A gram of healthy topsoil can contain several billion bacteria, a few million fungal cells, hundreds of thousands of protozoa and a smaller but still impressive number of nematodes, arthropods and other larger invertebrates. Counts at this scale have been known for decades, but the function of the organisms, and their interactions with one another and with the plants they support, have only recently become possible to study in detail. Much of the change comes from genetic sequencing, which has allowed ecologists to identify soil organisms that cannot be cultured in the laboratory. The old view of soil as a more or less inert substrate has given way to a picture of soil as the most biologically dense habitat on the planet.

At the base of the soil food web are the decomposers, organisms that break down dead plant and animal material into smaller molecules. Fungi are particularly important here. Their hyphae - long, branching, thread-like structures - penetrate decaying wood and leaf litter, secreting enzymes that digest material externally. Many fungi also form partnerships with living plants. Mycorrhizal fungi wrap around or penetrate the root cells of most land plants, supplying minerals and water in exchange for sugars manufactured by photosynthesis. A forest tree without its mycorrhizal partners would not survive. In agricultural fields, tilling and fungicide use can disrupt these partnerships in ways whose consequences are still being quantified.

Bacteria are more diverse still. They include nitrogen fixers, which convert atmospheric nitrogen into forms that plants can use; denitrifiers, which perform the reverse reaction; sulphur cyclers; and groups that degrade an enormous range of organic compounds. A single gram of soil may contain tens of thousands of different bacterial taxa, most of them not cultivable with standard laboratory media. Soil microbiologist Dr. Lucia Fontana, working at the University of Seville, has used environmental DNA sequencing to map bacterial communities across a range of Mediterranean soils, and has shown that agricultural management can shift these communities substantially over only a few years.

Soil fauna are less diverse than soil microbes but occupy higher levels of the food web. Protozoa and nematodes feed on bacteria, regulating bacterial populations and releasing nutrients that were locked in bacterial cells. Arthropods such as springtails and mites feed on fungi and on each other. Earthworms, though individually conspicuous, are far less numerous than the microbial populations; their importance lies not in mass but in the mixing of soil layers that their burrowing produces. Dr. Fontana has noted that earthworm numbers can be a useful general indicator of soil health, even though the full picture involves many less visible organisms.

Human activity interacts with soil biology in several ways. Heavy tillage breaks up fungal networks and reduces the pore spaces that soil fauna need. Synthetic fertilisers supply plant nutrients directly, but they can simultaneously suppress the natural communities that would otherwise perform the same job. Pesticides, particularly persistent ones, accumulate in soil and can reduce invertebrate populations over years. And contamination of various kinds - by heavy metals, by hydrocarbons, by salts from over-irrigation - can shift soil communities into states from which recovery is slow.

Regenerative agriculture, a loose set of practices that aim to build soil rather than merely use it, has drawn on this science. Cover crops planted between cash crops keep roots and root exudates in the soil throughout the year; reduced tillage preserves fungal networks; livestock rotation in grasslands returns nutrients in balanced form. A 2020 review by the Rodale Institute found that fields under regenerative management accumulated organic carbon at several times the rate of conventionally managed fields, a gain that benefits both soil health and climate. Critics point out that the definition of regenerative agriculture is loose and that not all claimed benefits have been independently verified.

The larger point is that soil is not a passive medium. It is an ecosystem, with its own structure, its own disturbances and its own capacity to recover. Treating it as such requires a longer-term view of agricultural management than the annual crop cycle that dominates industrial farming. Dr. Fontana has argued that the economics of agriculture will eventually adjust to reward the maintenance of soil biology, but that the adjustment is slow and, in the meantime, many of the world's most productive soils continue to lose carbon, microbial diversity and structural stability at rates that earlier generations of farmers would have found alarming.