Seed digging is one of the most widespread foraging behaviors on Earth, practiced by rodents, birds, ants, and marsupials that scratch, probe, or excavate soil to reach buried seeds. The behavior shapes plant communities, moves nutrients through soil, and creates microhabitats where new plants germinate. Whether an animal is recovering its own buried cache or raiding someone else’s, the act of digging for seeds connects animal survival to plant reproduction in ways that ripple through entire ecosystems.
How Animals Find Seeds Underground
A seed sitting a few centimeters below the soil surface is invisible, so animals that dig for buried seeds rely on a combination of smell, memory, and environmental cues. For rodents, olfaction is the primary detection tool for seeds they did not bury themselves. Seeds that have absorbed water release volatile compounds that rodents can detect through soil. In experiments with hamsters, researchers found the animals readily located buried seeds that had taken up moisture, as well as an artificial cocktail mimicking those seed volatiles, but could not find seeds with hard, impermeable coats or dry soft seeds that released no scent.1PubMed. Physical dormancy in seeds: a game of hide and seek? The implication is striking: a seed’s physical armor may have evolved partly as camouflage from mammalian noses, not just protection from the elements.
When rodents are recovering their own caches rather than searching for unfamiliar seeds, the picture changes. Scatter-hoarding animals like squirrels and chipmunks bury seeds across dozens or hundreds of locations and later return to dig them up. Research on small scatter-hoarding rodents shows that animals repeatedly caching and recovering seeds with weak odors developed stronger olfactory ability and better spatial memory over time, with measurable changes in the protein profiles of brain regions involved in smell and navigation.2PubMed. Proteomic evidence for seed odor modifying olfaction and spatial memory in a scatter-hoarding animal In other words, the act of caching literally reshapes the brain to make retrieval more efficient. Rodents also use their whisker-mediated touch system to gather information about the soil surface while foraging, particularly at night when vision is of little use.
There is also a clever visual shortcut. When a cached seed germinates and pushes a seedling above the soil, that sprout acts as a beacon. Studies of heteromyid rodents found that they associate emerging seedlings with potential cache locations and dig at the base of seedlings even when no seed cache is actually present underneath.3Springer Link / Oecologia. Seedling-aided cache detection by heteromyid rodents For a pilferer hoping to steal someone else’s buried food, a tiny green shoot is essentially a flag marking where to dig.
How Deep Is Too Deep
Burying seeds deeper makes them harder to detect but also more expensive to retrieve. This tradeoff is central to the economics of seed caching. The ability of foragers to detect stored seeds decreases with depth, so the optimal burial depth for a scatter hoarder is the point where the expected recoverable energy, minus the energy cost of making and retrieving the cache, is greatest.4PubMed. A model of caching depth: implications for scatter hoarders and plant dispersal Bury too shallow and a competitor will sniff it out; bury too deep and you spend more calories digging than the seed is worth.
Birds face the same equation but with different anatomical tools. Among sparrows in North America, species that are strong scratchers, like eastern towhees, song sparrows, and white-throated sparrows, can meet or exceed their energy needs when foraging on seeds buried at all depths down to about 1.5 to 2.25 centimeters. But a weak scratching species like the Savannah sparrow, which scratches at the same rate as the strong scratchers, runs an energy deficit when forced to forage on subsurface seeds.5Oikos. Interspecific variation in extraction of buried seeds within an assemblage of sparrows The difference is not effort or motivation but raw scratching power. A centimeter of soil separates a profitable meal from a losing proposition depending on the bird’s leg and foot anatomy.
In Argentine desert habitats, a similar pattern emerges. Rufous-collared sparrows found and consumed high proportions of buried seeds, though their success declined with increasing depth. Meanwhile, several other finch species in the same habitat could not find buried seeds at all. The rufous-collared sparrow’s advantage appears to come from a “double scratch” foraging method, a rapid bilateral kick that displaces litter and topsoil efficiently enough to uncover shallow seeds.6Journal of Avian Biology. Litter and seed burying alter food availability and foraging efficiency of granivorous birds in the Monte desert Even a thin layer of leaf litter on top of soil reduced seed consumption by about half for most bird species tested, though the double-scratching sparrow was far less affected. Litter acts almost like a second layer of burial that many birds cannot efficiently penetrate.
Seeds That Bury Themselves
Not all seed burial involves an animal doing the digging. Some plants have evolved mechanisms that let their seeds drill into the ground without any outside help. The seeds of plants in the genera Erodium and Pelargonium carry long coiled awns that respond to humidity changes. When dry, the awn is tightly coiled. When moisture rises, it uncoils and straightens. This cycle of coiling and uncoiling generates a rotational drilling motion that pushes the seed head into the soil.7Oxford Academic. Self-burial Mechanics of Hygroscopically Responsive Awns Researchers have shown that while the resistance of even relatively coarse soil would be enough to block a seed pushing straight down, the spinning motion dramatically reduces that resistance, letting the awn’s modest force do the job.
Self-burial neatly sidesteps the risks of relying on an animal. A seed cached by a rodent might be eaten by the same rodent or stolen by a competitor. A self-burying seed reaches the soil on its own terms, placing itself at a depth favorable for germination and protected from surface-level seed predators. The strategy is especially common in grasslands and open habitats where litter cover is thin and surface seeds are highly exposed.
Ants as Underground Seed Handlers
Harvester ants are among the most systematic seed diggers on the planet, but their version of seed digging is less about excavation and more about organized subterranean logistics. The Florida harvester ant, Pogonomyrmex badius, stores seeds in chambers deep underground. Forager ants collect seeds at the surface and deposit them only in the topmost nest chambers. From there, a separate group of workers rapidly transports the seeds downward in detectable waves, eventually placing them in dedicated seed chambers 20 to 80 centimeters below the surface.8PLOS ONE. Sequential Subterranean Transport of Excavated Sand and Foraged Seeds in Nests of the Harvester Ant, Pogonomyrmex badius Each individual worker usually handles only one leg of the journey, a form of task-partitioning that keeps the system efficient even as hundreds of seeds move through the nest at once.
These ants are not indiscriminate collectors. Surveys of Florida harvester ant nests have identified seeds from roughly 58 plant species, with more than a dozen appearing frequently and the rest showing up occasionally or rarely. Some seed species present in the surrounding habitat never appear in ant nests, pointing to either forager selectivity or differences in which seeds are physically accessible.9PubMed Central. An illustrated guide to seeds found in nests of the Florida harvester ant, Pogonomyrmex badius
The ants also face a mechanical problem: some seeds are simply too large or too hard to consume directly. When offered both germinating and non-germinating seeds, the ants preferentially fed germinating seeds to their larvae. Germination softens the seed coat and unlocks stored nutrients, essentially pre-processing the food. A comparison of seed sizes in storage chambers versus discarded husks confirmed that the consumed seeds were far smaller than the stored ones, suggesting the ants rely on the germination process to make large seeds edible.10PLoS ONE. The Florida Harvester Ant, Pogonomyrmex badius, Relies on Germination to Consume Large Seeds The ants are, in effect, farming germination underground.
Seasonal Rhythms of Caching and Consumption
Whether an animal buries a seed or eats it on the spot can depend on the time of year. In the fynbos shrublands of South Africa, the spiny mouse Acomys subspinosus shifts its diet dramatically across seasons. In winter and spring, it eats mostly insects. As summer and autumn arrive and the seed bank fills up, the mouse switches to a seed-heavy diet and begins caching seeds underground rather than consuming them immediately. By the following winter, the seed bank is so depleted that the mouse reverts to eating whatever seeds it finds rather than storing them.11South African Journal of Botany. Seasonal fluctuations in rodent seed caching and consumption behaviour in fynbos shrublands: Implications for fire management Another rodent in the same habitat, Rhabdomys pumilio, showed much less seasonal variation, maintaining a steadier diet year-round.
This seasonal swing matters for plants. When rodents cache seeds during periods of abundance, some of those seeds escape retrieval and germinate. When rodents switch to immediate consumption during lean months, every seed they encounter is destroyed. The timing of prescribed burns in fire-prone landscapes like fynbos may need to account for these patterns: a burn that triggers seed release during a period when rodents are caching rather than consuming could result in very different plant regeneration outcomes.
Cache Owners, Pilferers, and the Arms Race
Scatter-hoarding creates an inherent conflict. The animal that buries a seed invests energy in caching and expects a future payoff, but any other animal that can detect the cache might steal it. Research tracking three sympatric rodent species in a forest system found that all three had a substantial cache recovery advantage, meaning they retrieved their own caches at rates far exceeding the average pilferage rate over 30 days. The smallest species, Apodemus draco, showed the highest rate of scatter-hoarding and the largest recovery advantage compared to two larger co-occurring species.12PubMed Central. Does scatter-hoarding of seeds benefit cache owners or pilferers? The finding supports the idea that scatter-hoarding is a viable anti-theft strategy: by spreading caches across many locations, the owner makes it costly for any single pilferer to find them all.
Pilferers have their own tricks, though. As noted earlier, some rodents use emerging seedlings as visual cues to locate others’ caches. And in communities with many seed-digging species, the line between legitimate cacher and thief blurs. A seed cached by one species and pilfered by another might then be re-cached at a new site, effectively giving the seed a second dispersal event. From the plant’s perspective, pilferage can be just as useful as the original caching, because any movement away from the parent tree improves odds of successful establishment.
Digging as Ecosystem Engineering
The physical act of digging into soil to find or bury seeds has consequences that extend well beyond the animal’s meal. In arid and semi-arid landscapes, the pits left behind by digging mammals create microhabitats with dramatically different conditions from the surrounding surface. In Australia, foraging pits that mimic those created by bilbies and burrowing bettongs were found to have daytime temperatures 7 to 11 degrees Celsius cooler than the adjacent soil surface, and they retained significantly more moisture for up to five days after rainfall.13Journal of Arid Environments. Foraging pits, litter and plant germination in an arid shrubland Compared with flat ground, significantly more plants germinated inside these pits, whether or not leaf litter had accumulated in them. The cooler, moister microclimate alone was enough to promote germination.
This effect persists beyond the initial growing season. A study examining pits created by reintroduced bettongs found significantly more seedlings growing in the pits than in control areas, and the difference lasted beyond a full year, suggesting that the pits improved seedling survival through dry conditions. Native plant species showed a stronger positive response to the pits than exotic species, particularly in a wetter year.14Biodiversity and Conservation. A reintroduced ecosystem engineer provides a germination niche for native plant species The finding has real conservation implications: reintroducing digging mammals to landscapes where they have been lost could actively favor native plants over invasive competitors.
More broadly, bioturbation from digging mammals increases soil turnover, alters soil chemistry and structure, improves water infiltration, and reduces surface runoff and erosion. The organic matter that collects in diggings supports a range of soil organisms and accelerates nutrient cycling.15Mammal Review. Is the loss of Australian digging mammals contributing to a deterioration in ecosystem function? In regions where digging mammals have gone locally extinct, the loss of this soil disturbance may be contributing to declines in soil health and plant diversity that are difficult to reverse through other means.
When Exotic Species Enter the Picture
The introduction of non-native animals into an ecosystem can disrupt seed-digging dynamics in both directions. A systematic review of how exotic species affect native seed dispersal found that invaders can negatively alter dispersal mutualisms by changing the behavior and visitation rates of native seed dispersers, by directly preying on native dispersers, and by eating seeds that would otherwise have been cached or dispersed.16PubMed Central. Ecological Impacts of Exotic Species on Native Seed Dispersal Systems: A Systematic Review An introduced rodent that is a more aggressive seed predator than native species, for instance, can deplete the seed bank before native cachers have a chance to bury anything.
But the picture is not entirely negative. The same review found that some exotic species produce positive effects, including dispersal of native plants, facilitation of native forest tree colonization, enhancement of seedling survival, and increased seed rain of both early and late successional native species. Whether the net effect is beneficial or harmful depends heavily on the specific invader and the community it enters. A seed-dispersing exotic bird might help restore degraded forest patches while simultaneously competing with native frugivores for food. The complexity of these interactions makes blanket predictions about exotic species and seed digging unreliable.
Seed Dormancy as a Counter-Strategy
Plants are not passive participants in the seed-digging game. Physical seed dormancy, where a hard, impermeable coat prevents a seed from absorbing water and germinating, may serve double duty as a defense against seed predators. As the hamster experiments described earlier showed, seeds with hard coats release virtually no volatile compounds into the soil and are undetectable to rodent noses.17PubMed. Physical dormancy in seeds: a game of hide and seek? A physically dormant seed can sit in the soil for months or years, chemically silent, waiting for the right conditions to break dormancy and germinate. By the time it starts absorbing water and releasing scent, it may already be too deep or too well-established for a foraging rodent to bother with.
This reframes physical dormancy as something beyond a simple germination timer. It is also a stealth technology. Plants that produce physically dormant seeds may be investing in the molecular equivalent of odor suppression, keeping their offspring invisible to the most common underground seed predators until conditions are right for rapid establishment. Seeds without this armor, particularly soft-coated species that absorb water quickly, are essentially broadcasting their location to every rodent within sniffing distance.
The Fossil Record of Seed Caching
Seed digging is not a recent innovation. The oldest known food cache was discovered in lignite mining sediments near Cologne, Germany, in deposits dating to the late Early Miocene, roughly 18 million years ago. A fossil burrow system contained a stored food cache, and based on the size and shape of the galleries, the burrow was likely made by a large hamster or possibly a large ground squirrel.18Palaeontology. A Miocene rodent nut cache in coastal dunes of the Lower Rhine Embayment, Germany The animal had excavated from a soil surface down into dune sand in an interdune area, creating a storage structure recognizable even across geological time.
Finding a preserved cache is exceptionally rare. The conditions that allowed this one to fossilize, rapid burial in coastal dune sediments, are unlikely to capture most caches, which decay or get consumed long before preservation could occur. The Miocene cache suggests that the behavioral toolkit of digging, storing, and retrieving seeds was already well established in rodent lineages tens of millions of years before the modern species we observe today. The co-evolutionary relationship between seed-producing plants and seed-digging animals has deep roots, and the ecological dynamics playing out in today’s forests and deserts have been running, in recognizable form, for a very long time.