How Caves Actually Form: A Field Guide to Karst and Speleogenesis
Stand in a limestone cave passage and it's tempting to imagine some violent event carved it out — an earthquake, a collapsing void, a river bursting through rock. The truth is far quieter and far slower. Most caves you can walk into were dissolved, not blasted, by water that is barely more acidic than a soft drink, working for a period of time longer than our species has existed.
The chemistry: carbonic acid meets limestone
The process starts in the sky, or more precisely, in the soil. Rainwater picks up carbon dioxide as it falls and, more importantly, as it seeps through soil rich in decaying plant matter and root respiration, where CO₂ concentrations can run many times higher than in open air. That CO₂ combines with water to form carbonic acid — a weak acid, but a persistent one.
Limestone is mostly calcium carbonate, and calcium carbonate reacts with carbonic acid to form calcium bicarbonate, which is soluble in water and gets carried away in solution. Every litre of acidic groundwater that moves through a crack in limestone dissolves away a tiny amount of rock and carries it off downstream. It's an unremarkable reaction on its own; the geology comes from doing it billions of times over.
Following the joints
Limestone doesn't dissolve evenly. Water follows the path of least resistance: the natural joints, bedding planes, and fractures already present in the rock from its formation and from tectonic stress afterward. Early in a cave's life, these features are hairline cracks that happen to carry a trickle of acidic water.
Over time, dissolution widens the cracks that carry the most water, which lets them carry even more water, which widens them further — a feedback loop geologists call positive enlargement. Cracks that carry less water stay narrow or seal up entirely. This is why cave passages so often follow straight, joint-controlled corridors rather than meandering randomly: they're tracing the rock's original fracture pattern, expanded by solution over a very long time.
The water table and how passages form at depth
Where the dissolution happens relative to the water table matters enormously for the shape of the resulting passage. Below the water table, in the fully saturated zone, water fills every available crack and dissolves rock on all sides fairly evenly, tending to produce rounded, tube-like passages. Right at the water table, where a cave river runs with an open air space above it, dissolution concentrates at the waterline and can carve a wide, flat-bottomed canyon passage instead.
As regional water tables drop over geological time — often because a nearby valley cuts deeper — passages that once carried a full-flowing underground river are left perched above the new, lower water table. These abandoned passages dry out, and it's exactly these dry, fossil passages that most recreational cavers walk through today. The active streamway, still doing the work of enlarging new cave at the current water table, is often lower down and sometimes still flooded.
Surface clues: reading karst from above
Everything described so far happens out of sight, but a limestone landscape usually advertises what's happening underneath it, if you know what to look for. A doline — a bowl-shaped depression in a field, sometimes just a few metres across, sometimes large enough to hold a farm — marks a spot where dissolution has widened a joint enough that the ground surface has slumped or collapsed into the void beneath. Strings of dolines following a straight line are a strong sign of a joint-controlled passage running directly underneath, long before anyone digs a way in to confirm it.
Streams behave differently on karst too. A "disappearing stream" or swallet is exactly what it sounds like: a surface watercourse that runs along quite normally and then simply vanishes into a crack or a sinkhole, continuing its journey underground rather than to the sea by the surface route. It may resurface kilometres away and considerably lower in elevation, at a karst spring or resurgence, having spent the intervening distance carving exactly the kind of passage this article has been describing. Cavers and hydrologists confirm these underground connections with dye tracing — releasing a harmless, detectable dye at the point a stream disappears and watching for it to reappear downstream, sometimes days or weeks later, which is often the only proof a particular swallet and a particular spring are actually connected. A landscape riddled with dolines, swallets, and dry valleys that clearly once carried a surface river is a landscape practically labelling its own cave systems for anyone who learns to read it.
Not just limestone
Limestone caves (technically "karst" caves, after the Karst plateau region of Slovenia and Italy where the landscape was first studied scientifically) are the majority of the world's long cave systems, but dissolution caves also form in gypsum and rock salt, both of which dissolve even faster than limestone — gypsum caves can develop in a geological eyeblink compared to limestone, sometimes noticeably changing shape within a human lifetime. Separately, entirely different processes cut caves too: lava tubes form when the outer crust of a flowing lava stream cools and solidifies while molten rock continues flowing underneath and eventually drains out, leaving a hollow tube behind, sometimes for many kilometres. Sea caves are cut by wave action battering a weakness in a cliff face, widening it grain by grain the same way a river widens a joint, just with saltwater and much faster erosion. Talus caves form in the gaps between enormous jumbled boulders at the foot of a rockfall or landslide, no dissolution involved at all — just gravity stacking rock loosely enough to leave voids a person can move through. None of these involve carbonic acid at all — "cave" describes a shape, not a single origin story, and the shape alone tells you a great deal about which story produced it.
The slow architecture of speleothems
Once a passage exists, a second, gentler process often starts decorating it. Water seeping through the rock above a cave, having already picked up dissolved calcium carbonate on its way down exactly as described earlier, emerges into the open air of the passage and loses a little dissolved CO₂ to that air — the reverse of the reaction that dissolved the rock in the first place. Losing CO₂ makes the water slightly less acidic, which means it can no longer hold as much dissolved calcium carbonate in solution, so a tiny trace of it precipitates back out as solid calcite, left behind exactly where the drop of water paused. Repeated a hundred thousand times, drop after drop over centuries, that trace mineral builds into the stalactites, stalagmites, flowstone, and columns that make a decorated cave chamber so striking.
Different drip behaviour produces different shapes from the same basic chemistry. A slow, steady drip falling from one fixed point on the ceiling builds a stalactite growing downward and, from the splash below, a stalagmite growing up to meet it; given enough time, the two can join into a single column floor to ceiling. Water running as a thin film down a sloping wall instead builds flowstone, a smooth, rippled sheet rather than a point-fed icicle shape. Airflow deflecting a drip sideways, or capillary action pulling water through an impossibly narrow central canal against gravity entirely, produces helictites — formations that twist and branch in directions ordinary dripstone never would, and one of the clearer reminders that "cave formation" and "speleothem growth" are related but genuinely separate processes running on very different clocks. Typical growth rates for dense, actively-forming dripstone run on the order of a tenth of a millimetre a year, which is why a formation the size of a human arm routinely represents tens of thousands of years of essentially uninterrupted mineral deposition.
That slow, steady growth turns speleothems into unusually good scientific instruments. A stalagmite grows in visible layers, much like a tree's rings, and each thin layer of calcite locks in a tiny sample of the water chemistry and trace elements present when it formed — which, in turn, reflects the surface climate and rainfall pattern at that moment. Slicing a stalagmite and analysing its layers with modern isotope techniques can reconstruct a region's temperature and rainfall history stretching back tens or hundreds of thousands of years, filling in gaps that ice cores and tree rings can't reach on their own. It's a genuinely elegant bit of science hiding in plain sight in decorated passages worldwide, and it depends entirely on those layers surviving undisturbed — a single touch transfers skin oils that can halt a formation's growth in that spot permanently, and a broken stalagmite doesn't just lose its shape, it loses a climate record that took longer to write than human civilisation has existed.
Life in the dark: a cave's zones
Caves aren't uniformly "the underground" from a biological standpoint; they're usually described in zones. The entrance zone still gets direct daylight and behaves ecologically like the surface, just in a rock alcove. Moving inward, the twilight zone gets indirect light and a much more stable temperature, and supports specialised plants and animals that shelter there without living there permanently. Beyond where any daylight reaches at all is the dark zone, home to species that have adapted to permanent darkness over evolutionary time — true cave-dwellers, or troglobites, which often lose pigment and functional eyes entirely because neither does anything useful in an environment with no light to see by or to be seen in, and instead rely on other senses tuned to a food-scarce, perfectly stable environment.
That last point matters beyond curiosity. A dark-zone ecosystem is built around scarcity — nutrients mostly arrive from outside the cave, carried in by water, by bats, or by other animals moving in and out — and its species are often found nowhere else on Earth, confined to a single cave system or a small cluster of connected ones. That combination of specialisation and isolation is exactly what makes cave ecosystems fragile: a population with nowhere else to go and very little margin for disturbance is unusually easy to damage permanently, whether through pollution reaching the water table, a disturbed roosting colony, or simply enough foot traffic changing the passage's humidity and temperature. Knowing a dark zone exists as a real, occupied habitat, not just an empty tunnel, is part of understanding what's actually underground.
Why the timescale matters to a visitor
Knowing the process changes how you look at a passage. A smooth, scalloped wall tells you about the speed and turbulence of the water that shaped it — small, tightly-spaced scallops mean fast-flowing water, while large, gentle ones mean slower flow. A perfectly circular tube cross-section says the passage formed entirely below the water table; a canyon with a flat floor and parallel walls says it formed at a stable water table over a long period. The rock itself becomes a slow-motion recording of hydrology that predates any human visitor by tens of thousands of years, sometimes far more.
It also explains why cave conservation matters in a very literal sense: a speleothem or a passage wall that took ten thousand years to form can be damaged in a second and will not grow back in any timeframe that matters to us, and the specific caves and passages where these features occur are deliberately left out of anything we publish — access to a decorated or ecologically sensitive system is a matter for the landowner, the managing body, and the local caving club that already has a relationship with the site, not for a webpage. Understanding the process is the first step toward respecting the result; not touching, not naming, and going with people who already know a system's access rules is the second.