One of the best known of the natural megaliths to which traditions cling is the Agglestone, or Hagglestone, situated on the moors near Studland, in Dorsetshire (Fig. 8). This Agglestone is a huge inverted cone of indurated rock in direct connection with the Lower Bagshot Sands on which it rests; in other words, its shape and position cannot be artificial. It is a mass of sandy material, so thoroughly cemented by oxide of iron that it has resisted denudation with some degree of success. Yet the so-called sub-aërial agencies, principally wind and rain, have undercut its base, rounded its outlines, and scooped out the “rock-basins,” which the eighteenth century antiquaries ascribed to the labours of Druids[110]. It is noteworthy that the Agglestone belongs to a part of the country the inhabitants of which were pictured by Bede as confirmed pagans (paganissimi)[111]. From a review of the legends, as well as from a consideration of the name, Agglestone (most probably from A. S. halig = holy), and its alternative designation, Devil’s Nightcap, there is fair reason to believe that the stone had some significance to the heathen folk of Wessex, and that it was very probably a Christian preaching station.

The Agglestone doubtless proved too unwieldy and obdurate for the tools of those who set up the first Christian crosses, but this has not been the case with many other pillars, whether hewn or unhewn. Some of the upright “crosses” of Devon and Cornwall, for instance, are of extremely coarse workmanship, as the student may see for himself by inspecting the illustrations given in the works of Messrs A. G. Langdon and W. Crossing[112]. Nor need the simplicity of the early workmanship cause surprise, for the oldest Cornish crosses date from the seventh century. A like plainness is met in many other parts of England. At Fulstow, Lincolnshire, I noticed a crude churchyard pillar of hard, grey chalk, roughly squared, now mounted on a much more recent plinth. The stone is much pitted by weathering, and is clad with lichens of varying hues. If the monolith be not a pre-Christian relic, trimmed into a rectangular form, it is most probably a very early pillar, co-eval with the first Early English church. It may have been dug out of the boulder clay, like many of the stones with which the churchyard paths are paved; or, if we accept modern theories respecting the glacial drift on the East of the Wolds[113], it is not an ice-borne relic, but must have been brought to the alluvial plain by man. The original home of the pillar was in the hill-slope, several miles to the West. This Fulstow “cross” is typical of others scattered throughout the East of England. Reverting to Cornwall, it must be observed that the numerous inscribed monoliths of that county are believed, on a balance of probabilities, to be of a Christian character[114]. Specimens are frequently found in remote spots, or they may occur in proximity to the church itself. At Camborne, an example is seen under the communion-table; at East Cardinham, in the graveyard; at St Cubert, in the wall of the church[115].

The early pillar “crosses,” though accounted Christian when tested by inscription and decoration, may yet have an earlier origin. It is now a commonplace that many of the crosses and calvaries of Brittany, “with shapeless sculpture decked,” are merely primitive menhirs adapted by the Christian artificer[116], and anyone who, like the writer, has had the opportunity of comparing the Breton series with the kindred group of our English Brittany, will readily agree that a similar story may be told of Cornwall. Something has been written on this topic elsewhere[117], and one need now only call attention to a curious instance of reversion in connection with the allied subject of tombstones, to show how deep-seated and perennial is the habit of imitation. In the “Quaker’s Cemetery,” two miles from Penzance, the only tomb remaining within the enclosure is formed of a massive slab of granite (5´.7´´ × 2´.1´´ × 1´.1´´), resting on large pieces of the same kind of rock. The tomb is evidently a copy of the dolmens of the moorland, yet its date is so recent as A.D. 1677[118]. This illustration of the “past in the present” supplies a warning note, and is not so irrelevant as it may appear for the moment.

We may follow our work by inspecting some interesting cases of the occurrence of unshaped masses of stone in, or near, the fabric of the church. We must start, however, with the clear axiom that natural blocks of stone, where readily procurable, must, like the spoil heaps of Roman buildings, at all times have invited the attention of masons. Not more than fifty or sixty years ago, Sir A. C. Ramsay noted that the “greywethers,” or sarsens, of the Marlborough Downs, were so thickly strewn over the surface, that across miles and miles of country a person might almost leap from stone to stone, without touching the ground. Yet, in our own day, the preservation of the greywethers has become a serious task, because they have been found so useful for paving-stones during the interval that has elapsed since Ramsay wrote, and it has been difficult to stop depredations on those that remain. Not forgetting our warning, there is still a possibility that, should the examples of churchyard sarsens prove numerous, and should there be a co-operation of other factors which indicate early sites of pagan worship, these two series of circumstances may be in relationship. A solitary example might be declared accidental; two or three citations only might raise an incredulous smile; hence, it is the cumulative force of recurring details which can alone afford pretence for a theory.

Situated in a long, dry Kentish valley which runs upwards in a Southerly direction towards the escarpment of the Chalk, and at a distance of about 1½ miles from the railway station at Eynesford, one may see the forlorn wreckage of Maplescombe church (Figs. 9, 10). This church, which had a semicircular apse, still partially remaining, has been in ruins for three centuries. My attention was first called to the spot by Mr Benjamin Harrison, of Ightham, an archaeologist whose knowledge of his native district is unsurpassed. On visiting the ruins in 1904, I found a large, partially-sunken sarsen stone (3´.0´´ × 2´.0´´ × 1´.6´´) occupying what appeared to be the site of the ancient altar. A few smaller sarsens were also discernible, and other specimens, Mr Harrison states, have been carried off, at various times, by hop-pickers, to build hearths in the fields. In the field adjoining the church, the ploughshare has turned up

human bones and other relics[119]. This area was presumably the graveyard, and may have been originally unenclosed, but with this hypothesis we shall deal in a later chapter. Parenthetically, it may be explained to the non-geological reader that a sarsen is a hard mass of rock, which was once part of the Bagshot Sands or the Woolwich and Reading Beds, and which, having resisted denudation, remains on, or near, the present surface of the soil. The earliest record of a church at Maplescombe is A.D. 1291, but the building may, perhaps, be of Norman foundation, and the largest stone may possibly be a sacred relic which existed previously on the present site. The worship of “stocks and stones” died hard, and it is at least conceivable that the church builders adapted one or more megaliths to form an altar. Further than that we cannot go, seeing that sarsens are fairly common in the locality. Examples of churchyard sarsens are abundant in Kent. Mr Harrison informs me that there are specimens at Kemsing Halling and Trottescliffe; in the last-named village, the stone is built into the church wall. At Meopham, there are several blocks just outside the churchyard, but, as the ground is merely fenced in, we have again, doubtless, an instance where the demarcation between consecrated and unconsecrated soil is of modern date. Still further records from Kent have been supplied by Mr F. J. Bennett. The ruins of the churches at Punish and Paddlesworth (near Snodland) enclosed in each case a large sarsen; the nave of the dismantled church at Dode contains a good-sized specimen; several other blocks stand just outside the graveyard wall at Birling[120]. In passing, it may be observed that the other Kentish village named Paddlesworth, near Lyminge, contains a font, of which the base is a massive round stone, evidently of great antiquity.

We now examine other counties where the Tertiary beds are represented. Crossing the Thames, we find in the churchyard of Ingatestone, Essex, a large sarsen, which was formerly a part of the foundation of the church[121]. At Pirton, in Hertfordshire, a huge mass of conglomerate, or “puddingstone,” consisting of rounded flint pebbles cemented by a siliceous matrix, supports the North-Western buttress of the church. The block, as determined by my friend, Mr James Francis, F.G.S., measures 5´.6´´ × 2´.7´´ × 1´.4´´ above the ground. At the base of two other buttresses on the North side are further lumps of conglomerate, each about 3 feet in length. These “puddingstones” are vulgarly believed both to breed and to increase in size, and the superstition is put forward to account for a block of this material which projects from the foundation of Caddington church, Bedfordshire[122]. It is worthy of notice, in passing, that a pre-conquest church existed at Caddington.

Our observations would be incomplete were they limited to the Tertiary area of South-Eastern England. In Devonshire, built into the chancel wall of North Molton church, we have a large, heavy stone, which is said to be composed of material foreign to the district[123]. At Branscombe, in the same county, where the church bears marks of considerable antiquity, a rough pillar, about seven feet long, doubtfully described as a coffin lid, lies in the churchyard. Just outside the churchyard wall of Whatley, Somerset, is a huge rounded sarsen, and another is to be seen near the cross-roads 50 yards distant. When the London Geologists’ Association visited Whatley in 1909, a doubt was raised whether the stones were true sarsens. Some authorities pronounced the material to be millstone grit, which could be obtained a few miles away; while, on the contrary, no Tertiary rocks occur in the immediate district. In Cornwall, there was discovered, under the collapsed Western tower of Constantine church, a large, rounded boulder of Cataclew stone, weighing a quarter of a ton. The nearest locality from which this stone can be obtained, says the Rev. R. Ashington Bullen, is a quarry which is 1¼ miles distant in a straight line. Mr Bullen believes that the boulder marked a meeting-place for ceremonial observances in pagan times, and that consequently it was assigned a place of honour in the Christian building[124]. It will be recalled that the ruins are adjacent to a kitchen-midden (p. 31 supra). At Bolsterstone, near Deepcar, in Yorkshire, two large stones lie in the village churchyard. One of them has been adapted for receiving another stone by mortising. On the high ground above the church is a cairn known as Walderslow, and it is believed that the churchyard stones may have had connection with this monument. The diligent searcher will not fail to discover many other examples of these natural megaliths, but he will doubtless preserve considerable detachment of mind, and be wary in the acceptance of theories. The scarcity of suitable rocks in many localities, the difficulties of transport,—whether accomplished by ox-drawn sledges or by canal barges,—the saving of time, and, far more important, the lessening of expenditure, are factors which must receive full weight. Nevertheless, while maintaining due reticence, we shall find ourselves continually wondering whether the probabilities do not point to site-continuity. The pronounced liking for megalithic monuments exhibited by the primitive Britons must have strongly influenced all future comers for many a century. All analogy suggests that Mediaeval folk were still sufficiently pagan to treat such relics with a kind of “hyperdulia.” A sacred stone, or group of stones, may well have been embedded in the walls of the church, or set up as an altar, in order to propitiate those who gave up the old faith with reluctance.

When we examine megaliths which were indubitably placed in position by the labours of men, we find ourselves on surer ground. The building of churches near such memorials as these cannot always have been at haphazard. Moreover, we should bear in mind that all the evidence is not now producible. The hand of the spoiler has been busy, and the results have been lamentable. Utility has been the common plea for the removal of many ancient monuments, but other motives have also been at work. The famous “Longstone” which formerly stood a little to the East of St Mabyn church, in Cornwall, was broken up and carried away in order “to brave ridiculous legends and superstitions[125].” Happily, the well-known menhir in Rudstone churchyard, near Bridlington (Fig. 11), remains with us. This pillar, which is composed of fine-grained grit, stands about 4 yards from the North-East angle of the building. Its height is 25 feet, and it is believed by some authorities that an equal length is concealed underground. The monolith was first fully described by the Rev. Peter Royston, in 1873. The present Vicar of Rudstone, the Rev. C. S. Booty, informs me that Mr Royston’s measurements are accurate. The conjecture has been made that the village took its name from the menhir. This may well have been the case, but what the first syllable of the name means is another matter. The word is commonly said to signify Rood-stone. The Domesday form Rodestan (cf. 13th cent. Rudestone; 14th cent. Ruddestan, Rudston, etc.), leads Mr Bonner to suppose that a personal name, Rod, Rodd, or Roda, is indicated. If the monolith bore an incised or carved cross, Mr Bonner would admit the rendering “Rood-stone.” But it should be remembered that a simple pillar might have been called a cross, and that it may have been accepted as a preaching cross. To consecrate an existing stone would save much labour. On this view, “Rood-stone” may actually be correct. Countryfolk do not care for etymology or archaeology, but they have not been remiss in attributing the presence of the stone to diabolic agency. What concerns us at the present is, that the site of the church was probably selected because the spot had already some significance to the older inhabitants of the neighbourhood. The whole district of Rudstone is rich in prehistoric remains[126].

The “sacred chair” of Bede, at Jarrow, is considered by Professor Rupert Jones to be an ancient sacred stone, which has been chiselled into shape by modern masons[127]. The Coronation Stone, in Westminster Abbey, has also perhaps a notable genealogy, but its deposition in its present quarters took place long after the foundation of the Abbey, and hence the relic is not illustrative of our theory.

On the Greensand hill a little above Mottestone church, in the Isle of Wight, there is a huge, untooled monolith, known as the “Longstone,” but it is not certain that it was originally solitary. A smaller pillar lies at its base, and Mr W. Dale, the Hampshire archaeologist, supposes that the two stones represent a fallen dolmen, or the remnants of a cromlech[128]. Other writers have considered the relics to be ancient boundary stones[129], but I think this explanation not very satisfactory. The Rev. G. E. Jeans, who advocates the boundary-theory, declares against the view that Mottestone signifies “mote-stone,” and points out that the Domesday spelling, Modrestone, indicates a personal name, Modr[130]. Even allowing for possible approximations made by the Domesday scribe, the etymology given by Mr Jeans seems more reasonable than the older one. Another Hampshire village, Twyford, on the Itchen, is worthy of a visit in connection with megaliths. The church in this old-world nook was believed by Dean Kitchin to be built on ground once occupied by a stone-circle or a dolmen, and Mr Dale considers that the two large sarsens which lie by the side of the building represent the wreckage of this ancient monument[131]. A particularly fine yew in the graveyard will be noticed in a subsequent chapter. On the neighbouring hillside of Shawford Downs, there are also some linchets, or ancient cultivation terraces. These associations imply that Twyford was not only an inhabited site, but presumably a sacred site, at a very early period. Still another Hampshire example is furnished by Bishopstoke, the church which Mr Hilaire Belloc asserts was erected on the site of an old stone-circle[132]. Cobham church, in Kent, stands a little to the North of the remains of a stone-ring. Outside the North porch there is a large sarsen, another lies against the wall at the West end, while a third is built into the South wall[133]. Thomas Wright long ago pointed out that the church of Addington, in Kent, was in the immediate neighbourhood of numerous megalithic remains, though all of these were in a ruinous and disordered condition. In fact the area seemed to be a vast tribal cemetery. Professor W. M. Flinders Petrie, in 1878, was able, from a study of the monumental relics, to make an imaginary restoration of parallel avenues of stones as they once existed. At the North-Eastern extremity, there was a stone chamber which has unfortunately since been disturbed[134]. Some writers have believed that the hillock on which Addington church is built was artificial, but it is practically certain that it is purely natural; its existence being perhaps due to a protective capping of ironstone which has been proof against denudation.

The church of Stanton Drew, near Bristol, is placed within the precincts of a veritable Valhalla of monumental relics. Three stone circles are situated, as it were, within a stone’s throw of the building, the most distant being about one-third of a mile away, and the nearest only 150 yards. But besides these more perfect remains, there is a group close to the churchyard, towards the South-West. This group, called the Cove, consists of two upright blocks, 10¼ feet and 4½ feet respectively in height, and one prostrate stone, 14½ feet in length (Fig. 12). The original character of the monument cannot be decisively known. Mr C. W. Dymond contends that the stones hardly represent a ruined dolmen, because of the unusual height of two of the remaining pillars. Other speculations, hazarded, as it seems to the writer, without a vestige of proof, regard the Cove as a “druidical chair of state,” and, again, as a shelter for sacrificial fire. On the whole, it is safer to consider these monoliths as survivors of a cromlech or stone-ring. The material, which is unhewn, is a siliceous breccia of Triassic age, and was probably brought from Harptree-under-Mendip, about seven miles from the present position. The church, it should be added, retains portions of Norman work[135].

The vanished menhir of St Mabyn has been noticed (p. 42 supra), but, before leaving the English megaliths, we ought to glance at the smallest cromlech in Cornwall, that of Duloe, which is situated near Duloe church. Its longer diameter is 39 feet, and its shorter, 37 feet, so that the cromlech is slightly elliptical. The “circle” contains seven standing stones, and one fallen or broken stone. One of the pillars, which are very unshapely, is 9 feet in height. The finding of charcoal, together with a cinerary urn enclosing bones, near one of the pillars, is sufficient to show the sepulchral character of the circle[136]. Cornwall should indeed prove the touchstone of our theory, and I believe that both Cornwall and Devon would stand the test well, could we recall the witnesses. But these, sad to relate, are for the most part gone. Here a gatepost, there a tombstone, and yonder the hearth of a cottage, warn us not to expect the impossible. Sir Norman Lockyer, in his work on Stonehenge, asserts that many churches have been built on the sites of circles and menhirs, but he proffers no actual examples[137]. He gives, however, numerous instances from Cornwall, Ireland, and Scotland, of the juxtaposition of megaliths and sacred wells. Now, it will be shown in the next chapter, that churches were frequently built in proximity to holy wells, so that we have a triple relationship. Sir Norman Lockyer’s informant doubtless knew of other examples of church-megalith sites than those which have been adduced[138]. Such sites are said to be not uncommon in Wales. The church at Yspytty Kenwyn (or Cynfyn), near the Devil’s Bridge, in Cardiganshire, had the circle of stones built, at intervals, into the churchyard wall. There were also stone pillars at the Eastern entrance to the church, just as they are sometimes found near stone-circles. Large megaliths are also recorded from the churches of Tregaron, in Cardiganshire, and Llanwrthwl, in Brecon[139]. Cordiner, an eighteenth century writer, asserts that Benachie church, Aberdeenshire, is built within a stone-circle, and that the practice of thus building was not infrequent in that country. And Mr W. G. Wood-Martin has recorded at least two cromlechs in Irish churchyards[140].

There is also a scrap of linguistic testimony which is pregnant of ancient tradition, and which has been noted by several writers. Sir Daniel Wilson seems to have been the first to make the fact publicly known. The common Gaelic sentence, Am bheil thu dol d’on chlachan? (Are you going to the stones?) may be rendered alternatively, “Are you going to the church?” and is used in this second sense by the Scottish Highlander when addressing his neighbour. Primarily, chlachan (clachan) means a circle of stones, hence, a battle, or the scene of single combats. The interpretation “place of worship,” is, as might be anticipated, derivative, though not recent. So far back as 1774, Shaw, in the chapter which he contributed to the third edition of Pennant’s Tour in Scotland, observed, “From these circles and cairns many churches to this day are called clachan, i.e. a collection of stones[141].”

A word of caution is necessary to those who may be inclined to accept too hastily, and without examination, the claims of this or that megalith to a great antiquity. For instance, there stands at the South-Eastern gate of Binstead church, in the Isle of Wight, a grotesque figure, called by the villagers “The Idol.” This uncouth image has been thought by some to be a pagan object of worship. Little, indeed, is definitely known about the object, but it is asserted, with much credibility, that the gate once formed the door of the church, and that the image is merely a Norman keystone, or perhaps a corbel[142]. We note, however, that if it were a corbel, it could scarcely have been a portion of a doorway, though this matter is inessential. Our second illustration shall be given in order to show the danger of dating objects as pre-Christian, when they bear clear signs of Christian influence. In the churchyard at Penrith there is a large tomb which bears the nickname of “Giant’s Grave.” It happens that this name is often applied to prehistoric barrows and megaliths, and in this particular instance it has been proclaimed that the tomb is a cromlech—a “dolmen” being perhaps intended. Hutchinson, Pennant, and other writers, were greatly exercised concerning this ancient relic. But if the reader will turn to the beautiful engraving of the monument in the Victoria History of Cumberland, he will understand, even without the aid of the letterpress, that the tomb has features decidedly Christian. The monument really consists of the shafts of two pre-Conquest crosses, one being placed at the head and the other at the foot, while the space between is enclosed by three “hog-backs,” one of which has been split longitudinally[143]. Once again, in the churchyard of Chadwell St Mary, Essex, a large sarsen, concerning which fantastic theories were current, was observed by the Rev. J. W. Hayes to have a weathered concavity, or “pebble-hole,” within which were carved the letters “N. G.,” followed by the date 1691. Referring to the parish register, Mr Hayes found an entry, made during that year, recording the death of a churchwarden, Nathaniel Glascock. The inference was clear, and the lesson of caution was delivered with some force. These reservations about the nature of burial monuments lead us easily to the subject of grave-mounds, to which we must allot a special chapter.

CHAPTER II

CHURCHES ON PAGAN SITES (continued)

A few words must be devoted to an explanation of mottes or mound-castles. These hillocks were essentially low, flat-topped, truncated cones of earthwork, usually surrounded by a ditch, and placed in direct connection with a larger defensive enclosure. The mound was generally artificial, either wholly or in part: the entirely natural mound is the rarest kind[144]. Of these natural hillocks, an illustration is found in the chalk “monticle” on which Corfe Castle is built (Figs. 13, 14). This mound need not detain us, because it is still crowned by the ruins of what was once a solid structure of masonry, built during the reign of Henry I. Of its true character there can, therefore, be no doubt. The castle-mounds which we are particularly considering, in their earlier forms at least, are believed to have supported a kind of wooden guard-house (turris, bretasche, or keep), which was surrounded by a stockade. Not until a later period of fortification, when the material of the mound had subsided and become firm and solid, did a structure of stone appear on the summit, if indeed, the wooden structure were ever replaced by a more permanent keep or fortress. Stone keeps were built on mottes at Kilpeck in Herefordshire, Fewston in Yorkshire, and other places, but this does not appear to have been the more general custom. Many mounds, at any rate, were never capped by a superstructure of masonry.

The castle-mound, as already stated, was encompassed by a moat, which probably, however, was not intended to contain water, except in special cases (Fig. 15). Yet it is very possible that “puddling” was often an undesigned result of the constant trampling to which the ditch was subjected. It should here be explained that the Norman-French term, motte, which is constantly applied to the moated mound, is not related to the word “moat,” though, owing to a misunderstanding of the Latinized form, mota, it has often been so translated. Beyond the real moat, or ditch, was the larger enclosure to which reference has been made. This was the outer ward, the bailey or base court; it was of horseshoe or crescentic form, and was reached by crossing a wooden bridge. The bailey had its own moat, which, in its turn, was engirdled on the outside by a bank passing along the counterscarp[145].

This short description must suffice. The question which first arises is concerned with the age of the moated mounds. The older opinion, as expressed by Mr G. T. Clark, and to some extent accepted by later authorities, such as Mr I. Chalkley Gould, was, that some of the hillocks, at least, were of Saxon date[146]. Mr Clark was largely influenced by the belief, which most modern writers consider erroneous, that the word burh of old documents referred to these castle-mounds. This word burh, however, is said to stand always for a fortified town and to have never been applied to a motte-and-bailey castle[147]. Among quite recent writers who assign some of the mounds to an early date, may be mentioned Mr Willoughby Gardner, who considers that, on a balance of evidence, the simple form of moated mound may be said to have originated in Saxon times. This view is also shared by Mr Reginald A. Smith. Again, Mr T. Davies Pryce has brought forward evidence to show that the moated mound belongs to diverse races and periods, and he contends that some mottes are of much earlier date than the Norman Conquest[148]. The trend of modern opinion, as enunciated by Dr J. H. Round, Mr W. St John Hope, Mrs E. S. Armitage, Mr G. Neilson, Mr A. H. Allcroft, and others, places the castle-mounds within the Norman period[149].

So far as the moated mounds are artificial and of Norman construction, they are extraneous to our inquiry about pagan sites; they are the feudal strongholds of which the village church was often the religious appendage. This relationship of fortress and temple will be forced upon us in the next chapter, and will continue to suggest itself when we discuss other matters. But if we suppose that the Norman mottes had their Saxon forerunners, or even that the Norman mound-builders took advantage of pre-existing knolls of an artificial character, we are led to search for vestiges of an accompanying Saxon church. For, under these conditions, it is conceivable that we might have a Christian church built near a pagan mound. From the nature of the problem, satisfactory proof is difficult to procure. Certain moated mounds have yielded more than a hint of the adaptation by the Normans of earlier works. The flat-topped castle-mound near the churchyard of St Weonards, Herefordshire[150], has been claimed, on “the testimony of the spade,” as having been a prehistoric grave-hill. This was the view held by Mr I. C. Gould. Thomas Wright, who opened this mound in A.D. 1855, declared that, “beyond a doubt,” it had been used for sepulchral purposes, though the discoveries did not warrant his assigning its specific period. It may be mentioned that a decayed yew, of considerable age, together with other trees, adorned the hillock[151]. A similar defensive hillock, 50 feet in diameter, near the churchyard of Thruxton, Herefordshire, and known to the peasantry as Thruxton Tump[152], was also found to contain animal bones and pieces of crockery[153]. I can gather no details concerning the excavations of this last-named mound, and am inclined to accept the claims with great reserve, principally because other mottes have furnished similar relics, which have been proved capable of a more obvious interpretation. The first example of these supposed barrows is the castle-mound which is included within the present extended graveyard at Penwortham, in Lancashire. Careful sections cut in this remarkable hillock exhibited a profusion of remains, such as animal bones, mussel-shells, decayed timber, and objects of iron and bronze. These relics were disposed in layers, in such a manner as to show that the mound had been raised in height at two different periods[154]. Successive elevations of surface were also discovered in the moated mound adjacent to Arkholme church, Lancashire[155]. The castle-mound, again, at Warrington, situated about 100 yards from a church which stands almost within the fosse of the outer ward, has been raised more than once. The last occasion when the height was increased was during its occupation by the Parliamentary forces in A.D. 1643[156]. In all these cases the relics seem to indicate alterations which took place after the Norman period of mound-construction had set in. The bronze articles found at Penwortham, and the broken amphora which is recorded from Warrington, superficially suggest an earlier origin. But these relics were most probably scraped up with the soil when the motte was enlarged, or were picked up by the inhabitants somewhere in the neighbourhood, and were afterwards blended with the refuse-stratum of that particular period. These explorations, then, tend to discredit, in some degree, the statements made with respect to the Herefordshire mounds. At the same time, we must not rashly conclude that, in every instance, the workmen commenced their work on a perfectly level surface. The story of St Weonards teaches us caution. There were hundreds of early burial-mounds, as well as hillocks of other kinds, which may well have served as bases for mottes. An incidental fact, noted by Dr Round, is worth recalling. Moated mounds are to be seen in places where, so far as we know, the Normans never had a castle. It is clear that castle-mounds, with their appendant bailey-courts, were sometimes thrown up, and afterwards abandoned for other sites. Such a mound was raised by William at Hastings[157]. This opinion is quite accordant with what has been previously said about the absence of stone keeps on earlier mottes.

Seeing that the feudal baron dominated the village community, and that compliance with the claims of religion was deemed secondary only to the arrangements for personal security[158], one would naturally expect to find the Norman church not far distant from the castle-mound. And this is actually what one often sees: the church is either just outside the moated mound, or within the crescentic bailey-court. It would, I think, be an over-statement to assert, as do some writers, that the inclusion of the church within the entrenchments is typical of the arrangement of a Norman earthwork