nightmare fuel
Parasitic zombie-ant fungus thrives in mosses, too
This may help fungi survive host scarcity and explains why infected hosts prefer the mosses as death sites.
Credit:
Tales Alves-Júnior et al., 2026
Credit:
Tales Alves-Júnior et al., 2026
The parasitic fungi in the Cordyceps genus are a favorite subject of nature documentaries. They’ve also infiltrated popular culture via the zombie-apocalypse video game The Last of Us (2013) and its TV adaptation, in which a parasitic fungus mutates to infect humans. Scientists are keen to study Cordyceps to learn more about the origins and intricate mechanisms underlying these pathogen-based diseases.
The latest finding: DNA analysis revealed the same fungus in both parasitized insects and surrounding mosses, according to a paper published in the journal IMA Fungus. (Yes, really.) This suggests a second life stage during which the fungus lives inside the moss, possibly an evolutionary adaptation to survive when insect hosts may be scarce. It could also explain why the host species seems to prefer biting into mosses during their final death throes.
As we’ve previously reported, there are more than 400 different species of Cordyceps fungi, each targeting a particular insect species, like ants, dragonflies, cockroaches, aphids, or beetles. The spores attach to the target insect, such as a carpenter ant, and germinate, spreading through the host’s body via long tendrils called mycelia. Cordyceps essentially turns its host into a zombie slave, compelling the ant to climb to the top of the nearest plant and clamp its tiny jaws in a death grip around a leaf or twig.
The fungus then slowly devours the ant, sprouting through its head in one final indignity. The bulbous growths on the ends of the mycelia burst, releasing even more spores into the air to infect even more unsuspecting ants. It’s not a great way to go: The entire process can take four to 14 days. Prior research showed that the zombification might be due to the release of a special chemical that causes the muscles in the infected ants’ mandibles to contract forcefully for that death-grip bite.
A 2017 study found that the fungal cells formed an elaborate, interconnected 3D network, enabling them to communicate and exchange nutrients. They essentially cut the brain off from the rest of the ant’s body so the networked cells can control its behavior. And in 2019, scientists found that the fungus doesn’t actually directly attach to an infected ant’s brain. Rather, it breaks apart the membrane that covers jaw-muscle fibers, causing contractions strong enough to damage or destroy the muscle filaments that slide past each other when the muscles contract.
A hidden stage
There is growing evidence that these kinds of fungal pathogens can expand their host range to adapt to a scarcity of hosts or environmental stress. For example, one species colonizes plant roots and can also parasitically infect cyst nematode eggs. Most relevant to this latest paper is recent research on insect-parasitizing fungi that have evolved to exploit both plants and arthropods.
This includes a 2020 Chinese study in which the “Himalayan gold” (O. sinensis) fungal species (which targets ghost moth larvae) was found lurking in plants in alpine regions. There have also been reports of ant-infecting Ophiocordyceps species showing a seeming preference for specific plants, like understory palm trees and moss carpets, as death sites for their unfortunate hosts.
For this latest study, the authors collected samples from “ant graveyards”—sites littered with the corpses of Ophiocordyceps-infected ants—at the Adolpho Ducke Forest Reserve in Brazil’s Central Amazon region. There were four sample categories: fungi emerging from infected insects, mosses located where the ants were bitten, mosses adjacent to biting sites, and a control group of mosses at least 10 meters away from biting sites with no evidence of infected insect bodies. All samples were photographed and cleaned of debris, placed in tubes, and stored in a freezer for DNA extraction and analysis.
The authors identified 19 Ophiocordyceps-infected insects, all but three of which were infected with the “zombie-ant” fungi in the first group—mostly ants, but also a weevil, a plant hopper, and a firefly larva. The majority (11 of 19) of the infected insects were found attached to mosses; one had bitten into liverwort as well as moss, and a handful of specimens were embedded in carpets made of several different kinds of mosses and liverworts. The team also identified Ophiocordyceps in the actual mosses, suggesting that the ants’ preference for those mosses when choosing the site for their “death grip” might be another form of manipulation by the parasitic fungus.
“We have always thought of this fungus (Ophiocordyceps) as a parasite exclusive to insects, especially the ones it manipulates,” said co-author Tales Alves Jr. of Instituto Nacional de Pesquisas de Amazonia in Brazil. “What our data suggest is that the story is much larger: The same fungus appears to be ubiquitous in the surrounding moss communities. Our results illustrate a far closer link between the lifecycles of the plants and fungi involved in the behavioral manipulation of Amazonian ants.”
IMA Fungus, 2026. DOI: 10.3897/imafungus.17.196998 (About DOIs).
