Ants Have Been Farming Fungi Underground for 66 Million Years, Long Before Human Agriculture

Around 66 million years ago, a group of ants in what is now South America made a deal with a fungus. The ants would carry organic material into sealed underground chambers. The fungus would break it down into something edible. Neither could thrive without the other, and neither has broken the arrangement since. Human agriculture, by comparison, is roughly 12,000 years old. The ants had a 65-million-year head start.
This is not a loose relationship or a coincidence of habitat. It is obligate mutualism, a biological term for two species that have become so dependent on each other that separation means death for both. The fungus cultivated by leafcutter ants, primarily from the genus Leucoagaricus, no longer exists in the wild in the form these colonies grow. It has been domesticated, shaped by millions of years of ant selection, just as wheat was shaped by human selection over millennia. Except the ants did it first, and longer.
Leafcutter ants, the most studied of the fungus-farming species, do not eat the leaves they cut. The leaves are raw material. Worker ants slice fragments from plants, carry them back along established trails, and hand them off to smaller workers inside the nest who chew the material into a pulpy substrate. The fungus is then cultivated on this substrate in dedicated garden chambers, some the size of a football.
The fungus produces structures called gongylidia, swollen, nutrient-dense tips that the ants harvest and feed to larvae and the queen. The colony does not eat the fungus body itself. It eats these cultivated tips, which the fungus produces specifically because the ants have, over evolutionary time, selected for strains that produce more of them. This is selective breeding. The ants did not choose to do it. It happened because colonies that tended more productive fungal strains survived better, and the fungus that produced more gongylidia got tended more carefully. Evolution ran the experiment for 66 million years and produced an agricultural system of remarkable efficiency.

There are over 200 species of fungus-farming ants, all within the tribe Attini, found across the Americas. Leafcutter ants represent the most advanced end of this spectrum. Their colonies can contain up to eight million individuals and maintain dozens of fungal garden chambers simultaneously.
Any monoculture, a single crop grown in large quantities, is vulnerable to disease. Human farmers know this well. Ant farmers solved it before humans existed. The ants carry a specific bacterium, Pseudonocardia, in specialised crypts on their bodies. This bacterium produces antifungal compounds that suppress Escovopsis, a parasitic mould that attacks the fungal gardens. The ants essentially grow their own antibiotic on their skin and apply it to the crop.

Worker ants also perform continuous sanitation inside the garden chambers. They remove dead fungal material, weed out contaminated sections, and carry waste to dedicated refuse chambers located far from the garden. A 2011 study published in the journal PLOS ONE found that leafcutter ant colonies maintain a separation between garden zones and waste zones with a consistency that rivals sterile laboratory protocols. The colony does not wait for disease to spread. It manages the garden as a living system that requires constant intervention.
The exchange is genuinely mutual. The fungus gains a sealed, temperature-regulated underground environment, a continuous supply of organic substrate it could not gather on its own, and protection from competing moulds and insects. In return it provides nutrition the ants cannot synthesise themselves. The fungus has become, over evolutionary time, a domesticated organism, its reproduction now depends almost entirely on the ants carrying fragments of mycelium to new colonies when a young queen leaves to found one.

A founding queen carries a small pellet of fungus in a special pouch in her mouth. She lands, digs a chamber, deposits the fungus, and begins cultivating it before her first workers even hatch. The entire agricultural operation restarts from that single inoculum. It is the ant equivalent of a farmer carrying seed grain to a new field, except the queen has been doing this since before the extinction of the non-avian dinosaurs.
Ant fungiculture is one of the few examples of non-human agriculture that is genuinely analogous to human farming, not just food caching or passive mutualism, but active cultivation with substrate preparation, crop selection, pest management, and waste disposal. Researchers studying Attine ant colonies have found chemical signals that ants use to assess fungal garden health, suggesting a form of crop monitoring that operates through scent rather than sight.
The system has also attracted interest from agricultural scientists studying how the ants manage monocultures without the collapse that typically afflicts them. The Pseudonocardia-derived antifungals have been examined as potential leads for new antibiotic compounds, since the bacterial strains the ants carry have co-evolved with their target pathogens for tens of millions of years, a duration no human drug trial can replicate.
What the ants built underground, in the dark, without tools or intention, turns out to be a model of agricultural resilience that human farming is still learning to approximate. The colony does not know it is farming. The fungus does not know it is being farmed. The system works anyway, and has worked longer than anything humans have ever built.