A kept place whose resident is itself the practice, and cannot leave it — sphagnum moss, which builds its own wet ground grain by grain, holds its place only while the conditions that let it build hold, and cannot follow when those conditions move
R870 closed on the beaver, which can at least judge a valley: leave a bad one, stay in a good one, and spend the extremes nightly where it has chosen to live. Its closing turn asked for a resident that can do neither — rooted where it stands, spending a limit it cannot flee and cannot follow when the limit moves. That resident is sphagnum, the peat moss. Two corrections are owed up front. First, sphagnum is not literally immobile over all time: its spores travel, and people can carry fragments to a damaged bog and spread them. What cannot move is the thing the moss has built — a peat body metres thick, raised over millennia — and a new start elsewhere begins again from zero, at peat speed. Second, sphagnum has no true roots at all: it is a non-vascular moss, anchored only loosely, taking water in through its leaves and moving it upward by capillarity. Rooted here is not a metaphor for grip. It is a statement about timescale: the practice is a place, and places do not walk.
The moss makes the wetness, not the other way round. A living sphagnum carpet is mostly water held in large, empty cells and in the spaces between its branches — up to about twenty times its own dry weight, in the accounts used here. As it grows, it releases hydrogen ions and takes up mineral nutrients in exchange, steadily acidifying the water around it; compounds in its cell walls, and the low-oxygen, acidic water it maintains, slow decay almost to a standstill, including its own. So the plant grows upward at the head while its lower parts die and do not rot away: they compress, year on year, into peat, accumulating at roughly a millimetre a year. Given centuries to millennia of that arithmetic, the peat lifts the surface into a gentle dome — raised bogs can stand metres, in places up to about ten metres, above the ground they started on — until the centre is cut off from groundwater altogether and is fed only by rain. The bog is then a perched lens of water held up by the moss and the peat beneath it: flood is absorbed and released slowly, drought is buffered by the stored water in the carpet, carbon captured at the surface is locked into the saturated mass below. Peatlands cover only about 3% of the land, yet their peat holds carbon on a scale the sources compare to twice that of all the world’s trees. None of this is service performed for anyone. It is a moss keeping itself wet, acidic and nutrient-poor enough that competitors and decomposers cannot take its ground — and the valley-scale effects arrive, as with the beaver, as side effects of that self-keeping.
And then the terms move. The dome holds only for exactly as long as the water stays near the surface. Dig a drainage ditch through it — as bogs have been drained for fuel, farming and planting across whole regions — or let a hotter, drier climate pull the water table down season after season, and the resident cannot respond the way a beaver would. It cannot abandon the drying dome and dam a wetter valley. It cannot follow the wet ground downhill, or north, at the speed the wet ground is moving. The carpet dries in place, and drying reverses every term of its own engineering: air enters the peat, microbes wake, and centuries of stored carbon begin to oxidise away as carbon dioxide; the surface subsides as the peat that held it up literally leaves as gas; shrubs and trees, which the acid water had excluded, root in the drier ground and shade the moss out; and dry peat, which wet peat would never do, can burn — deep, smouldering burns that the fire literature warns can flip a moss peatland into a shrub-and-grass system that no longer accumulates carbon at all. The moss spent flood and fire a little at a time for thousands of years, but it spent them here, in a body it cannot carry.
The return is at the moss’s speed, not ours. Restoration states the same asymmetry from the other side. Blocking the drains and rewetting a bog is necessary, and the sources are blunt that it is often not sufficient: a drained surface can sit disconnected from the raised water table in drought, sphagnum recolonises slowly against faster mosses, shrubs and trees, and full recovery of its cover and composition may take decades — one Scottish programme’s work puts a minimum of about ten years before even a revived bog stops emitting carbon dioxide, let alone resumes storing it at the old rate. Where recovery succeeds it usually needs the moss itself carried back in, fragments spread by hand: people doing, slowly and at small scale, the relocating the resident could never do for itself, and then waiting at a millimetre a year for the new carpet to become a bog again. That is the distinction this page exists to keep. The beaver’s practice is portable because it lives in an animal that chooses. The Martu’s practice was portable because it lived in people who could walk home. Sphagnum’s practice lives in the ground it has made, and the ground cannot be walked home. A keeping like this survives only if its place survives: protect the water where it stands, because there is no second venue the practice can be moved to on any timescale that answers the threat.
One moss, and the ground it made. Schematic, compressed from the published record.
Walk it in order.
Schematic only: the four stages compress millennia of peat growth and decades of drainage into one bog, and real bogs differ — blanket bogs, fens and tropical peat swamps are built by different plants and different hydrology, and not every drained bog burns or converts. The water-holding figure is an upper figure from the accounts used here; published values vary by species and method, from several times to over twenty times dry weight. The roughly one-millimetre-a-year figure is a long-run peat accumulation average, not a yearly promise. Carbon totals for peatlands are published global estimates. The restoration timescales are from northern, managed peatlands; they are a measure of slowness, not a schedule for any particular site.