Investigating the Physical Evidence
The investigation of Ackling Dyke began with a straightforward principle: examine the surviving physical structure independently of the historical label assigned to it.
For nearly two centuries, archaeologists have identified substantial sections of the Old Sarum–Badbury Rings route as a Roman road. This identification appears in archaeological inventories, Historic Environment Records, and the scheduling of surviving earthworks [1–3].
However, identifying a feature as Roman does not automatically establish its original construction date, engineering purpose or relationship with neighbouring structures.
Our investigation therefore examined approximately 64 kilometres between Old Sarum and Dorchester, including the traditionally recognised Ackling Dyke alignment and its continuation beyond Badbury Rings [4].
Rather than assuming that every surviving section represented the remains of a single road, we investigated the physical characteristics of each identifiable feature and its relationship with the surrounding landscape.
The study combined modern digital terrain mapping with historic archaeological records, measured earthwork profiles, aerial photography and landscape analysis [4–6].
The purpose was not to begin with a preferred interpretation, but to establish what the surviving evidence could demonstrate.
LiDAR — Revealing the Surviving Landscape
Our principal mapping tool was LiDAR (Light Detection and Ranging), which provides detailed measurements of ground-surface elevations.
By displaying these measurements as shaded-relief maps, features that can be difficult to recognise at ground level become more apparent.
These include raised banks, abandoned trackways, ditches, cuttings, terraces, extraction hollows and the remains of former watercourses.
LiDAR is particularly useful for examining linear earthworks because it lets you follow individual structures over long distances and compare them with the surrounding terrain.
It also reveals changes in alignment, width and construction form that may be overlooked when a route is represented simply as a continuous line on an archaeological map.
For this investigation, the LiDAR evidence was used to trace the conventional Old Sarum–Dorchester corridor and identify the surviving morphological characteristics of individual sections [4].
Where possible, the results were compared with historic maps, published excavation plans and existing archaeological records [1–6].
However, LiDAR does not independently establish an earthwork’s age or original purpose.
A ditch visible in a terrain model could represent drainage, quarrying, a boundary, a transport channel or another activity.
Its value lies in measuring what exists, not automatically determining why it was constructed.

Route-Wide Morphological Classification
To compare the corridor systematically, the mapped alignment was divided by surviving physical characteristics.
Four principal classifications were used [4]:
1. Wheeled-width road
A surviving surface or formation broad enough to support wheeled transport.
This classification identifies physical width and form; it does not automatically establish Roman construction.
2. Narrow surface
An identifiable track or raised formation narrower than the principal wheeled-width classification.
These features may represent narrower roads, tracks, paths or surfaces subsequently modified by later activity.
3. Ditch or channel
A surviving linear depression whose morphology is more consistent with a cut feature than a raised running surface.
The classification does not assume the feature was originally navigable, used for drainage or constructed for any particular purpose.
4. No identifiable surviving route
Sections where no corresponding physical feature could be confidently identified in the available terrain mapping.
Such gaps may reflect erosion, agricultural disturbance, later development or the absence of original construction.
The classification produced the following results across the approximately 64-kilometre Old Sarum–Dorchester corridor [4].
| Classification | Percentage |
|---|---|
| Wheeled-width road | 24.3% |
| Narrow surface | 21.7% |
| Ditch or channel | 6.2% |
| No identifiable surviving route | 47.8% |
| Total | 100% |
Within the conventionally recognised Old Sarum–Badbury Rings section, only approximately 22.1% was classified as wheeled-width surface, while 63.9% lacked an identifiable surviving route [4].
These figures describe the present visibility and morphology of the surveyed corridor.
They do not establish the original form or date of sections that have disappeared.
Nevertheless, the results demonstrate that the surviving remains are far from physically uniform.
The important distinction is between a route represented as continuous on archaeological maps and a continuous surviving piece of engineering on the ground.
Section 4 presents the full geographical distribution of these classifications.
Measuring Banks, Ditches and Road Surfaces
Where sufficient terrain detail was available, the investigation compared the dimensions of surviving features using transverse profiles and longitudinal elevation measurements [4].
The principal measurements included:
- Width and height of raised banks.
- Width and depth of adjoining ditches and cuttings.
- Width of identifiable road-like surfaces.
- Differences in elevation between neighbouring features.
- Changes in gradient along individual alignments.
- Relationships between raised surfaces, surrounding ground levels and nearby palaeochannels.
These measurements matter because a metalled running surface and the much larger earthwork beneath or beside it are not necessarily the same construction.
A relatively narrow road surface could have been installed over an existing bank, or a bank could have been constructed specifically to support that road.
Both possibilities could produce superficially similar remains today, despite implying very different construction histories.
Historic England records substantial surviving sections of Ackling Dyke between Oakley Down and Handley Down, including raised construction approximately 13 metres wide and 1.2-2.5 metres high [2].
These dimensions can be compared with excavated road surfaces elsewhere along the corridor.
At Thorncombe Wood, for example, archaeological investigations recorded a metalled running surface approximately 3.3–3.6 metres wide, with evidence of wheel tracks [6,8].
The contrast raises an important engineering question:
Are we examining variations of one road-building method, or physically different structures that have subsequently become associated with the same route?
Measurements alone cannot answer this question, but they allow us to examine competing explanations quantitatively rather than through visual impressions.

Historical Excavations and Archaeological Records
Modern terrain mapping was compared with published archaeological excavations, historical surveys and official monument records.
Particular attention was given to General Pitt-Rivers’ nineteenth-century investigations at Woodyates and Bokerley, published in Volume III of Excavations in Cranborne Chase [5].
These excavations are especially valuable because the original drawings record the physical relationships between banks, ditches, drainage features and road-like construction.
Unlike a modern map showing a single line labelled Roman road, Pitt-Rivers’ excavation plans distinguish individual structures and provide recorded section levels.
These include the Fore Drain, Boundary Drain and Cross Drain, together with evidence relevant to the associated road and earthworks [5].
The recorded differences in ditch-floor elevations provide an opportunity to investigate possible drainage relationships and construction sequences.
One comparison identifies an approximately 6.6-foot difference between recorded ditch-bottom elevations [5].
This does not independently demonstrate that the earthworks were canals, but it does provide measurable evidence that different components occupied different levels within the same landscape.
The investigation also examined archaeological work in Puddletown Forest, particularly the detailed landscape survey undertaken by Fletcher and the earlier Thorncombe Wood excavation associated with W. Putnam [6,8].
These records are important because they demonstrate that metalled wheeled surfaces existed along parts of the traditionally recognised route.
However, the presence of such surfaces does not automatically establish the original date or purpose of adjoining banks, ditches, terraces or extraction features.
Historic England scheduling descriptions and Dorset Historic Environment Records were also consulted to establish the conventional interpretation, recorded alignments and recognised archaeological characteristics of the surviving monuments [1–3].
The investigation therefore distinguishes between what earlier archaeologists physically excavated and the broader historical interpretations subsequently applied to those discoveries.

Testing Alignments and Landscape Relationships
Beyond measuring individual structures, the investigation examined how the route interacted with its surrounding landscape.
This included relationships with prehistoric monuments, extraction areas, valley systems, palaeochannels and changes in terrain elevation [4].
At Wyke Down, for example, researchers identified and measured two contrasting alignments.
Track 3 turns approximately 17.8° towards the Sixpenny Handley landscape, while Track 4 continues on a considerably straighter course towards Gussage All Saints [4].

Their geometry was compared with the surrounding prehistoric monuments and the relative elevations of the two possible terminal landscapes.
At Moor Crichel, near The Rookery, attention focused on a substantial cut-and-bank feature descending from an extraction landscape towards a palaeochannel.
At Sovell Down, the investigation examined the recognised route diversion around a substantial extraction feature [4].
We treated these relationships as physical observations requiring explanation, rather than automatic evidence of a particular construction date.
Where an alignment appeared to approach a former watercourse, we considered its elevation and surrounding landforms in relation to the proposed hydraulic transport model developed in The Post-Glacial Flooding Hypothesis [7].
However, identifying an apparent connection with a palaeochannel does not, on its own, establish that the feature contained navigable water at the time of construction.
That would require additional evidence concerning water supply, channel gradients, sedimentary deposits and chronological relationships.
The purpose was therefore to identify where the surviving landscape provides opportunities to test competing interpretations.
Separating Observation from Interpretation
Throughout this investigation, we distinguish evidence into four categories.
Observed
Physical features visible in LiDAR, aerial photography, ground photographs or measured terrain profiles.
Examples include identifiable banks, ditches, changes in route alignment and the dimensions of individual earthworks.
Documented
Findings recorded in published excavations, archaeological inventories, historic maps and official records.
Examples include Pitt-Rivers’ measured excavation sections and the recorded Thorncombe Wood metalled running surface [5,6,8].
Inference
An explanation reasonably derived from physical or documentary evidence but not independently demonstrated.
For example, two differing alignments may suggest more than one construction phase, but their geometry alone cannot establish which was earlier.
Working Hypothesis
A proposed construction history or function that requires further testing.
The suggestion that some substantial Ackling Dyke earthworks formed part of a prehistoric extraction and water-transport system falls into this category [4,7].
This distinction prevents treating the presence of a ditch as proof of a canal, just as it prevents treating a metalled surface as proof that every associated earthwork was originally Roman.
Both interpretations must be evaluated against the available evidence.
Applying the Same Test Across the Entire Route
The final methodological requirement was consistency.
The investigation applied the same classification principles and physical measurements to the conventional Ackling Dyke route north-east of Badbury Rings and to its recognised continuation towards Dorchester [4].
This comparison is particularly important because the southern section includes archaeological evidence for metalled road construction and wheeled traffic at Thorncombe Wood [6,8].
By examining both sections under the same criteria, it becomes possible to determine whether the surviving structures exhibit comparable engineering characteristics or major differences.
The investigation does not assume that such differences necessarily prove separate construction periods.
Instead, it asks whether terrain, geology, maintenance, erosion, and known Roman engineering practices can adequately explain the variations— or whether an inherited, multi-phase infrastructure corridor provides a more coherent explanation.
The objective is to determine what the surviving physical evidence actually demonstrates before deciding which historical interpretation best accounts for it.
Section 4 brings together the complete route-wide classification, providing a geographical overview of the corridor before the investigation proceeds from Old Sarum towards Dorchester.
References
[1] Royal Commission on Historical Monuments (England). (1975). An Inventory of Historical Monuments in the County of Dorset, Volume V: East Dorset. London: HMSO, pp. 33–34. Associated Dorset Historic Environment Record MDO2228.
[2] Historic England. Part of Ackling Dyke (Roman road), including Roman road on Oakley Down. National Heritage List for England, Scheduled Monument 1003309.
[3] Dorset Council. Roman Road from Old Sarum to Badbury Rings. Dorset Historic Environment Record, MDO6400.
[4] Langdon, R. J. (2026). Ackling Dyke: Old Sarum–Dorchester Route-Wide LiDAR and Morphological Investigation. Original research dataset, route classifications and measured profiles. Unpublished; supporting tables and figures accompany the investigation.
[5] Pitt-Rivers, A. H. L. F. (1892). Excavations in Cranborne Chase, near Rushmore, on the Borders of Dorset and Wilts. Volume III: Excavations in Bokerly and Wansdyke, Dorset and Wilts, 1888–1891. Privately printed.
[6] Fletcher, M. (2004). The Ackling Dyke Roman Road in Puddletown Forest, Dorset. English Heritage Archaeological Investigation Report 37/2004; Historic England Research Report 103/2004.
[7] Langdon, R. J. (2026). The Post-Glacial Flooding Hypothesis: The Evidence. Version 3.4. ABC Publishing Group.
[8] Putnam, W. (1968). Thorncombe Wood road excavation, discussed in Fletcher (2004). Full primary publication details remain to be verified.
Continue Exploring Ackling Dyke Reconsidered
Follow the investigation as we examine the surviving earthworks, their engineering characteristics and the evidence for different construction phases.
⬅️ Previous Page — 2. The Question — Is Ackling Dyke Really One Road?
➡️ Next Page — 4. The Whole Route at a Glancee
