Stourpaine

46 STOURPAINE (8609)

(O.S. 6 ins. ST 80 NE, ST 81 SE)

The parish, of 2,300 acres, lies entirely upon Chalk on the E. bank of the R. Stour. The Iwerne Brook flows through the western part of the parish to join the Stour a little S. of the village. To the E. the land rises steeply from the Iwerne valley, at about 130 ft. above sea-level, to an altitude of 435 ft. on the Chalk upland; to the N.W. rises Hod Hill, its summit nearly 500 ft. above sea-level. There were certainly three and perhaps four original settlements, each with its open field system (see map, p. 260). In the N., on the E. bank of the Iwerne, stood Lazerton, once a separate parish with its own church; adjacent, to the S. and also on the Iwerne, was the hamlet of Ash. The present village, further S., stands near the confluence of the Iwerne and the Stour. Lazerton, Ash and Stourpaine are all mentioned in Domesday (Vol. I, ff. 82a, 83, 84a, 85), with a total recorded population of 28. France Farm in the S. of the parish was probably another early settelement; from mediaeval documents quoted by Hutchins (fn. 1) it appears to have been one of the Nutfords, but its relationship with the Blandford Nutfords, now in Pimperne (Dorset IV), is obscure.

The most important monument in the parish is the Iron Age hill-fort on Hod Hill; it contains abundant occupation remains and also a Roman fort. The church is mainly of the mid 19th century, but it has a 15th-century tower and an interesting wall-monument of 1670.

Ecclesiastical

(1) The Parish Church of the Holy Trinity, at the S. end of the village, was rebuilt in 1858 with the exception of the West Tower which is of the second half of the 15th century. Three three-light windows from the old church were restored and reset in the N. wall of the new nave; the two to the W. were of 15th-century origin; that to the E. was probably somewhat earlier but very little of the original stonework remains.

Architectural Description—The West Tower is of Greensand ashlar and has three main stages with a moulded and battered plinth and an embattled parapet. The stages are defined by weathered string-courses; the parapet string-course is weathered and hollow-chamfered and has a grotesque gargoyle at each corner. In the two lower stages the tower has four diagonal buttresses, each of three weathered stages; those of the N.E. and S.E. corners are incorporated in the W. end of the nave and are visible inside. The top stage has shallow angle pilasters which continue in the parapet as pinnacles with crocketed finials. On the N. is a polygonal vice turret, of three stages with a pyramidal head. The tower arch is two-centred and of two chamfered orders; the inner order dies into the responds and the outer order is continuous. The external vice doorway has a chamfered segmental-pointed head with continuous jambs. The west window is two-centred and of three cinquefoil-headed lights under vertical tracery in a chamfered head with continuous jambs; it has a hollow-chamfered label with square stops. The second stage has a single square-headed light in the W. side. The top stage has, in each side, a belfry window of two trefoil-headed lights under a central quatrefoil, in a two-centred head with a moulded label with square stops.

Fittings—Bells: six; 2nd, by T. Purdue, inscribed 'T.P., E.M.R.M. ano domni 1624', on dome 'IF'; 4th inscribed 'Richard Gould Als Courage Church Warden, William Rosiear cast mee in the Year of Our Lord 1721'; tenor inscribed in black-letter 'In Ter Sede Pia Pro Nobis Virgo Maria', mid 15th century; others modern. Chair: of oak, with turned front legs and arm-rest supports, guilloche ornament on rail, plain seat, scrolled arm-rests, panelled back with guilloche enrichment, carved top rail with rounded head, and three turned finials; 17th century, finials later. Monuments: In nave, on N. wall, (1) of Robert Maurice, 1818, sarcophagus-shaped marble wall tablet with draped urn and arms, by Hiscock of Blandford; (2) of Thomas Bower, 1787, and Elizabeth his wife, 1809, slate tablet with incised arabesques and urn-shaped surround to inscription. In tower, on N. wall, (3) of John Straight [d. 1680], vicar, wall-monument erected during his lifetime, comprising round-headed niche flanked by Tuscan columns supporting entablature (Plate 35); above, painted inscription panel in moulded surround with scroll supports; within niche, kneeling effigy in black gown with gathered sleeves and tall collar, hands missing; monument dated 1670. Plate: includes Elizabethan silver cup and cover-paten without maker's or date marks.

Secular

(2) Lazerton Farm (86381032), house, ⁵/8 m. N. of the church, is two-storied with attics and has walls of ashlar, rubble and brick, and roofs covered with modern slates; the E. range (Plate 57) was built in the first half of the 17th century and there are 19th-century additions on the W.

The original E. front, of ashlar, is symmetrical, with five bays on the ground floor and three on the first floor; at the base is a chamfered plinth. The central doorway has an added porch with two stone Ionic columns and side walls supporting a flat wooden hood with a moulded cornice. The ground-floor windows are of four square-headed lights, with chamfered and hollow-chamfered surrounds, hollow-chamfered stone mullions and hollow-chamfered labels; a few of the lights have been blocked. The first floor has two four-light windows flanking a central two-light window. The gabled N. wall is of coursed rubble, partly rendered; the ground-floor window is modern but the attic has a small original light with a chamfered and hollow-chamfered surround. The gabled S. wall is of ashlar, with a chamfered ashlar plinth continuing from the E. front. On the ground floor is a mullioned three-light window, as before described, and the first and attic floors have windows from which the stone mullions have been removed. The N. and S. gables have moulded and weathered copings. The W. elevation is partly masked by 19th-century additions but the W. wall of the original stair bay, which projects from the main block, is exposed; it is gabled and has ashlar quoins; a stone two-light window occurs at mezzanine level and there is a single light above.

Inside, the E. range comprises two ground-floor rooms and a central through-passage. The S. room has late 18th-century fielded pine panelling with a moulded dado rail and a moulded wooden cornice. The N. room has, in the W. wall, an open fireplace with a four-centred hollow-chamfered stone head with sunk spandrels in an ogee-moulded rectangular surround; the mouldings are continuous on the jambs and end in shaped stops. A two-light window to the N. of the fireplace now opens into a room of the 19th-century W. wing; to the S. a square-headed stone doorway, with an old plank door, opens into the stair bay. From the stair bay a doorway with a chamfered four-centred head, originally external, now leads into the W. wing. The doorway from the through-passage to the stair bay has chamfered stone jambs; the head is concealed. At the centre of the stair well is a cylindrical oak newel-post, continuous from ground to attic; from the first floor to the attic the original winders radiate from the newel, but in the lower storey the newel stairs have been removed and an 18th-century straight flight with turned balusters and a moulded handrail has been substituted. The first-floor rooms have panelled dadoes, and one room has a moulded ceiling cornice and a bolection-moulded fireplace surround; in the attic is a small stone fireplace with a chamfered four-centred head.

(3) Stourpaine House (86030956), 200 yds. N. of the church, is of two storeys and has rendered walls and pantiled and slated roofs. The house is of two periods; the main range, on the E., is of the early 19th century and has a symmetrical E. elevation of five bays with sashed windows. A lower wing on the W. probably dates from the late 18th century; it has a square-headed front doorway with a reeded architrave. A covered way with a coved lead roof, supported on small columns with moulded caps, connects the front doorway with a doorway in the garden wall. Inside, the 18th-century part of the house has ceiling-beams with beaded edges. The stairs, with a moulded mahogany handrail and plain uprights, are of 18th-century origin but were remodelled in the 19th century.

(4) Cottage (85930954), 225 yds. N. W. of the church, is two-storied and has brick walls and a thatched roof; it is of the late 18th century. The S. front was originally symmetrical and of five bays, with a central doorway flanked by two-light casement windows, and similar openings on the first floor. The doorway has been transferred from the central bay to one of the lateral bays. The middle first-floor window is enriched with scrolled wooden cheek-pieces, probably reset.

(5) Cottage (85960950), 150 yds. N.W. of the church, is single-storied with an attic and has a thatched roof and walls that are partly of brick and partly of timber-framing; it is probably of the 17th century. In the attic storey of the timber-framed N. gable is a three-sided bow window of five casement lights with ovolo-moulded mullions; the window projects from the wall-face and rests on shaped brackets.

(6) France Farm (86770819), house, ¾ m. S.E. of the church, is two-storied, with rendered walls and tiled roofs; it was refronted and remodelled in the mid 19th century but it probably incorporates an earlier nucleus. A barn to the W. is of the mid 19th century.

Dispersed in the village are twelve Cottages of the late 18th and early 19th century. They have walls of rubble, flint and brick, sometimes rendered; the roofs are thatched, or of modern materials. Three such cottages are located ¼ m. N. of the church, at 86050983, 86050975 and 86070977; three are grouped together, 300 yds. N.W. of the church, at 85800960, 85800958 and 85820960; six others are located at 86020960, 85990949, 85930955, 86020942, 86130948 and 86010951.

Mediaeval and Later Earthworks

(7) Settlement Remains of the former village of Lazerton (864106) survived until recently in the extreme N. of the parish, along the E. side of the R. Iwerne.

Lazerton was an independent parish and was one of the 'Iwernes' of Domesday Book; it has been identified by Eyton (137–8) as the three-hide manor of Edwinus (D.B. Vol. I, f. 84a), with a recorded population of seven. In the late 12th century a charter states that the church was so poor that it was released 'from all payments except synodals' (Hutchins I, 311; W.A.M. IV (1858), 68). The parish was not taxed in 1428, having fewer than ten inhabitants (Feudal Aids, 1284–1431, Vol. II, 94); in 1431 it was joined to the parish of Stourpaine because 'the church of Lazerton had so small profits that it had been and was then destitute of a chaplain' (Hutchins loc. cit.). The village was certainly without inhabitants when the 1662 Hearth Tax returns were made.

The remains, covering 10 acres, were destroyed in 1962; before this the site comprised two distinct parts. The N. third was a roughly rectangular area, centred on the site marked on O.S. maps as 'St. Andrew's Church' and now divided into two fields; it was covered by numerous scarps, ditches and banks and the area marked as the church was a roughly rectangular platform (50 ft. by 25 ft.), orientated E.-W. When the area was bulldozed in preparation for ploughing a limestone coffin-lid with chamfered edges and traces of an incised cross was found; fragments of flat unglazed roof tiles also came to light.

The southern two-thirds of the site contained at least seven closes, orientated E.-W., 20 yds. to 35 yds. wide and defined by low banks. Fragmentary platforms and sunken areas, probably the sites of buildings, lay at the high, or eastern ends of the closes.

(8) Settlement Remains of the former hamlet of Ash (865102) lie immediately S. of Ash Farm on the E. side of the R. Iwerne.

The recorded population in 1086 was seven (D.B. Vol. I, f. 82a), fifteen taxpayers were listed in 1333, four households were recorded in 1662 (Meekings, 68) and there were still ten houses late in the 18th century (Hutchins I, 309). At the beginning of the 19th century six or seven buildings stood S. of the farm (O.S., 1811), but none of them corresponds with the existing earthwork remains; until quite recently a group of cottages stood to the W. of the earthworks (O.S. 6 ins., 1902).

The fragmentary remains comprise at least four closes, orientated E.-W., 70 yds. long, 25 yds. to 30 yds. wide and bounded by low banks. Areas of rubble at the western ends of the closes indicate former house-sites.

(9) Moat and Enclosure (860093), lie S. of Stourpaine Church, near the place where the Iwerne joins the Stour.

The moat (class A2(b)), (fn. 2) formerly with a wet ditch surrounding a rectangular island, occupies the N. third of the site. The N. side and the N. third of the E. side are almost completely filled in and their alignment is only marked by a scarp, 2½ ft. to 3 ft. high, along the edges of the island. The rest of the E. side and all of the S. and W. sides are bounded by a deep wide ditch, now dry. A pond covers the N.W. corner. An external bank about 1 ft. high occurs along the W. side. The interior of the island is flat and has no features apart from a modern spoil heap in the S.E. corner; no entrance or causeway is visible.

Adjoining the S. side of the moat are the remains of a rectangular banked enclosure of about 2 acres. It is bounded on the E. by a low spread bank and on the S. by a bank 3 ft. high, with an external ditch. The W. side is no longer traceable on the ground but the 1st edition of the 25-in. O.S. map shows about 50 yds. of bank at the S. end; it probably was destroyed in 1863 when the railway was built (B.M. Add. MS. 37784, nos. 50, 62). At the N. end of the enclosure, and lying against the S. side of the moat, is an earthwork which Hutchins (I, 305) describes as a 'Cockpit'. It is a roughly circular area with an internal diameter of about 100 ft. bounded by a low bank and an external ditch. There is a gap in the N.W. part of the bank. This earthwork is certainly of later date than the moat.

(10) Cultivation Remains. Contour strip lynchets of the open fields of Lazerton (865106–874113) cover some 15 acres; they are now almost wholly destroyed. The fields were probably enclosed at about the same time as Lazerton was deserted, presumably in the 15th century.

Contour strip lynchets of the open fields of Ash (around 872101) cover some 70 acres, mainly on each side of a dry valley which runs E.S.E. from Ash Farm. The strips are up to 320 yds. long with treads 30 yds. wide; some have risers up to 20 ft. high. Further traces of strips, now ploughed out, can be seen on air photographs (R.A.F. CPE/UK 1944: 1321 & 3324) on the flatter ground to the N., N.E. and S.W. The date of enclosure of these fields is unknown, but until late in the 18th century the glebe land remained in strips within enclosed fields (late 18th-century map of Ash and Lazerton, D.C.R.O.).

Contour strip lynchets of the open fields of Stourpaine, finally enclosed in 1856 (map in D.C.R.O.), remain in a number of places. To the S. of Hod Hill (853104–857103) are the ploughedout remains of contour strip lynchets; some of these had already been enclosed by 1841 (Tithe Map) but others remained in Walls Linch Furlong. Some 500 yds. to the S. are two partly lynchetted ridges; they were formerly headlands, 30 yds. wide and 2 ft. high, running E.-W. along the slope, roughly parallel and about 350 yds. apart. The headlands lay between three furlongs of strip fields, no longer visible but shown on the Tithe Map as Farmacre, Middle and Lower Furlongs. To the E. of the village (863095–871092), extensive remains of contour strip lynchets cover some 40 acres on each side of a dry valley running E.-W.; in 1841 they were part of the open fields. Further remains of strip lynchets occur, to the N.E. of the last named, on the S. side of another dry valley (875098); in 1841 they lay beyond the limits of the open fields and within an area of common pasture. (See also 'Celtic' Fields (61), p. 341.)

Strip lynchets S.W. of France Oaks Coppice (867084) are perhaps the remains of the open fields of France Farm, a mediaeval settlement (Fägersten, 58).

Roman and Prehistoric

(11) Hod Hill Iron Age Hill-fort and Roman Fort (856107). In area the hill-fort is the largest in Dorset, its multiple ramparts enclosing 54 acres. It occupies the domed top of an isolated chalk hill rising to 471 ft. above O.D. at the N.W., where the later Roman fort was built in the angle of the Iron Age earthwork. Except on the S. there are steep slopes on all sides, particularly on the W. where the ground falls abruptly to the R. Stour. The interior of the hill-fort was formerly covered with Iron Age occupation remains, of which some 7½ acres survive in the S.E. quarter (Plate 198 and folding plan in pocket at back of volume); the rest has been destroyed, mostly by cultivation, since the mid 19th century. Ploughing has produced many finds, notably those in the Henry Durden collection in the British Museum (J. W. Brailsford, Hod Hill I, 1962). Sir William Boyd-Dawkins dug earthen circles and pits in 1900 (Arch. J. LVII (1900), 52–68) and Sir Ian Richmond directed excavations for the British Museum from 1951 to 1958. These researches indicated a sequence of defences on the same alignment from a late Iron Age 'A' boxed rampart to the Iron Age 'C' remains which are still seen. The tested occupation remains extend from an Iron Age 'A/B' phase down to the storming of the hill-fort by the Romans in A.D. 44 (I. A. Richmond, Hod Hill II (1968)).

The Iron Age Hill-fort (Plate 198) comprises two ramparts with an outer counterscarp bank on the N., E. and S., and a single rampart and ditch with a counterscarp bank on the W. On the N. and E. the defences measure 140 ft. to 150 ft. across, overall, with a main inner rampart up to 30 ft. high above its ditch bottom. On the S., where the external slope is less steep, the defences are bigger and measure up to 180 ft. across, with a main rampart 36 ft. high. The relatively slight defences on the W. are compensated by the very steep natural slope falling to the river. On the N. side of the hill-fort is a natural shelf above a steep fall; an unfinished N.W. outwork extends the defences to dominate this shelf at its widest point, but it ends 150 ft. short of the fairly sharp natural shoulder of the slope. There is no access to this outwork from the hill-fort interior. The enclosed area of 1½ acres is bisected by a low lynchet, perhaps dividing gardens associated with the Roman garrison. E. of this outwork there has been much disturbance, particularly by quarrying. On the S. and E. slopes of the hill are the remains of strip lynchets and later ploughing.

Of the two Iron Age entrances, to N.E. and S.W., only that to the N.E., the Stepleton Gate, has been examined by excavation. Some form of entrance may have existed here in the initial Iron Age 'A' phase but the present entrance represents an insertion which involved the building of a causeway across the original Iron Age 'A' ditch, and ultimately involved the construction of a hornwork 220 ft. long in the final Iron Age 'C' phase. The hornwork channelled all approach into an enfiladed position, after which entry was between inturned arms which extended back 60 ft. from the main rampart. The S.W. gate also has inturned arms but here the outwork, unfinished, consists of the outer (counterscarp) bank of the hill-fort, splayed away from the main defences; it has a causewayed and presumably unfinished outer ditch. The Hanford Gate at the N.W. angle of the hill-fort is a Roman insertion and the Ashfield Gate in the middle of the E. side is also Roman. The Home Gate on the S.E. is mediaeval or later; the lane leading to it from Stourpaine village is named 'Hod Drove' on the Tithe Map of 1841.

An almost continuous line of quarry pits, to provide additional material for the defences, lies everywhere behind the inner rampart, except along most of the W. side and a small part of the S. side. The pits cover some 6 acres and vary much in form and size; many of them are about 5 ft. deep. There are at least two phases of pit digging, one associated with an extension of the inner rampart in the Iron Age 'A/B' phase, and the other associated with the extension in the Iron Age 'C' phase.

The interior of the hill-fort was eventually covered by remains of Iron Age settlement, varying in type and density; they include hut circles, some with annexes and 'yards', 'working platforms', pits, mounds, and boundary banks and scarps. All the hut remains but one, No. 60, are circular. Hexagonal enclosure No. 36, which also contains a hut circle, is the only complete angular enclosure. In the 7½ unploughed acres that could be planned in detail there were forty-nine embanked circles or circular platforms that are likely to have been hut sites; a further sixteen hut sites might have existed. Some of the hut sites clearly cut into others, representing a sequence and also suggesting that conditions were crowded. A count of probable hut circles visible on various air photographs of the ploughed areas indicates a reasonable minimum for the whole hill-fort of about two hundred. The well established huts are represented by circles, generally defined by low banks about 9 ft. across, with a shallow outer ditch 6 ft. or so across, both broken by a simple gap entrance on the S.E. side. The circles vary in diameter from 22 ft. to 36 ft., measured from crest to crest of the banks; many are about 30 ft. across.

In the unploughed area, some of the hut circles have annexes marked by hook-shaped enclosures with ditches, running out from the parent circles. Two of these are associated with other features. Hut No. 168, a circle with an annexe on the N.E., has to the S. of it an open area of about ¹/8 acre bounded by a low scarp. S. of this is Hut No. 121, with an annexe on the S. containing two small pits. Immediately E. of Hut No. 121 a low scarp runs S.W.-N.E.; beyond lies a notable concentration of big storage pits. N. of the hut are three curved ditch or palisade lines, and W. of it, for 120 ft., is an open area, unusually clear of remains, as has been confirmed by proton-magnetometer. To the S. of this area is an unusual circle, 30 ft. across and defined by a narrow ditch or palisade trench, broken by opposed openings to W. and E. A third hut with an annexe, No. 56, fully excavated by Richmond, was slightly ovoid in plan and measured 19½ ft. by 17½ ft. internally; it yielded a substantial amount of Iron Age 'C' pottery, and a hoard of 117 sling-stones was found near the entrance. The annexe contained an area of heavily trampled subsoil, perhaps indicative of tethered animals, and post-holes which suggested a rough shelter. Hut No. 60, a sub-rectangular hut measuring 28 ft. by 18 ft. internally, had an annexe associated with a storage pit while the hut itself yielded horse trappings from near the entrance, one and a half 'currency-bars' and Iron Age 'Second B' pottery; the trench around the back of the hut and its annexe contained a palisade. Hut No. 43 had a small external enclosure at the N.W. and also was apparently defined by a palisade; it yielded both 'B' and 'C' pottery, iron sickles and a hoard of 218 sling-stones. The hexagonal enclosure No. 36 had a palisade trench with an entrance at the S.W., to the N. of pit No. 38. In the N.E. corner of the enclosure a ditch defined a hut circle 30 ft. across with an entrance on the S.E.; it yielded Iron Age 'C' pottery, often matching fragments that were found in the surrounding enclosure. This pottery is comparable with that from Hut No. 37, which impinges upon the palisade of the enclosure and must be regarded as an extension of the same complex. The hut within No. 36 and Hut No. 43 had hearths; Huts Nos. 37, 56 and 60 had none.

The importance of enclosure No. 36 is suggested by the roadway which, as air photographs show, apparently led to it from the Stepleton Gate. The road, one of two or possibly three from the Stepleton Gate, all built over (see Plate 198), had already been put out of use in Iron Age times. The concentration of ballista bolts found in Hut No. 37 and also ranging over the enclosure indicates the importance of the area as a Roman target.

In the unploughed area one hundred major pits are traceable. Some of them are marked by depressions as much as 2½ ft. below the surface of the ground. Not all of them were for food storage, as the puddling pit for clay noted below clearly shows. The proportion of pits to huts (say 2:1) is very small indeed. Some pits lie inside huts (e.g. No. 107) and these are usually on the side opposite the entrance, perhaps indicating that they belonged to those huts. Pits examined by excavation included No. 60c, used twice, No. 38, used for puddling clay as were two others in the Roman fort area, No. 17 and Nos. 15a, b and c. The last mentioned closely resemble Boyd-Dawkins's site No. 6, which also contained 3 pits, in one of which were found burials. Pit No. 15a produced Iron Age 'A' and 'B' pottery, No. 15b produced burials of 'A/B' date and No. 15c produced 'B' pottery and another burial. There are a number of working platforms and enclosures and it is clear that the quarry ditches were also extensively used as working areas. Working platforms lay to the E. of the twin shallow curves of a palisade which sheltered them from the prevailing wind; on excavation they yielded a bone weaving-comb, an iron bucket handle, the hook and socket of a sickle, and also Iron Age 'C' pottery.

None of the mounds has yet been dug but some of them are probably composed of heaps of refuse or perhaps of unused chalk from pits, as on site 15. The long banks are nowhere found complete; the slight bank and ditch which runs N.E. from the gap in the quarry pits on the S. side is destroyed, after 250 ft., in heavily ploughed ground. A similar feature, cut by later pits, can be seen on air photographs in the N.E. quarter of the hill, running for 200 ft. or so, almost parallel with the N. defences; there are also traces of other similar banks. The associations noted suggest a relatively early date, and their form suggests an analogy with the banks which divide the interior of Eggardon hill-fort (Dorset I, Askerswell (12)).

The Roman attack appears to have been limited to a bombardment of enclosure No. 36 with ballista-bolts; the bolts were confined to its immediate environment. There is no evidence of fighting at the gates, as at Maiden Castle. The attack was followed by the evacuation of the hill-fort and by the demolition of the huts.

The Roman Fort was built on the highest part of the hill, in an unusual position in the N.W. corner of the Iron Age hill-fort. The early date suggested by the ditch system of 'Punic' type, by the plan, and by resemblance to such forts as Valkenburg in Holland, was confirmed by the finds; these indicate that the fort was built c. 44 A.D. and was evacuated a few years later. The arrangement of the interior implies that it was garrisoned by a legionary detachment of some 600 men (a cohort) and by an auxiliary cavalry unit of about 250 men. After the site had been ploughed up in 1858 and in subsequent years, H. Durden formed the large collection of finds that is now in the British Museum. The fort was tested by excavation in 1897 by Sir T. Baker and it was thoroughly excavated in 1951–1958 by Sir Ian Richmond.

Including the defences the fort covers 11 acres, the interior having an area of 6.86 acres. It was defended on the S. and E. by a rampart of packed chalk 10 ft. wide, revetted at back and front with vertical turf cheeks and probably laced with horizontal timbers; the rampart now survives as a bank 15 ft. wide and 4 ft. high. The Iron Age defences were retained on the N. and W. Triple ditches on the S. and E. were interrupted by causeways, narrowing as they led to the gates in the centres of these sides. The inner pair of ditches lay immediately outside the rampart and returned on either side of the causeways, the innermost ditch returning yet again after a distance of 48 ft. and continuing as the outer ditch. Returns of the second and smaller inner ditch were also made where the ditch-system ran out into the native defences. At the S.E. angle an extra curved length of ditch was interposed between the middle and outer ditches to provide additional defence. A length of ditch and bank (tutulus) was formed beyond the outer end of each causeway as an obstacle to direct entry.

Of the three gates, that to the E. was the porta praetoria, with twin passageways 10 ft. wide flanked by towers 10 ft. wide and 20 ft. deep; that to the S. was the porta principalis dextra, with a single passage covered by a tower; that to the N.W. was similarly planned, and was cut through the Iron Age rampart and approached by a causeway across the Iron Age ditch. The anomalous position of the N.W. gate was presumably dictated by the location of the water-supply. Six post-holes for a tower 30 ft. wide, with angled front, were found at the S.E. corner. The E. and S. gates were flanked, to N. and E. respectively, by ballista-platforms attached to the back of the rampart.

Before the ploughing of 1858, lines of structures and roads were seen as earthworks, but by 1951 only the roads could be discerned. The buildings were separated from the rampart by an intervallum 50 ft. wide and lay in three groups, divided by the E.-W. via praetoria and by the N.-S. via principalis; they were constructed of timber framing and wattle-and-daub and their plans are indicated by post-holes and trenches for sleeper-beams. A fire destroyed the buildings near the S. gate, but even before that occurrence some buildings had required partial renewal, probably in consequence of green timber decaying.

The simple headquarters building was in the W. half of the fort, and to the W. of it stood the house of the centurion in command of the legionaries; both buildings had internal courts surrounded by porticoes. To the S. of these buildings were six legionaries' barracks of semi-permanent design, and store sheds; to the N. were six stables for the mounted section of the garrison. The cavalry barracks lay in the E. half of the fort, or praetentura; they comprised two buildings to the S. of the via praetoria, associated with a hospital, store buildings, a granary and latrines, and buildings to the N., associated with the large house of the praefectus equitum. A compound to the N.E. enclosed a timberlined water tank with a capacity of 1,500 gallons; it was constructed in the quarry ditch of the N. rampart.

Finds from the excavations and from the 19th-century ploughing included legionary and cavalry equipment. That the occupation was intentionally of short duration is shown by the semipermanent character of the barrack plans, by the sparse yield of pottery (exclusively Claudian) and by the early Claudian character of the coin list. Excavations in other parts of the hillfort showed that Iron Age huts had been destroyed by the Romans, and that part of the area was probably used as a parade ground. Three coins of Trajan and a little late 3rd or early 4th-century pottery from inside the hill-fort indicate a later, non-military occupation of a minor nature. Most of the finds are in the B.M. but some are in the D.C.M. and others are in the Pitt-Rivers Museum at Farnham. (Arch. J. LVII (1900), 52–68; Brailsford, Hod Hill I (1962); Richmond, Hod Hill II (1968)).

'Celtic' Fields, see p. 341, Group (61).

Round Barrows

A barrow at 88701191 has been destroyed in recent years. An unlocated barrow on Stourpaine Down yielded a whetstone and two bone pins (Durden Cat. Nos. 122, 137, 138).

(12) Bowl ? (85531076) lies at 470 ft. O.D. on the summit of Hod Hill, within the Roman fort and under the remains of the via praetoria. It is poorly preserved, having been ploughed in the past. Diam. about 30 ft., ht. about 1 ft.

Undated

(13) Enclosure (875094), visible only on air photographs (St. Joseph OAP. AT 40), lies on top of a small rounded hill at about 420 ft. O.D., immediately S.E. of Down End Farm Buildings. It is a roughly circular area of 8 acres bounded by a ditch 5 ft. to 10 ft. wide; there are traces of an inturned entrance on the S. side but there are no internal features. The site is one of a group of similar enclosures in Pimperne and Tarrant Hinton, to the E. (see Dorset IV, Pimperne (16–18), Tarrant Hinton (20–21)).