Place-Names and Early Settlement in Kent

PLACE-NAMES AND EARLY SETTLEMENT IN KENT By P. H. REANEY, LITT.D., PH.D., F.S.A.1 APART from a few river-names like Medway and Limen and the names of such Romano-British towns as Rochester, Reculver and Richborough, the place-names of Kent are of English origin. What Celtic names survive are too few to throw any light on the history of Romano- British Kent. The followers and descendants of Hengest and Horsa gave names of their own to the places where they settled and the early settlement with which we are concerned is that which began with the coming of the Saxons in the middle of the fifth century A.D. The material for the history of this conquest is late and unsatisfactory, based entirely on legends and traditions. The earliest of our authorities, Gildas, writing a full century after the events, was concerned chiefly with castigating the Britons for their sins ; names and dates were no concern of his. Bede was a writer of a different type, a historian on whom we can rely, but his Ecclesiastical History dates from about 730, nearly 300 years after the advent of the Saxons. He was a Northumbrian, too, and was careful to say that the story of Hengest was only a tradition, " what men say ". The problems of the Anglo-Saxon Chronicle are too complicated to discuss in detail. It consists of a series of annals, with dates and brief lists of events. At times we can see behind the bald statement of facts traces of ancient poems, whether oral or written we cannot say. In the form in which it has come down to us, it is a compilation of the time of Alfred the Great and has a strong bias towards the history of Wessex, carrying back the ancestors of Alfred and the kings of Wessex to Adam. Thus for some 400 years the early history of Saxon England is based on oral traditions finally committed to writing by a Briton, a Northumbrian and Wessex annalists, none of whom was particularly concerned with Kent. In recent years, fresh light has been thrown on this dark period by archaeology and place-name study but both disciplines have their limitations. Archaeology can deal only with the results of actual discoveries and many problems are at present unsolved. Place-name study is based on the forms of names found in early documents many of which have disappeared whilst others survive only in medieval copies which are less reliable. Neither archaeology nor place-names can 1 I am indebted to our member, Sergeant H. Bostock, R.A.F., for the unselfish devotion of his time, care and skill in re-drawing the distribution maps. 58 PLACE-NAMES AND EARLY SETTLEMENT IN KENT provide a definite dating. They can give us a relative chronology but cannot give an accurate date for the making of a pot or the naming of a place. It has long been recognized that certain place-names are of an ancient type and that their distribution throws light on the earhest areas of settlement. Names ending originally in -ingas, such as Meallingas, now Mailing, were originally names of people, " the followers of Mealla ", and this came later to be used of the place where they lived. Of a similar type are names like Essex and Sussex, really " the East Saxons " and " the South Saxons ", and Norfolk and Suffolk, the north folic and the south folic of East Anglia respectively. But many of the modern place-names ending in -ing are found only in medieval documents and it is difficult to decide whether they were originally singular or plural, whilst the Old English material proves conclusively that we have a number of names originally ending in -ing which were singular and here the formative suffix sometimes denotes a river, a hill or a wood and sometimes appears to mean merely " place "; such names, though often early, throw no light on early settlement. Occasionally, the singular -ing name appears to have given rise to a derivative in -ingas. Stowting is found in 1044 as Stuting, a hill-name from an unrecorded O.E. stut. But in Domesday Book it occurs as Stotinges " the dwellers at Stuting ", or " the dwellers on the hill ". The original name was the name of a natural feature ; the habitation-name developed later but must still have been early, though later than such a name as Mailing. In Kent we have three examples of the survival of the archaic -ge " district, region, province ", in Eastry, Sturry and Lyminge (Fig. 1). This term was so ancient that it is never found in independent use in literature, though it has survived in the German gau. It survives also in Surrey and Ely and must have been used to denote a very early stage in the organization of the Saxon kingdoms. The early obsolescence of the term proves its age, but such provinces could not have been created in the early confusion of the migration. They imply a more settled and more peaceful period and must be later than names in -ingas. Sturry " the Stour province " is first recorded in 605, Lyminge " the province on the Limen " in 697, whilst Eastry " the eastern province " is not mentioned before 788. All three names must be much older and their late occurrence is a warning that many names must have been old and well-established long before the date at which they chance to be mentioned in a surviving document. Occasionally we can fix the earhest or the latest date at which a place can have been named. A few miles south of Milton Regis, an ancient possession of the Icings of Kent and the head of a hundred, lies Wormshill, a hill where once the heathen Woden was worshipped, as 59 PLACE-NAMES AND EARLY SETTLEMENT IN KENT also at Woodnesborough, near Eastry, which, as we have seen, was the centre of a Saxon province and where the king of Kent once had a palace. Wye, south-west of Canterbury, also a royal province and the centre of a lathe, was also once a place of heathen worship, taking name from O.E. wig " an altar ", whilst Alkham, about five miles north-east of Folkestone, owes its name to a heathen temple which once stood near the village. The same element, O.E. ealh, occurs in the lost ealhfleot, the name of a creek flowing into the Swale between Faversham and Graveney, which must have owed its name to a heathen temple on its banks, in or near Faversham, another royal town of great antiquity. Somewhere near Minster in Thanet was once a place called Thunorhlaw, the sole evidence in Kent for the worship of Thunor. All these names, widely distributed in Eastern Kent (Fig. 1), are closely associated with places known to be sites of ancient settlement and early centres of administration. They probably all date from well before 500 A.D. and cannot be later than the introduction of Christianity by Augustine in 597. The fact that four of them survive as parish names proves that they were so old and so firmly established that they successfully defied all the hostility and antipathy of the Church against everything heathen, although their meaning must have been known to peasant and priest alike. On the other hand, no name connected in any way with Christianity can be older than about 600 A.D. Both Minster in Thanet and Minster in Sheppey were names of Saxon monasteries founded in the middle of the seventh century. Names like Eastchurch and Ivychurch must date from the foundation of the particular church. Cheriton means " the farm by the church ". Monkton, recorded in 961, perpetuates a possession of the monies of Canterbury. Much of our knowledge of early Kent place-names comes from Anglo-Saxon charters, the earhest of which, preserved only in a twelfthcentury copy, is dated 604. In one of 605 we have reference to the " villain nomine Sturigao, alio nomine dictam Cistelet". This is of interest and importance for the parishes of Sturry and Chislet are not today conterminous. Sturry, originally the name of the province, came to be used specifically as the name of its chief village. Some time during the sixth century a new settlement grew up, inside the province, but lower down the Great Stour, and took the name of the mother settlement. But the inconvenience of having two places so close together bearing the same name led to the substitution of Chislet for Sturry, a process which had begun before 605, though both alternatives were in use in that year. In another charter dated 738, Stoke is referred to as " Stokes quae antiquitus vocabatur Andscohesham ". No other reference to Andscohesham is known. It is an early habitation-name, '' the homestead of Hondscioh ", a rare personal-name found in Beowulf but occurring in no other English place-name. Stoke, used alone, was a 60 PLACE-NAMES AND EARLY SETTLEMENT IN KENT common term for a secondary settlement. Here, there must originally have been two separate places, Andscohesham the mother-settlement and Stoke the dependent daughter-settlement. For some reason unknown, well before 738, the secondary settlement had become more populous and more important; Andschohesham had decayed, its name had been replaced by Stoke, and but for this solitary reference we should not know that it had ever existed. There is also evidence that Swalecliffe was once called Bi Northanuude " the place to the north of the wood " (it is north of an extensive wood once part of Blean forest) and that a lost Wilmington near Lympne was once known as berdelhames wic. Such evidence for early changes of name is rare but it suggests that some of the many documents which have been lost or destroyed may have contained other examples and that some of the many unidentified places in existing documents may still be represented on the map by quite different names. Of the place-names which from the first denoted habitation-sites, the earhest and most common are those ending in -ham and -tun, as Adisham " Eade's homestead " and Wootton " the farm by the wood ", recorded in 616 and 687 respectively. Side by side with these habitation-names the Saxons must, from the first, have found it necessary and convenient to give names to woods and hills and other natural features and in course of time, as population increased, new settlements were founded and these often took their name from some neighbouring natural feature. Some of them we can recognize for the Anglo-Saxons distinguished such settlements from pure nature-names by prefixing the preposition cet, e.g. " the place called Mt Hoe " (697), whereas Stretleg (697) was the name of a wood. As this preposition was lost early, such names are of little value in discussing early settlement for it is often impossible to decide whether a nature-name denoted a settlement or not. A further difficulty is that by the seventh century when our material begins all these types of names are found. For the chronology and development of our place-names between 450 and 600, the documents fail us. Heathen place-names and names containing archaic terms obsolete before our written literature began are clearly early. So, too, are those containing personal-names found on the Continent but unknown or rare in England. Erith, " the muddy landing-place ", first recorded in 695, is identical with Earith in Huntingdonshire, both places in areas where early settlement is likely. The first element is found in O.E. only as the name of a rune and in the Runic Poem it probably denotes " earth". Birling (Kent), Barling (Essex), Biding (Sussex) and Barlings (Lines), all names in -ingas, all in areas known to have been settled early, all contain a personal-name Bcerla never found in independent use. The first element of Ightham is probably an unrecorded personal-name Ehta, a mutated form of Ohta, found only as the name of 61 PLACE-NAMES AND EARLY SETTLEMENT IN KENT a king of Kent in 645. Lenham is probably'' the homestead of Leana '', a personal-name unknown in England but cognate with O.H.G. Launus. The interpretation of Kent place-names is full of difficulties. For many names no definite etymology can at present be offered and we are reduced to discussion and speculation. Sound generalizations are impossible when so many problems are still unsolved. But here we are concerned with the light thrown on early settlement and both the habitation terms used and their meaning are known so that they can be used for this purpose even though the element with which they are compounded cannot be definitely explained. As we shall see, there are other difficulties. Many early names have not survived. Some we can locate near some known spot or in a particular parish. Many others are still to be discovered. All early names in -ham and tun, for example, whether they survive or not, are of importance in estimating the extent of early Saxon settlement. The distribution maps are not complete or final but will give a broad impression which is probably correct. Doubtful examples are excluded ; unidentified places are included only when they can be located at least within a particular parish. If we are to achieve anything approaching accuracy in the distribution of these early place-names, it is essential that we should be quite certain of the original form of the final element. The modern spelling is no safe guide, in fact it may be definitely misleading. We must distinguish between singular and plural names in -ing, and separate from these those in which the suffix is -ling, from an earlier -linch. Particularly important is it to keep apart names ending in -hamm from those ending in -ham. This is, at times, difficult, even in Saxon charters, for many of them have come down to us only in medieval transcripts. Ham is a habitation-term, indicative of early settlement and was probably obsolete before the days of Edward the Confessor. Hamm may mean " enclosure " or " low-lying land on the banks of a stream or in a marsh ", " a water-meadow ". It does not denote a habitation-site and was in common use after the Conquest. It undoubtedly occurs in a number of Kent place-names, clear examples of which are Ham Green, Ham Street, Bettenham, Frogham, Waterham, Witlesham and some 20 or more others. Some names ending today in -ham are found only with that spelling but are in districts where early settlement was impossible or very unlikely and the situation of these places near streams makes it almost certain that they were water-meadows and derive from hamm. The term was particularly common as a medieval field-name in Romney Marsh and in the marshes along the lower Stour. With names ending in -tun, we have problems of a different kind. A combination of assimilation of consonants and a slurred pronunciation completely conceals the true origin of such names as Baxon, Foxen, Hereson, Lewson and Jesson, all from tun, and gives a false impression 62 PLACE-NAMES AND EARLY SETTLEMENT IN KENT of the real etymology of Nackington, Poppington and Stuppington, all of which were names of hills and ended originally in -dun. On the other hand, Boyden Hill in Chislet is identical in origin with Boyton Court in East Sutton and Boyington Court in Swingfield. A further complication is well illustrated in Lulhngstone. This name occurs in Domesday Book as both Lolingeston and Lolingestone and in the Textus Roffensis as both Lullingestuna and Lullingestana. It is obvious that we have two distinct names, Lullinges-tun " Lulling's farm " and Lullinges-stan " Lulhng's stone", two separate names for two distinct places, both later parishes with separate churches, both named from the same man. As tun regularly becomes -ton or -tone, and as stdn normally becomes -ston, the two names coalesced in form as Lullingstone. Many placenames ending in -stone derive from -tun. Elmstone, Goodnestone and Goldstone were farms named from their former owners iEthelmaer, Godwine and Goldstan, whilst Cuxton and Teston were named from stones. Place-names ending in -ing have long been recognized as ancient and as pointing to the site of the earliest Anglo-Saxon settlements but a false impression has often been given by a failure to distinguish between the different types of such names. We must keep apart names which were originally plural and ended in -ingas and those which were singular, ending in -ing, and this is not always easy. Reading Street in Tenterden and the same name in St. Peters are both from the common term rydding " a clearing ". Hawkinge and Ruckinge are both singular names, " hawk wood" and " rook wood" respectively. In some names the suffix is really -ling, from an earlier Mine, in the sense " hill ": Bramling " bramble hill ", Ratling " hill with rough growth ", Sandling " sand hill ". All these are nature-names and throw no light on early settlement. But names like Halhng, Heallingas " the people of Healla ", Bobbing, Bobbingas " the people of Bobba " and Yalding, Baldingas " the people of Ealda ", were originally names of communities which later became the names of the places where these communities lived. They are among our oldest place-names and are found in the eastern and south-eastern counties which bore the first brunt of the Saxon invasion but this type of name fell out of use early. It gradually becomes less common as we pass west through the Midlands and is not found in the north-west or the south-west. A further proof of the age of these names is that they are based on personal-names rarely or never found in independent use, names difficult to explain, and for parallels we must go to the Continent. The leaders of these bands of invaders brought their names from their ancient homes beyond the seas and within a generation or two their descendants ceased to use them. But for their accidental survival in early place-names we should not have known of their existence. 63 PLACE-NAMES AND EARLY SETTLEMENT IN KENT Some 25 of these names of communities still survive on the map of Kent (Fig. 1), some near the Medway, but more in eastern Kent. There are none in the marshy districts near the Thames and its estuary or in Romney Marsh or the Weald. These early communities settled on the more attractive sites. There is less evidence in Kent than in some other counties of the way in which these early communities established dependent settlements, Basingstoke from Basing, Navestock from Nazeing, Wokingham from Woking and Ellingham from Eling. These new settlements were still community settlements but their names were from the first names of places, in which the name of the mother-settlement was combined with some term denoting habitation, ) \ \ VA \ \ i. \ MILES '-ge 5 • - ingas « \ < " l KILOMETRES x heathen names FIG. 1. Distribution of the earliest OE place-names of Kent. Names of provinces, community-settlements and sites of heathen worship. " the stock or dependent settlement of the Basingas ", " the village of the Woccingas ", etc. Occasionally we get a ghmpse of a similar extension of settlement in Kent. Eastling about four miles south-west of Faversham, is first recorded in Domesday Book, where we find both Eslinges and Nordeslinge, two distinct vills. But already in the middle of the eighth century we have mention of Mslingaham, Islingham in Frindsbury, some distance away, north of the Medway. There can be little doubt that this was " the village of the Eslingas ", a dependent settlement founded by the men of Eastling. Eslingas means " the people of Esla ", a personal-name found only in the West Saxon genealogies of Icings and related ultimately, perhaps, to O.E. os " god, divinity ". 64 PLACE-NAMES AND EARLY SETTLEMENT IN KENT Just south-west of Eastling is Otterden, Otringedene in Domesday Book, a form recalling the Domesday spelling of Wateringbury, Otringeberge. A similar form occurs in an eleventh-century Anglo- Saxon charter, but the tenth-century charter forms all begin with W, Wothringaberan, etc. The second element varies considerably and it is impossible to be certain whether the reference was originally to a fort, a hill, a cowhouse or a swine-pasture ; the latter is, perhaps, most likely ; the first element is undoubtedly the name of a community, explained by Elcwall as that of the Ohtheringas, the people of Ohthere, •\. \ \<\ i \ MILE.S £ * I* • -ham KILOMETRES • -ingaham FIG. 2. Distribution of place-names in -ham and -ingaham. the development of the initial W being a feature of the old Kentish dialect. The Domesday forms Otringedene and Otringeberge suggest that the first elements of both Otterden and Wateringbury are identical. Who were these Ohtheringas 1 We cannot say with certainty but it is possible they were the people who gave name to Ottridge in Bearsted. This place has been noticed only once, in 1285, in the form Wotring. It is rash to draw deductions from such slight evidence, but we may hazard the suggestion that, in the light of the early forms of Otterden and Wateringbury, this may have been originally Ohtheringas, the people of Ohthere, and that they had swinepastures to the east at Otterden and to the west at Wateringbury, and that these developed into independent settlements and ultimately became parishes. The mother-settlement at Ottridge must have decreased in population and importance, perhaps because of the growth of Maidstone where the Medway, still navigable, was crossed by 65 10 PLACE-NAMES AND EARLY SETTLEMENT IN-KENT a bridge. It is even possible that Maidstone was originally within the territory of the Ohtheringas and that the name does, after all, mean " the people's stone ". Kent has some 10 place-names which ended originally in -ingaham, including Farningham, Gillingham, Rainham, Wingham and Mongeham (Fig. 2). This combination of ham with the genitive plural of a folkname proves the antiquity of ham as a habitation-term, a conclusion confirmed by its combination with personal-names not found in independent use in such names as Beckenham, Meopham, Teynham and others. Further proof of the age of names in -ham is the frequency of Higham " the high homestead or village ", found at least six times in Kent and in some nine eastern counties. It is a formation of an early Old English type, heah-hdm, where the adjective is joined to the noun in its uninflected form. When compounded with terms like dun, leak and tun, it appears in the normal weak form, in such names as Hendon, Henley, Hanley, Henton, Hampton, etc. Twice, in Essex and in Suffolk, we find Henham " high homestead ", the weak form of Higham, whilst when compounded with new it is always found in the weak form. Such names as Newnham and Nuneham, " new homestead " or " new village ", could not have been so called at the time of the earliest settlements. Thus we get the chronological sequence, -ingas, ham, tun. But there was overlapping, for ham is found compounded with -ingas and was still in use in the weak form after tun had come into use, whilst tun itself, as we shall see, could be compounded with -ingas. Ham had ceased to be a formative suffix before the Norman Conquest ; it is never found in medieval field-names, whilst tun continued in use long after the Old English period. Of the modern Kentish names in -ham, some 30, at least, derive not from ham, " homestead, village " but from hamm " enclosure ", " watermeadow " and these must be carefully excluded in discussing the distribution of early settlements. Some of these Aamm-names are old and are now parishes, as Cobham, Harrietsham, Petham and Wittersham. Names in -ham number about 60 (Fig. 2), whilst two or three others have disappeared. They are more widely distributed than the -ingas names, especially in the west, along the Medway and between that river and the Darenth. They occur, too, in Central Kent and along and to the east of the Stour. There is none in the Weald. More common than any of the elements so far discussed is O.E. tun which meant originally " a fence ", then that which was enclosed by the fence, " an enclosure ". Quite early, it came to be used of "an enclosure with a dwelling ", " a farmstead ", and later of the village which grew up around the farm. Still later came the meaning " estate, manor ". It is thus not always possible to decide the exact meaning of the word when found as a place-name. The common Garston was a 66 PLACE-NAMES AND EARLY SETTLEMENT IN KENT grass-enclosure, a paddock ; O.E. osppeltun was used of an orchard, burhtun, now Burton, of a fortified or stockaded enclosure, hamtun, the modern Hampton, of the enclosure in which a homestead stood. Beretun, literally " barley-enclosure ", came to be used of a corn-farm and later of an outlying grange as in Wingham and Wingham Barton. Many compounds of tun are now names of villages but very many of these must have originated as farms, especially those compounded with a personal-name, as Elmstone " iEthelmaer's farm", Goodnestone " Godwine's farm ", Lewsome " the farm of Leofric ", Guston " the farm of Guthsige ", and numerous others. • j •\ \ M \ \ MILES is K LOMETRES fun FIG. 3. Kent place-names in -tun. The period during which tun was a living element was very long, stretching throughout the whole of the O.E. period until at least the thirteenth century. There are undoubtedly a few names in which tun, like ham, was compounded with the genitive plural of a folk-name and denoted an early dependent settlement ; Winterton, e.g., earlier Wintringatun, was a dependent settlement of Winteringham. But very many of the -ington-n&mes show no sign of the genitive plural ; some occur also in the genitive singular, whilst others are compounded with women's names and women would hardly have been leaders of the communities which estabhshed the early folk-names in the unsettled period of the migration. Elcwall explains all these ington-na,mes as " the tun of so and so's people ", but it seems safer to reserve this explanation for names with clear evidence of the genitive plural. The personal-names compounded with tun are of a less archaic type than those compounded with hdm, names found in common use, of the 67 PLACE-NAMES AND EARLY SETTLEMENT IN KENT compound type like Goldstan in Goldstone and Wighelm in Wilmington, which replaced such short names as Offa in Offham and Wulda in Wouldham. Compounds with women's names are found, too, and these could have arisen only in later, more settled and peaceful times : Bilsington " the farm of Bilswith", Chilverton, that of Ceolwaru, Garrington, Liverton and Ringleton, the farms respectively of women named Garwynn, Leofflozd and Hringwynn. Blackmanstone was named from the Blcecmann who held the manor in the reign of Edward the Confessor, whilst a few place-names preserve the name of a post- Conquest lord of the manor. Filston Hall in Shoreham occurs as Vielston in 1203 and contains the French name Viel, from Latin Vitalis. The place is named from the family of Simon Vele of Shoreham who is mentioned in 1292 but the family must have been here a century earlier. Jesson Farm in St. Mary-in-the-Marsh is Geffreyeston in 1254 and derives from some earlier owner bearing the French name of Geoffrey. Vincent Farm in Garlinge is Vincheston and Fincheston in the thirteenth century and is named from the family of Adam le Vinch who is mentioned in the same document. In addition to the tam-names on the map there is a considerable number of similar names which have disappeared and more will probably be discovered as the Survey of Kent Place-names proceeds. There was a Ravenstune in Thanet in 1312 which suggests a place near Ramsgate named from the man who gave his name to the way down to the sea. Others of the thirteenth century, are compounded with O.E. masculine and feminine names, one with a medieval occupation-name, Bakerestun. Particularly interesting is the large number of compounds of tun found in Bilsington and Romney Marsh in a fifteenth-century survey, some obviously late, such as Halletone, named from Symon Hall, Prowdyston, Tholytounys, held by John Tholy, and one, Jon Stefenestoun, consisting of only half-an-acre and four perches, held by John Stefene of Brookland. It is clear that in this district tun was in living use as late as the fourteenth and possibly the fifteenth century for small parcels of land often named from the tenant. The distribution map of the tuns (Fig. 3) shows a marked increase of settlement in the centre and east of the county. Its most noteworthy feature is the signs of settlement in Thanet (which, curiously enough, contains no early place-names) and in Romney Marsh. The Weald is still blank. In 893 the Weald is named and defined in the Anglo-Saxon Chronicle as " the great wood which we call Andred "; it is said to be 120 or more miles long and 30 miles broad and in the same entry is described as a weald or forest. It stretched from the New Forest of Hampshire to Lympne in Kent. Its thick woodland and heavy clay soil were unattractive to early settlers but the eighth-century charters reveal that it was abeady being used for keeping swine and pasturing sheep 68 PLACE-NAMES AND EARLY SETTLEMENT IN KENT and also as a source of timber. It would appear that in Kent, in the earliest Saxon period, this forest-land was common to the ancient regiones and, later, to their successors, the lathes. We have references to the Limenwearawalda " the forest of the dwellers on the Limen ", the ancient regio of Lyminge, and to the Cozstruuarouualth or the Ccestersozta walda " forest of the men of Rochester ", which certainly included part of Brenchley, Lamberhurst and Hadlow. But, by the time our records begin, there had been a change. In the eighth century we find the ) \ \ VA \ \ MLES \ t i* \t KILOMETRES denn FIG. 4. Distribution of place-names in -denn. woodland of Kent divided into two classes, the silva regis and the silva communis, according as it was reserved to the principal demesnes of the king or open to the common use of the countryside and these royal and common woods were divided among the villae regales and the lathes. In the ninth century, it became usual for the king, in granting land, to allot certain denns or swine-pastures as appurtenant to the estate. These are named but then position is only vaguely described as " in Andred " or in the weald and it is frequently impossible to locate them exactly. These swine-pastures are commonly called denns, sometimes denbmra or wealdbozra " woodland swine-pastures", Latinized as pascua porcorum. Denn is a common term in place-names and must be carefully distinguished from denu " a valley ". It is chiefly found in the woodland areas of Kent (Fig. 4) and Sussex and also in Surrey and probably meant originally " a forest, a haunt of wild-beasts ". It is usually, but not invariably, used of swine-pastures, but such names as 69 PLACE-NAMES AND EARLY SETTLEMENT IN KENT Cowden, Lambden and Oxenden point clearly to the pasturing of cows, lambs and oxen, whilst Hinksden was where the hengest or stallion was kept and Marden was a " pasture for mares ". The term is often compounded with names of trees : Hazelden, Hollanden " holly ", Iden " yew tree ", Oakenden, Riseden " brushwood " and Haysden " brushwood ". Moorden was marshy, whilst the nature of the wealden soil is commemorated in five compounds of O.E. scearn " dung, filth, mud ": Shardens Farm, Sherenden, Shirrenden and Shernden (twice). Smarden derives from O.E. smeoru " fat, grease, butter " and probably denoted a pasture where there was rich grazing. Compounds of denn with a personal-name are not uncommon : Bethersden " the swine-pasture of Beaduric ", High Halden, that of Heathuweald and Rolvenden, that of Hrothwulf. In the charters we have a number of names of denns compounded with a personal-name plus -ing, as Billincgden, Wilfrethincgden, Hildgaringdenn, with others first recorded in medieval documents, but probably much older. There is no possibility that such names can contain a genitive plural -inga-. They must be parallel to names like Suithhunincg lond and Seleberhting lond, the names of two sulungs at Graveney granted to the Archbishop of Canterbury between 811 and 814, which must denote the land or sulung of peasants named Swithhun and Selebeorht respectively. Curiously enough, the latter personal-name is the base of Silverden in Sandhurst, probably the meeting-place of Selbritten Hundred, " Selebeorht's swine-pasture ". It is unlikely that these denns should preserve the name of a swineherd. They were appurtenant to a distant estate and, were our records complete, we should probably find that Mlhfrethincden, a swine-pasture of Warehorne in 830, was paralleled by Mlhfrethinclond, both named from their owner, one Ealhfrith of Warehorne. Not all swine-pastures have names ending in -denn. Many of them are named from some natural feature and are compounded with elements common in woodland districts, Petteridge " the ridge of Psetla", Lindridge " lime-tree ridge"; Ashour "ash-tree bank"; Hever " the high bank "; Tapners " Taeppa's brushwood "; O.E. hyrst " wooded hill" is common as in Maplehurst, Hawkhurst, Ewehurst "boar-wood"; leah "wood" occurs in Pluckley, Brenchley, and Friezley ; Lossenham and Broxham contain O.E. hamm " enclosure ", Hawkenbury and Bedgebury O.E. beer " woodland pasture ", " swinepasture ". The first element of Speldhurst and Spilshill is O.E. speld " splinter, piece of wood ", Speldhurst being a wood from which pieces of wood or timber could be got and Spilshill a wooden shed. The denns were frequently at a considerable distance from the estate to which they were appurtenant, those of Bromley around Edenbridge and those of Bexley near Hever. Tenterden was a swine- 70 PLACE-NAMES AND EARLY SETTLEMENT IN KENT pasture of the men of Thanet. Lenham had 13 denns in Cranbrook, Frittenden and Staplehurst, Milton Regis 42 in the hundred of Marden, whilst the woodland pastures of Wye lay in the parishes of Hawkhurst, Cranbrook, Biddenden, High Halden, Woodchurch, Smarden, Pluckley and Bethersden.1 With such distances between estates and then dependent denns, the swineherds would tend gradually to form small communities of then own. They would need food and shelter for themselves ; they would cut timber and build sheds and huts, in the small clearings they made they would begin to dig and sow and to grow food to improve and vary then monotonous diet. And they would be faced with the problem of driving their pigs and conveying then timber to the distant villages. A grant of king Offa of land in Ickham, Palmstead and Ruckinge included the rights in three denns in Andred, the pasture of one drove of swine at Denge Wood, and wood for building, burning and salt-making in the forests of Buckholt and Blean, together with the right of way for two carts carrying 100 loads. Lenham had the right of passage for three waggons in the wood called Blean and Little Mongeham, in addition to the right of feeding pigs, herds and cattle, and licence for one cart in the wood at Singledge. The swine were driven from the denns to then parent villages along the droveways which must, in the main, have followed the higher ridges, hence, perhaps, the names of denns ending in -ridge. Sometimes boats were used, probably, for example, from Tenterden, with its harbour at Smallhythe, to Thanet. In the thirteenth century the tenants of sulungs at Northbourne had to find a ship to carry timber from the Weald to the granges as far as Greistonehende and thence to carry it in their carts to the court without breakfast. The cottars had to carry only from the sea. Our distribution maps provide no evidence of early settlement in the Weald but a growing population would tend to increase the number of swineherds and woodcutters with a consequent increase of small clearings and buildings, preparing the way for some simple farming. We get some hints of this from both charters and place-names. Denns began to be divided—or else new denns were created and named from one near by. In 850, e.g., Lenham had two denns named Maplehurst and three named Babbingden. As early as 762 denns could change hands, for in that year St. Augustine had confirmation of an exchange by which Wye was to have the use of half a mill at Chart in return for rights of pasture in the wood of Andred. Clear indication of wood- 1 Fig. 5 is a tentative and incomplete attempt to show the general distribution of the swine-pastures of certain upland manors. Many denns cannot be definitely located. In addition to those mentioned in the text, the map includes the denns of Lewisham (in Cowden), Hailing (Speldhurst and Rusthall), Wouldham (Hever), Snodland (Whetsted, Lamberhurst and Pembury), Chart Sutton (Sutton Valence, Headoorn and Staplehurst), and indicates those of Westwell, Ickham, Eastry and Northbourne. Others could be added, but this would destroy the purpose of the map—to indicate distances and distribution. 71 PLACE-NAMES AND EARLY SETTLEMENT IN KENT cutting is provided by the names of Staplehurst, Yardhurst and Bredhurst, woods where staples or posts, poles or spars, and boards were got. Spunley was a wood where wooden shingles were made. O.E. (ge)sell meant" shelter for animals " or " herdsmen's huts " and is found in the names of a number of lost swine-pastures, often in the plural. It survives in Bowsell " hut by the beech-tree ", Hensill " high huts " and Nizels " the new huts ". Domesday Book throws little light on the development of settlement in the Weald. Only 48 denns are mentioned—a very small proportion of the whole—and these not by name. They are included in the parent manor, usually with a statement of the number of swine paid as pannage. Dartford had three large denns of woodland, Milton Regis woodland to j \ \ \ \ MILLEES 'j f«*2 IO KILOMETRES FIG. 5. Some early Kent settlements and their swine-pastures in the Weald. render 220 swine, Wye 300. Identified Saxon denns are clearly indicated by the render of 100 swine at Bromley, 100 at Bexley and 150 at Hailing. But when we find that Orpington received 50 swine from five denns and that the pannage at Elham was 100, at Barham 150, and at Wrotham no less than 550, it is clear that there are serious gaps in our knowledge of Saxon Kent. When we find that Lenham, with its 13 denns in 850, in 1086 received only a total of 73 swine as pannage, we must suppose that the number of swine had decreased as agriculture increased, a conclusion confirmed when we learn that more than one third of these denns were in Cranbrook which, by 1086, had become a parish with a church. There was a church, too, at Benenden, assessed in 1086 at half a sulung, with three ploughs, four villeins and nine 72 PLACE-NAMES AND EARLY SETTLEMENT IN KENT borders. Other indications of agricultural settlement are found at Pluckley, with 13£ ploughs, 16 villeins and seven borders, and at Newenden, assessed at one sulung, with 25 villeins, four borders and five ploughs. Rolvenden had a denn of half a yoke, with two villeins and half a plough and had abeady become a hundred by 1086. From the Domesday Monachorum we find that in the Weald there were churches at Biddenden, Bethersden, Cranbrook, High Halden, Marden, Pluckley, Rolvenden and Smarden. In spite of the almost complete absence of signs of settlement in the Weald on the map of the Domesday vills (Fig. 6), it is clear that a gradual change had long been talcing place. i ) t\t \ \A \ \ MIIL LEESS i f — i b — i s Lu,9 i.'P.-if 2P KILOMETRES Domesday vills FIG. 6. The Domesday vills of Kent. The place-names of the Weald are nature-names, a mass of names of swine-pastures, woods, hills and brooks. The only names pointing to settlement are those of a few still insignificant places where herdsmen had their sheds and shelters. Many of the swine-pastures became centres of small settlements but kept their nature-names and by 1086 some 12 of them had become parishes with a church and one the centre of a hundred. The settlement of the Weald did not cease with the Conquest but, as this can scarcely be called early, we cannot pursue the subject here. From medieval documents we learn much of the customs of the denns ; many of them were clearly ancient and some, at least, are probably pre-Conquest. The medieval sumerhus silver was a rent paid in place of building a house for the lord on his summer visit to the denn for hunting or general supervision. By a charter of William I the 73 PLACE-NAMES AND EARLY SETTLEMENT IN KENT men of the Weald owed a sumerhus to Bishop Odo or paid 20s. in lieu. St. Augustine had a chief permanent house at Wespherhawk in Smarden and others in temporary use at the time of pannage. Tenants had to attend the court of the denn called a " parrock ", a court-like kind of meeting not unlike the forest swainmote, where account was taken of what hogs or swine had been taken in to feed and fatten in the previous year. Such a court was probably held at Westpherhawk, whilst Paddock Wood probably owes its name to the court of the denn of the men of Rochester who had three denns in this area in 747. The hght thrown by place-names on the early settlement of Kent is thus selective and incomplete but, though we cannot give precise dates, we do get a general impression which is intelligible and probably accurate. We still find on the map the names of three of the early provinces of East Kent ; four place-names mark sites of heathen worship. The earhest settlements were made by communities of which some 25 can be identified with certainty, as at Cooling and Postling (Fig. 1), whilst nine villages preserve the names of slightly later communities who settled at Gillingham, Terlingham, etc. Names in -ham number 62, all certainly pre-Conquest (Fig. 2), whilst those in -tun are considerably more numerous (185), some of them post-Conquest (Fig. 3). These names do not form a chronological sequence as there was undoubtedly overlapping. Some hams may be contemporaneous with or earlier than some names in -ingas and some tuns may be earlier than some hams. The distribution maps show clearly the steadily increasing number of the successive types and then distribution. The place-names of the Weald provide a striking contrast, providing no evidence of early settlement. They are, almost without exception, nature-names. The modern map contains 132 names in -denn, for all of which there is early evidence (Fig. 4). To these should be added many pre-Conquest names now lost with many modern names so far undocumented. In addition, we have countless names in -hurst, -ley, -hill and -ridge, all names of natural features, characteristic of a high, thickly wooded district. This survey is not and never will be complete. Places like Sarre and Bishopsbourne are known to have been occupied in the eighth century, but as to the date when such natural sites came to be centres of habitation the documents are silent. 74