle on the march, conquer as much of Wales as they could, and hold as their own what they had conquered ; so that their high spirit might be thus diverted into a country where they had plenty to do, and little opportunity to make themselves dangerous to their king. To the native Welsh chroniclers the invaders of their country were ' Frenchmen,' who fought in the ' Gallic ' style. The

MORRIS B t^ mailed horseman and the stone castle were always in evidence, the grim Norman fighting-man and the grim Norman fortress. Like their ancestors who conquered Normandy, like their kinsfolk who pushed on to Italy and Sicily, the Norman lords who settled on the borders of Wales — lords marchers they came to be called — domineered and grasped. They conquered for their own profit, and they satisfied the ambition which was characteristic of their race. They had their fill of the fighting which they loved, they had excitement, and they were their own masters. They were much more free than the barons of England. Or it might be that some lord had both fiefs in England and lands on the borders of Wales: in the one case he was fettered in various ways, might not build castles at his will, and had to perform the duties imposed on him by the conditions of feudal service : in the other he was free to fight his way into Wales, and to build castles so as to hold what he had won. In this way the work of ' Normanisation ' was more complete on the march than elsewhere in our island ^.

One can hardly say, however, that the early Norman kings deliberately devised this method as the cheapest way of conquering Wales and also of giving to their barons plenty of that hard fighting which would keep them too busy to rebel. The ' custom of the march,' namely, that the lords marchers could wage private war at will, grew up rather as a natural consequence of the state of affairs which prevailed after the Conquest. None of the kings could spare the time which was necessary for a permanent conquest of Wales. They had to organize and govern in England, look after their possessions in Normandy or Aquitaine, gratify their ambition by the more highly attractive task of crusading, or crush baronial risings. Thus for more than two centuries no king put himself resolutely to subdue Wales until Edward did. Royal expeditions were, indeed, frequent, but spasmodic and rarely successful. Meanwhile, the resistance of the Welsh was being gradually worn down by the lords marchers, who had a direct

^ Morgan, E71 gland under the Norman Occiipatioti, p. Z"] ; Ejtg. Hist. ReviLW, Jan. 1900, p. 74.

THE NORMAN LORDS MARCHERS 3

profit in the work, the kings rather allowing than actually authorizing them to do so. The lords who were originally planted by the Williams on the flanks of the Welsh, their direct descendants^ or the new men who by marriage with heiresses came in to carry on their work, must have the sole credit for the successful wearing down.

But in saying this we have to make a strong distinction between the fortunes of the Norman invaders on the northern and middle marches and those of their more lucky comrades of the south. There are two separate chapters in the history of Wales up to the reign of Edward. North Wales and South Wales were quite distinct both politically and geographically, the great block of central mountains forming a barrier to keep them apart. The Princes of Gwynedd, a district roughly equivalent to Snowdonia and outlying valleys, claimed at times overlordship over the Princes of South Wales, Cardigan, and the Vale of Towy. But rarely did they actually penetrate to the south with an armed force. Similarly, the Earls of Chester and their neighbours had little in common with the lords marchers of Glamorgan and Gwent. And as war was waged independently in either direction, so the results varied. In the two centuries the Norman conquest from Monmouth to Pembroke and Cardigan was steadily successful, following the line of the coast and striking up the valleys of the Usk and Wye, and this region was studded over with Norman castles. On the other side very little advance was made beyond the present boundary lines of Cheshire and Shropshire, the efforts to conquer along the north coast were unsuccessful, and a belt of fortresses rather protected the western parts of England than threatened the north-east borders of Wales. When Edward came upon the scene he found South Wales more or less securely held by the marchers, but he had to start from the very confines of Chester itself to carry out his great invasion of Gwynedd.

William I began by creating two counties palatine — counties whose earls were practically independent administrators, — namely, Chester over against the Princes of Gwynedd, whose power stretched out from Snowdon to Flint ; and Shrews-

B3 bury against the Princes of Powys, who were settled on the upper Severn and upper Dee. There was also an outwork at the mouth of the river Clwyd, where rose the castle of Rhuddlan, some three miles from the sea on the right bank. The earls palatine had to hold the border, and to conquer across the border as they could. William was, in fact, making a dangerous precedent. He, whose great aim in England was to avoid the creation of powerful feudal baronies which might be too strong for the crown, was on the Welsh border founding two such possible sources of danger. So he had to choose his men carefully. The first Earl of Chester was his own nephew, Hugh of Avranches. Hugh held Rhuddlan of the king, and kept half of the borough in his own hands in dominio, while a certain Robert was his man and held the other half. The Earl of Shrewsbury was Roger of Montgomery, who in 1066 had been left behind in Normandy as one of the regents. Personal loyalty was the tie which was expected to bind them to the crown ; for, if to lands in Normandy and England they should add by conquest large possessions in Wales, and then aspire to be altogether independent, they would be accomplishing the very thing which it was the Conqueror's life-work to prevent. Roger was especially powerful ; he held fiefs in Normandy and Picardy through his father and his wife, was Earl of Arundel and Chichester in England, and now founded a new Montgomery in Wales in remembrance of his Norman home. On his death, the hereditary possessions passed to his eldest son Robert, the acquired possessions to his second son Hugh.

A quantity of valuable evidence has been collected by Miss Bateson \ which shows that a group of lords marchers cooperated in the work of Normanisation on the Welsh borders. Certain burgess rights were introduced from Breteuil in Normandy to Hereford, Shrewsbury, Ludlow, and other places in the west of England ; hence they were passed on to Rhuddlan and Montgomery, to Cardiff and Haverfordwest. Some of the evidence is direct and undoubted, coming from Domesday Book ; some is indirect, and the existence of the rights

THE 'LAWS OF BRETEUIL' 5

at the end of the eleventh century is inferred from their existence later. The various lords marchers who introduced them were connected with each other, most of them also being kinsmen of the Conqueror, and it would follow that there was a conscious co-operation under the sanction of the king himself. Near a castle was formed a bourg, a Frankville^ where Frandgenae biirgenses lived in enjoyment of certain privileges as tenants of some great lord : these would be Norman adventurers attracted to settle on favourable terms alongside an existing English borough, or to form an entirely Norman nucleus in a new foundation. William FitzOsbern, cousin of the Conqueror, and already seneschal for him of Breteuil, then created Earl of Hereford, must have first introduced the ' laws and customs ' : the phrase in Domesday Book is expressive, ^ leges ei constietiidines quae sunt in Hereford et in Bretidlll and FitzOsbern's connexion with both places points to him as the originator. Under Hugh of Avranches and Robert of Rhuddlan the laws were extended to Rhuddlan ; under FitzOsbern or, more likely, under Roger of Montgomery, to Shrewsbury and to Montgomery, where they are undoubtedly found in the thirteenth century ; under Roger de Lacy, ' son of Walter de Lacy, who was FitzOsbern's man,' to Ludlow ; under Arnulf, son of Roger of Montgomery, to Haverfordwest ; and under Robert FitzHamon, a son-in-law, to Cardiff. The cumulative evidence thus adduced by Miss Bateson is very strong, even if the connexion in each case cannot be proved directly. A Norman colony with its own rights was planted at many of the most important posts on the border. So valuable was this Norman element, or so welcome were the rights conferred by the * laws of Breteuil,' that they spread to many new places later on in Wales and in Ireland as fast as the Normans conquered in each country. Though it is true that the laws were introduced to other places than the march, still these other places were in the west of England and under the influence of members of the same group of families ; for instance, Roger le Poitevin, another son of Roger of Montgomery, gave them to Preston, and a kinsman of FitzHamon gave them to Bideford.

Hugh of Chester, Roger and Hugh of Montgomery, and Robert of Rhuddlan all contributed towards the work of conquering Wales which had been assigned to them. Success seemed to be certain, and even the island of Anglesey was for a time occupied. But then the Welsh roused themselves, and with the aid of Norse Vikings regained the island. Hugh of Montgomery was shot by an arrow in the eye, ' as though paying the wergeld for England's fallen king^.' His death was of great importance. Not only did the Welsh roll back the tide of Norman advance, but also the county palatine of Shrewsbury, becoming vacant, reverted to the elder brother Robert. An apology is necessary for the introduction of so well known a figure, yet it is not out of place to discuss his position. He now succeeding to Hugh's inheritance could control the resources of the county palatine, the Sussex fief and earldom of Arundel, both his father's and his mother's lands across the Channel, with his wife's possessions in Ponthieu ; in history he is generally known as Robert of Belleme, the county which was his earliest inheritance through his mother ^. He acted as his wealth and influence tempted him to act. On the chance of finding Henry I weak on his accession, he aspired to play the part of an independent baron, was unexpectedly crushed, and driven out of England. It was a critical moment in the history of the Welsh march, Henry acted with vigour. Not content with crushing Robert he also crushed another brother, Arnulf of Montgomery, who was settled in Cardigan, for it was imperative to extirpate the family which had dared to dictate to the King of England. Montgomery became a royal castle : Shrewsbury ceased to be a county palatine, and was joined to Staffordshire as a county of the ordinary type under the same sheriff. Thus, on the one hand, was overthrown a power which would have been a bulwark against the Welsh of Powys and a base for the gradual conquest of Central Wales ; but, on the other hand, a serious attempt to utilize the peculiar position of a palatine earl towards overawing the crown of England was

^ Norman Conquest, vol. v. p. 113. ^ Diet. Nat. Biog., vol. iv. p. 183.

CHESTER AND SHREWSBURY. 7

thwarted. No family rose to take the vacant place. The Fitzalans obtained the lands of Clun in the south-west of Shropshire and Oswestry in the north-west, and succeeded in course of time to the earldom of Arundel ; but they held by knight-service, and were hampered by the ordinary feudal restrictions. The Mortimers became an important family on the middle march : Ralph, their founder, was both a sub-tenant of Robert of Belleme and a tenant-in-chief of the crown. But, as a rule, they were either loyal or prudent, and did not yet aspire to independence.

The Earls of Chester, the other county palatine, were loyal ; at least, they did not set themselves in opposition to the crown. They must have been engaged in constantly fighting the Welsh. But the lords of Gwynedd, when once they had recovered Anglesey, more than held their own against the Normans of the march. Frequently we find Rhuddlan occupied by the invaders, but as frequently it was conquered or destroyed by the Welsh. Diganwy, on the right shore of the estuary of the Conway, changed hands in the same way. Neither Rhuddlan nor Diganwy was a centre for gradual conquest and Normanisation. The ' Four Cantreds,' the district lying between the Conway and the Dee, that is to say between Gwynedd and Chester, were often the battle-ground of Welshman and Norman ; but, when at last they were thoroughly reduced by Edward I, they were still quite Welsh at heart.

There was a line of Norman castles built westwards, and therefore forming a sort of advanced screen ahead of Chester and Shrewsbury. Such were Mold, Hawarden, Oswestry, Caurs, Bishop's Castle, Ludlow, and of course Montgomery. But, as was suggested above, they rather formed a belt of fortifications to protect the western borders of England than were bases for the gradual absorption, under Norman influence, of the Four Cantreds or Powys. The result was that the western parts of the present counties of Cheshire and Shropshire were securely held, and were acclimatized to the Anglo-Norman system of government, the inhabitants being very largely Welsh, until a line of demarcation was at last drawn between England and Wales. But hardly any impression was made further westwards beyond this belt. The reason is probably to be found in the suppression of Robert of Belleme by Henry L The crown had had a lesson, and could not afford to let any other family rise to a similar position of independence. Where there was no incentive to private adventure, there was no powerful irruption of Norman adventurers into the heart of Wales. But an exception may be made in the case of the Mortimers, who conquered along the upper Severn.

The failure to conquer along the coast, and thence up the Clwyd and Conway, is also explained by geography. There is a very narrow strip between the sea and the mountains, and the mountains were then covered with dense forest. The Welsh had cover close upon the flanks of an invading force. Not being able to maintain themselves on this narrow strip, the Normans had no base from which to advance inland up stream. To venture with inadequate transport to cross the watersheds, so as to conquer down stream, was out of the question. Lastly, Chester was not a good naval base. When we come to consider what difficulties were surmounted by Edward, how with a genius for organization and great resources at his back, he could only by great effort cut his way into Wales, hewing paths through the forests and maintaining his lines of communication by sea with ships brought up from Kent and Sussex, we can then understand how, before his time, natural obstacles defied invasion.