Click slide to view full size and download

Timpendean Tower and its Environs: an Historical Study

Hear the melody begin
At the door of Cleikiminn,
East a mile or so of Timpendean.
Drown the echo of your grief
In the joy of Lilliesleaf,
Or the merriment of Redfordgreen.
William Landles, Border Clachans
Situation and General Description
Timpendean Tower, a tower house [636 227 1] of sixteenth century date and described as a ruin by
1821 (Ainslie), stands on the NW slope of Lanton Moor, the high ground between the River Teviot and the Jed Water, at a height of around 135 m above sea level and 80 m above the Teviot, and at a distance of about 2 kilometres NW of Jedburgh. It is associated with conspicuous earthworks in large part of
earlier but uncertain date. The site lies at the NE end of 
the level Parkhead shelf where it terminates at the Timpendean Burn. The shelf itself runs some 400 m in a NEly direction from 632 223 and is about 30 m in width although it broadens considerably as it approaches the tower. The Parkhead shelf would have formed a natural approach to the site, being easily accessed from the broad, gently rising ground toward Lanton and leading as it does to the entrances both of the earthwork and the hollow way associated with the tower; this aspect will be considered more fully below.

The tower commands a wide view over the portion of Teviotdale between Ancrum and Minto and its occupants would have been well placed to monitor movements both along the well-drained
fluvioglacial terraces on either side of the valley and on the Salters' Road which ascends the NW side of the valley opposite Lanton, as well as being able to maintain visual contact with fortified sites at Ancrum, Lanton, Minto Crags and Barnhills. 

The site is most conveniently approached on foot from Timpendean Farm [628 232] on the A698, where vehicle parking unfortunately is not practicable. A farm track ascends the valley side for some 750m before making an acute turn to the NE to approach the environs of the tower along the Parkhead shelf for a distance of about 100 m. On meeting the ladder style over the drystane dyke which forms the boundary of Cow Park the track resumes its uphill course, away from the site.

Although level to the SW the site falls away to the NW, rises to the SE and to the NE is abruptly truncated by the steep sided course of the Timpendean Burn. This burn eventually joins the Teviot some 250 m downstream from the confluence of the latter with the Ale and can be traced to its source in the vicinity of the tower. This, a spring marked on the OS map,2 lies on a National Grid bearing of 140º from the tower, some 30 m or so short of
Fig. 1 Timpendean Tower (above) from the north and (below) from the south, with the site of the demolished extension in the left foreground. (DC)
the drystane dyke bounding the field to the SE and in an area of very disturbed ground 
largely obscured by mature whin (Ulex europaeus); this spring has been led off in the past by a red clay field drain, numerous fragments of which are scattered about the area. The Parkhead shelf to the SW
Fig. 2 Portion of 1791 Estate Map showing Timpendean Tower. Parkhead Shelf lies along the top of Mid and West Parks, the upper limits of which now lie 25 m to the north, i.e. downhill, of where indicated here. The bottom of the map faces 30° W of Grid North and measures approximately 750 m.       (Lothian Estates.)

is drained by a ditch which runs parallel to the short section of farm track mentioned above; on entering Cow Park this water course is culverted for approximately 100 m before springing in a marshy area within the earthworks about 7 0m from the tower on a grid bearing of 185° and becoming in effect a second source of the burn (fig. 3).

Drainage in the environs of the tower is extremely poor in places and is exacerbated by cattle tramp
so that the terrain is quite difficult in places.


The underlying geology consists of Upper Old Red Sandstone lithofacies, gently undulating and dipping at a low angle, generally in a roughly southerly direction,3 and largely overlaid by a thick deposit of chocolate-brown Devensian boulder clay4 deriving substantially from the underlying country rock. In the stream section, proceeding up the Timpendean Burn gully from its commencement [635 228], fissile red sandstone, unsuitable for building, is exposed in situ both in the stream bed and in the right bank of the gully where it is overlaid by boulder clay. The left bank, however, consists largely of clay mixed with pieces of pale buff stone, apparently quarry spoil, discussed again below, and deriving from a higher geological horizon than that exposed in the stream section and right bank. Some 200 m or so to the E of the tower there is a small doleritic intrusion of Jedburgh type, of Lower Carboniferous age;3 its presence is not conspicuous on the ground, however, and apart from the very occasional piece of field stone none of this relatively intractable rock type appears to have been used in the construction of the tower.

Fig. 3 The environs of Timpendean Tower, Cow Park. Orientation is as in Fig. 2.

Legend: d-d, drainage ditch; h-h-h, hollow way; l-l, line of culvert; r-r, rectangular feature; s-s-s, drystane dyke; t-t, turf dykes (selection); T, tower; E, site of demolished extension.


The name derives most likely in part from the Brittonic5 din, 'fort', and pen, 'hill', a probable reference to the series of earthworks stretching for some 2.5 kilometres NE along the ridge, on the NW slope of which the tower is situated, from Lanton Hill [622 205; 627 207; 629 208] via Lanton Moor [638 218] to a site [641 227] which is roughly equidistant from Timpendean Tower and Monklaw [648 227], lying due east. With Anglian immigration and ensuing acculturation from the sixth century onwards, the Old English denu, 'valley', would have become affixed, doubtlessly referring to the small, deep gully of the Timpendean Burn immediately NE of the tower. Timpendean has appeared as, inter alia, Tympenden (1540, in Paul & Thomson, 1883), Tempindene (1600), Timpendein (Blaeu, 1654), Typenden and Timpenden (douglashistory.co).

Brief Historical Outline

Timpendean Tower was originally part of the estate of the Douglas Lairds of Bonjedward. In 479 George Douglas, 4th of Bonjedward, with the agreement of his older son, James, made  Timpendean over to Andrew, his younger son. In 1540 William Dowglas (sic), 5th of Bonjedward, was granted a charter from James V which included ". . . the manorial lands of Bonjedward with the tower and woodland thereof, the estate and fields of B[onjedward], 21 agricultural fields and four adjacent fields and appurtenances, with corn and fulling mills, the lands of Timpendean, with tenants etc., in the regality and demesne of Jedforest . . ." (Paul & Thomson, op. cit.).6  In 1545 Timpendean was burned by an English force under the Earl of Hertford, during the Nine Years' War.7 In this context it is perhaps worth noting that, if in existence at this period, the deep depressions and spoil heaps across the burn and eastwards from the tower would have afforded both plenty of cover for an attacking force and advantageous positions for its artillery. In 1843 the lands of Timpendean were sold off by George, 12th of Timpendean, to William, 8th Marquess of Lothian, along with lands at Broomhall and Langton, and now form part of Lothian Estates (douglashistory.co, op. cit.).8


Discounting the probable quarry and spoil heaps already mentioned, and discussed below, these
consist of a conspicuous subrectangular system of ramparts and ditches with a central enclosure, a
hollow way and a system of turf dykes following a generally rectilinear layout.

(a) The system of ramparts and ditches would originally have measured some 130 m from SE to NW by between 100 and 130 m from NE to SW, enclosing an area 80 m from SE to NW by between 55 and 65 m from NE to SW, i.e. some 0.78 ha. The defensive earthworks originally consisted of double ramparts with an intervening ditch on three sides of the structure but with three ramparts and two intervening ditches on the SW side. Since the uphill, SE side would have been at least as vulnerable as the SW side, it is suggested that three ramparts were constructed to impress persons approaching along the Parkhead shelf rather than for any serious military reason. The innermost of these three ramparts has a short return to the NW side; whether this is all that was constructed, or whether it is what is left of a now levelled third NW rampart, is not clear; this side, facing as it does downhill, would be most easily defended so that an extra line of defence would be of dubious advantage. Much of the inner NE rampart and all of the outer one have been levelled, presumably in association with the construction and occupancy of the tower. Similarly, the outer rampart to the SE has been obscured by the construction of the hollow way.

The structure is one of several in the Borders which are difficult to assign to a precise period. Dates ranging from the early middle ages or, simply, the mediaeval period (RCAHMS, 1956), the 13th century (Salter, 1994) and pre-1545 (Cruft et al., 2006) have been mooted. What may be dismissed is the suggestion (douglashistory.co, op. cit.) that the structure could have been
Roman, with its stonework subsequently utilised in the construction of the tower; certainly no reused Roman masonry is anywhere evident on the site. It may be confidently stated that the
rampart and ditch system is the earliest structure at Timpendean but until archaeological investigation is undertaken a precise date can not be assigned, although a suggestion (Cruft et al., op. cit.) that the earthwork could have incorporated the house destroyed in 1545 is not implausible. Just discernable on the ground and hinted at by RCAHMS aerial photographs there is a rectangular feature, the only complete edge of which measures about 40 m, in the uphill portion of the enclosure and in a similar orientation to the main system of ramparts and ditches; this feature could conceivably date from the pre-1545 phase.

(b) Hollow ways frequently evolve as a result of the passage of people and animals over long periods of time, but the feature described here shows every sign of having been constructed
deliberately, probably at the same time as the tower. Cut into the hillside to the SE and contained by a prominent earth rampart to the NW which encroaches upon the rampart and ditch system described above, it runs for some 120 m initially from SW to NE but swinging round ever more sharply, ultimately to face the SE elevation of the tower. At first very shallow, by the time it faces the tower it is constricted between steep embankments some 2 m high. Any party approaching the tower from the Parkhead shelf would thus be confined very closely in full view of the SE gun loop which, based as it is upon a 26" equilateral triangle,9 would have a 60° field of fire, effectively deterring further movement until intentions were ascertained. It is significant that, of the four gun loops of the original sixteenth century tower, this is the only one to have survived later alterations, discussed briefly below; the SW one was obscured during the existence of the seventeenth century extension, with the NW being destroyed by the insertion of a window and the NE by that of a new doorway.

(c) The modern field, extending to slightly more than 9 ha, in which the tower and earthworks 
are situated, is traversed by a complex, largely rectilinear, system of turf dykes, almost certainly dating from the period during which the tower was occupied but not necessarily all of the same age, representing the boundaries of fields and other enclosures. It is possible that some of these were sighted on Gersit Law [612 268], the summit of which is now obscured by afforestation, but this may be merely fortuitous. Some, notably of the garden enclosure in which the tower itself stands, are indicated on the 1791 estate map (fig. 2).

The Tower

The tower stands on the NE side of the earthwork and has been laid out accurately to a rectangular plan. The NW elevation is exactly 29 feet in length and the SW 24 feet. The SE elevation is 29' 2" in length, an error of merely + 0.57 %, the NE elevation being 23' 9½", an error of - 0.87 %. The greatest angular error is at the N corner, 88° 53' being only 1° 7' short of a right angle, i.e. only - 1.24%.

During the seventeenth century the tower was extended, as indicated by projecting bond stones on the SW gable, as well as a recessed portion of that wall which would have accommodated a newel stair in the E angle of the addition (fig. 1). Evidence on the ground would suggest that this extension had a footprint similar to that of the original structure, so that the tower at its maximum extent would have occupied an area of 24 feet by almost 60 feet. Construction is of random pale buff sandstone rubble with dressed quoins of the same material; angular fragments of similar stone are to be found both in the spoil forming much
Fig. 4 Area of quarrying east of Timpendean Tower. (DC)
of the left bank of the Timpendean Burn and in the bed of the burn as far downstream as the point where it enters a lower field to follow an artificial course along a field boundary of eighteenth century date. It seems certain that the deep hollow and disturbed ground across the burn and in an approximately easterly direction from the tower (fig. 4) is where building material has been quarried. 

The original entrance to the tower was on the NW side, the elevation farthest from the end of the hollow way. This was by an arched doorway, 2' 8" wide, which had provision both for an inner door, 3' 1" wide, and an outer one, most likely a wrought iron yett, 3' 3" wide, the check for which shows evidence of systematic and deliberate damage. During the seventeenth century modifications a second doorway, 3' 0" in width and which would have accommodated a door of width 3' 5", was inserted into the NE side of the tower. This carries a 1", 45° chamfer on its entire margin and stands ½" proud of the wall, presumably allowing for rendering, now no longer in existence (fig. 5).was extended, as indicated by projecting bond stones on the

The seventeenth century extension to the tower was subsequently dismantled so that only the original sixteenth century structure now stands. It is not clear when this occurred; the representation of the tower in the 1791 estate map, while stylised, would suggest that demolition had not yet taken place, but it certainly had by the time of the 1859 Ordnance Survey. The stone was almost entirely removed from the site and it would appear to have been reused in the construction of Timpendean Farm (fig. 6), the successor to the tower, which had been built by 1821 (Ainslie, op. cit.). 

Text Box

Fig. 5 Details of doorways: (right) 16th c., and (left) 17th c. The damaged portion is indicated at X. (DC)

The detailed architecture of the tower has been described adequately elswhere (RCAHMS, Cruft et al.)

Summary and Conclusions

The oldest structure at Timpendean, the approach to which was from the southwest, is an earthwork comprising a system of ramparts and ditches. Opinion is divided as to its age but it seems likely to be mediaeval. It is possible that it formed the defences of the settlement destroyed in 1545 by an English force; it certainly appears to contain traces of a structure which merits further examination.

During the later sixteenth century the present tower was constructed, a considerable portion of the earlier earthwork probably being levelled at this
Fig. 6 Detail of masonry at Timpendean Farm, showing typical reused rubble stone from Timpendean Tower, with occasional dolerite fieldstone and quoins of pink Belses freestone. (DC)
stage and a number of turf dykes constructed. The 
approach to the tower was again from the southwest by an artificially constructed hollow way which opens directly on to levelled ground covered by the tower's southeastern gun loop. In the seventeenth century the tower was doubled in extent, a new wing being added to the southwest. A new doorway and windows were inserted and the only one of the four gun loops to be retained was the strategically important southeastern one. 

A map of 1791 seems to show the tower still with its southwestern extension intact. The tower was ruinous by 1821 and by 1859 the extension had been dismantled to leave only the sixteenth century portion still standing. The demolished masonry was almost entirely removed from the site, to be reused in new building projects such as the nearby Timpendean Farm.


The writer is grateful to Lothian Estates for access to an estate map dating from January 1791, a portion of which is reproduced in this paper. Thanks are also due to Fiona Morris of Lothian Estates and to Jim and Gavin Millar of Timpendean Farm, as well as to my colleague, Ian Thompson, and my son, Hugh Cain, both of whom uncomplainingly held the ends of surveying tapes.

1. 1:25 000 Outdoor Leisure Sheet 16, 1995.

2. ibid.

3. Geological Survey Sheet 17, Solid Edition, 1982.

4. Geological Survey Sheet 17, Drift Edition, 1982.

5. The term Brittonic has been used in preference to the terms Old Welsh, Cymric, Cumbric, British, Continental Celtic or P-Celtic. Similarly Old English has been used instead of Anglian or Anglo-Saxon.

6. i.e. this author's translation. The original entry, quoted in Paul & Thomson, reads ". . . terras dominicales de Bonejedburgh cum turre et nemore earundem, villam et terras de B., 21 terras husbandias et 4 terras contagias continen., cum granorum et fullonum molendinis, terras de Tympenden, cum tenentibus &c. in regalitate et dominio de Jedburgh-forest . . ." It is significant that this charter makes no mention of a tower at Timpendean.

7. The Nine Years' War, later to be known romantically as the Rough Wooing, was conducted with unparalleled savagery under Henry VIII of England and subsequently his son Edward VI, in an attempt to coerce the Scots into agreeing to a dynastic marriage between the latter and the infant Mary, daughter of the late Jam es V.

8. The now vanished farm of Broomhall, appearing in Ainslie's 1821 map and which was situated approximately at [635 236], 1 km N of Timpendean Tower, and Lanton, about 1.7 km to the SW and known as Langton into the 19th century, are indicated. The Broomhall on a bend of the Tweed opposite Maxton and the Langton (Langton Mill, Langton Burn) near Duns, may be discounted.

9. A North German foot of 335 mm, or 13.2", was introduced into these islands in the fifth century but some time around the latter half of the thirteenth century either Henry III or Edward I of England instituted the modern foot, exactly 10/11 of its predecessor, i.e. 304.5 mm, or 12" exactly. When the tower is surveyed it is clear that its builders were using this modern foot and its subdivisions so it will be more meaningful to retain its use in this context than to give metric equivalents.

Ainslie, J., Map of the Southern Part of Scotland, Macreadie Skelly & Co., 1821.

Cruft, K. et al., The Buildings of Scotland: Borders, YUP, 2006.

Paul, J. B., & Thomson, J. M., The Register of the Great Seal of Scotland, Vol. 3, 1513 - 1546, HM Register House, Edinburgh, 1883.

RCAHMS, An Inventory of the Ancient and Historical Monuments of Roxburghshire, Vol 1, HMSO, 1956. Some of this article is reproduced at http://canmore.rcahms.gov.uk/en/site/57087/details/timpendean+tower/ along with useful aerial photographs.

Salter, M., The Castles of Lothian and the Borders, Folly Publications, 1994.

David Cain Ancrum and District Heritage Society

To navigate to the site you can download a gpx file below, for instructions see the Visiting Sites page.
Border Archaeology,
5 May 2016, 06:17