THROUGHOUT most of this period the countries of Western Europe had little or no respite for the establishment of permanent fortifications. Spain was overrun by the Saracens, who, but for their decisive defeat at Poitiers in 732, would have overrun France also. Italy, Germany, France and England were all engaged in a continuous and desperate struggle for existence—either with other sections of their own peoples or against wave after wave of barbarous invaders. From the fifth to the tenth century the military architecture of those countries consisted largely in the repair of existing Roman fortifications or the building of others on the same plan. Even the military works of Charlemagne were designed on the Roman model.

Contemporary chroniclers confine their notices of buildings mainly to those of ecclesiastical or civil character, but here and there references to fortresses occur. In a poem written in the sixth century Fortunatus, Bishop of Poitiers, describes a castle built on a precipi- tous eminence on the banks of the Moselle, the foot of the hill being washed by the river on one side and by a stream on the other. The curtain was strengthened by thirty wall-towers; and a tower con- taining a chapel and armed with ballistae guarded the approach. The palace (aula) stood on the summit of the hill, and was of con- siderable size and magnificence .

Other references occur, but they are mostly vague in character; and, having regard to the paucity of remains and the facility with which the Normans overran northern and western France in the ninth century, the military works there at that period could not have been of great strength. Indeed so little faith was placed in fortifications by Charlemagne and his immediate successors that permission was given to some bishops to pull down the walls of their cities, and use the material in building their cathedrals and churches.

But the raids of the Normans demanded that France should put itself in a state of defence. In 862 Charles the Bald, King of France, ordered the construction of fortresses at all points to resist the invaders. The response to this order was immediate. The bishops began to repair and rebuild the walls of their cities, and nobles to build private castles. The multiplication of private strongholds became so great a menace to the authority of the crown that in 864 Charles issued the Edict of Pistes, ordering the destruction of all fortresses which had been built without royal licence. But by a further edict of 869 instructions were issued for the fortification of all towns between the Loire and the Seine.

In England, following the departure of the Roman legions the inhabitants continued to defend themselves behind the Roman fortifications of London, Colchester, Lincoln, York and elsewhere. But the successive raids of the Jutes, Saxons, Angles and Danes, the continuous and desperate strife between these peoples after their arrival and penetration of the country, and the constant passing of power into the hands of fresh nations of rovers were all factors inimical to the establishment of fixed defences. The conflict was ceaseless; and it was a war rather of pitched battles than of sieges. Both the Saxons and Danes constructed earthworks here and there throughout the country; but it was not until after the Peace of Chippingham in 879 that Alfred the Great and, following him, his son and daughter were able to turn their attention to systematic fortification.

It was essential for the progress and development of the people that each community should be provided with adequate defence, and be secure at least from sudden attack. To this end Burns, or fortified towns, were built in suitable positions in various parts of the country. From 910 to 924 Edward the Elder and his sister, Ethelfleda, Lady of the Mercians, built over twenty such fortified towns, including Hertford, Tamworth, Stafford, Warwick and Towcester. They also repaired and strengthened existing defences, including those at Huntingdon and Colchester.

Of the character of these fortifications we have very little reliable evidence. The illustrations in Anglo-Saxon manuscripts which show towns surrounded by stone walls and towers are not helpful in this respect, because the illustrations themselves have been deliberately and meticulously copied from more ancient sources, and can have no local application. The Psalter Harl. M.S. 603, for example, dating about A.D. 1000, contains illustrations which have been reproduced again and again as illustrative of Saxon buildings and life. But these drawings were closely copied from those in a manuscript of the ninth century, now in the Library at Utrecht and known as the Utrecht Psalter; and they again were copied from earlier classical types. And this is a typical example. The Anglo-Saxon Chronicle states tersely in respect of the towns that they were built or repaired, as the case may be; except in respect to Towcester, which the Chronicle states was encompassed with a stone wall . Florence of Worcester speaks of old towns, constructed of stone, which by orders of Alfred the Great were moved from their old sites and handsomely rebuilt in more fitting places . Therefore it may reasonably be inferred that where stone could be procured the fortifications of this period were of masonry. But where wood was plentiful and stone scarce they were probably of timber.

With the advent of the Normans to the court of Edward the Confessor, in the middle of the eleventh century, a type of fortification much employed in Normandy was introduced into England. In France as early as the ninth century, as we have seen, the nobles were building private castles. By the middle of the eleventh century a special form of fortification had developed, now known as the motte and bailey type, and consisting of a motte, or mound, varying from 10 ft. to 100 ft. in height and from 100 ft. to 300 ft. in diameter, and of one or more wards, or baileys, the motte and baileys being surrounded by ditches. Fortifications of this type are found in Germany, Italy and Denmark, but they are most profuse in Normandy and England.

The mounds are of three kinds: natural hillocks, partly natural and partly artificial, or wholly artificial, according to the nature of the sites they occupy. And it should be noted here that mounds are often described by writers as artificial before any investigation as to their composition has been made. The mound at Launceston Castle, hitherto referred to as artificial, was found, on excavation being made a few years ago, to be a natural hillock throughout.

Occasionally the mound stood alone surrounded by its ditch, but normally it was protected by one, two or more outworks, or baileys, the baileys being arranged in such order, dictated by the character of the site, as would defend the keep on the mound. If there were two baileys, in some cases they were in line in front of the mound, in others on either side of it, and in others again side by side in front of the mound. A ditch was carried all round the fortification and also between the mound and the baileys and between the baileys. In some cases the mound stands entirely within the bailey, as at Bramber in Sussex and Skenfrith in Monmouthshire. At Arundel the mound stands exposed on one side in the centre of a long and relatively narrow bailey. At Lewes and Lincoln there are two mounds, and at Hedingham and Old Basing there are no mounds.

Castles such as these, placed in strategic positions, while being a source of danger to a weak ruler when held by truculent and disaffected lords, were of great value to a powerful king, such as William the First, when held by trusted vassals. A few castles were built in England during the reign of Edward the Confessor, as at Ewias Harold and Richard's Castle, both in Hereford; but following the Norman Conquest they were spread profusely throughout the country.

If the mound was artificial, then the defences on its summit must have been of timber; and such a fortification is depicted on the Bayeux tapestry and described by Jean de Colmieu, writing about 1130.

Speaking of the flat open country south-west of Calais, Colmieu says: " It is the custom of the nobles of that neighbourhood to make a mound of earth as high as they can and dig a ditch about as wide and deep as possible. The space on the top of the mound is enclosed by a palisade of very strong hewn logs, strengthened at intervals by as many towers as their means can provide. Inside the enclosure is a citadel, or keep, which commands the whole circuit of the defences.

The entrance to the fortress is by means of a bridge which, rising from the outer side of the moat and supported on posts as it ascends, reaches to the top of the mound . Recent investigations have disclosed such a castle of the smaller type at Abinger in Surrey. Here on the summit of the mound there is a close row of post holes for the palisade and other posts, outlining a rectangular site in the middle for the keep.

Another chronicler, Lambert d'Ardres, writing of a wonderful timber keep built in the eleventh century (also in the flat country south-west of Calais) describes an elaborate structure of three storeys containing halls, chambers, a guardroom and a chapel. The ground storey was occupied by storerooms .

In districts where conditions were favourable the keeps were built of stone. The keep of the castle of Brionne in Normandy, built about 1045, was of stone, as described by William of Poitiers, writing in the same century . The keep at Langeais on the Loire, built at the end of the tenth or beginning of the eleventh century, is of stone. Of the Norman castles represented on the Bayeux tapestry, one is clearly built of stone, and stands on the ground; while two others are represented much in the same manner as is the Westminster Abbey of the period, known to have been of stone. In some cases, as at Corfe Castle, Dorset, though the keep was of stone, the baileys were defended for many years by stockades.

At Durham Castle, built first about 1072, the keep on the mound was remodelled in the fourteenth century and largely rebuilt in 1838-40. It has been stated frequently that the original keep was of timber, but there is no evidence to that effect. The shape of the existing structure is similar to that of many shell keeps built in the eleventh and twelfth centuries, as at Tickhill and Lincoln; while it bears no sort of resemblance to keeps known to have been built in the fourteenth century, as at Dudley Castle. Such parts of the eleventh century curtain at Durham as remain, including that portion running from the chapel up to the top of the mound, are strongly built of stone; and there is every reason to believe that the original outer wall of the keep, parts of which are incorporated in the existing structure, was of stone also.

In Prior Laurence's rhapsody on Durham Castle, written about 1150, he uses terms in his description of the keep (arx) which might well apply to many shell-keeps with outer walls of stone and internal divisions of timber . He says that the level of the ground inside the keep was three cubits higher than that outside, and that the keep was embellished by noble projections gradually dying into the grim wall—a poetic description of buttresses which, as they rise, fall back by offsets into the general face of the wall. Owing to a mistranslation, intus enim cubitis tribus altius area surgithas been held to refer to the three terraces on the mound. But, apart from the fact that the three terraces on the mound were laid out by Bishop Cosin in 1670, what Laurence does say is that the ground level inside the keep was three cubits higher than that outside.

The work of rebuilding in 1838-40 was extensive; but on a close examination of the outer wall of the keep made recently I found that, while the upper part of the wall had been rebuilt entirely, the lower part, though extensively repaired, had not been rebuilt. Here and there are large patches of original masonry, having stones deeply pitted with weatherworn holes and some stones set upright against the quarry bed, all similar to a wall in Barnard Castle in the same county, by common consent Norman work. On some sides of the keep are vertical joints where new masonry, differently coursed, abuts against the old work. The original outer wall of the keep was undoubtedly of stone, as one would expect. For it is most improbable that, while during the twelfth and thirteenth centuries ponderous and elegant buildings were being raised on the north and west sides of the bailey, the opulent bishops of Durham would have been satisfied with a timber keep, their last line of defence, until the middle of the fourteenth century and then build a structure three centuries out of date.

The perishable character of timber and its liability to destruction by fire and decay render it an unsuitable material for permanent fortification. From the earliest times fire has been one of the principal weapons of offence; and although even stone buildings are not proof against its attacks the damage done to them is partial and often reparable. With timber the destruction is total. The Normans must have been well aware of this fact, and the paucity of remains of castles of the early Norman period may be due in some measure to the pillage of stone for building purposes. At Topcliffe, Yorks., every stone has vanished from the site of what was the principal castle of the Percies for six centuries; and of the stone keep on the mound of Richard's Castle, Hereford, all but small fragments have disappeared. In some cases masonry may still lie buried beneath the soil. A mound on a hill at Lydney, GIos., hitherto considered an earthwork, was found, on investigation being made in 1929, to cover the foundations of a stone-built castle, dating from the twelfth century ; and the existence of hidden masonry on other sites is not improbable.

Timber, however, was of great value for temporary camps and forts which had to be constructed in haste. The Norman castles in England were founded to overawe and govern the districts in which tney were built, and a large number of them must have been con-