OUCC Proceedings 4 (1966)

See also article in OUCC Proc 11, 1983

Swallet holes and other solutional features occur on two different Jurassic limestones where these outcrop in the Oxford area.

Coral Rag

The first group of features to be described occurs at the junction of the Coral Rag (Upper Corallian) and the Kimmeridge Clay where the latter feathers out onto the underlying limestone.

Such conditions are met in the area between Forest Hill and Shotover, two hills immediately to the east of Oxford, and in the area further to the west on the other side of the Thames (see map). Here, water from springs at the base of the Lower Greensand runs down the clay slopes of Cumnor Hurst and Pickett's Heath onto the Corallian plateau, promptly to disappear into fissures in the limestone. These fissures are known locally as "swilly" or "gulley" holes, and have been described by W.J. Arkell in his 'Geology of Oxford' and referred to in 'British Caving'.

While it may be difficult for the speleologist to find as much enthusiasm for gulley holes as Arkell seems to manage, it cannot be denied that they show several points of interest. As he says, the quantity of water that manages to disappear down them after a sharp rainstorm is certainly visual testimony of the amount of solution that must have taken place underground. Accordingly, it seemed that further investigation was necessary.

Excavation of a gulley hole situated at NGR SP 483039 at 395' OD was undertaken by the O.U. Cave Club. The result has proved interesting.

The entrance to the hole lies at the bottom of a ditch. This has a permanent flow of water, nourished by a spring, that is of quite large proportions in wet weather. When first observed, the entrance consisted of two tubes, connected and running vertically downwards, which were filled to within three feet of the surface with mud and debris. The larger of the two tubes was 1'6" in diameter. Removal of some of the surface soil showed that the entrance tubes are only two of several running along a 'joint' whose axis is approximately N-S. The surface of the limestone at this point is smooth and unfragmented.

The mud has now been removed from the hole to a depth of some 15'. Most was of recent origin but there was evidence of an earlier fill in alcoves and tubes in the walls. Many pieces of wood in the lower portion of the debris were slightly calcited, while occasionally a piece was completely petrified. A few animal bones were found mostly of sheep.

The hole descends vertically for about 9'. Below the entrance tubes there is a considerable widening. At the bottom of the vertical section it is just possible to squeeze under a rock bridge into a very small chamber. From this chamber a way leads off in a northerly direction, silted to within a few inches of the roof, which takes the stream. The total depth of the hole is about 15'.

Apart from its gross structure, the most interesting feature of the hole is the way in which, in the walls, the structure of the reef limestone is revealed. The hard corals, pale cream in colour and crystalline, stand out from the walls in relief where the interstitial filling has been etched out. Some fossils can be seen in the filling.

Further attempts to deepen the hole are unlikely to be successful. The indications are that the water table has been reached.

Arkell mentions a good example of a Coral Rag sinkhole near Forest Lodge, Shotover. This is situated in a copse 200 yds. south of the A 41 at NGR SP 5676 0726 at 340` OD and appears to be the only one in the area despite similar conditions further to the east. Likewise, the holes to be found between Cumnor Hurst and Pickett's Heath are developed in a very restricted area considering that conditions appear similar along the whole of the inner edge of the Corallian plateau.

The indications are that the answer to this anomaly probably lies in the nature of the Limestone itself. This, as with most reef limestones, is of two distinct types: the actual reef masses with coral still in the position of growth, and a bedded detrital deposit laid down in the gaps between them. This latter stone is known separately as Wheatley limestone. The maximum extent of the reefs is seen at Cumnor, while eastwards the proportion on Wheatley stone increases towards Wheatley where it reaches its maximum thickness. Development of discrete channels by solution appears to occur only in the coral masses of the reefs.

In the coral masses, the matrix is softer and more porous than the corals embedded in it. It is clear that the water has been able to seep into this filling and after removing it, enlarge spaces in between the coral beds in a manner analogous to joint enlargement in some bedded limestones. The Wheatley stone is porous throughout the whole mass and, like the Oolite, seems to be subject to decalcification by water soaking through as if through a sponge without preferential enlargement of cracks, joints, or cavities.

It is certain that many "fossil" holes exist that have long been choked with clay. In wet weather, water has been observed to sink in places where there is not the slightest surface indication of a solution channel, while there are several obvious depression at the bottom of shallow valleys where water must once have sunk but has now been diverted. An indication of the amount of mud and debris which is carried underground is obtained by examination of the active swallet at NGR SP 476035 at 390` OD.

Local tradition has it that the corpse of a duck dropped into one of the Henwood Farm sinks later reappeared at Sunningwell, but this tale seems unlikely to be true. Firstly, the chances of such a bulky object being carried so far in what is undoubtedly a phreatic system are very small and, secondly, the sinks if they were anything like they are today would have far too choked to permit its entry. The story seems no more than an interesting variation of the "shepherd and chaff" tale, recounted for so many of the sinks and resurgences of the Yorkshire Carboniferous. The destination of the water still remains a pr4oblem. The use of a dye or isotope tracer has to be vetoed on account of the possible pollution of wells and springs in the area. Other methods, such as the use of lycopodium, would, on account of the number of springs that are potential risings for the sinking water, involve very laborious sampling. The boggy nature of many of the springs makes propagation slow and presumably effectively filters the emerging water, while dye detectors become irreversibly stained after only a few hours of immersion. Pin-pointing of risings then becomes a matter largely of intelligent guess work.

None of the solution hollows on the Coral Rag can be described as dolines in the usual meaning of the term. This is probably due to the thinness of the limestone (about 30`) and to its generally unjointed character, explaining the absence of karst features one might be led to expect. The presence over much of the Corallian plateau of a tenacious clay means, in the majority of cases, the blocking of incipient solution pipes. Thus wide, shallow hollows and not funnel-shaped shakeholes are the rule.

Arkell suggests that the narrow, steep sided valleys that dissect the plateau may be the result of the collapse of caverns developed at the base of the limestone, but it seems certain that the valleys were in fact formed by spring sapping. At NGR 488043, the long valley which runs down towards the Hinkseys is, for the first part of its course, shallow and dry with no sign of recent stream flow. On reaching the 350` contour however, the valley abruptly deepens and widens, the sides become precipitous, and a stream emerges from a bog on the valley floor. This seems likely to be the water that sinks in the hole that was excavated. The steep sides of this and like valleys are accounted for by the relative resistance of the hard limestone to erosional processes compared to the soft underlying Calcareous Grit (Lower Corallian), while their narrowness is due to the tendency of the sandy clay of the latter to be rapidly entrenched, thus concentrating the extension of the valley headwards. The sharply convex slopes to be observed above the steep side scarps are due to cambering. The result of cambering on a large scale can be seen on the neighbouring Wytham and Beacon Hills, where the settling of a previously horizontal limestone bed into terraces down the slopes has been plotted by Arkell and others.

Cornbrash

Another group of swallets occurs at the junction of the Oxford Clay and the Cornbrash, under conditions very similar to those at Cumnor, about 3 miles NW of Oxford at North Leigh (see map). .

There are, however, several significant differences. Unlike the reef limestone at Cumnor, the Cornbrash is a very coarse, impure limestone, but which has the distinction of being very well jointed. Indeed in places it might almost be described as shattered. It is very much thinner than the Coral Rag at Cumnor, being at the maximum 10-15` thick. It is underlain by a bed of clay 2½` thick. The problem of interest is to understand how, in the formation of sinkholes, water has been able to penetrate this clay of the Wychwood beds, and through several other formations as well, down to the Oolite where, one presumes, it reaches the water table. The possibility of perched water tables seems ruled out, there being no spring line in the several places that the junction is exposed.

In contrast to the average Coral Rag sinkhole, those on the Cornbrash are rather more characteristic of the genus, the water sinking in well-defined and fairly steep-sided depressions. The flow of water, originating in a drain, that flows into one of the holes is quite impressive.

It is interesting to note, in reference to the map, how closely the line of sinkholes follows the edge of the clay but always a little way in, although it must be admitted that the boundary of the clay is very ill defined. Probably the best guide is the changing soil character, which is the reason for including field boundaries on the sketch map. Fields on the clay are smaller, sport ponds, and are generally laid down to permanent grass, while the lighter soil over the limestone downslope is devoted to cereal growing. The decreased necessity for drainage here due to the water sinking upslope allows fewer ditches and a larger field size. This situation is in interesting contrast to parts of the region to the North where heavy soils are usually of alluvial origin and restricted to the valley bottoms.

A clue to the formation of the swallets is seen in Breakspear's Quarry, East End, at the top of the steep hill that descends from the latter towards Ashford mill. Here, to quote, "are exposed remarkable gulls in which vertical Cornbrash, Kellaways, and Northern Drift are let down into the Forest Marble apparently through slipping of the strata". This conclusion needs qualification. Closer examination of the quarry reveals a small "cave" entrance in the east wall, about 3`6`` high, which narrows rapidly to a fissure and descends steeply a few feet in. This is clearly of solutional origin, and there are stalagmite flows on the walls. Examination of the Cornbrash exposed ten feet up the section shows a small channel running in at the base of the exposure, immediately above the clay bed. Clearly penetration of the clay has been effected in a region where it is thin, the two openings being part of a now dry sinkhole of the type seen still active at N. Leigh.

The destination of the water from these swallets is a mystery. Although the water is most likely to flow down dip, the clay cover to the South is continuous. Some rather more intensive work is about to be started by the Oxford Dept. of geography under the direction of Miss Sweeting which should throw light on this interesting region.

F.E. Sanders
Exeter College
Oxford
11th March 1966

References

Arkell, W.J., The Geology of Oxford, pp245-246 (1947)
Blake, J.H., G.S. Memoir The Water Supply of Berkshire p4 (1902)
Craven, S.A., J. Craven Pothole Club, 3 [5], 234-235 (1965)
Hull, E. & Whitaker, H., G.S. Memoir, Geology of parts of Oxon. and Berks., (13) pp3-9 (1861)
Richardson, L., Arkell, W.J. & Dines, G.E., Geology of the Country around Witney (236) pp. 74-80 & 85-98 (1946)
Warwick, G.T., British Caving, p.179, Cullingford Ed., Routledge and Kegan Paul (1962)