The behaviour of Offa’s Dyke on Rushock Hill, north of Kington in Herefordshire, is crucial for understanding the monument’s broader military, political and socio-economic significance. I recently walked the dyke on Rushock Hill and tried to understand its behaviour as it dramatically shifts alignment as it ascends and encircles Herrock Hill and then sweeps eastwards, way off its usual path, to conduct an angle turn at the highest point on Rushock Hill. In doing so, it funnels movement to a point that enjoys long-distance views both north and south. It then heads south-east, then east, before turning north-east and descending to Kennel Wood where it is lost. As a result, there has been intense discussion by those few commentators who have, over the last century, attempted to understand Britain’s largest human-made feature from before the modern era.

Looking west, Offa’s Dyke skirts the southern slope of Herrock Hill through thick fern-cover. In doing so, it is acting remarkably different from its behaviour in relation to previous prominent hills on its N-S route. If it had acted ‘predictably’ it might have continued to the left to the neighbouring Bradnor Hill, keeping its west-facing stance. Instead it heads east, giving it a south-facing stance at this crucial point of intersection between the Wye valley to the south and the Walton Basin to the west.

Subsequently, I’ve been trying to get my head around what I saw in relation to what I’ve read. I cannot say it all fully matches up. Here’s my provisional understanding of what has been said, followed by what I think is going on. First, let’s review what each previous commentator has said about how the Dyke behaves on Rushock Hill, focusing on the key issue of how the dyke leads off Rushock Hill in an exceptional direction: to the north-east.

Looking south-west, Offa’s Dyke runs west-east from Herrock Hill to Rushock Hill, heading for the highest point on the latter
Annotated section of the 1890 OS map showing Offa’s Dyke heading south, then skirting Herrock Hill before its unprecedented journey eastwards to Rushock Hill and even more unprecedented turn north-east downslope into Kennel Wood where its course is lost


Cyril Fox felt he could trace the Dyke off Rushock Hill to the east/north-east, but has nothing to say about the striking nature of this north-easterly turn:

From the ‘Three Shepherds on Rushock Hill the Dyke is of normal dimensions … and with a S. ditch, for 320 yards. The Dyke then suddenly becomes small … and irregular, and the ditch shifts from S. to N. These characters persist for over 150 yards; then the Dyke becomes again clearly defined with a well-cut N. ditch…. So it continues, on a steppening slope, to the Kennel wood boundary: through the wood it is present as a low rounded bank with N. ditch. In the meadow beyond, the levelled Dyke can be traced nearly to the parish boundary-a small streamlet; there is no sign of it on the ploughland forming the N. flank of the little valley. The level at which it fades out is 810ft (Fox 1955: 175).

Crucially, Fox dismisses an earthwork – the Scutch Ditch – as part of the Offa’s Dyke monument, lying only a short distance across the valley on the opposite south-facing slope. This is despite stating that:

Its form is that which Offa’s Dyke, or any earthwork of defensive type, might take on such a slope-upper bank, steep-scarped ditch: slight bank on counterscarp (Fox 1955: 177).

In short, I’m left confused regarding what Fox thinks of the dyke in this area.

Looking south-west to Hergest Ridge, Offa’s Dyke ascends Rushock Hill



Tracing it from south to north, rather than from north to south, Noble has a different take. Regarding the route from Flintsham to Kennel Wood as ‘very uncertain’, he suggests that if it had crossed to incorporate Scutch Ditch, it would give a ‘more commanding line than that of the stream, but scarcely sufficient to justify the extra detour’ (Noble 1983: 29). He cannot therefore explain why it drops off Rushock Hill to the north-east:

The course and construction of the recognisable Dyke above Kennel Wood along the top of Rushock Hill to the curious angle which it makes at the summit are, however, equally difficult to explain in terms of any of the considerations which appear to have determined the siting of the Dyke elsewhere (Noble 1983: 29).

The adjusted-segmented nature of Offa’s Dyke as it ascends Rushock Hill, looking south-west

If it had stayed heading east, it would have had a strategic view over the Hereford plain. Instead:

From almost any point along it a stronger line towards the Lyonshall earthworks could be laid out than any which can be suggested in relation to the weak boundary bank to the north side of Rushock, which turns away down into the Eyewood valley as a feeble bank with a ditch on its north side: a course and construction which represents a reversal of the normal characteristics of Offa’s Dyke. It is tempting to suggest that the original builders may have missed their way, perhaps almost as badly as some later people, such as Fox, following their work down from the north, have gone astray when confronted with the problems of the plain of Hereford (Noble 1983: 29).

Looking north-east, the dyke ascends to the highest point on Rushock Hill before angle-turning south-east

The dismissive take on the earthwork a ‘feeble’ and the suggestion that the dyke-builders just got confused is difficult to match up with what I saw on the ground. Moreover, Noble argues for a timber palisade across the Herefordshire plain, dismissing Fox’s idea of inpenetrable woodland. But he isn’t clear as to which trajectory he prefers other than reference to Fox’s ambiguous comments.

Hill and Worthington

Hill and Worthington are keen to address the Rushock Hill section because it is pivotal to their argument that Offa’s Dyke did not run from ‘sea to sea’ but extended from Treuddyn in Flintshire to Rushock Hill and no further. In describing the overall route, they state:

The Ordnance Survey maps show an earthwork extending into the woods on a north-east alignment, but our fieldwork and excavation have shown this to be a false line following a field boundary (Hill and Worthington 2003: 50).

The phenomenal views southwards from Rushock Hill over the Herefordshire plain

In characterising Fox’s view, they seem to misread him, suggesting he is in support of an idea that Scutchditch was part of the Dyke. As we reviewed above, he does not make clear where he thought Offa’s Dyke went.

They then outline their own fieldwork. They made an intervention (site 90) at the base of Kennel Wood and no signs of the Dyke were found. Geophysical survey and visual examination failed to link the postulated Offa’s Dyke in Sheepwalk and the Scutchditch or beyond the Scutchditch to the plain.

The angle turn at the top of Herrock Hill at a point affording views north, west, south and east.

They then tested another option: that Offa’s Dyke actually continued east. They report briefly on their interventions to ascertain what happens to it further east. First, for site 140, they test the possibility in response to aerial photographs by Chris Musson that the Dyke continued beyond ‘Box Folly’ and instead found geological features. They then explored an unclearly located site 141 to ‘test Fox’s suggestion that the Dyke had turned into Kennel Wood’. They concluded from their trench that they:

‘failed to locate any feature consistent with the ditch of Offa’s Dyke. The apparent northern ditch and southern bank described by Fox would seem to be caused by a field boundary of the now deserted Box Folly Farm. A ditch of Offan proportions is clearly visible on the main alignment of Offa’s Dyke on Rushock Hilll and it would seem therefore that the Dyke does not continue directly down the hill to the Herefordshire Plain, nor, unless it dramatically changes its character, does it continue into the Kennel Wood. In fact it appears to stop on the hillside.

Fox had already attempted to continue Offa’s Dyke to the south, beyond its apparent end on Rushock Hill. His suggestion was that the Dyke turned east to run diagonally down the western slope of the Eyewood Valley though [sic] Kennel Wood and the Fox map shows traces of the Dyke in the wood as far as the valley bottom in Sheepwalk, although he found no trace here. His suggested line for Offa’s Dyke would then proceed across the valley floor of Sheepwalk to join with the short length of the Scutchditch on its eastern side before it was presumed to run down onto the Herefordshire Plain proper. As in Sheepwalk, Fox found no trace of an earthwork in the kilometre and a half between Scutchditch and the next point to the south that he identified as Offa’s Dyke.

From this evidence, they conclude that the dyke stopped on Rushock Hill at a point where it entered into the tree-line of Kennel Wood.

Looking south-east, Offa’s Dyke runs away from the highest point on the hill

Ray and Bapty

Ray and Bapty discuss this section as follows. From Burfa it:

…crosses the Hindwell Brook before, half a mile (0.8km) further on, climbing nearly vertically up the side of a spur of a hill directly opposite. Here it turns abruptly south up the narrow spin of the spur itself and then winds upwards around the precipitous west-facing slope of Herrock Hill. The Dyke traces a prominent half-mile course below the summit of Herrock Hill but rather than crossing a deep valley and ascending around Bradnor Hill immediately to the south as it might be expected to do, it heads due east and crosses a saddle that links Herrock Hill and Rushock Hill. From the top of the latter hill it unusually provides a view both northwards and south (Ray and Bapty 2016: 46).

Looking west, the dyke checks itself back on an easterly route

Unfortunately, Ray and Bapty’s text omits discussing where the dyke goes next on Rushock Hill itself and their map (Figure 1.24) provides insufficient detail to describe its route either, providing a conjectural dotted line for the dyke’s progress towards the River Arrrow that is counter to its direction of travel north-east off Rushock Hill. This is made more confusing by a small gap in their maps’ coverage around the River Arrow: i.e. between Figure 1.24 and 1.26, making it difficult for the reader to navigate what is going on in this area.

The crucial craziness of Offa’s Dyke on Rushock Hill. Rather than heading straight on and keeping the ridgeline, or tacking right to descend to the River Arrow directly, Offa’s Dyke falls off northwards toward Kennel Wood, making it look uphill and ‘blind’ to the south.

They return to Rushock Hill in Chapter 4 to regard it as a zone where it adapts its ‘major stance’ from facing W to facing SW as it navigates the Herefordshire plain to face the Black Mountains (Ray and Bapty 2016: 128):

Having made its extraordinary loop to overlook the Walton Basin and link onto Rushock Hill, the  Dyke faces in part south-westwards towards another major upland massif: this time the east-facing scarp of the Black Mountains.

From Rushock Hill, the course trending south-east is resumed (if the Dyke did once exist continuously here), towards the lower slopes of Burton Hill/Ladylift in north Herefordshire.

Offa’s Dyke descending off Rushock Hill to the north-east, looking south-west uphill onto Rushock Hill.

They assert that:

The consistency of the sighting here suggests that even if the  Dyke was not completed across the north Herefordshire landscape, it was planned to be continuous across this area. Along most of this stretch towards the north bank of the Wye, it would have faced south-westwards towards the east-facing scarp of the Black Mountains.

Offa’s Dyke adjusting itself as it hits Kennel Wood

I need to read further through Ray and Bapty’s book to see if they address this point elsewhere. However, if we put these different accounts together, I think I’m justified in remaining somewhat confused regarding what might be going on here with this early medieval linear earthwork.

Some thoughts

It’s clear that, to date, no one has satisfactorily worked out what is going on with the dyke as to descends from Rushock Hill to the north-east. Did it stop, at a point that coincides with a post-medieval forestry boundary? Or did it continue? If it continued: where did it go?

As Noble observes, it is strikingly irrational for Offa’s Dyke to depart from its line and vantage to the west or south to drop down north-east with higher ground to its south-east. Heading north-east is just plain bizarre at this point rather than heading along the top of the ridge to the east to keep its position of observation to the south and then east. This suggests very strongly something that previous commentators have long argued: something special is going on here!

The dyke descends to obscurity in Kennel Wood

What needs to be asserted quite clearly is that the north-east turn of the Dyke as it heads down to Kennel Wood is not ‘feeble’. It is made of straight sections in the very same way as the Dyke is composed across Herrock to Rushock. In other words, there’s no doubting this is indeed the original line of the Dyke.

What might we conclude from this strikingly unusual behaviour following a key angle turn at the highest point on Rushhock Hill? One explanation was that this in-turn does indeed represent the southernmost end of the monument. It curves east and then north-east to conclude itself and defend Rushock Hill as a southernmost vantage point with views in most directions. Putting Herrock and Rushock Hills together, the dyke defends a pairing of hilltops dominating landscapes in all directions: to the north, west, south and east.

An alternative view is that it is indeed controlling these hilltops from all sides, but it was negotiating them in order to push east before hitting the River Arrow at a key strategic point of equal significance. From there it could head south to the Wye, incorporating the stretches of dyke at Lyonshall. The north-east turn therefore mirrors its behaviour immediately to the west, but whereas it turns north-east to reach the highest point at Rushock Hill, it is now turning north-east to head towards the next hillside across the Sheepwalk.

Current evidence doesn’t allow us to conclude between these options, despite Hill and Worthington’s interventions and Ray and Bapty’s recent evaluations. I’m minded to suggest that it might well have continued and that its push north-east and then east before turning south-east to continue its stance. This adaption may have been intended to control the ultimately significant hilltops of Herrock and Rushock as well as to configure its approach to the River Arrow. It might have been intended to strategically control approaches to a specific point on the River Arrow where the Dyke can dominate the river for a long stretch, but that is for a separate discussion. Only further fieldwork and perhaps interrogation of LIDAR might tell…