The Long Demise of the Wantsum Sea Channel: a Recapitulation based on the Data

249 THE LONG DEMISE OF THE WANTSUM SEA CHANNEL: A RECAPITULATION BASED ON THE DATA dave perkins The long slow degradation of the Wantsum Channel, from the wide inundated valley and sea channel of prehistory, to its relic existence in the fourteenth century and today as muddy brook adjoining the Stour, was an enormously complex process. One of which much is still unknown, and one that would require a volume to properly explore. However, the broad framework of the process, both geographical and historical (at least from the Roman period) has been established, and set out in various contributions in the volumes of Archaeologia Cantiana. This being so, a recapitulation of the evidence would be unnecessary, except that by being presented piecemeal in a number of papers, with factual evidence sometimes accompanied by tendentious conjecture, a degree of confusion seems to have arisen. The writer here offers a concise guide to the data. A main cause of past confusion stemmed from an interpretation of the prehistoric configuration of the South Mouth of the Wantsum proffered by George Dowker (1897) which had currency until the 1940s (Fig. 1). Dowker rejected Ebbsfleet in Thanet as the landing place of Hengest and St Augustine. He believed that ancient tradition, and the account given by Bede and subsequent writers to be wrong in this attribution. The actual landing-place he thought to be Stonar, near Sandwich. His thesis was that the southern entrance to the Wantsum had been narrow almost since the formation of the English Channel, and situated at Stonar. Thus Ebbsfleet in Thanet would have been landlocked at the time of Augustine and could not be the landing site. His theory was based on three assertions: - that during the initial filling of the English Channel the dominant tidal current ran southward through the Dover Strait from the North Sea, this process for some unknown reason reversing once the Channel had filled. - that the southward current brought great quantities of eroded flints from the Thanet cliffs to form the Stonar Bank. - that the Stonar Bank extended southward from Thanet as it was deposited. DAVE PERKINS 250 Fig. 1 An interpretation from George Dowker’s text (1897) of his reconstruction of the Wantsum Channel in prehistoric times. THE LONG DEMISE OF THE WANTSUM CHANNEL: A RECAPITULATION 251 Since these three conditions had occurred just after the end of the last glaciation, by Roman times he thought that a situation would have existed as shown in Fig. 1. So strongly did Dowker, a Fellow of the Geological Society, put this case that subsequent studies of the Wantsum accepted all or part of it for many years, and it even surfaces occasionally today. Hardman and Stebbing (1940-41) in their history of Stonar were cautious about rejecting it in its entirety, although their paper contained data that invalidated it. Rebuttal of Dowker’s three-part theory can be made as follows: - the strong southward current through the Dover Strait, subsequently reversed, is purely hypothetical, and has no shred of supporting evidence. He did not know that there is in fact a net residual southward current, although this has no effect on the deposition of shingle which is brought north along the Channel coast by strong inshore currents. - flint bands in the Thanet Cliffs are too thin and few to have supplied the Stonar Bank flints, whatever the scale of erosion. The origin of the Stonar flints is obscure although they contain erratics thought to come from East Anglia (Baden-Powell 1942), and, unlike other flint in the area, the Stonar flint has special qualities causing its large scale exportation to the Staffordshire potteries (Pearce 1937). Further, borings through the flints found at a depth of 12.0m a channel floor scoured of Thanet Beds sand down to chalk (Hardman and Stebbing 1940, 74); precisely demonstrating the one-time existence of the wide Wantsum South Mouth denied by Dowker. - there is no evidence that the Stonar Bank extended southward from Thanet. To the contrary, it becomes attenuated the further north it goes towards Thanet, and disappears 1200m south of Ebbsfleet. From that point on, as revealed by a deep pipeline cut in 1992, only alluvial and estuarine measures are present, capped thinly with beach deposits. While Dowker’s theory as to the formation of the Stonar Bank can no longer be accepted, the process is still a matter for speculation, of which that of Robinson and Cloet (1953) postulating the on-shore migration of an off-shore bank has some support among geologists. The Wantsum Channel when fully inundated; dimensions and tidal effects At the time of Caesar’s raid into Kent in 55 bc the Wantsum Channel had probably reached its maximum width and depth, and geophysical effects leading to its deterioration had commenced. As shown in Fig. 2 (based on the extent of the alluvium) the northern and southern mouths of the channel were about 4,000m and 3,000m across respectively, narrowing to about 1,800m between Wall End and Sarre (the traditional Medieval ferry crossing) and 1,500m between Sarre and the Stourmouth island. There DAVE PERKINS 252 Fig. 2 The likely configuration of the Wantsum Channel at full inundation, c.2,000 years ago. THE LONG DEMISE OF THE WANTSUM CHANNEL: A RECAPITULATION 253 were a number of islands in the channel, as shown in Fig. 2, and probably others that, having been ploughed or eroded to the level of the alluvium, have escaped notice. Borings carried out in 1923 (Hardman and Stebbing 1940, 70) indicated a maximum mid-channel depth of 40ft (12m), although later borings on which the cross-section shown on the OS Geological Map for the area is based indicate that it might once have been much deeper. Along the shores and round the islands depths would have been much less, large areas probably being exposed as saltings on the low water as is the case today on the Swale. Lunar effects on the North Sea give it a slightly different tidal clock to that of the Dover Strait, and its flow into the Strait on falling tides is more powerful than the reverse (Admiralty data in 1996 Channel Pilot). This, combined with the release of the pent-up Thames flow on falling tides and the deflected flow of the Stour would have made for a stronger north to south tidal current within the Wantsum on the ebb than the reverse on the flow, a factor leading eventually to dire consequences for the channel (see below). The navigation of the Wantsum Channel; historical and archaeological sources In 55 bc Caesar ran his fleet ashore on the shingle storm beaches west of the Wantsum instead of anchoring at Richborough, thus courting destruction in a subsequent storm (Frere 1967). According to Sheppard Frere (1967, 30), the near disaster that ensued resulted from Caesar’s lack of reconnaissance information. The Claudian invasion fleet of ad 43 is generally believed to have made its landfall at Richborough and a camp was constructed there. About ad 80 Richborough was expanded into a great supply base and fort (Rutupiae), and it was probably Agricola who set up the triumphal monument at its centre, ‘marking the conquest of all Britain, at the gateway to the island’ (Frere 1967, 118). The writer doubts that such developments would have taken place on a tortuous and muddy estuary. Late in the third century, in response to the threat from Saxon sea raiders, Richborough and Reculver (Regulbium) were rebuilt in stone as Saxon Shore forts. Richborough became headquarters of a Legion, Legio II Augusta, and in 386 Theodosius used the fort as the bridgehead for his pacification of Britain. The Wantsum was therefore of fundamental importance to the Romans throughout the whole of their occupation, in respect of both trade and defence. In the Dark Ages Sarre was one of the two principal ports of the Kentish kingdom and was under the command of a port reeve (Hawkes 1982, 76). The Danes wintered in Thanet in 851 and 860, and seem to have made the island a base, so that King Edgar ordered the island to be harried in DAVE PERKINS 254 969 and 980. In 1046 the Danes raided Sandwich with twenty-five ships and sailed round Thanet (AS. Chron. E 1045). Harold and his father, Earl Godwin, with a great fleet sailed from Dover to London via Sandwich and North Mouth in 1052. The Thames became frozen for several weeks in 1269, so that goods from Sandwich had to be transported overland (Stow quoted in Scott-Robinson 1878). Thereafter there are few explicit references to maritime traffic in the Wantsum. Scott-Robinson believed that this was because the route was so well known as to be taken for granted, the whole waterway from London to Sandwich being occasionally spoken of as the ‘Thames’ (Scott-Robinson 338). He pointed out that London had claimed jurisdiction over Stonar as a subject town, and mentions also a survey of the Thames banks by commissioners in 1374-5, which went ‘from Gravesend to Sheppey, and thence to Reculver and so to Sandwich and Dover’[author’s italics]. After this date most references to the Wantsum relate to its deterioration, although implying some sort of traffic was possible until about the year 1460 (see below). Thomas of Elmham’s ‘map’ of Thanet (dated about 1414) shows a passenger being carried out through the shallows to the ferry at Sarre on the shoulders of a boatman in what is depicted as a broad channel. Not much reliance can be placed on this schematic drawing unless perhaps as an indication that the ship channel and ferry crossing were both viable at that time. The little archaeological work carried out to date on the Wantsum alluvial plain and ancient shore lines consists of pipeline watching briefs in 1987 and 1994, with an evaluation in the latter year. The writer participated in all these operations. In 1987 a deep pipeline trench was cut between Wall End and Sarre parallel to the line of a post-medieval flood defence and road, the ‘Sarre Wall’ (Perkins 1989). Along the whole length, about 1,000m, the trench cut revealed saturated alluvial deposits with bands of marine and estuarine shells to a depth of 4.0m, demonstrating the width of the Wantsum after full inundation. In two places roughly midway in the channel ‘Red Hill’ sites were found during topsoil stripping. They consisted of mounds 20m and 35m in diameter rising about 0.80m above the surrounding land surface. The soil of which they were composed had been fired to a bright red, and in some places was of sub-ceramic consistency. Finds included briquetage and pot sherds in a local shell-tempered medieval fabric, dated c.1150-75. Unfortunately the trenching technique adopted under the extreme conditions of saturation did not allow the writer to determine whether these salt-making sites had been established on islands in the Wantsum, or on consolidated alluvium. In 1994 the pipeline watching brief and evaluation were carried out in relation to construction of a wastewater treatment works between Ebbsfleet and Weatherlees Hill (Hearne and Perkins 1995). Evaluation trenching THE LONG DEMISE OF THE WANTSUM CHANNEL: A RECAPITULATION 255 and topsoil stripping on the west side of the Ebbsfleet peninsula (see ‘Shoreline Area’ in Fig. 3) exposed a steeply shelving marine shoreline buried under alluvium. Tideline deposits on the shore contained remains of marine rather than estuarine creatures, and water-rolled chalk boulders were present, bearing the bore holes of piddocks (Pholas dactylus); also the shells of sand gapers (Scrobicularia plana). It should be noted that these species are not edible, and so are likely to be in situ rather than midden material. Just east of this shoreline and running either into it or parallel to it, were a number of ditches, presumably to drain off flood water. These contained midden materials including many marine shells, and medieval pot sherds ranging in date from the eleventh to the early fourteenth century. This is good evidence for the west side of the Ebbsfleet peninsular constituting the sea shore up to the fourteenth century, presumably as the eastern side of the Wantsum South Mouth. The silting and inning of the Wantsum by natural process and human agency The silting of the Wantsum Channel commenced as a natural process fairly late in prehistoric times with the establishment in its South Mouth of the Stonar Bank, which is presumed to have arrived there by off-shore migration (Robinson and Cloet 1953). The Stonar Bank did not join the shore of mainland Kent near the site of modern Sandwich, probably being scoured away by the force of several small rivers draining the North Downs. Of these the Delft and North Stream are still extant, and others can be traced. After the bank had formed, great quantities of flint shingle were carried north by inshore tidal currents in the English Channel, and deposited by gravity in the area of tidal slack caused by the Stonar Bank (Robinson and Cloet 1953). The timing of these events is uncertain, although both areas were habitable by or after ad 50 from the evidence of Roman settlement remains (Fig. 3). After the Norman Conquest the Wantsum starts to attract attention presumably as deterioration was noticed. In 1267 there was an official perambulation by the lord warden of the Cinque Ports, who inspected both the Thanet and mainland shores (Boys 1792, 70, reproduced by Scott-Robinson 338). It is worth quoting this itinerary: First beginning at the Stone Cross, at the west part of the town, near the causeway or common road between Sandwich and Ech, which cross is within the liberty; and from thence going along close by the river side, to Northmouth, everywhere by the line of high water mark, at spring tide; and then returning along the other margin of the river on the opposite side, through Sarr and Boxley in Thanet, to the shore at the passage directly against the cross of Hennebergh; and from that cross straight on DAVE PERKINS 256 Fig. 3 A reconstruction of the Wantsum in c.ad 1200 with the shingle ‘storm beaches’ advancing north and much ‘inning’ taking place. THE LONG DEMISE OF THE WANTSUM CHANNEL: A RECAPITULATION 257 the opposite side to the sea; and thence along the sea shore to Stonore, including the whole town of Stonore and the marshes within Hennebergh which are within the precinct of the liberty aforesaid; and on the other side of the river, crossing over to Peperness, and thence to a stream that runs into the river called the Gestling, by the thief downs, where persons condemned within the liberty are buried alive, and so going along that stream to a marsh called Holburgh, belonging to the lord of Poldre .... The two references (italicised) in Boys’ translation clearly demonstrate that at that time there were two southern mouths to the Wantsum, and it is quite surprising that their significance went un-noticed by later writers. The first has the lord warden’s party on the Thanet shore, then crossing the passage opposite the cross at Hennebergh (later known as ‘Little Joy’). The position of Hennebergh under its later name can be found from eighteenth-century maps to lie opposite or just north of Stonar Cut, this coinciding with the extreme northern extension of the Stonar Bank shingle. The party having crossed one southern mouth of the Wantsum then proceed along the shore (of the Stonar Bank) to Stonar (the town), and then cross over the river (Sandwich Haven) to Peperness on the mainland. The name Peperness has often been used to refer to the shingle point on the east side of the mouth of the Stour estuary (now called Shell Ness). If Peperness was at that time opposite Stonar (the town) then this is a case of the name migrating with the increase of the shingle. Obviously, at the time of the lord warden’s perambulation a situation existed similar to that given in Fig. 3. Worse was to come, however. During the later part of the thirteenth century the Ebbsfleet channel had silted so badly that merchant shipping could not negotiate it and had to enter the Wantsum via Sandwich Haven (Hardman and Stebbing 1941, 50). This progressive silting meant that the Haven now had to handle the major share of river flow and Wantsum tidal current, and itself started to silt badly. The people of Sandwich reacted with fury to this, believing that the Ebbsfleet channel was closing as a result of ‘inning’ and other activities carried out by the Abbot of St Augustine’s. In 1266 men of Sandwich and Stonar burnt two water mills at Ebbsfleet and Stonore (Hennebergh) that obstructed the Ebbsfleet channel (Davis 1934). The Abbot commenced legal action against the people of Sandwich in 1280, and it appears from this that he had built a sea defence wall from Stonar to Cliffsend that was being vandalised by the Sandwich men (Boys 1792, 660). This sea wall does not appear, however, to have entirely blocked the Ebbsfleet channel, at that time known as ‘Minster Flete’. Boys’ map of the Stour made in 1775 shows it as ‘Minster sewer’ (sluice) running into the Stour east of Weatherlees Hill (Hardman and Stebbing 1940, facing 69). So ended the Ebbsfleet mouth of the Wantsum, of which it was said in 1313 that it ‘used to be so wide that two cogges (cogs) might turn therein clear of one another’ (Boys 1792, 665). With the Wantsum from North Mouth DAVE PERKINS 258 to Sandwich Haven now silting fast, and subject to extensive inning and draining operations from both banks, its fate as a navigable channel was sealed. The final demise of the Wantsum as a navigable waterway; historical sources By an Act of Parliament of Henry VII, in 1485 a bridge was allowed to replace the ferry at Sarre, since the channel was so silted there that neither the ferry or other vessels could pass except on spring tides. Even so it was stipulated that the arches had to be big enough for boats and lighters to pass, in the hope that ‘the water shall happen to increase’ (Scott-Robertson 1878). Of this bridge the writer can find nothing more recorded. In order to fulfil the above condition it was presumably built of wood and long ago rotted. The present structure is a small stone bridge carrying the Sarre Wall road across the cut that joins the Stour to the Wantsum brook and the drainage system of Birchington marshes. It would seem to have been constructed in the late eighteenth century. John Twine writing in 1590 (quoted by Scott-Robertson 1878) claimed that: there are eight worthy men still living who have seen not only the smallest boats but larger barks pass and repass between that isle and our continent. A manuscript of about 1650 in the archives of Quex Park, Thanet, has one of the Crispe family deploring the fact that a boat can no longer be sailed from Stourmouth to North Mouth, ‘as the Wantsum has been let to be overgrown with trees’. Seventy years on however, John Lewis heard local farmers talking of ‘crossing over to England’ on their way to Canterbury Market (Lewis 1723, 11). By this time the whole area of the one-time Wantsum Channel had become an alluvial flood plain cut by the Stour making its way via oxbow bends to the relic of Sandwich haven and thence north to the sea. Even so, in its convoluted course the river approaches so close to the ancient Ebbsfleet entrance that in the eighteenth century Smeaton and others almost re-established it with the ‘Stonar Cut.’ This was not however an attempt to allow river craft to bypass Sandwich, but to use pent-up river water released through a sluice to scour the lower reaches of the river estuary (Hardman and Stebbing 1940, 69). Summary It can be seen therefore that Thanet was an island throughout the Neolithic THE LONG DEMISE OF THE WANTSUM CHANNEL: A RECAPITULATION 259 and Bronze Ages, and continued so in the generally understood sense of the term until late in the medieval period. Moreover, that the body of water separating it from mainland Kent, the Wantsum, was for most of that time of such magnitude as to be un-crossable except by boat, and of such depth as to be navigable by the largest vessels of those days. With a wide Wantsum South Mouth exposing that sea channel’s western shore to the eroding effect of south-easterly gales, it is possible to see the collapsed state of Richborough’s eastern defences and the steep fall of the bluff as being the result of wave erosion. Certainly, Ebbsfleet, demonstratably a sea-girt peninsular until the early fourteenth century, is reinstated as a prehistoric and historic entrepôt of the first importance. references Baden-Powell, D., 1942, ‘Report on the Erratics from Stonar’, in Hardman, F., and Stebbing, W. Boys, W., 1792, History of Sandwich, 536. Dowker, G., 1897, ‘On the landing place St. Augustine’, Archaeologia Cantiana, xxii. Frere, S., 1967, Britannia, Book Club Associates, London. Hardman, F., and Stebbing, W., 1940, ‘Stonar and the Wantsum Channel’, Archaeologia Cantiana, liii ; see also Part Two (1941), Archaeologia Cantiana, liv. Hawkes, S., 1982, ‘Anglo-Saxon Kent, c. AD 425-725’, in Archaeology in Kent to AD 1500, CBA Research Report No. 48, 64-78. Hearne, C., Perkins, D. and Andrews, P., 1995, ‘The Sandwich Bay Archaeological Project, 1992-94’, Archaeologia Cantiana, cxv, 239-354. Lewis, J., 1723, A History of the Isle of Thanet, 1st ed. Pearce, B.W., 1937, ‘Medieval Discoveries at Stonar’, Archaeologia Cantiana, xlix, 278-79. Perkins, D., 1989, ‘The Selling to Thanet Trunk Water Main’, Archaeologia Cantiana, cvii, 267-79 (270). Robinson, A. and Cloet, R., 1953, ‘Coastal Evolution in Sandwich Bay’, Proceedings of the Geological Association, 64, 2. Scott-Robertson, W., 1878, ‘Thanet’s Insulation’, Archaeologia Cantiana, xii.