A possible Saxon canal around Woodmill and Gater's Mills in the historic manor of South Stoneham?

Introduction

Fieldwork connected with the local Saxon charter bounds has located the remains of a very substantial water channel, up to 1.5 kilometres long and 15-20m wide, following the southern boundary of Riverside Park on the outskirts of Southampton. These are very overgrown in places, and have been obscured on the edge of woodland called Marlhill Copse at their eastern end near Gater's Mill. They relate closely to the conjectured position of a feature called the 'new river' on the Saxon charter of 1045 for South Stoneham (Grundy 1927, 249). This suggests the possibility that this channel may be the remains of the feature mentioned in the Saxon charter. This essay discusses this possibility, and looks at the history of the river in the vicinity of Woodmill and Gater's Mills.

The charter evidence

A full analysis of the bounds of the charters relevant to this study is given in Currie (1993, forthcoming), to which readers are referred. This present work follows this study, but restricts itself to discussing only the passages in the charters directly related to the presence of a 'new river'.

The earliest charter for South Stoneham dates from 990 x 992, and records a grant of land to an unnamed party by King Ethelred (Sawyer 942, Kemble 712). The bounds of this estate appear to cover roughly the same area as the later charter of 1045. It is this later charter that mentions the 'new river'. This is not given in the earlier charter. Instead the bounds start on the Itchen, and move along the king's boundary to the Bitch's Pole. From here they move on to 'Wadda's Stake'. On the charter for 1045 the same apparent land is granted by King Edward to the Old Minster at Winchester (Sawyer 1012, Kemble 776). The bounds here start at Swaythling, and probably move down the contemporary equivalent of the Mansbridge Road to the Itchen. The first mention of the river in 1045 refers to the 'Old Itchen' (Ealden Icenan). From here, it moves along the top of an orchard to the 'New River' (Niwan Ea), then along the boundary to the 'claypits' (Lampyttas), and along the boundary again until it comes to 'Wadda's Stake'. The 'boundary' referred to is probably the undefined 'king's boundary' of the first charter, showing that the boundary itself does not appear to have changed, but that three extra points have sprung up between the original 'Itchen' and 'Wadda's Stake'.

Although it is possible that the second charter is elaborating on the first by giving extra bounds, it is more probable that the additions have been made because the landscape between the original points has changed. The local topography is such that it is unlikely that a change in the boundary is the cause of the additional points, as any boundary following the approximate line of these charters would have to cross the Itchen valley. Therefore it is a more plausible solution to suggest that the 'new river' may not have existed in 990 x 992, but had come into being by 1045, rather than the boundary has changed.

The same argument can be made for the appearance of the 'claypits' on the second charter, and the disappearance of the 'Bitch's Pole' mentioned on the first. The likelihood is that the clay pits have been dug in relation to the making of the new river to provide clay to seal the bottom, or some other functional task. In digging them, the feature, the Bitch's Pole' was probably removed. It is noteworthy that the copse to the immediate south (the direction in which the charter bounds are moving) of the recently discovered channel is known as Marlhill Copse; 'marl' being a term used for earth dug out of the ground as a fertiliser. This name was probably given by later generations to explain the existence of pits in the area, and they may have subsequently been used for agricultural purposes.

The 1045 charter gives a list of other features after the bounds. These probably belong to the estate of South Stoneham, but for reasons not given fall outside of the bounds. Three of these features appear to be associated with the Itchen. They are 'an eyot at Port's bridge' (se iggath aet Portes Bricge), 'half a sea weir' (healfe saewaere), and 'the millstead at Mansbridge' (se mylnstede aet Mannaes Bricge). Both the terms 'Port's bridge' and 'Mansbridge' suggest bridges over the river at these points. The latter still exists today as an eighteenth-century stone bridge over which the Mansbridge Road passes, and was probably in existence by the tenth century at the latest (Currie 1993, forthcoming).

The 'eyot' at Port's bridge is probably the island, shown in the tidal portion of the river on the 1810 one inch Ordnance Survey map, opposite to the site of Roman Bitterne. 'Half the sea weir' is problematic. Initially this seems to be a reference to the well-recorded fishery at Woodmill, at the head of the tidal portion of the Itchen. Reference to two fisheries in South Stoneham in the Domesday Survey has been given as evidence for the probable existence of this fishery by 1066 (Grundy 1908, 481). However, a curious reference in a mid-fifteenth-century Inspeximus to 'half a weir and half a crossing over the Itchen...' at Bitterne (Greatrex 1978, 80) hints that a fish weir may have existed between the island and the mainland at Bitterne that may have been on the site of the 'seaweir' of the 1045 charter.

The identification of the 'mill at Mansbridge' is not clear cut either. The initial impression that Mansbridge mill is Gater's Mill (as this is nearer to the modern Mansbridge than Woodmill) is called into question by the mention of a mill at 'North Mansbridge' on a charter for North Stoneham dated 932 (Sawyer 418, Birch 692) in which King Athelstan grants the estate to a man named Alfred.

This mill is recorded after the bounds, and is almost certainly Gater's Mill. The boundary for this estate comes very close to this mill site, but does not include it. It is thought that the lands given after the bounds on this charter are those between the boundary for South Stoneham given in 1045 (approximately the modern Mansbridge Road) and the 1810 parish boundary for North Stoneham (Currie 1993, forthcoming). Therefore, like the lands given after the bounds of the 1045 South Stoneham charter, these are lands that come with the estate, but fall outside of the bounds. Currie (ibid.) argues that this suggests that Gater's Mill was the mill for the joint estate of North and South Stoneham before it was divided, probably in the Middle Saxon period. This suggests that Gater's Mill is an earlier mill than that at Woodmill. It is possible that the mill mentioned at 'Mansbridge' on the 1045 charter is a more recently built mill than that at 'North Mansbridge', and might be equated with that at Woodmill. Alternatively, the descriptive 'North' has merely been dropped, and both charters refer to the same mill. As will be seen below, the inability to resolve this question causes problems when trying to explain both the reasons for the creation of the 'new river' and its later history.

The present remains

As late as 1940 the 25 inch Ordnance Survey map (sheet LXV.3) showed a substantial channel heading ESE from Woodmill to the southern corner of Riverside Park. This channel was parallel to the substantial levee bank that then followed the course of the river up to its upper tidal limits. Fragments of further banks also survived further north along the east bank of the tributary stream, Monk's Brook. The latter is considered to have been a possible Saxon creation to prevent the flooding of a conjectured open field associated with the village of South Stoneham (see Appendix). The latter has recently been excavated ahead of development near the Montefiore Halls of Residence (Crockett forthcoming).

The levee banks along the Itchen are probably of some antiquity, and were constructed to prevent flooding. The fields of the lower Itchen have probably always been highly prone to this, especially when high spring tides coincide with periods of heavy flow. This has remained a problem in the area until recent memory. That the levee banks should have reached some 250m beyond Woodmill to the junction of Woodmill Lane with Manor Road attests to the former size of the channel it followed.

This channel is thought to be part of the remains of the 'new river' mentioned in 1045. As late as the nineteenth century it was equal in width to the main river. It has been gradually backfilled over the period 1940-75. It is still marked today by a substantial hollow along the edge of Woodmill Lane, with a bank to the south, up to 1.2m high, representing the former levee. Near its junction with Manor Road, Woodmill Lane clearly cuts across the line of the channel, indicating that its present course is a more recent feature. This lane has been in existence since at least 1810, and it has all the appearances of having been an old routeway from South Stoneham to Bitterne, and may have Roman origins. Its present course seems to date from the period after the 'new river' had fallen into disuse.

On the east side of Woodmill Lane (SU 4421 1505), the old channel is continued by a a deep cutting 15-20m wide. At the bottom of this cutting is a small stream, representing the local catchment of water flowing off Town Hill. The channel has been clearly utilised in the post-medieval drainage of the area, and may have been recut. This continues until crossed by a bridge leading into Riverside Park at SU 4456 1529. Stonework in the side of the bank adjacent to this structure has tool markings characteristic of c. 1840 on it (Bob Thompson pers. comm.).

ontinuing NE from this point the channel is no longer contained within a cutting as deep as previously. Instead the channel becomes increasing overgrown and stagnant as it follows the base of the slope of the steep-sided hill along the edge of Marlhill Copse. The channel here is about 15-20m wide depending on the extent of silting and other natural factors. There are many fallen trees and alder and willow scrub in the channel as far as SU 4576 1545, about 750m from the above mentioned bridge, when the channel turns north towards Gater's Mill, leaving the edge of Marlhill Copse.

At a point approximately SU 4475 1542 on the south side of the channel is an earthwork bank, up to 1.5m high and about 10m across its base. Cut into the hillside between the bank and the hill is a ditch, up to 1.5m deep. This is about 100m in length, although its extent has not been accurately measured. It enters the main channel at by cutting across the bank at right angles. It follows parallel to the main channel, and terminates abruptly in a dead end. There is currently no drainage flowing into it from the hill, and no immediate explanation for its existence. It may be contemporary with the main channel, or a subsequent feature, possibly of an ornamental nature. However, the present evidence is so scanty that no serious hypothesis can be put forward until further work is done to try to explain its purpose.

After the channel has left Riverside Park to continue eastwards through a scrubby piece of former meadow, the remnants of an old hedgeline follows the line of the channel on its north side. This stands on a very degenerate bank, but there is no trace of a ditch. This hedge appears to stand some five metres or so north of the conjectured line of the north bank of the channel, as if leaving a deliberate gap between itself and the channel. The present footpath does not respect the hedgeline, and cuts across it on a number of occasions. The channel itself is probably in its best condition in this stretch, is over 20m wide in places, and apparently little disturbed by later alterations.

Just before the main channel reaches the main Mansbridge Road near Gater's Mill, it narrows suddenly, and is crossed by a trackway crossing a concrete pipe (SU 4524 1552). This feature seems to be the terminal of that part of the channel that is marked on the Tithe Map for South Stoneham as a 'Lake' of about 3.2 acres, and owned by Edward Gater, the lessee of Gater's Mill. The terminology suggests that this was a recent adaptation of the earlier channel as an ornamental feature or fishpond. In the case of nearly all older ponds and lakes the term 'pond' or 'fishpond' is used; the usage of 'lake' as a body of still water only became common from the eighteenth century.

Although it is possible that the channel is a late post-medieval linear pond, this goes against the evidence of the earlier 1810 one inch Ordnance Survey which shows it as a substantial water course at least equivalent in width to the main river, running continuously from Gater's Mill to Woodmill. On the Tithe Map only that part between the bridge leading into Riverside Park and the concrete crossing at SU 4524 1552 is marked as the 'Lake'. That this alteration had occurred after 1810 is supported by Bob Thompson's observations about the stonework revetting the bank at the aforementioned bridge.

After the crossing of trackway the channel degenerates into a shallow water filled pool. Although it is still shown as reasonably substantial ditch when the Ordnance Survey last mapped this area in 1967 (25 inch edition, plan SU 4515), it has been largely filled in today, and is not traceable where the channel on the 1810 map meets Mansbridge Road. However, a substantial diversion still exists on the north side of this road, taking the main River Itchen around Gater's Mill.

Saxon river engineering

Recent work at Glastonbury, Somerset, has uncovered evidence for a Saxon 'canal' there dating from the tenth century, possibly used for transporting stone and other materials to the site of the abbey (Hollinrake and Hollinrake 1991). From the number of mills recorded in Domesday, it is clear that the Saxons must have undertaken a large amount of river engineering before the Norman Conquest. At nearby Titchfield there is a long artificial leat, nearly 800m long, feeding the existing village mill. If this mill can be assumed to be on the site of that mentioned in Domesday, then this artificial watercourse is one of a number of suspected Saxon date which suggests that they were involved with major river alterations. This suggestion is supported by a reference in a charter of 948 (Sawyer 1968, 535) for the estate of Segensworth (later in Titchfield) which refers to a meadow which lies between 'the Meon and the mill ditch' (Hare 1992, 119).

Mills do not necessarily need such leats on larger streams and rivers as a matter of course. They are usually constructed because the siting of a mill across the main stream is a major obstacle to access up and down stream. On larger rivers, this would include access for boats undertaking local trade, but even on smaller rivers the blockage of the main stream was a frequent cause of litigation, as it prevented salmon and other migratory fish access to the upper reaches to spawn. Salmon have always been important in the economy of any river system, and rights to fish were jealously guarded. The provision of diversion leats to prevent mills blocking rivers was therefore important, and it can be assumed that the Saxons would have constructed them.

It is with this in mind that the mention of the 'old' and 'new' river in the vicinity of Woodmill and Gater's Mill on the Saxon charter for South Stoneham for 1045 is significant. This appears to suggest that river engineering associated with both mills, almost 1.5 kilometres apart, had been undertaken by this date. The lack of mention of the 'new river' in the bounds of an earlier charter of 990 x 992 seems to suggest that the work may have been undertaken between those dates and 1045. Biddle and Keene (1976, 270) quote the 1045 charter as evidence for alterations to the Itchen in the early medieval period, and note that although this may have been carried out to facilitate navigation, there is no further mention of such possibilities until the episcopate of Godfrey de Lucy (1189-1204).

Although Roberts (1985) has shown that the conjectured medieval canal built by de Lucy to Alresford was an antiquarian myth, it would seem that boat traffic could pass beyond Gater's Mill in the eleventh century. The 1810 map seems to show that a channel once existed that would have once enabled access to the upper river by-passing Woodmill and Gater's Mill (the latter the 'Allington or Up Mill' of medieval documents). The remains of this channel show that, where it has survived unmodified, it was at least 15m wide and 1.5 kilometres long, and would have allowed barges to pass each other without trouble. However, only excavated evidence could prove its age beyond question, it appears that this feature is artificial, and may be of Saxon date.

In a recent synthesis of the reasons for the decline of Hamwic, Morton (1992, 75) has suggested that many of its functions had migrated upstream to Winchester by the early tenth century at the latest. If the construction of the artificial river was to facilitate the moving of supplies into that town, it might be expected that it would have been undertaken by that date. The evidence, however, suggests that the work was carried out between 992 and 1045, and there is no clear evidence that boats could have reached Winchester until the building of the Itchen Navigation in the post-medieval period.

That the 'new river' was designed, at least partly, to take boat traffic around obstructions in the river seems highly likely, but the destination of that traffic must remain conjectural. A mill already existed at 'North Mansbridge' in 932, probably on the site of Gater's Mill. By 1086 there are two mills and two fisheries in the two Stonehams (Munby 1982, 3-16, 6-8), all of which would have probably caused obstructions on the Itchen. A mill and a 'sea weir' are first mentioned at South Stoneham in 1045, but they are not mentioned in 992. Could the need for the artificial river have been the building of a substantial new mill, with an important fishery, at Woodmill between 992 and 1045? Or was it simply that the problem had existed for much longer, but it had been tolerated until some unknown factor came into play, forcing the hand of the authorities to carry out what would have been a very substantial undertaking?

The later history of the Itchen, with reference to possible navigation

A scan of some of the more obvious sources, both in the Hampshire Record Office and elsewhere, has failed to find any certain mention of this conjectured 'new river' after 1045. This need not be surprising because written records for the area do not generally resume until the early thirteenth century. With the passage of 150 years it is quite likely that the local circumstances had changed, and the channel had fallen into disuse. Certainly the national prominence of Winchester had declined considerably between the mid-eleventh and the thirteenth century when it had been replaced by London as the administrative centre of the realm, and the passage of boats in the former's direction may have become less urgent.

Roberts (1985) has examined the tradition of a canal or navigation for boat traffic along the Itchen in some detail. It would seem that the tradition for Bishop de Lucy constructing a canal from Southampton to Alresford c. 1200 can be dismissed as antiquarian fiction (Roberts 1985, 135). However, the argument against possible navigation as far as Winchester is not so clear cut. Records show that it was considered to be a desirable object from at least the thirteenth century. An inquisition was called in 1275 to examine this possibility. This concluded that a number of mills would need to be removed to allow passage. The Victoria County History argues that this implies that the jurors were attempting to improve 'an existing canal' (Hewitt 1912, 451-2), but Keene (1985, 57-9) argues the reverse, and states the canal was proposed, but never built.

Another enigmatic document is a copy of a charter giving Bishop de Lucy the right to take tolls on goods carried between Winchester and Southampton on a canal he is supposed to have made. This document first occurs in the register of Bishop John de Pontoise in 1282, but there are commentators who consider it to be spurious (Deedes 1924, ii, 741; Keene 1985, 57-9).

There are certain anomalies, both in the documents themselves, and in the arguments for using them against the idea of a canal. The statement by the jurors that in order to get boats to Winchester, a number of mills must be destroyed is not true.

Firstly, it would be more convenient to build a by-pass around a mill than to destroy it. Secondly, it was normal practice to build these in association with mills anyway. Thirdly, there is considerable evidence to suggest that by-pass channels around a number of the mills mentioned as obstructions in 1275 did exist. Admittedly this evidence post-dates the inquisition, but there is a long tradition of building by-pass channels around mills. That they can be shown to exist after 1275 implies that most of them probably existed from the creation of the mills themselves.

In a 1401 survey of the precincts of St. Cross Hospital, two rivers are mentioned in the vicinity of St. Cross Mill, the 'main river, known as Ichenstreame', and 'the old river' (Kirby 1899, 532-3). John More's map of the River Itchen between Woodmill and Winchester dated 1618 explains these references. This map shows the 'Ichenstreame' and names it so as the main river by-passing St. Cross Mill. Another river parallel to it and feeding the mill is known as the mill stream (HRO 102M71/P1). Today the mill stream is the main river, and the 'Ichenstreame' has degenerated into a minor carrier for former water meadow ditches.

More's map shows similar diversion streams, some of them of considerable length, around all the mills on the Itchen north of Gater's Mill except possibly Brambridge Mill, where an exceptionally large mill-pond is shown downstream of the mill. It can be concluded therefore that apart from Woodmill and Gater's Mill, and possibly Brambridge Mill, small boats could have got around all the mills mentioned in 1275. It is most unlikely that all these diversions and mill leats were made between 1275 and 1618 as this was not how mills were generally made. It is possible one or two could have been built against tradition without the by-pass, but to have a whole river system with such anomalies is stretching credulity.

Further, it is certain that both Woodmill and Gater's Mill had a by-pass channel at some stage in their history. It almost certainly did not exist in 1618 or it would have been shown on a map that goes to such lengths to show the water arrangements of all the other Itchen mills. A scan of the evidence has found no reference to it being made subsequently despite a very good series of almost continuous records for Gater's Mill from 1433 to the present. The weight of the evidence suggests that the Woodmill/Gater's Mill diversion is earlier, and that it had fallen out of use by 1618. The only possible reference is that of 1045 charter referring to 'the new river'.

It is therefore likely that it existed in 1045, but later fell out of use. It would be tempting to suggest that it was disused by 1275, if the jurors evidence for the other mills on the Itchen did not appear to be misleading. It is even possible that the jurors' statement that passage as far as Bishopstoke might be allowed might remember a tradition that navigation was once possible to this point. Their qualifying statement that Woodmill would need to be destroyed to bring this about suggests that the existence of the conjectured by-pass around this mill had been forgotten by 1275. It is additionally curious that a number of early sources, such as William Dugdale, claim that de Lucy 'restored' the navigation of the river, implying that it had been possible before his time, but had fallen out of use (Hewitt 1912, 451).

If Keene and Deedes (op. cit.) can argue that the charter in de Pontoise's register is a forgery by propagandists wishing to discourage the citizens of Winchester wanting the canal (the bishop's right to collect tolls being the discouraging factor), it is possible the jurors of 1275 were also part of the same deceitful game. They seem to be giving the inquisition false information. At the worst some of the by-pass streams would need to be dug out because they might be too full of silt to allow navigation. Possibly some of the mills were without the diversions shown in 1618. This is unlikely, but it would not have required all that much effort to dig them. Instead the jurors stubbornly state that all the mills on the river would need to be destroyed, a statement that is untrue. The only realistic conclusion to be drawn from this is that the jurors were either being bribed, or were, by inclination, against the idea of navigation. The evidence of the inquisition can not therefore be accepted as objective.

The only conclusion that can be drawn from this curious chapter is that in the thirteenth century there was a desire to have a navigation up the Itchen to Winchester, and that there would appear to be a faction that was prepared to give misleading information to prevent it. The ability to show this must negate Keene's argument that navigation never existed. All we can say is that the evidence both for and against a thirteenth century passage is unreliable. Although this does not speak for the existence of an earlier passage, it seems to imply there could have been a historical tradition for navigation that may have had origins earlier than the thirteenth century. This does not suggest that the Saxons had created passage to Winchester, but the evidence seems to suggest they could have reached at least Bishopstoke. Even partial navigation at one time could have been the fuel required for the imaginations of the pro-canal faction.

Perhaps the more realistic argument against the existence of a canal constructed by Bishop de Lucy is the fact that most of the materials leaving Southampton for Winchester in the medieval period seem to have gone by road. It is notable that heavy items such as building stone for Winchester Castle seems to have gone by cart in 1220, and again in 1258 (Keene 1985, 58). Even more damning are the building slates that were brought by boat to Woodmill in 1289, and then taken on to Winchester overland (Keene 1985, 58). Certainly the Brokage Books for the port of Southampton seems to indicate traffic in the fifteenth century to Winchester was exclusively by cart.

However, the desire for navigation refused to die. In 1538 the mayor and aldermen of Winchester concluded that the main cause of their present poverty was the lack of transport on the Itchen. The departure of Bishop Gardiner soon after this allowed part of Woodmill to be demolished, but there is no evidence that this was connected with navigation. Only the migration of salmon appears to have benefited from this action (Keene 1985, 59).

In 1617 another attempt was made to revive interest in a navigation. A survey was commissioned to be carried out by John More (Course 1983, 6), resulting in the map of the river of 1618. Finally in 1665, work began on that 'canal' now known as the Itchen Navigation. This was supposed to have been completed by 1671, but work does not seem to have finished until 1710. This watercourse took an entirely different route to the conjectured earlier 'canal'. A lock was built at Woodmill to by-pass the mill, thereafter the course followed the main river to a point just beyond Mansbridge before heading north along the western side of the Itchen valley. This enterprise was never entirely successful, and was much hampered by competition for water from water meadow carriers and mills. With the coming of the railways, traffic almost ceased. The last commercial barge to use it was in 1869, from which time it has been neglected, and is now dry for considerable sections (Course 1983, 5-7).

The later history of the mills at Woodmill and Gater's Mill

There is some discrepancy amongst commentators as to which manors these mills belonged after the Norman Conquest. Even the pre-Conquest evidence is not clear. The inability to establish with any degree of certainty which mills are referred to in the Saxon charters for North and South Stoneham has been discussed above, and need not be repeated here.

The Domesday Survey adds to the confusion. There are no mills mentioned for South Stoneham, but two fisheries worth 39p (Munby 1982, 3-16). However, there are two mills listed at 30/- in North Stoneham (Munby 1982, 6-8). According to the boundaries of the Saxon charters for these two later manors, Woodmill would appear to be in South Stoneham, and Gater's Mill was certainly within that manor by the later medieval period.

The Victoria County History does not resolve this discrepancy by offering the simplified solution that the boundary must have subsequently changed leaving Woodmill and Gater's Mill, which it argues are in North Stoneham in 1086, in South Stoneham at a later date (Grundy 1908, 479). This argument rests on the error that Woodmill was within the Saxon charter bounds of North Stoneham, which is highly unlikely. Further the VCH seems to contradict itself by stating under South Stoneham that the fishery at Woodmill was probably that mentioned in Domesday Book. It is most unlikely that the fishery was in South Stoneham if the mill was in North Stoneham. Although the boundary between the two manors has clearly changed, the solution suggested by the VCH is incorrect.

The answer seems to be that Woodmill was simply not mentioned in 1086, although it probably existed. Such omissions are quite commonly found in Domesday, as is evidenced by the number of churches containing Saxon fabric which are not recorded therein. Furthermore, it would appear that Gater's Mill is recorded in 1086, but under the manor of Allington, and not under North Stoneham.

From the later medieval period right through into the eighteenth century, Gater's Mill was known as Upmills or Allington Mill. Two mills are recorded in Allington in 1086 (Munby 1982, 49-1), and it is quite likely that one or both are on the site of Gater's Mill. An early eighteenth-century document records at least three mills being formerly on this site before they were converted in paper mills (HRO 8M56/116). Furthermore, it is possible that one of the Domesday Mills in North Stoneham was at Barton Peverel, where there was a large mill on the River Barton. The sub-manor of Barton was included in the bounds of North Stoneham in 932 (Currie 1993, forthcoming), along with Eastleigh and Boyatt. These lands were later transferred to South Stoneham, but it is not known when this occurred. It is possible this occurred after 1086. The two mills both could have been at Barton, as it was common practice in the past to refer to the number of sets of mill stones as being the number of 'mills'. Alternatively, the second mill has been lost. There are no known mills at all on Monks' Brook, which flows for a considerable length through North Stoneham. The failure to utilise a moderately substantial stream such as this one would appear to be odd.

It would appear therefore, from the confusing evidence that is available, that the mill at North Mansbridge, outside the 932 boundaries of North Stoneham, but included with the estate, had become lost to the emergent sub-manor of South Stoneham at Allington by 1066. The indications are that this was once a shared mill between the North and South Stonehams, and that this situation was a relict of the time when these two portions probably formed one large estate. The fishery at Woodmill appears to be mentioned in 1045, and again in 1086, under the lands of South Stoneham. Its possible mention after the bounds of the 1045 charter could suggest that it was additional rights to lands outside of the boundary of the South Stoneham estate, as North Mansbridge mill was to North Stoneham. However, whereas North Stoneham lost its additional lands to South Stoneham at a later date, South Stoneham was able to bring the additional lands mentioned in its Saxon charter within its bounds. When this occurred is uncertain, but there are strong suggestions, that if the Woodmill fishery is that mentioned in Domesday, the mill would have been in the manor also, but was somehow exempt from tax. It is possible that Woodmill did not exist at this time, but this seems unlikely in the circumstances.

The fishery at Woodmill has a well-documented history. The present fishery, known as the Salmon Pool, is almost certainly an artificial creation made at the confluence of the Itchen with its tributary, Monk's Brook (formerly known as Stirbrook). This pool acts as a collecting point for catching salmon attempting to enter the two rivers. It may have been the 'seaweir' mentioned in the 1045 charter, as it stands at the head of the tidal portion of the River Itchen, thereby being technically in 'sea' water, although in reality the water here is brackish. Although the manor of South Stoneham was reserved to the monks of St. Swithun's Priory in 1086 (Munby 1982, 3-16), the bishop appears to have disputed these rights in the twelfth century. A document in the Winchester Cathedral cartulary records a dispute between the monks and the bishop over 'Stanham' in 1171 (Goodman 1927, 3, 12). Although the bishop seemed to concede to the monks at this time, by the early thirteenth century he had regained the manor. From hereon through to the early sixteenth century, records are kept fairly regularly in the bishopric pipe rolls of the annual catch of salmon taken here.

A mill in the bishopric manor of Bitterne is mentioned as being repaired in the first pipe roll of 1208-9 (Hall 1903, 31). It is presumed that this is the mill at Woodmill. It is further mentioned in an Inspeximus of the later fifteenth century when the bishop granted to Thomas Ederigge of Swaythling the farm of two mills 'which are conjoined and are called lez Wodemyllis'. Included in the lease is a pasture of three acres adjacent to the mill called 'Hereslese' together with fishing rights in the bishop's waters there with permission to use a 'Rackis' and Gynnys' belonging to the mill according to ancient custom. Ederigge had to bear the cost of repairs to the mill 'causeway', and had the right to take clay and gravel for that purpose. He could also take alders for making piles and twigs for making hurdles for the mill (Greatrex 1978, 168-9).

These enigmatic references record age-old techniques used to make fish-traps and mill-dams. The clay and gravel for the causeway is self-explanatory. Research into the historic techniques for making mill and fishpond dams show that the clay in the dam was often rammed in behind wooden revetments held in place by piles. The hurdles were often placed across the river to prevent fish passing, and to channel them into the traps set for them. These were often wicker nets at the head of a 'weir' of hurdles, and examples have been excavated on the River Trent near Nottingham (Losco-Bradley and Salisbury 1988). Examples were still used in Ireland into this century, as they are on the River Severn below Gloucester. The traps themselves are often referred to as 'gins'.

An explanation of the 'Rackis' requires a knowledge of the habits of eels to fully appreciate, but they were commonly associated with mills. The Winchester mill at Durngate had an eel rack that still operated to within living memory. Eel Racks are still recorded at Woodmill on early twentieth century 25 inch Ordnance Survey plans.

These ingenious traps relied on the eels' ability to leave the water and crawl overland when encountering obstructions to their migration from freshwater to spawn in the sea. They comprise an iron rack, probably a wooden hurdle in earlier times, that is placed in the entrance to a sluice or where the water flows under the mill-wheel. Any place where the water is channelled into narrow confines will suffice. This rack is placed so that it is at an angle of 45 degrees to the bottom of the river, with one end firmly placed in that bottom, and the other end protruding out of the water. The eels finding their passage blocked will usually wait until nightfall, and then crawl up the rack. The top of the rack is blocked in some way to prevent them dropping over the edge. Again finding their passage blocked, the eels will find a deliberately cut hole in the side of the channel wall through which they can pass. From here they can drop down into a water tank, which they mistake for the river. There is no way out of this tank, and all the miller has to do is collect the trapped eels in the morning. On a night when the eels are moving en masse down river to spawn, one of these traps can take over a hundred fish per night. The 'Rackis' mentioned at Woodmill in the fifteenth century is undoubtedly one of these traps.

As mentioned above, part of the mill was demolished in the mid-sixteenth century, probably to enable salmon to reach their spawning grounds more easily. At this time the fishery appears to have been prolific. In 1538, the fishery of river between Winchester and Southampton is recorded as being so rich in salmon, that the local people were neglecting their normal occupations to steal the fish (L & P Henry VIII 1538, i, 1240).

The bishop of Winchester claimed the fishery, not only of Woodmill, but of the entire Itchen. From the post-medieval period, the 'Itchen fishery' was held in conjunction with lease of Woodmill (HRO 102M71/T157-9). This led to a lawsuit in 1896 between the Flemings of North Stoneham and Sir Samual Montagu, on behalf of the bishop, and John Gater of Gater's Mill, on behalf of Winchester College, over the right of Allington manor to piscary in the river where it passed through that manor's lands (HRO 85M88W/12).

The two mills at Woodmill, with the fishery and Horselease Close continue to be mentioned in a series of leases from the bishops of Winchester starting from February 1641 with a lease for three lives to Ann Clerke of South Stoneham (HRO 102M71/T143). The lease was renewed by members of the same apparent family until 1697-98 when a new lease was granted to Edward Fleming of North Stoneham (HRO 102M71/T146). A detail of Fleming's lease of the mill to Edmund Pitman in the same year refers to the two mills being 'under one roof'. Furthermore the fishery is mentioned along with the 'Racks' and 'Ginns' pertaining to it (HRO 18M67/227). From 1741 a lease for three lives was grated to the Sloane family of South Stoneham House (HRO 102M71/T147), who are last recorded renewing their option in 1792 (HRO 102M71/T159).

Gater's Mill is first recorded after the Norman Conquest in a late thirteenth century grant of William Alis to the Priory of St. Denys' of his rent of 'Allington' mill (Blake 1981, 6). By 1360 a fulling mill had been created here, probably alongside the earlier corn mill. In this year John Wodelock granted Robert Torold his share of this mill, called 'Upmill', with the fishery passing under the mill wheel, and his share of the adjacent piece of land called 'Rakkeley' for drying cloth in the shed there (Himsworth 1984, 1545). From hereon the greater share of the mill appears to be held by Winchester College. A good series of leases is available from fifteenth century through until 1853 in the College Muniments (Himsworth 1984, 25037-56). In 1433 John atte More granted the College a moiety of the watermill there, with the pond and ground immediately adjacent on either side (Himsworth 1984, 1546).

Those with shares in the mill in 1453 included an annual rent to Winchester College of £6, 3/- a year to St. Denys' Priory, and 2/- a year to the honour of Wallingford (Himsworth 1984, 25037). Up until 1517 the lease appears to have been of relatively short duration, seven and ten years being recorded, but after this date longer leases were issued. In 1517 Thomas Fysher leased the mill for 40 years (Himsworth 1984, 25039). In 1549 a lease was issued for 60 years (Himsworth 1984, 25040). In 1612 Arthur Blomfield obtained a 40 year lease on the Winchester College moiety on two corn mills and one malt mill 'under one roof called Upmill' (Himsworth 1984, 25042).

There are two surviving inventories of the mills and their contents. The first dates from 1696, and records four 'mills' on the site. These are an 'old decayed mill' called the upper mill, a second mill called 'Salisbury mill', 'Ye middle mill' and 'Ye 4 mill near ye Lodging house'. The contents suggest the site is being used as a paper mill, and the stones and 'tackle' for the 'cornemills' are kept in a warehouse. Another inventory of 1723 records only the corn mill and middle mill by name, and does not make it clear what use the mills are put to (HRO 8M56/116).

In 1724-25 Thomas Dummer purchased rights to the mills. These are referred to as paper mills on the site of 'fulling mills', plus a moiety of another paper mill formerly corn or malt mills. The lease refers to two corn mills and one malt mill 'called Upmill in Allington' (HRO 8M56/116). Exactly how many 'mills' were once here is difficult to ascertain, but it appears that the number varied at different times from between two and four. The present layout at Gater's Mill, although largely dating from the nineteenth century, does indicate that there was more than one mill here. On the site today it is still possible to recognise the channels feeding at least two separate wheels. There are buildings either side of each wheel with evidence that each wheel turned axles on both sides of it, thereby having the capacity to operate four 'mills'.

This evidence demonstrates that both Woodmill and Gater's Mills were complex industrial units in their own right, and were both more substantial than many more simple country mills. Both operated a complex system of fisheries and fish traps, and Gater's Mill operated at least two large wheels capable of turning at least four sets of stones. Although both mills probably started their life in the Saxon period as corn mills, by the fourteenth century at the latest they were operating other complex functions. There is good evidence to suppose the important fishery at Woodmill was operating before the Norman Conquest, thus making both sites major obstacles to boat traffic. It is probably in the context of a by-pass around these busy points that the 'new river' mentioned in the 1045 charter for South Stoneham should be seen.

A map of the manor of Allington in 1770 in the Winchester College Muniments gives some additional information about the mills. This states that the 'left hand stream L runs through the College Corn Mill, the main river, half of which is College property runs through the Paper Mill' (HRO 28M61/2). At the present day there are two channels running under the mill buildings, emptying into the same pool. However, the present main river by-passes the mill buildings on the west side so it is difficult to interpret these comments exactly. It is most likely that the term the 'main river' was incorrect, and that the 1770 arrangement reflects that currently surviving. However, it is possible that changes may have been made subsequently. In a map of the Itchen of 1618 by John More, nearly all the Itchen mills are shown with by-pass streams. Allington Mill is not shown to have one (HRO 102M71/P1), and it is possible that by the post-medieval period the buildings were straddling the entire river.

In 1785 the lease of the mill was granted to John Gater of Swaythling (Himsworth 1984, 25049). Members of his family renewed the lease until at least 1853, when the last Winchester College lease is recorded (Himsworth 1984, 25056). It is from this family that the mill takes its present name.

General discussion

The two Stonehams may have once formed part of a large estate, centred on an important mid-Saxon villa regalis of Hamtun from which developed the trading port of Hamwic (Hase 1975, 142-43). The subsequent division of this estate is discussed elsewhere (Currie 1993, forthcoming). Disruptions caused by viking raids, and the rise in importance of Winchester are thought to be amongst a number of reasons responsible for the decline of Hamwic (Morton 1992, 75-77). The charter evidence, which shows the granting of the Stonehams to church estates in Winchester, may be partly reflecting the shifting of power in the region northwards (Yorke 1984, 66). As the economic ties of the Stonehams with Hamwic declined from the mid-ninth century, so the urge to reorganise them and their boundaries must have been felt. There appears to be some delay before the process is completed, probably in the first half of the eleventh century. The apparent date of c. 992-1045 for major river engineering on the Itchen in the vicinity of Woodmill suggests a possible connection with this development.

Navigation in the tidal and lower reaches of rivers seems to have had an important role in the local economy. Timber from Botley was moved down the River Hamble to Bursledon and Portsmouth in the post-medieval period (Currie 1991), and it is suggested that Southampton may have been served in the same way at an earlier date. Saxon charters refer to a 'new river' near Swaythling, but this is probably a late Saxon response to the obstruction of traffic that had been plying the lower river for generations.

The evidence given above, however, shows that the arguments for and against early passage for boat traffic between Southampton and Winchester are extremely complex. The discovery of a substantial water channel, apparently by-passing Woodmill and Gater's Mills, has prompted a need to assert that the arguments put forward by Keene (1985, 57-9) and Roberts (1985) against a thirteenth-century and later passage up-river are not used as arguments against an earlier passage, and to suggest that the strong local tradition for navigation on the river may have had historic precedence.

It would seem that the water channel recently discovered can be linked with a reference on the charter for South Stoneham of 1045 for a 'new river'. This strongly implies that passage could be obtained at least as far as Bishopstoke at this time. It also implies that passage had existed earlier, but had been probably blocked by the expanding milling and fishing interests on the river over the course of the tenth and eleventh centuries. It is not possible to argue on present evidence that navigation was available as far as Winchester in the eleventh century, although barges could have off-loaded at Bishopstoke, an important episcopal manor, before continuing by road to Winchester.

It might be suggested that the bishops had once gained revenue from this passage, but it had fallen out of use by the episcopate of de Lucy in the late twelfth century. Around this time a spurious charter appears that gave the bishops the right to levy toll on the river, and claims to have cut a new passage to Winchester. Could this possibly be because there is talk of making a passage in the later thirteenth century, and the bishops are keen to revive some old right they may have had in advance of the intentions of the citizens of Winchester?

It is possible that the long-standing tradition of navigation along the Itchen was founded in an actual passage that existed in the Late Saxon period. The absence of a mention of a new river in the South Stoneham charter of 990 x 992 indicates that it may not have existed at this time. This may have been because the passage of boats had not yet been impeded. There is tentative evidence that implies Woodmill did not come into being until after this date. Its possible creation, with its important fishery between 992 and 1045, could have blocked an earlier passage causing a new river to be cut to keep a vital link with a town of national importance open. However, there is no proof in the evidence examined here that this passage went by river beyond Bishopstoke, although passage around the mills upstream of this point can be shown to have been possible.

Unfortunately, the period when this conjectured passage was probably operating, in the eleventh and early twelfth centuries, is one of the poorest documented periods in English history. It would be expected that growing activity in the Itchen valley between 1045 and 1204 increased the demand for milling facilities, thus further impeding navigation. The 'new river' would seem to have fallen out of use by 1275 at the latest, if not well before the time of de Lucy. It is unlikely that any further evidence will now be forthcoming to clear up the mystery, and the conclusion forced on us is that man-made navigation on the Itchen may have existed briefly, but this was short-lived, and possibly of a limited extent (to Bishopstoke?). Nevertheless, this existence created a folk-memory that refused to die. Even the threat of reviving it seems to have prompted the bishops of Winchester to produce a forgery to lay claim to rights on the river, probably more as a safeguard against vague existing rights that could then only be dimly remembered, rather than as an attempt to highjack revenues from any new proposals.

Appendix: Middle Saxon earthworks of field systems surviving at Mansbridge?

Other earthworks noted to the north of Riverside Park are worth commenting on here. Both Saxon charters for South Stoneham suggest that there were open fields within this manor in the tenth and eleventh centuries. These are suggested by the reference to the 'gate of the ploughlands', between Southampton Common and the later medieval town.

There are nine acres of arable recorded in the North Stoneham charter which appears could partly coincide with a rare survival of about two acres of possible ridge and furrow in a public open space on the north side of Mansbridge Road at SU 447157. There are two largish arable fields on the 1844 Tithe Map for South Stoneham (HRO Tithe Map and Award: South Stoneham) known as Great and Little Stoney Fields (SU 444157) on the other side of this road. The proximity of these fields to the recently discovered Saxon 'village' of South Stoneham (SU 439156), and their names may indicate that they were once part of the fields of that settlement, a physical characteristic from which it took its name. It is possible that this field may have once included the nine acres later recorded in North Stoneham, possibly as a result of subsequent division of the estate into its north and south components. Although this land appears to be in North Stoneham in 932, it is included in land outside the bounds of that estate, and may represent detached land which was later reincorporated into South Stoneham by the time of the first known parish boundaries shown on the 1810 one inch Ordnance Survey map (details of these arguments are given in Currie 1993, forthcoming).

There is some suggestion in the Saxon charter of 1045 for South Stoneham that the Mansbridge Road from Swaythling to the ford over the Itchen may have formed the northern boundary of this field. After leaving Swaythling to head to the 'Old Itchen', the bounds move 'along the king's row'. It has been suggested that this could have been a hedgerow (Della Hooke pers. comm.). If the area in the vicinity of the 'Stony Fields' on the Tithe Map was a large open field such a feature at its boundary with a road would have been appropriate to keep unwelcome feet from trampling the crops.

There is further evidence that this field was enclosed as there was a particularly large former hedgebank surviving parallel with Monk's Brook along its conjectured western edge at SU 442155 until recently. Only the far southern fragment of this bank now survives as the result of development by the Swaythling Housing Society in the late 1980s. Nevertheless, its full extent can be seen on the 1967 edition of the 25" Ordnance Survey map (Plan SU 4415), where it is shown as a substantial earthwork. This bank would have further acted as a flood bank to keep the notoriously unreliable Monk's Brook from damaging the crop in times of heavy flow. This river has a long history of being prone to flooding because of its former sinuous course, and has recently been canalised by the River Authorities as a response to this recurring problem.

The survival of ridge and furrow beyond the apparent 1045 boundary of this field suggests it had been divided before 932 (when the area covered by the earthworks appears to have been in North Stoneham). This suggests that both the ridge and furrow, and the surviving remnant of the western hedgebank, both date from before the division of the Stoneham estate into its north and south components, thus making them possible rare survivals of the Middle Saxon period.

Acknowledgements

The author is grateful to Dr. Della Hooke, George Watts and Dr Barbara Yorke for their suggestions, and to Harold Barstow for imparting much useful information about the history of the area.

Bibliography

Manuscript Sources

Hampshire Records Office (HRO):

HRO 8M56/116: documents relating to Allington Mill, 17/18th century.

HRO 28M61/2: photocopy of 1770 map of Allington manor.

HRO 102M71: papers relating to estates of the Fleming family.

102M71/P1: John More's map of the River Itchen, 1618.

102M71/T143-59: documents relating to Woodmill and the Itchen fishery, 17/18th century.

HRO 18M67/227: Lease for Woodmill, 1697.

HRO 85M88W/12: Documents relating to the case of Fleming v Gater, 1896.

Maps and Awards in the Hampshire Record Office:

Tithe Map and Award: South Stoneham.

Ordnance Survey one inch map, 1st edition, 1810, sheet XI.

Ordnance Survey 25 inch maps, editions 1865-66, 1910, 1933, 1940, sheet LXV.3.

Manuscript sources in print:

Blake, E. O. (ed.), The cartulary of the Priory of St. Denys near Southampton, 2 vols., Southampton, 1981.

Deedes, C. (ed.), Registrum Johannis de Pontissara, vol. II, Oxford, 1924.

Goodman, A. W., (ed.), The cartulary of Winchester Cathedral, Winchester, 1927.

de Gray Birch, W. (ed.), Cartularium Saxonicum, 3 vols. and index, London, 1885-1899.

Greatrex, J. (ed.), Register of the Common Seal, Winchester, 1978.

Grundy, G.B., 'The Saxon land charters of Hampshire with notes on place and field names', Archaeological Journal, 2nd series, xxxiv (1927), pp. 160-340.

Hall, H. (ed.), The pipe roll of the bishopric of Winchester, 1208-09, London, 1903.

Himsworth, S. (ed.), Winchester College Muniments, 3 vols., Chichester, 1981.

Kemble, J. M. (ed.), Codex diplomaticus aevi Saxonici, 6 vols., London, 1839-48.

Kirby, T. F. (ed.), Wykeham's Register, vol. II, London, 1899.

Munby, J. (ed.), Domesday Book: Hampshire, Chichester, 1982.

Sawyer, P.H., Anglo-Saxon charters, London, 1968.

Secondary Sources

Biddle, M. and Keene, D. 'Winchester in the eleventh and twelfth centuries', Biddle, M. (ed.), Winchester in the Early Middle Ages, Oxford, 1976, 241-448.

Course, E., The Itchen Navigation, Southampton, 1983.

Crockett, A., 'Excavations at Montefiore New Halls of Residence, Swaythling, Southampton, 1992', Wessex Archaeology typescript, forthcoming.

Currie, C. K., 'Botley Manor Farm's woodlands in the eighteenth century', Hampsh. Field Club & Archaeol. Soc. newsletter new series 15 (1991), pp. 20-24.

Currie, C. K., 'Saxon charters and landscape evolution in the south-central Hampshire Basin', Proceedings of the Hampshire Field Club and Archaeological Society, 49 (1993, forthcoming).

Grundy, A.R., 'Mainsbridge hundred' in Page, W. (ed.), The Victoria history of the county of Hampshire and the Isle of Wight, vol. 3, London, 1908, pp. 462-89.

Hase, P.H., The development of the parish in Hampshire particularly in the eleventh and twelfth centuries, 1975, unpublished PhD thesis, Trinity College, Cambridge.

Hare, M., 'Investigations at the Anglo-Saxon church of St. Peter, Titchfield, 1982-1989', Proc. Hampsh. Field Club and Archaeol. Soc., 47 (1991), 117-44.

Hewitt, E. M., 'Industries: introduction', Page, W. (ed.), The Victoria History of the County of Hampshire and The Isle of Wight, vol. V, London, 1912, 451-461.

Hollinrake, C. & N., 'A late Saxon monastic enclosure ditch and canal, Glastonbury, Somerset' Antiquity, 65 (1991), 117-8.

Keene, D. (ed.), A survey of medieval Winchester, vol. I, Oxford, 1985.

Losco-Bradley, P. M. and Salisbury, C. R., 'A Saxon and a Norman fish weir at Colwick, Nottinghamshire', Aston, M. (ed.), Medieval fish, fisheries and fishponds in England, BAR British series 182 (1988), 329-351.

Morton, A. D. (ed.), Excavations at Hamwic: volume 1, London, (Council for British Archaeology research report no. 94), 1992.

Roberts, E., 'Alresford Pond, a medieval canal reservoir: a tradition assessed', Proc. Hampsh. Field Club Archaeol. Soc., 41 (1985), 127-138.

Yorke, B.A.E., 'The foundation of the Old Minster and the status of Winchester in the seventh and eighth centuries', Proc Hampsh. Field Club & Archaeol. Soc., 38, (1982), pp. 75-83.