The Castles of Kent No. 5

The Castles of Kent No. 5: Sandwich Castle

I am the first to concede that I know very little about individual castles in East Kent or indeed outside the Medway area. Certain castle sites however, I find intriguing and although no doubt I will be taken to task by local archaeologists and historians of the Sandwich area I like to believe that the following ideas (rather than 0 'facts') may move our knowledge forward.

Figure 1. Location plan of published Trenches on the site of Sandwich Castle

Although mentioned in Kent Castles (Guy 1980, p205-6) it was a brief 1983 article that really brought Sandwich Castle to my attention (Tatton Brown et.al. 1983, p.243-7). It was suggested that the Castle was a thirteenth century motte and bailey (mound and courtyard) fortification. At this time I was writing my dissertation (Ward 1985) and this was only the second reference to a thirteenth century rnotte and bailey castle for southern England that I had come across. The other reference being for Castle Toll near Newenden, overlooking Romney Marsh, to which I shall return below. For southern England such structures are usually regarded as being of late eleventh to midso .__ twelfth century date. Only in Scotland and ___ ...._ __ __. Ireland were motte and bailey castles still commonly constructed in the thirteenth century. The latest securely dated examples being mid-thirteenth century for Scotland and c.1220 for Ireland, although in each instance others may be later, perhaps considerably so (Higham and Barker 1992, p67; McNeill 1997, p.71-3).

Sandwich Castle was apparently situated on the east side of the town on what is known as 'Castle Mead', near to the late medieval Sandown Gate. However, even its position was apparently at one time (and for all I know may still be) contested for John Guy records that there were those who believed it had been constructed in the south-west comer of the town within the area later enclosed by the town defences. Since 1983 the Canterbury Archaeological Trust have undertaken several evaluation trenches and watching briefs on the Castle Mead which is marked as being the castle site on earlier Ordnance J Survey maps. Other than the short 1983 article only one of these trial trend1 projects, undertaken in 1995, has to date been published (Herdman 1996, p.36).

No detailed drawings have been published of either the 1983 or 1995 trenches, there is only a location plan available for consultation which has been traced with permission of the Canterbury Archaeological Trust and amended (Fig. 1). Based on the textural description my own schematic section drawing of the 1983 excavation also appears (Fig. 2.); the numbering on the latter and referred to in the text below is my own. In 1970 a Mr. A. Southam excavated three trenches of which apparently little, other than their position, is known (op.cit. 1983, p245). In the 1983 excavation two trenches were excavated. The longer one picked up evidence of a curving ditch at both its north and south end. This ditch was described as being approximately 14.25m wide and 4.5m deep. The shorter trench was within the ditched enclosure. The 1995 trench found evidence of other ditches c. 100m to the east, the largest of which was 3.2m in width.

Figure 2. Schematic section of deposits recorded In 1983.

The schematic section drawing shows the results of the 1983 excavation as described in the text (op.cit. 1983, p.245-6). The reader should take note that this sketch will have no similarity to what was actually drawn it merely depicts the stratigraphic order of the archaealogical layers and features described. Archaeological section drawings across deep recut ditches are seldom simple. The work of Philip Barker at the motte and bailey castle of Hen Domen in Montgomeryshire has shown how complex ditch infilling and recutting can be. At Hen Domen of five achaeological trenches cut across a 20m length of ditch not one produced the same results as another (Barker 1979 p.42, Figs. 8-11) and most of the soil deposits probably produced no dating evidence.

The sequence described within the text of the 1983 article is relatively easy to understand. Two topsoil deposits were found. The lower, undisturbed deposit (6) contained eleventh century sherds and the upper disturbed deposit (4) was regarded as being the product of agricultural activity. As no pottery datable later than c. 1200 was recovered from this upper topsoil and as the documentary evidence does not mention a castle before 1290 (op. cit. 1983, p.243) it was suggested the castle was not constructed until the late thirteenth century. The 1995 2 excavation trench produced pottery of c. 1150-1200 within the upper fill of another ditch. However, as this latter feature was excavated to a depth of no more than 30on and as it is almost 100m to the east it may have no relevance to the ditch found in 1983. Indeed it is possible that it does not form part of the castle defences but may merely be a large drainage ditch of which, I suspect, there are many around Sandwich.

Whilst the stratigraphic evidence for a foundation date in the late thirteenth century seems outwardly secure, there is a major problem in that there is no published section drawing. The ditch apparently cut the disturbed topsoil (4), but depending how far back towards the enclosed area it began this cutting may have been the product of the observed recut or, as at Hen Domen, there may be several recuts only identifiable if enough trenches are excavated. Alternatively (and this is an archaeologist's nightmare) a recut might destroy all sign of the original ditch and the latter would not therefore appear in the archaeological record. It is stated within the text that due to waterlogging the base of the earliest ditch was not reached. Consequently the thirteenth century pottery that� was found within the ditch infill might not represent silting belonging to the original ditch; it may be considerably later.

The pottery recovered from the disturbed topsoil is described as 'of later (my emphasis) twelfth-century' date rather than late (op. cit. 1983 p.246). There is an important chronological difference between the two words; late would imply the last quarter of the twelfth century (at the most), whereas I would (and have) used the word later to describe the last half of a particular century. The meaning of both words can be subjective. Also no details of the pottery date range was given. Whilst the pottery may be 'no latter than c1200' that does not necessarily represent the time of occupation which, as stated (op.cit. 1983, p246), is perhaps indicated by the post holes, hearth and stake holes (5). If we were to obtain a pottery date of (say) 1170 plus or minus 30 years we would be doing (exceptionally) well. This means that the occupation an� the suggested ploughing could be as late as 1200, buf_ 1/ alternatively as early as 1140. This discrepancy makes all the difference between making it more acceptable for the fortification to be a motte and bailey style castle or not.

Furthermore the disturbed soil may not be the result of ploughing. Apparently Mr. Southam in his 1970 trench believed he had found a bank on the north (op.cit. 1983 p245). If this interpretation can be trusted (which is perhaps doubtful) then the bank may have formed a ringwork enclosing an open courtyard area. Unless this area was cobbled it would be prone to disturbance and deposition of rubbish by humans and perhaps animals (pigs can create a wonderful 'disturbed soil' deposit). If the earliest end of the pottery date range is as early as 1140 then the disturbed topsoil could have been created in a year or two. The silty loam (3) possibly forming a 'thin compact ... turf horizon', whilst it may have been in existence for a century before the construction of the castle may equally have been formed in a relatively short time. All of these deposits were covered by the (suggested) mound (2) and the whole process . could easily occur within a decade. The reign of King Stephen (1135-1154) was of course regarded as 'The Anarchy' and saw the construction of many earthwork and timber castles, large numbers of which, but by no means all, were to be abandoned or demolished in the reign of Henry II (1154-1189). Whilst it is not provable that the motte and bailey castle was constructed in the midtwelfth century rather than a century or more later this type of structure would certainly be more in keeping with the earlier period.

According to E. W. Parkin (1984, p.199) the castle was in existence by 1260 and the insertion of this date on his Figure 4 implies that he regarded this as being not too far removed from the time of actual construction. Apparently brattices (a temporary breastwork or parapet) were erected in 1266 (op.cit. 1983, p.243) and the first reaction is to conclude they were at the castle. However, they may equally have formed palisades across the town streets at the time of the Barons War against Henry III when Richard de Leyboume captured the town on behalf of the r"royalist forces (Powicke 1962, p.207). According to Tim Tatton Brown the date of construction of the castle may have been as late as 1290, or alternatively this may have been the period of rebuilding of an earlier structure (op.cit. 1983, p243). Documentary records for many royal buildings of the reigns of both Henry III (1216-1272) and Edward I (1272-1307) are relatively good. Whilst it is possible that a royal castle could have been built in this period without any reference to its foundation it is far more likely that such a lack of evidence would belong to the twelfth century (earlier rather than later). Alternatively an unrecorded temporary structure may have been constructed by Richard de Leybourne to dominate the rebellious town. Reluctantly I concede the stratigraphic evidence points in the latter direction and indeed this date is reinforced by the construction of Castle Toll in the same century.

However, both Henry III and Edward I were lovers of l"'display and I find it hard to accept that either monarch - would revert back to an old fashioned and out-of-date form of defensive structure as a permanent work especially in an urban centre such as Sandwich. If such a structure had to be hurriedly constructed in a time of emergency then (in my view) either monarch would bring it up to date as quickly as possible. The castle (probably) found in 1983 and that apparently constructed in the late thirteenth century may, other than their position, have nothing to do with each other. The motte may have been levelled before the probable construction of a stone castle. This latter structure is perhaps best regarded as one of a small group of fortifications being constructed in the course of the thirteenth century in Kent and Sussex. The most famous and best preserved of these is Ypres Tower at Rye which seems to have been in existence by 1249 (Renn 1979, pl93-8), but there was also (probably) a stone castle at Winchelsea dating from c. 1290 (ibid. p.198-200). Castle Toll at Newenden seems to have been the earliest of this group dating from the early thirteenth with a later, mid-century, phase. This fortification was constructed in earth and timber with 'a 3 low motte like mound in the north-east' of the enclosed area (Davidson 1972, pl24). Both of its structural phases were short lived and that neither was of stone need come as no surprise for there was no one to impress in this rural area and it was probably only viewed as a temporary expedient in a time of emergency (1204? 1216-7? 1240's?). It is one of the few, indeed as far as I am aware the only, certain thirteenth century motte in southern Britain. On other sites earlier mottes were altered (Kenyon 1990, pll) and, admittedly, a mound was constructed at Lydford in Devon (ibid. p.43-5), but this encased an earlier free standing stone tower and therefore can hardly be regarded as a true motte. That such structures could be built is obvious but I doubt if Sandwich, one of the most important and wealthiest towns in England at that time, would be the correct setting for a permanent structure of this type at that late date.

The 1983 excavation produced no evidence of the small central keep, by implication a stone tower, mentioned by Parkin in his text (1984, p.199) and shown on his Figure 4. Identifiable structures on both stone and timber are notable by their absence in the published excavation trenches. However, apparently Edward Hasted did report seeing 'foundations' on Castle Mead in the late eighteenth century (Guy 1980, p.205; I have been unable to find the relevant passage within Hasted's work). The position of a stone structure very similar in plan to Ypres Tower is shown on maps of Sandwich and would be in keeping for a mid to late thirteenth century date. However, whether this shape has been based upon Ypres Tower and therefore created by antiquaries rather than representing what actually existed is of course open to debate.

The construction of the town defences creates further problems. Although there appear to have been earlier defences (Gardiner 1954, p.35, 102) it was not until 1385 that a stone wall along the river frontage with a rampart and ditch on the remaining sides of the town were ordered to be constructed. The ramparts on the west were apparently not completed until 1436 when the grand total of six labourers supervised by all the aldermen and constables were to complete the earthen defences without delay (ibid. p. 135). The ditch alignment shown on Figure 1 shows that part of the slope of the town rampart would have been constructed within the castle ditch, in other words the latter would be (or already was) infilled. It is perhaps rather surprising that the castle, whether we are writing about the motte or the suggested stone tower was left outside of this defensive perimeter. To leave a fortification of any strength, whether clay mound or stone tower (or both) overlooking the new rampart and (presumably) an early version of the Sandown Gate would mean the new defensive perimeter was a fiasco even before completed. The solution would be to level the castle. However, a problem with this latter suggestion is that at this date the castle was apparently being repaired and strengthened with turrets, gates and a drawbridge (op.cit. 1983, p.243) both carpenters and masons being present (Gardiner 1954, p. 121).

According to Parkin (1984, p. 201) in 1383 the ships of the Onque Ports captured a French floating wall supposedly 3000ft long and 20ft high and having integral turrets. lf this monstrosity actually existed, of which I am doubtful, it was apparently brought ashore and its timbers used to construct a two-storied gun platform known as the 'Bulwark' erected next to the Sandown Gate. It seems far more plausible that the traditional (albeit not so exciting) interpretation offered by Dorothy Gardiner is the correct one (1954, p.120). According to her this wall was in prefabricated sections being carried on two French ships and was intended to act as a protection for an army once it had landed. When first mentioned the Bulwark, Parkin states that it was erected 'subsequently' i.e. after 1383 (1984, p.201) and later (p.211) that the Bulwark was built six years before the great French raid of 1451. Dorothy Gardiner states that the wood was used in the late fourteenth century defences but she makes no mention of the Bulwark in the context.

The date of 1451 seems to be the accepted date for the construction of the Bulwark (Guy 1980, p.205; Kenyon 1994, p.149) and it appears to have been a purpose built two-storey artillery platform. In the early fifteenth century (c. 1417) a (still standing) stone two-storied gun platform with three storey tower had been constructed at Southampton as the first (known) purpose built artillery fortification in the country (Platt and Coleman-Smith 1975 p.62-7). There has to be the suspicion (but as far as I am aware there is no evidence) that the Sandwich structure may also have been of the same type. However, there is apparently some debate as to whether the Bulwark and the castle are in fact different structures, the latter may have been updated (Guy 1980, p.206). To date I have come across no details as to what this debate entails other than the statement that the Bulwark was constructed in the south-east corner of the town (Gardiner 1954 p.136). This could mean within the perimeter of the town walls or it could mean the area marked on early Ordnance Survey maps as being the site of the stone castle. The latter is certainly the simplest interpretation.

Let us assume for the moment that this artillery platform was a new tower rather than merely an updating of the castle, and was in the position shown by Parkin on his Figure's 4 and 9. This may have been an attempt to create a more up-to-date defence system within the defensive perimeter on the most vulnerable side of the town. However, we enter into a 'catch 22' situation. Such a tower would only be secure if the castle tower (and/ or the mound if it still survived) were themselves strongly defended, but if the castle was updated and cannon inserted the construction of the Bulwark becomes superfluous. lf the castle (whether mound or tower or both) was not strongly defended it would have given cover to an enemy advance and when captured it would have provided a fighting platform from which to enfilade the town rampart and the roof of the Bulwark and Sandown Gate. At the time of the raid of 1457 the dangers from French missile weaponry were well known. It is unlikely that the French army (one can hardly call 4000 men a raiding party) had heavy cannon (although Gardiner does mention artillery; 1954, p.137), with them, but they would undoubtedly have handguns, crossbows and longbows. The latter would have an effective range of 250-300 yards (Hardy 1994, p.180) and, by this date the French were probably as proficient in their use as English armies; crossbows by this time had an even greater range (ibid. p.162), although the range of the handguns would be considerably less. lf the stone castle shown on maps of Sandwich was at the position indicated its capture and/or that of the mound (if the latter still existed) would provide superb positions for the French to sweep the ramparts with their missile fire.Not to have captured the castle first would mean they themselves could be bombarded in flank and rear.

As the castle dominated the town rampart the military solution in 1385 would be to demolish the stone tower, level the mound and infi1J the ditches. It is noticeable that Parkin does not show the castle on his Figure 9 portraying the town at the time of the French raid, which might imply this is what had happened, but unfortunately he does not discuss the point. Not to undertake this destruction would be (and probably was ) military suicide.

The fourteenth century pottery found within the infill of the recut ditch excavated in 1983 might indicate that the mound was indeed levelled at the time of the construction of the town defences. The rampart and ditch around the town may have been ahead of their time for relatively low earthen defences protected by wide ditches were to provide a far better defence against cannon fire than stone walls. The failure to hold this defensive perimeter in 1457 may have been due to lack of manpower or surprise. However, if the castle (updated and perhaps renamed the Bulwark) was still standing its capture would have put the defenders at a distinct disadvantage. Withdrawal into the town would be the only tenable military option.

As far as I can see the published evidence does not allow us to advance much further and the following phasing of the castle site is put forward more as a series of ideas for. discussion rather than statements of fact:

a. c.1100. Area unoccupied.

b. A motte and bailey castle was probably constructed in the mid-twelfth century (before 1150). The pottery evidence within the disturbed topsoil may represent the first phase of this fortification (perhaps a ringwork) and the ditch cutting this deposit is a recut. The lack of evidence for the existence of a castle in the twelfth and much of the thirteenth century is explained by its almost total destruction, as well as the very limited amount of excavation and the comparative rarity of documents for the earlier part of this period. The castle may have gone out of use, which would explain the earlier infi11ing of the Castle Mead ditch. This ditch was perhaps recut in the time of the Barons War when Richard de Leyboume refortified the site as a temporary fortification. Alternatively ( and supported by stratigraphic evidence) a motte and bailey may have been constructed in the last half of the thirteenth century perhaps by Richard de Leyboume as a temporary structure. The mound being levelled and the ditch infilled soon after, or in c. 1385.

c. ln the late thirteenth century a stone tower, perhaps in the same style as that at Rye was probably constructed. Apparently Hasted (quoted by Guy-see above) records the presence of stone foundations but as far as I am aware no excavation has yet taken place on the traditional site of these remains. As both Henry ill and Edward I were known for prestige building projects, it would be surprising if any castle at Sandwich remained an earth and timber structure at this late date, especially as the town was one of the wealthiest in the kingdom.

d. c. 1385 the town defences were constructed and (if not before this date) the mound pushed into its ditch. The stone tower and presumably its outer defences were ,... retained outside the new defensive perimeter.

e. 1451 'The Bulwark' constructed. In all the readily available publications no evidence has been produced for the bulwark having been constructed within the town defensive perimeter adjacent to. the Sandown Gate. The simplest (but not necessarily correct) scenario is to regard this structure as the stone tower of the castle, which has been updated for the deployment of cannon. If this latter idea we� correct then, when captured by the French in 1457 its fall would automatically lead to the sacking of the town.

f. The Bulwark was repaired and another brick defensive structure built near the Fishergate (Gardiner 1954, p.140). In 1471 Thomas Nevill The Bastard of Fauconberg held the town and castle against Edward IV. This seems to be the last reliable reference to the structure.

g. Presumably Henry VIII demolished the Bulwark and/ or the castle. As stone is in short supply in the area the masonry and even eventually the foundations were taken away. It seems likely that all that will remain of the structure(s) will be 'robber' trenches filled with soil and rubble.

As stated in my opening paragraph this scenario will probably be torn to shreds by those with more local knowledge than myself, but I think all would agree there are serious problems in understanding the chronology and typology of this interesting site. I would be surprised (and delighted) if documentary studies advanced our knowledge. It seems probable that only large scale, unhurried archaeological excavations undertaken in summer may solve the problems. Such work would ideally take place not only outside the town defences, but 5 also within their perimeter in the area immediately adjacent to the castle site.

Alan Ward

Part-time lecturer in archaeology at Christ Church University College.

References

1954 Historic Haven: the story of Sandwich. 1980 Kent Castles. 1994 The Longbow, Arms, Armies and Fortifications in the Hundred Years War; ed. A. Curry and M. Hughes; p.161-182. 1996 Manwood Road, Sandwich, Canterbun/s Archaeologtj 1994- 1995, p.36. 1992 Timber Castles. 1990 Medieval Fortifications. 1994 Coastal Artillery Fortification in England in the Late Fourteenth and Early fifteenth Centuries, Arms, Armies and Fortifications in the Hundred Years War; ed. A Curry and M Hughes: p.145-151. 1997 Castles in Ireland. 1984 The Ancient Cinque Port of Sandwich, Archaeologia Cantiana, Vol. c, p.189-216. Platt C and 1975 Excavations in Medieval Southampton 1953-1969, Vol.1. Coleman- Smith R Powike M 1962 Renn D 1979 Tatton Brown T 1983 Bennett P and Blockley P Ward A 1985 The Thirteenth Century 1216- l30l. The castles of Rye and Winchelsea, The Archaeological Journal, Vol. 136, p. 193-202. Sandwich Castle, Archaeologia Cantiana, Vol.xcix, p.243-247. The Three Castles of the Stockbury Valley (unpublished B.A. dissertation)