Old Romney: An examination of the evidence for a lost Saxo-Norman port

OLD ROMNEY: AN EXAMINATION OF THE EVIDENCE FOR A LOS T SAXO-NORMAN PORT* MARK GARDINER It has long been supposed that Old Romney may have been the predecessor to the medieval Cinque Port of New Romney. Fieldwalking to the east of Old Romney located a number of medieval sites, but produced no evidence for the presence of a former town or port. Documentary research confirmed that the Domesday town of Romenel was very probably that of New Romney. Old Romney is identified as the site of a shrunken medieval village. Since at least the sixteenth century historians and antiquarians have speculated that the hamlet of Old Romney might be the predecessor of the medieval Cinque Port of New Romney.1 The very names of the two places seem to suggest that it must have been so. In recent years that theory has been revived by Tatton-Brown who has argued that due to the silting of the estuary on which Old Romney stood, the town moved before the mid twelfth century to a new location nearer the sea.2 The evidence for such an estuary has been established by R.D. Green who was able to trace the line of a watercourse, presumably the river called the He or Rhee, from as far inland as Brenzett to New Romney.3 The • Published with the aid of a grant from English Heritage. 1 W. Lambarde, A Perambulation of Kent (1826 edn.), 178; E. Hasted, The History and Topographical Survey of the County of Kent, 8, 440. F o r an example of early seventeenth-century views on the relationship of Old and New Romney, see A. Roper, The Church of Saint Clement, Old Romney (5th edn., 1983), 30. 2 T.W. Tatton-Brown, 'The Towns of Kent', in (Ed.) J. Haslam, Anglo-Saxon Towns in Southern England (1984), 26-8; 'The Topography of the Walland Marsh Area between the Eleventh and Thirteenth Centuries', in J. Eddison and C. Green, Romney Marsh: Evolution, Occupation, Reclamation (1988), 106; 'The Anglo-Saxon Towns of Kent', in (Ed.) D. Hooke, Anglo-Saxon Settlements (1988), 229-31. 3 R.D. Green, Soils of Romney Marsh (Soil Survey of Great Britain, bulletin no. 4, 1968). A.F. Butcher, 'The Hospital of St. Stephen and St. Thomas, New Romney: The Documentary Evidence', Arch. Cant., xcvi (1980), 19 citing Magdalen College, Oxford Muniments, Deeds, Romney Marsh, 53. 329 MARK GARDINER river passed close to Old Romney, where it was joined by the River Rother flowing along an earlier course, 4 and broadened into a wider estuary before passing to the south of New Romney. One of the grounds for the identification of Old Romney as the site of a Saxo-Norman port is the notable concentration of churches around the former estuary of the Rhee. In addition to the standing church of St. Leonard's, there were two other churches dedicated to St. Michael and St. Lawrence. The church of Hope All Saints, which still survives in part as a ruin, stood further along and on the north side of the estuary. The discussion of the relationship of Old and New Romney has rarely been substantiated by a detailed examination of the evidence. Little fieldwork has been carried out in the area to determine whether traces of the supposed town survived. In 1991, English Heritage and the Romney Marsh Research Trust jointly funded a new survey by the Field Archaeology Unit (Institute of Archaeology, London) of the Old Romney area and a review of the historical evidence. The purposes of the work were to identify the location and character of archaeological remains, to determine whether action was necessary to ensure their preservation, and to undertake a study of the historical sources relating to Old Romney. PREVIOUS EVIDENCE In addition to the supposed site of the former town of Old Romney near to the church, a number of other sites are known in the area: 1. Two oval ditched enclosures or moats lie on either side of Five Vents Lane. The eastern one has been virtually levelled by ploughing, but trial excavations in 1970 revealed building remains and burials, including a stone-lined grave. Evidence of a stone wall with footings over 1 m. in width were found. The site is marked on the tithe map as St. John's Field. The interior of the western enclosure has also been ploughed and trial work exposed building footings.5 2. The sites of the churches of St. Michael and St. Lawrence are indicated by the Ordnance Survey, though the former is shown in the wrong position. Trial trenching on the probable site of St. Michael's church identified a chalk floor and the south-east corner of a building. Burials in coffins were found aligned east-west to the north of the 4 C. Green, 'Palaeogeography of Marine Inlets in the Romney Marsh Area', in (Eds.) J. Eddison and C. Green, Romney Marsh: Evolution, Occupation, Reclamation (1988), 167-74. s C(entre for) K(entish) S(tudies) U1045/043; J. Bradshaw, 'Investigations and Excavations during the Year: Ashford Area', Arch. Cant., lxxxv (1970), 179. 330 AN EXAMINATION OF THE EVIDENCE FOR A LOST SAXO-NORMAN PORT building. Building debris has also been found on the site of St. Lawrence's church.6 FIELD EVIDENCE A programme of field-walking was undertaken by the Field Archaeology Unit during October 1991 to define the extent of medieval activity around the present hamlet of Old Romney. The fields immediately adjacent to the church were under permanent pasture, but those beyond the hamlet were mostly under arable. Access was not possible to most of the land to the north of the Wallingham Sewer due to a change in tenancy shortly before fieldwork and an area to the east was substituted instead. The land walked lay in three blocks. The eastern area (Fig. 1, Area C) lay between the Rhee Wall and Wallingham Sewer. The second area lay immediately east of the permanent pasture next to the hamlet of Old Romney and covered the land south of the site of 'St Lawrence's church' (Fig. 1, Area B). The third block lay alongside Five Vents Lane between the supposed sites of St. Michael and St. Lawrence churches. It included the enclosure in St. John's Field to the east of the road (Fig. 1, Area A). The area was examined intensively. Lines 5 m. apart were walked and the finds from the field surface were collected in 25-m. squares aligned to the Ordnance Survey National Grid. All artefacts were collected. Ceramic building material, slate and shell were counted, weighed and subsequently discarded. The pottery was retained, washed and marked. The pottery was then sorted into 50 fabrics and was quantified by a simple sherd count. The results were entered into a database and the fabrics grouped into five period categories. The means and standard deviations for each period group were then calculated. The distributions were examined by plotting the sherd counts for each square in terms of the number of standard deviations above the mean, a method which allows significant distributions to be identified. The dating of pottery in the Romney Marsh area is poorly understood, but the pottery fabrics may reasonably be attributed to the five broad periods. The periods used were Roman, early medieval (1000-1250), later medieval (1250-1500), early post-medieval (1500-1700) and later post-medieval (1700-1850). Table 1 shows the quantities of pottery recovered by period. The numbers of Roman and early post-medieval pottery found were very small. A total of 3494 6 J. Bradshaw, 'Investigations and Excavations during the Year: Ashford Area', Arch. Cant., lxxxiv (1969), 235. 331 MARK GARDINER Fig. 1. Old Romney showing the areas field-walked and the location of glebe land. sherds of pottery were recovered from the 791 squares walked covering an area of approximately 49 ha. The finds of ceramic building material, excluding brick, West Country slate and shell were recorded by weight rather than numbers of pieces. Ceramic building material is a broad category which comprised mostly roof-tile, but also included floor-tile and field-drain fragments. It is not always easy to separate abraded pieces of medieval tile from similar material of post-medieval date or from field-drain fragments. All the ceramic building material, except brick, was therefore treated collectively and a total of 155 kg. was recorded. West Country slate was a common roofing material used along the south coast of England and inland near to river valleys during the medieval period. Its use ceased about 1450 and when slate was used again in the nineteenth century and subsequently, quarries in Wales rather than England were exploited. West Country slate may be differentiated fairly reliably by eye from Welsh slate. Slate needed to be renewed at intervals as it did not weather well. A total of 13 kg. of West Country slate was recorded. 7 Shell, effectively oyster shell since 7 E.M. Jope and G.C. Dunning, 'The Use of Blue Slate for Roofing in Medieval England', Antiq. Jour. xxxiv (1954), 209-15. For usage of Welsh slate in Sussex see, E.W. Holden, 'Slate Roofing in Medieval Sussex', Sx Arch. Coll. ciii (1965), 67-78; E.W. Holden, 'Slate Roofing in Medieval Sussex - A Reappraisal', Sx Arch. Coll. cxxvii ( 1989), 73-88. For an example of the use of slate on a medieval building at New Romney, see S.E. Rigold, 'Two Kentish Hospitals Re-Examined: St. Mary, Ospringe and Sts. Stephen and Thomas, New Romney', Arch. Cant., lxxix (1964), 56, 58. 332 uJuJ uJ Early Medieval Pottery 256 03, '! ···········-·········:············-···-·····= :. • : • • = • • . 􀀆- • • i -􀀄·· 􀀉. -•:;t-􀀑-:-;1.-:-! -i -- r---- T :• . . ; . \ . 􀀅 Later Medieval Pottery ; 􀀅 . . 􀀆 . . . 􀀈 • 1 • • ; • i. • • ' I· • 1 · : • i t N I O 100m Sd above mean 1 • 2 • 3 • 4 • 5 vvest Country Slate ··(·· 􀀃 .... i ·---􀀛-:--t---;-...... ---\- -·' ---·-··!·· -·-+- Ceramic Building Material ·f ·• ........... ; ··-·········;···. ·········· i ··-· .. 1:--;-:/\-. +---􀀘 .-!-:--: • l; · I􀀄 : . ! ------' ..... 1 • ;··· • i Fig. 2. Area A. Finds distribution plotted as standard deviations above the mean. 􀀇 􀀈 0 z􀀉􀀊 􀀋 8 > Cl'l >-l Cl'l 􀀌 􀀍 "Cl 0 MARK GARDINER these are the largest, was measured by weight. Concentrations of shell were also noted during field walking and their positions recorded. Note was also made of concentrations of other finds, variations in soil colour and areas of shingle present on the field surface. The greater part of all the early medieval pottery was found in Area A (Fig. 2). Three concentrations could be isolated, though a scatter of sherds was found across the whole of that area of land. One of the concentrations lay within and extended to the south of the enclosure in the area called St. John's Field. The site of the enclosure is marked by a slight rise. Building stone with mortar was noted on the field surface on the southern half of the enclosure and West Country slate was picked up to the east. Tile was discovered mainly to the south. Some bone was picked up in the area and, though it was not determined whether these were human or animal, the presence of butchery marks on many pieces suggests the latter. Ploughing is continuing to degrade the site and groups of conjoining pot-sherds with fresh breaks were noted on the surface where the plough had bitten into deposits. Two other concentrations of early medieval pottery were noted in Area A both set a little way back from Five Vents Lane. An area of darker soil was noted around the one to the north-east. Later medieval pottery was found particularly in Area A. Pottery of that period lay in two concentrations, though neither coincided with the St. John's Field enclosure. A scatter of ceramic building material was found with the southern pottery concentration. Area B produced few finds of later medieval pottery. In Area C, though greater quantities were found, the distribution is not immediately clear. The mean and standard deviation of pottery of the period are sufficiently high that the concentrations with lower densities do not appear. The statistical method masks real patterns within the distribution. These may be brought out in a less formal way by marking with tone groups of squares with a pottery density at least 1.5 times the mean (Fig. 3). Six areas of pottery then emerge, all towards the north side of the block of land. That trend to locations nearer the Wallingham Sewer is not apparently a result of biases in field collection, but is representative of finds location. When the pottery distribution in the area is compared with that of other finds, further patterns appear. Many of the pottery concentrations lie near to scatters of shell. West Country slate is found mainly towards the north-west of Area C and lies generally just beyond the area of pottery (Figs. 3, 4). The patterning of ceramic building material, though clear in itself, cannot be readily understood in relationship to other finds. Only a small number of early post-medieval sherds were found and no pattern could be discerned in their distribution. The density of finds of that date in the three blocks of land walked was approximately 334 l,.) l,.) UI Later Medieval Pottery l•.. . . ······-! ',,:-··,: ---- r· - - 0 t N I 100 Shell -····---·): ·· ·······- ·•- ----•··=i ·•-·- I . l :1: r- ---!􀀊::::T ·::;:i.􀀑 ::]; - -, i􀀂 • • • • i • 1 • "" 􀀓 􀀄 􀀅 . . 􀀃 . ; ·······􀀜···············•··•···!····················, ! I -·-·····1·······1··􀀈 : ; Sd above mean 1 • 2 • 3 • 4 • 5 Fig. 3. Area C. Finds distribution plotted as standard deviations above the mean. Tone indicates groups of squares with a pottery density at least 1.5 times above the mean. 􀀆 I 􀀇 􀀈 i @ > t"' 0 en ..,i en "' 0 w w 0\ Ceramic Building Material i 􀀏 m/-•·-·•• ! ·--· J> ! l = l i J :1 I . · .j ,c••··· ------ ·····i-··-···-··-··--·r··········· i • -· 1 . . 􀀄 . =f·-r·---·--·􀀐·-·Y·--- i ' 2u/ ·······-·····--l----·····-·---L---, • • j·--t 0 t 1r' . 􀀅 . -----'--􀀏 / Sd above mean I • 1 • 2 • 3 • 4 • 5 Fig. 4. Aiea C. Finds distribution plotted as standard deviations above the mean. 􀀁 :;s: I AN EXAMINATION OF THE EVIDENCE FOR A LOST SAXO-NORMAN PORT equal. Late post-medieval pottery tended to occur in fields adjacent to the Rhee Wall. That is apparent in Area B (Fig. 5) and in Area C. It is possible that eighteenth- and nineteenth-century pottery was being spread on the fields with nightsoil and manure brought from New Romney along the main road on the Rhee Wall and dumped on the immediately adjoining land. A scatter of ceramic building material beside the Rhee Wall in Area B indicates the site of a recent structure (Fig. 5). The interpretation of the finds patterns is by no means easy. It is notable that the finds of West Country slate occur within and partly to the west of the St. John's Field enclosure, but the early medieval pottery lies on either side of it. It may be that the enclosure did not surround an occupation site and therefore little pottery was used, or that the pottery was removed from the site when the middens were cleared to manure the surrounding fields. The distribution of ceramic building material is less clear. It lies around the enclosure rather than within it, and may be unrelated to the building there. Little settlement activity was identified in Area B, but Area C produced a greater number of finds. West Country slate was found in the same areas and around concentrations of later medieval pottery. Ceramic building material too lies just beyond the medieval pottery scatters. These finds suggest one, or more probably a number of medieval settlements on the north side of Area C. Shell, however, has a quite different distribution. It lies in a band across the centre of the area and no interpretation is readily apparent. A notable result of fieldwalking in Area C is the lack of medieval finds on the south side nearer to the Rhee Wall. The most significant result from field-walking, from the point of view of the research aims, was the relatively low densities of early medieval finds from the Area A. The quantities of material found certainly do not support the contention that it was the first site of the town of Romney. The evidence suggests that the areas examined were occupied by a number of farms and there was no nucleated settlement. The supposed churches of St. Michael and St. Lawrence, which lay just outside the areas walked, evidently stood in isolation. These, like most of the medieval sites located in field walking, lay close to roads. The sites indicated by the finds scatters in Area A were set back a little way from Five Vents Lane in the fields. They occupied a similar position in Area C in relationship to the track beside Wallingham Sewer, the medieval road between Old and New Romney (see below). No sites were identified adjoining the Rhee Wall, though the banks of the wall may not have served as a routeway in the medieval period. 337 uJ uJ 00 Late Post-Medieval Pottery ... .., . . . 253􀀌··········􀀍··················-·· i • ; J____i._ _ __ I _ i ! I 251/----····f······················!······················i······················1· I I ! = t N I 􀀓 .................... i--􀀔----·-· ......... 􀀕􀀔-----· .............. 􀀖 ................... : 􀀂 . ; . 0 Ceramic Building Material and Slate 1---1----·.r i······-·· -i- __ J. - I-... --l-··-·--l- _J ._ CBM Slate ij ! • 1 "' 1 • 2 • 3 Fig. 5. Area B. Selected finds plotted as standard deviations above the mean. I i AN EXAMINATION OF THE EVIDENCE FOR A LOS T SAXO-NORMAN PORT DOCUM ENTARY EVIDENCE Documentary sources before the mid twelfth century do not distinguish between Old and New Romney, but refer simply to 'Romney'. The first reference to Romney may be a charter of 741, which granted to the church of Lyminge fishing rights at the mouth of the Limen with part of the estate on which the oratory of St. Martin and fishermen's houses were situated. 8 The oratory has been identified with the later medieval church of the same dedication in New Romney, and the reference to the mouth of the Limen, the former name for the River Rother, seems appropriate for that location. It is possible that the land granted in the eighth century formed the core of the later manor of St. Martin's. The lands of Lyminge monastery passed to Christ Church, Canterbury, in the ninth or tenth century and St. Martin's manor was later in the estate of the archbishop of Canterbury. There is no evidence for a town at Romney before the eleventh century when a mint was established there.9 The port played an important role in the struggle between Godwine and Edward the Confessor and, later, burgesses were punished for killing some of the men of William I. Four entries in Domesday Book refer to the town of Romney or manors in the vicinity: (i) 85 burgesses in Romney belonged to the Archbishop of Canterbury's manor of Aldington (DBi, 4a); (ii) Robert de Romenel held 1 ½ sulungs at Langport of the Archbishop of Canterbury. Twenty-one burgesses in Romney belonged to that manor (DBi, 4b); (iii) Robert de Romenel held 1 sulung at Afettune in Langport hundred of the Bishop of Bayeux (DBi, 4b); (iv) Robert de Romenel had 50 burgesses in the borough of Romney (DB i, 10b, lla). It was presumably from the first of these manors that the archbishop at a later date received a rent of £6 10s.3¼d. recorded in a rental of c. 1285 and payments from stallage, fair and market tolls and port dues from ships arriving there. 10 It is convenient next to treat the third of these manors. The descent of Afettune or Offeton has been traced in part by J.H. Round who has 8 P.H. Sawyer, Anglo-Saxon Charters: An Annotated List and Bibliography (1968), no. 24. 9 S.E. Rigold, footnote 7, in E.W. Parkin, 'The Ancient Buildings of New Romney', Arch. Cant., lxxxviii (1973), 118. 10 Lambeth Palace Library ED 2068. 339 MARK GARDINER shown it was held by the serjeanty of keeping the king's hawks.11 Robert's lands were inherited by Lambert de Romenel and before c. 1159 they came into the hands of David de Romenel who drowned in Romney Marsh shortly before 1168.12 David had two daughters, Emma and Wimarc. Afettune was inherited by Wimarc who was the 'ancestor' of Aubrey de Old Romney who was holding the lands by 1180. 13 She married William de Jarpenville and had a daughter, Alice.14 Aubrey outlived her husband and died an elderly widow in 1226.15 Alice, her daughter married Thomas fitz Bernard who held the manor on his death in 1238.16 He was succeeded in turn by his son John who died in 1259 leaving his son Ralph who was a minor. 17 Ralph fitz Bernard died in possession of the land in 1305/6, but had been predeceased by his son, John and the manor was inherited by his grandson, Thomas.is On his death the manor was held for a life term by John fitz Bernard and in 1361 after his death came to Thomas de Vere. During the thirteenth century part of the land of Afettune was granted to Roger de Romenel, whose son, Stephen held it in 1284-5.19 The second of the manors, Langport was held of the archbishop for the service of three knights' fees. It followed the same descent as Afettune until c. 1168 when, on the death of David de Romenel, it was divided between Emma and Wimarc. Emma, or one of her four daughters and heiresses, seems to have married Peter de Langport who is found holding 1½ fees in Langport in 1171,20 In 1210-12 that half of the manor of Langport was in the possession of Jocelin de Spain.21 It was held by Robert de Septvans before his death in 1239 and subsequently descended in that family until about 1422 when it was sold to John Writtle.22 The remaining 1½ fees were held by William de 11 J.H. Round, The King's Serjeants and Offices of State (1911), 303-10. 12 Pipe Roll of 31 Henry lll, 64; Pipe Roll of 14 Henry II, 154. 13 Curia Regis Rolls 9, 14. 14 Cartae Antiquae (Pipe Roll Soc.) 1, 135 (no. 282); British Library Cotton MS, Faustina A i, f. 228v. (modem foliation), no. 368. 1s Excerpta e Rotulis Finium in Turris Londinensi (Record Commissioners (1835)) 1, 152. 16 Round, The King's Serjeants, 305. Round cites no source for that date. 17 Calendar of Inquisitions Post Mortem, 1, 130 (no. 468); Book of Fees 2, 1381. 1s Calendar of Inquisitions Post Mortem 4, 255. 19 Feudal Aids 3, 3; Rotuli Hundredorum, (Record Commissioners (1812)), 1, 226; British Library Campbell Ch. XXVII, 25. 20 East Sussex Record Office GLY 954, printed in 'A List of the Archbishop of Canterbury's Tenants by Knight Service in the Reign of Henry II', (Ed.) H.M. Colvin in Documents Illustrative of Medieval Kentish Society, (Ed.) F.R.H. Du Boulay (Kent Records 18 (1964)), 1-40; Curia Regis Roll 9, 14. 21 Red Book of the Exchequer (Rolls series), 2, 725. 22 R. Tower, 'The Family of Septvans', Arch. Cant., xl (1928), 128. 340 AN EXAMINATION OF THE EVIDENCE FOR A LOST SAXO-NORMAN PORT Jarpenville, the husband of Aubrey de Old Romney and descended with the manor of Afettune. The manor of Afettune is not mentioned after the later thirteenth century, but seems subsequently to have been known as the manor of Old Romney. On the death of Ralph fitz Bernard the manors of Langport and Old Romney were apparently separated from the other lands of fitz Bernard. The other part of the estate, including the advowson of Old Romney church, descended to Margaret, the daughter of Ralph who married Guncelin de Badlesmere. Those lands were inherited by her son Bartholomew de Badlesmere and subsequently, grandson Giles. Giles' land passed to his sisters, but by an agreement of 1350 the manors of Langport and Old Romney were to be received by one of these sisters, Matilda de Badlesmere. She married John de Vere, but both died before John fitz Bernard who held the manors for life.23 The purpose in tracing these descents has been to identify the location of the lands of the Domesday manors. The moiety of Langport held by Peter de Langport in 1171 came to be known as the manor of New Langport. A rental of that manor of 1393/4 when it was in the possession of William Septvans provides a detailed description of the lands, which lay mostly in West Brooks and towards Lydd.24 The other half held by the fitz Bernard and later de Vere families was called the manor of Old Langport. A detailed rental of that manor of 1552 based on earlier records, including perhaps the now lost rent roll of 1403,25 shows that the lands of Old Langport lay around and to the south of the church of Old Romney. Beyond the church to the north and east were the lands of the archbishop of Canterbury's manor of St. Martin's. In the thirteenth century St. Martin's was a large manor with assized rents of £34 19s.0½d. and lay within the archbishop's bailiwick of Aldington. It was divided into two parts, Northee and Southee, presumably so named from their position either side of the Rhee estuary.26 The 1552 survey of Old Langport contains sufficient detail to allow the location of the holdings to be reconstructed schematically (Fig. 6). The Rhee Wall, the road between Old and New Romney which followed the Wallingham Sewer, the church and the glebe are mentioned and these provide fixed landmarks to which other places may be related. The survey records ten messuages within the liberty of Old Romney and mentions the sites of a further eight others which had 23 W.A.S. Robinson, 'Romney, Old and New', Arch., Cant., xiii (1880), 370. 24 CKS U442/M72. 25 CKS U1043/M4. The rental of 1403 is mentioned in Robinson, 'Romney, Old and New', 372n. 26 Lambeth Palace Library ED 2068; Public Record Office SC 6/1129/3. 341 w .j:s. N Old Romney 1552 Road to f Culve-rpit 0 Fee of Nort hee Wm. Barowe 11 T hos. Harlakinden 5 ac 6 ac 1 rood 3 roods Wm. Epce 11 Wm. Epce Culverton demesne 1 ac 1 ac ½ rood watercourse Henry Whateman Preysroun I Henry , rood Whateman -t2½ac Wm. Epce 5ac Glebe Glebe Fee of Prior of Dover Old Romney to New Romney Road Fee of Northee Wm. Epce demesne 2'2 ac Priory of Leeds' land Fig. 6. Schematic reconstruction of Old Romney in 1552 (based on CKS Ul043/M4). 􀀁 AN EXAMINATION OF THE EVIDENCE FOR A LOS T SAXO-NORMAN PORT been deserted since the date of an earlier survey, now not extant. The pattern of settlement in the mid-sixteenth century was very similar to the present with a cluster of houses to the north and north-west of the church and close to Romney Cross, the junction of Five Vents Lane and the road by Wallingham Sewer. In addition, there was also a large house to the west of the road from Romney Cross to the Rhee Wall which is now farmland. The glebe is mentioned only as a boundary in the survey, but its location is shown on the tithe map. There is good reason to think that the land of the glebe in the nineteenth century was more or less the same as in the sixteenth and probably even in the mid thirteenth century. It lay to the north and south of the Rhee Wall and it is very probable that it had been bisected when the 'wall', or more correctly the channel, had been extended after 1258.27 Tatton-Brown has shown that when the first length of the Rhee Wall was constructed the land of the Hospital of Sts. Stephen and Thomas, New Romney, which lay to the west, had been similarly divided.28 DISCUSSION It has been possible to identify the manors of the Old Romney area and trace their descent from Domesday Book. The area immediately around Old Romney church and all the land in liberty of Old Romney, except the glebe, lay within the manor Old Langport. It is possible that the glebe land had been granted to the church by the lords of Langport (later Old Langport) manor who also held the advowson of the church.29 That area around the church and within the liberty is the most likely situation of the town of Romenel, if it lay at Old Romney. The wording of the Domesday entry makes clear, however, that the manor of Langport was separate from the town of Romenel. Twenty-one burgesses in the town belonged to the manor (ad hoe manerium pertinent xxi burgessos), but it is one of numerous examples in Domesday Book of urban tenements in the tenure of nearby rural manors.30 Another part of the town of Romenel was held by the archbishop's manor of Aldington. That is of considerable importance 27 For the extension of the Rhee Wall, see Calendar of Patent Rolls 1247-58, 635-6; Green, Soils of Romney Marsh, fig. 20. 28 T. Tatt on-Brown, 'The Topography of the Walland Marsh A rea between the Eleventh and Thirteenth Centuries', in (Eds.) J. Eddison and C. Green, Romney Marsh: Evolution, Occupation, Reclamation (1988), 106. 29 Calendar of Inquisitions Post Mortem 4, 255. 30 H.C. Darby, Domesday England (1977), 309-13. 343 MARK GARDINER for the late eleventh-century town can not have been around the church at Old Romney, which was entirely in the manor of Langport. That may suggest that the Domesday town of Romenel should be identified with New Romney. One further strand of evidence supports that tentative conclusion. Domesday Book records that the archbishop had the largest number of burgesses within Romenel. It has already been suggested that it was the rent from the descendants of these which was noted in the survey of c. 1285. The level of income received at that time and the reference to money from market and port dues show that these were not coming from a decayed town, but an active urban centre. It must be concluded that in the late thirteenth century the burgesses of the archbishop were in the only flourishing town in the Romney area, New Romney. CONCLUSION The evidence from archaeological and historical work has refuted the suggestion that Old Romney was the site of the decayed predecessor of the Cinque Port town of New Romney. The sixteenth-century hamlet of Old Romney was the rump of a larger medieval village which had developed around the church, though much of the settlement in the parish was in dispersed farmsteads scattered among the fields. Fieldwalking undertaken by Anne Reeves in the north-east of Romney Marsh has shown a very similar picture. She has discovered medieval sites scattered across the whole of the area surveyed. The sites she located, like those at Old Romney, were typically beside trackways or at the edge of fields. The greatest number were dated to the early medieval period (defined as 1050 to 1250), with a smaller number of medieval sites (1250 to 1400/50) and significantly fewer later medieval sites (1400/50 to 1500/50).31 The period distribution of finds in the Old Romney area is rather different, though it may reflect the more limited area surveyed. There remains a small possibility that Old Romney might have been superseded by New Romney before the late eleventh century. There is no documentary evidence in favour of that interpretation and fieldwork failed to find evidence of a substantial settlement at Old Romney. A number of concentrations of pottery from the period 1000-1250 were discovered in an area to the north of Romney church, but these have been interpreted as isolated farmsteads. Two problems remain if Old 31 A. Reeves, 'Romney Marsh: The Fieldwalking Evidence', in (Ed.) J. Eddison, Romney Marsh: The Debatable Ground (forthcoming). 344 AN EXAMINATION OF THE EVIDENCE FOR A LOST SAXO-NORMAN PORT Romney was not the Saxo-Norman port. Firstly, the names Old and New Romney suggest that the latter superseded the former. Secondly, the supposed churches of St. Lawrence and St. Michael, both of which have produced evidence for masonry structures and for burials, are difficult to explain. How was it possible for a small settlement to sustain these two churches and the surviving church of St. Clement? Neither of these problems can be explained entirely convincingly and further research is required. The study of Old Romney has emphasised the potentiality of combined documentary and field research. A number of problems still remain in the understanding of the area, but the first steps have been taken towards comprehending the nature of medieval settlement on the estuary of the Rhee. ACKNOWLEDGEMENTS The Field Archaeology Unit is grateful to English Heritage and the Romney Marsh Research Trust for jointly funding the study of Old Romney. We are indebted to those landowners who allowed us access to their fields. The field-walking was supervised by Luke Barber and a large part of the pottery classification and entry of figures for the data analysis was undertaken by Tessa Machling. I am also grateful to Luke Barber for suggestions about the finds distributions. TABLE 1 POTTERY FOUND BY PERIOD Period Roman Early Later Early Post- Later Post- Unidentified Medieval Medieval Medieval Medieval Sherd 1 677 2486 65 200 65 number Per- 0 19 71 2 6 2 centage 345