Collins New Naturalist Library. Philip Chapman

South Wales. In the Northern Pennines alone, there are 1800 or more documented cave entrances and around 350 km of explored and mapped cave passages – while current estimates give the total length of mapped passages in Britain and Ireland at somewhere around 800 km. Cave exploration is still in a very active phase in these islands, and significant new discoveries are still being made. For example, recent explorations in Ogof Daren Cilau beneath Mynydd Llangattwg in South Wales have yielded over 20 km of new discoveries in less than two years, including the largest passages yet found in Britain.

      By any standards, 800 kilometers of open cave passage represents a significant habitat, worthy of the attention of naturalists – yet passages of explorable size must form but a tiny proportion of the total cave habitat, the vast majority being of mesocavernous dimensions. In the absence of data, I would guess that the habitable surface area within the mesocaverns of limestone terrains must run to at least two or three orders of magnitude more than that within explored caves.

      So what is it which makes limestones so spectacularly cavernous? To understand the process in which caves are formed, we must begin by examining the origins, nature and structure of the limestone rock itself. Limestone is a sedimentary rock, that is, it began life as suspended particles in an ancient sea, gradually settling to the ocean floor millimetre by millimetre over millions of years. During this inconceivably long period, there were intervals when conditions changed enough to interrupt the steady downward rain of lime, allowing some other type of deposit to intervene briefly in the sequence of otherwise pure calcium carbonate. Aeons later, and now hardened to rock, these geological glitches have become ‘fossilized’ as bedding planes – horizons of weakness between the solid layers of limestone.

      The Carboniferous Limestones of the British Isles were laid down somewhere in the tropical seas of the southern hemisphere. Pushed along on a northward-drifting chunk of continental crust, they have had a bumpy ride. Some, like those of the Yorksire Dales, have survived their 340 million year journey the right way up, though somewhat fragmented by massive vertical faults. Others have fared less well. The Mendip limestones lie like a wrecked car, buckled and perched at a steep angle, so that the bedding planes dip downhill at an average gradient of 50° or so. In all cases, the rough ride has produced vertical stress cracks, called joints, which link with faults and with the original bedding weaknesses of the limestone to form a boxwork of crevices reaching from the highest hilltop to beneath the deepest valley.

      Limestone is a strong rock and so frequently forms upland regions. Solid limestone is impervious to water, but water is able to flow through the cracks within it. It is these cracks which are the key to understanding the origins of caves. Limestone caves form principally by means of a simple chemical reaction in which hydrogen ions from groundwaters, acidified with dissolved carbon dioxide, act on the relatively insoluble carbonate ions in the limestone to produce soluble bicarbonate 10ns which are then flushed away. The reaction renders the limestone 25 times more soluble than it would be in pure water and the result is holes.

      Some of the carbon dioxide in groundwaters is collected by raindrops falling through the atmosphere, and some from the breakdown of organic material picked up as the rain then trickles down through the soil. Immediately beneath the soil, the weathered surface layers of rock are more fractured than those at a greater depth and the acidified, aggressive soil waters have their maximum impact here – so that at any one time, up to 15% of the volume in the top three metres of limestone may be occupied by air-, or water-filled spaces (Stearns, 1977 – figures for Central Tennessee, USA). These mesocaverns act like the guttering beneath the roof of a house – collecting soil water and quickly conveying it to natural drainpipes, often developed on the intersections of major joint fractures, or steeply inclined bedding planes.

      The flow of water downwards into the limestone carries with it sediments which contain particulate organic material, dissolved organic acids and microorganisms (bacteria and fungi). Decomposition continues way below the soil in the cracks and crevices of the limestone. There, to paraphrase Hoover’s famous advertising slogan, groundwater micro-organisms ‘eat-as-they-seep-as-they-clean’, mopping up the organic impurities and excreting CO₂ – in effect arming their liquid medium with chemical teeth. In the larger conduits it may take up to 50 days and several kilometres of flow before the bacteria finish mopping up the water-borne food, and even longer and further before the chemical aggression of the cave water is finally spent. All this time, limestone is being steadily corroded, and the cracks along which the water travels widened, resulting in the slow opening of a complex drainage network reaching deep below the water table.

      The initial pattern of flow within the flooded cracks is dictated by the structural geology of the limestone and the shape of the land surface. Between them, these two factors determine where water will escape from the rocks as a spring, and lay down the blueprint for the caves to come. If the rock strata lie horizontally, as they do in the Yorkshire Dales, water is forced to follow a rectangular course down vertical joints and along short sections of horizontal bedding, producing the characteristic stepped profile of a ‘pothole’ system. If the strata are inclined, as in the Mendip Hills, drainage will alternate between short joint-controlled shafts and longer bedding-plane slopes, producing caves with a steep profile, but few vertical sections.

      At or below the water table, the course taken by percolation water is determined by the direction of the hydrological gradient between where the water goes into the permanently-flooded system of cracks known as the ‘phreas’ and where it comes out again as a spring. Within the phreas, water is free to follow along, down or up the 3-D maze of cracks to produce the smoothest possible overall flow. Where the rock beds lie horizontally the smoothest profile may be along just one bed, producing a horizontal cave which may run for several kilometers (Yorkshire has many such systems). Where the rocks dip steeply in a direction which does not coincide with the hydrologically-determined direction of flow, the groundwater may be forced into a series of vertical z-bends, down joints and up the bedding, to tack its way to the spring (a classic example is Wookey Hole in the Mendip Hills). Where the limestone beds are trapped in a syncline or U-bend beneath impermeable rocks, water may be forced to travel down to great depths following the configuration of the rock strata. Chinese geomorphologist Yuan Daoxian has recently reported the discovery of a substantial cave at a depth of 2900 m below the water table in the Sichuan Basin in China. When perforated by drilling, the hole gushed water. Caves at this depth are, however, extremely rare and probably have little or no biological significance.

      As the profile of the cave takes shape, under the twin controls of hydrology and geology, the ‘best route’ is inevitably favoured with a greater rate of flow, which promotes more rapid corrosion of its boundary walls – which in turn results in a still greater flow capacity. In this way, the initially diffuse drainage within the limestone is gradually simplified with increasing time (and depth) into a pattern of coalescing collectors of increasing diameter. Over the aeons, the vagaries of geology and hydrology may conspire to favour one particular route in preference over all others, opening it out to form a major trunk conduit which drains the entire subterranean aquifer to its spring, or ‘resurgence’.

      I have suggested so far that cracks in the limestone can be enlarged to form mesocavernous conduits by corrosive water trickling gently down towards the water table under the influence of gravity, or by slow creeping but aggressive groundwaters draining towards a resurgence. However, chemical solution is not the full story – once the cave becomes large enough to be an efficient drain, fast-moving waters can carry sharp-edged grit, adding physical claws to chemical teeth. Grains of quartz (a mineral much harder than calcite) provide a particularly effective scouring agent, and are easily collected by streams running over the Millstone Grit which butts onto much of our cavernous limestone. Where such a stream is captured by a well-developed Yorkshire joint, the outcome is a vertical pothole; on Mendip, an inclined swallet cave results.

      Erosion of the river valleys which drain a limestone block may lower the water table within it, so that a conduit formed within the flooded phreas eventually becomes emptied of water and filled with air. One such system, GB Cave in the Mendip Hills, has received intensive study in recent years – in turn by Drs Tim Atkinson, Pete Smart and Hans Friederich of Bristol University. Their work has demonstrated the extent and importance of the system of mesocavernous conduits which overlies