The Earthwork
Charles Darwin's last book was about earthworms. He published The Formation of Vegetable Mould Through the Action of Worms in October 1881, five months before he died. It was, by the standards of his career, a minor work — no grand theory, no dangerous implication, just forty years of patient observation about invertebrates in English soil. It outsold On the Origin of Species in its first year.
The central finding: earthworms produce the soil they live in. An acre of English pasture contains roughly 50,000 worms, and those worms process between ten and eighteen tons of earth per year. They ingest soil at depth, extract organic matter, and deposit castings on the surface. The castings are chemically different from the ingested material — richer in nitrogen, phosphorus, and potassium, with a pH closer to neutral. Over decades, the process buries surface objects under a steadily accumulating layer of processed earth. Darwin measured the rate at his own property: a layer of cinders spread on a field in 1842 had been buried under seven inches of worm castings by 1871.
The worms were burying the Roman Empire. Darwin observed that ancient ruins in England — tessellated pavements, stone foundations, scattered coins — lay beneath a uniform layer of fine-grained earth. The burial was not geological. It was biological. The worms had processed the soil above the ruins, carried it through their bodies, and deposited it on the surface, gradually sinking the artifacts under a blanket of their own manufacture. The process that maintained the soil's fertility was the same process that erased the visible past.
The standard evolutionary arrow points from environment to organism: the desert selects for water conservation, the canopy for epiphytes, the deep sea for bioluminescence. Darwin's earthworms inverted the arrow. The worms did not merely survive in the soil. They produced it. The topsoil of English pasture — its texture, its chemistry, its fertility — was a worm product. The environment that selected for earthworms was an environment that earthworms had built.
In 2003, John Odling-Smee, Kevin Laland, and Marcus Feldman formalized this inversion as niche construction theory. Organisms do not merely adapt to environments. They modify their environments, and those modifications change the selection pressures acting on future generations. The modification is not incidental. It is a second inheritance system alongside genetic inheritance. Organisms pass two things to their offspring: genes and an environment. Both accumulate across generations.
The beaver dam is the canonical large-scale case. A single dam raises the local water table, creates a pond, transforms a section of stream into wetland. Clive Jones, John Lawton, and Moshe Shachak named this role in 1994: ecosystem engineer. What distinguishes niche construction from ordinary habitat modification is the feedback loop. The beaver's dam creates conditions — still water for dam maintenance, deep water for food caching, cover from predators — that benefit the builder. And the construction is inherited: beaver kits are born into a dam complex built by their parents and grandparents. The kit inherits a constructed environment alongside its constructed genome.
When a beaver colony is removed, the dam degrades within years, the pond drains, and the wetland reverts to stream. Remove the builder and the building disappears.
Sphagnum moss engineers its environment more subtly and more completely. Sphagnum acidifies the water it grows in — it exchanges hydrogen ions for mineral cations, dropping the pH from near-neutral to below 4.0. At this pH, most vascular plants cannot compete. The few that survive — sundews, pitcher plants, cranberries — are specialists adapted to the acidic, nutrient-poor conditions that Sphagnum created. The bog is not an environment that Sphagnum adapted to. The bog is Sphagnum's construction project.
The feedback loop compounds across millennia. Sphagnum's acidification slows decomposition, because most decomposer fungi and bacteria cannot function at pH 4. Undecomposed Sphagnum accumulates as peat. The peat holds water, maintaining the saturated conditions that Sphagnum requires. The waterlogging excludes trees whose roots need oxygen. The treelessness maintains the open light conditions that Sphagnum requires. Each consequence of Sphagnum's chemistry reinforces the conditions for more Sphagnum.
The result is a self-maintaining system that can persist for ten thousand years. The Sphagnum growing at the surface today is growing on a column of its own ancestors' undecomposed remains. It is standing on the niche its lineage built.
The standard objection to niche construction theory is that the effects are real but the framing is redundant. The beaver builds a dam; the dam changes the environment; the changed environment selects differently. You can describe the entire process in standard evolutionary terms without invoking a second inheritance system. The environment changed — it just happened to change because of an organism rather than because of geology or climate.
Odling-Smee's response is that the redundancy objection misses the accumulation. A geological change is a one-time event. Niche construction is a directed, cumulative process. Each generation of earthworms inherits the soil processed by previous generations and processes it further. Each generation of beavers inherits a dam complex and extends it. The modifications compound in a direction determined by the organism's biology, not by physical accident. The niche has a lineage. It evolves.
This matters because it changes which outcomes are reachable. Without niche construction, the organism is constrained to adapt to whatever environment it finds itself in. With niche construction, the organism can modify the environment toward conditions that favor its own phenotype — even conditions that never existed before the organism created them. The Sphagnum bog is not a natural environment modified by Sphagnum. It is an environment that did not exist before Sphagnum built it. The bog is an artifact.
Darwin knew this. The final paragraph of The Formation of Vegetable Mould reads: "It may be doubted whether there are many other animals which have played so important a part in the history of the world, as have these lowly organised creatures." He had demonstrated, with four decades of measurement, that the surface of England was a biological product — manufactured by organisms so small and so numerous that their collective action was invisible until someone spent forty years on hands and knees measuring the rate at which cinders sank.
The earthworm does not choose to build soil. The construction follows from its ordinary metabolism — eating, excreting, growing, dying. The niche is the cumulative consequence of the organism doing what it does. The organism that modifies its environment does not need to know it is modifying its environment. The modification happens because the organism exists. The niche is built by living, not by planning.