How Symbiosis Works: Mutualism, Commensalism, and Parasitism Explained

Your body holds roughly 38 trillion bacteria, a number on the same order as your 30 trillion human cells. Beneath a forest, fungal threads run through the soil and trade minerals for sugar with the roots of most plants. Inside a coral, algae supply most of the energy the animal needs. These are symbioses: long-term associations between different species that shape evolution and ecology. This article explains the three main types, shows the evidence behind famous examples, and covers why the categories are blurrier than textbooks suggest.

What is symbiosis?

The word comes from the Greek sym (together) and bios (life). The German biologist Heinrich Anton de Bary used it in the late 1870s for any intimate, persistent association between two different species, regardless of whether the relationship helped or harmed either one. That broad definition is still the one most ecologists use.

Everyday usage has narrowed the word to mean "a relationship where everyone benefits," which is only one of the possibilities. The useful way to think about it is as a ledger. For each partner, ask whether the association raises, lowers, or leaves unchanged its fitness, which is a measure of how many offspring it leaves behind.

Relationship	Effect on species A	Effect on species B	Is it symbiosis?
Mutualism	Benefits	Benefits	Yes
Commensalism	Benefits	Unaffected	Yes
Parasitism	Benefits	Harmed	Yes
Competition	Harmed	Harmed	No, not an intimate long-term association
Amensalism	Unaffected	Harmed	Sometimes, but rare

The ledger is a simplification. Costs and benefits are measured in a particular environment, over a particular period, and they can change. A fair summary is that mutualism, commensalism, and parasitism are points along a continuum, not three sealed boxes. Later sections show how often a partnership slides along that line.

Mutualism: both partners benefit

Mutualism gets the most attention, and for good reason. It underpins forests, reefs, farms, and your own digestion. Even so, "both benefit" does not mean "both are generous." The better-studied examples show partners that cooperate because cheating is costly.

Mycorrhizae: the root-fungus bargain

Most land plants partner with fungi. In a 2018 synthesis in New Phytologist, Mark Brundrett and Leho Tedersoo estimated that about 72% of vascular plant species form arbuscular mycorrhizas, a type in which fungal threads enter root cells, and that only around 8% are completely nonmycorrhizal. The fungal threads (hyphae) reach farther into soil than roots can, collecting water and minerals such as phosphorus. The plant pays with sugars made by photosynthesis.

Think of it as outsourcing. A plant rents a long-range delivery network that it could never build from its own roots, and it pays in the currency it makes best. Plants that rely on these partnerships often grow better than non-partnered plants in nutrient-poor soil.

The more famous story is the "wood wide web," the idea that trees share food and warning signals through shared fungal networks. A 2023 perspective in Nature Ecology & Evolution by Justine Karst, Melanie Jones, and Jason Hoeksema challenged this. They found that only two studies showed common fungal links among trees, and that no single peer-reviewed field study supported the claim that adult trees preferentially send resources or warning signals to young trees. They also reported that unsupported citations had almost doubled over 25 years. The authors do not dispute that the fungi matter. They argue that the evidence for tree-to-tree sharing is far thinner than popular accounts suggest.

Legumes and nitrogen-fixing bacteria

Soybeans, peas, beans, and clover house bacteria called rhizobia in root nodules. The bacteria pull nitrogen gas (N₂) out of the air and convert it into ammonia (NH₃), a form the plant can use. In return, the plant supplies carbohydrates and a low-oxygen home.

What makes this partnership stable is policing. In a 2003 experiment published in Nature, E. Toby Kiers and colleagues forced a normally cooperative rhizobium strain to stop fixing nitrogen by giving it an atmosphere without N₂. The soybean plants penalized the non-cooperating bacteria, whose reproductive success fell by about 50%. The proposed mechanism was reduced oxygen supply to the cheating nodules. The result suggests that mutualisms can persist because hosts punish partners that take without giving.

The scale is large. Synthetic nitrogen made by the Haber-Bosch process now feeds a big share of humanity, with estimates in the range of 40% to 50% of the world population, according to a review by Our World in Data of work by Erisman and colleagues (2008) and later studies. The review stresses that such estimates are hard to pin down, because breeding, irrigation, and farm management also raised yields. The biological route that legumes use came first, and crop rotation with legumes is an old way of restoring soil fertility. The chemistry of the industrial route is covered in the article on the Haber-Bosch process.

Corals and their algae

Reef-building corals house single-celled algae called zooxanthellae inside their tissues. The algae photosynthesize and pass sugars to the coral, which in turn supplies shelter and nutrients. The algae supply a large share of the coral's energy, and the exact share varies by species and conditions.

This is why bleaching matters. When heat stress breaks the partnership, the algae leave or die, the coral turns white, and if the loss lasts long enough the coral can starve. A reef is therefore a symbiosis that people can see from space. For more on reef biology, see the article on coral reefs.

Lichens: two kingdoms in one body

A lichen looks like a single organism but is a partnership between a fungus and a photosynthetic partner, an alga or a cyanobacterium and sometimes both. The fungus provides structure and collects water and minerals. The partner makes sugar. About 20,000 lichen species have been identified.

Lichens also show how the categories blur. Some descriptions treat the association as a mutualism, others note that in certain species it looks closer to controlled parasitism of the alga by the fungus, since the photosynthetic partner can often live without the fungus but not the reverse.

Gut bacteria: the partnership inside you

The best-known human example is the gut microbiome. A 2016 analysis by Ron Sender, Shai Fuchs, and Ron Milo in PLOS Biology estimated about 3.8 × 10¹³ bacteria and 3.0 × 10¹³ human cells in a 70-kilogram reference man, a ratio of about 1.3 to 1. The older claim of ten bacterial cells for every human cell was not supported. Most of those bacteria live in the colon, not on the skin or elsewhere.

What do they do? Several groups of gut bacteria break down dietary fibers that human enzymes cannot digest, producing short-chain fatty acids. One of these, butyrate, is a main energy source for the cells lining the colon, and research reviews link it to gut barrier integrity and lower inflammatory signaling. Gut bacteria also make vitamin K and several B vitamins. Whether those vitamins make a meaningful contribution to what the body absorbs is still being worked out, because the colon is not the main site of absorption.

The microbiome also trains the immune system and competes with disease-causing microbes for space. When the community is disrupted, for example by antibiotics, the balance can shift in ways researchers are still mapping. For a full tour, see the article on the gut microbiome and health and the shorter explainer on gut microbiome basics.

Figs and fig wasps: no exit

Some mutualisms are locked in. There are hundreds of species of fig trees, and in the textbook model each is pollinated by its own species of tiny wasp. The wasp lays eggs inside the fig's flowers, and its larvae feed on some of the seeds, while the wasp's travels between trees pollinate the rest. Neither partner can reproduce without the other.

The textbook pairing is not absolute. Researchers have documented figs pollinated by more than one wasp species, and wasp species shared among closely related figs. Obligate mutualisms are real, but nature is messier than the diagram.

Clownfish and anemones

Clownfish live among the stinging tentacles of sea anemones, protected by a mucus coating. The benefit to the fish is clear: shelter. For a long time the anemone's side of the bargain was less clear. Field studies have reported that anemones hosting anemonefish can grow and reproduce better than anemones without fish, and nutrients from the fish's waste are one proposed mechanism. The relationship is usually described as a mutualism, though the dependence is not equal. Clownfish need an anemone, while many anemones occur without fish.

Commensalism: one benefits, one is unaffected

Commensalism is harder to find than it sounds. A relationship may look neutral for one partner until someone measures carefully. Many biologists regard clean examples as rare and treat commensalism as a label for "no effect detected so far."

Epiphytes. Orchids, ferns, bromeliads, and mosses grow on tree branches, gaining light and a perch without drawing nutrients from the tree. For the host, the effect is usually negligible, though a heavy load of epiphytes can stress a branch.

Remoras. These fish attach to sharks, rays, and other large animals with a modified dorsal fin that works as a suction pad. They gain transport and scraps. Whether the host is helped, by parasite removal, or harmed, by a small increase in drag, depends on the pair, so the relationship sits near the border of commensalism and mutualism.

Barnacles on whales. Barnacles attach to whale skin and filter-feed as the whale swims, gaining a ride and a feeding position. Whether the host pays any cost, such as drag, is rarely measured well.

The takeaway is that commensalism is a hypothesis about a missing effect, and the effect might be too small to detect, or hidden in a part of the animal's life that nobody has studied.

Parasitism: one benefits, one is harmed

Parasitism is likely the most common symbiosis by sheer number of species. In a 2008 analysis in PNAS titled "Homage to Linnaeus," Andrew Dobson and colleagues estimated that approximately 75% of the links in food webs involve a parasitic species. Parasites include viruses, bacteria, fungi, protists, worms, insects, and even plants such as mistletoe and dodder.

The arms race

Parasites and hosts are locked in what biologists call the Red Queen dynamic, named for the character in Lewis Carroll's Through the Looking-Glass who must run as fast as she can just to stay in the same place. Hosts evolve defenses, parasites evolve ways around them, and the cycle repeats. Much of the complexity of the immune system is thought to come from this contest, and some of the diversity of immune genes in populations is believed to be maintained by parasite pressure. The mechanics are covered in the article on how the immune system recognizes threats.

Mind control, with caveats

A few parasites change host behavior in ways that serve the parasite's life cycle:

• Toxoplasma gondii, a single-celled parasite, completes its sexual cycle only in cats. In experiments, infected rats lose their normal aversion to cat odor, and in some experiments become attracted to it, which raises the chance of being eaten by a cat. A 2007 study in PNAS by Vyas and colleagues found that the effect was largely specific to cat odor, with other fears intact.
• Ophiocordyceps fungi infect ants and drive them to clamp onto vegetation before dying. The fungus then grows out of the ant's body to release spores.
• Lancet liver flukes (Dicrocoelium dendriticum) cause infected ants to climb grass tips and hold on, which makes it more likely that grazing sheep or cattle eat them and complete the fluke's cycle.

Toxoplasma also infects people, and reviews describe about one-third of the world's population as latently infected, with wide variation by region and method. Humans are a dead end for the parasite, not part of its normal cycle. Studies have reported statistical associations between infection and various behavioral outcomes, but association is not causation, and no firm cause-and-effect case has been established.

Parasites of parasites

Hyperparasitism, in which a parasite is itself parasitized, is common. Bacteriophages are viruses that infect bacteria, so they are parasites of organisms that may themselves be parasites. Some parasitoid wasps lay eggs in other parasitoid larvae that are developing inside caterpillars, which stacks three layers of parasitism in one system. For how viruses jump hosts, see the article on how viruses cross from animals to humans.

When the labels change: symbiosis as a moving target

The position of a relationship on the continuum is not fixed. It depends on the environment, the partner's state, and the time scale.

Gut bacteria illustrate the point. Under normal conditions they are partners. If the gut wall is breached, say by a burst appendix, the same species can spill into the abdomen and cause a dangerous infection. The bacteria did not change. The context did.

Herpesviruses offer another case. Many people carry herpes simplex virus for life in nerve cells. The virus reactivates in some people and never in others, so whether it behaves as a harmful parasite depends on the person and the time. The article on herpesviruses explains the latent state.

Wolbachia: from reproductive parasite to disease blocker

The bacterium Wolbachia is one of the most widespread symbionts on the planet. A statistical analysis by Zug and Hammerstein in 2012 put the share of terrestrial arthropod species carrying it at about 40%, down from an earlier estimate of 66%. In many insects it behaves like a reproductive parasite, manipulating its host's reproduction to spread itself through the population.

In some mosquitoes, though, it does something useful to people. Aedes aegypti mosquitoes carrying Wolbachia are less able to transmit dengue virus. In a randomized trial in Yogyakarta, Indonesia, published in the New England Journal of Medicine in 2021, releasing these mosquitoes was associated with a 77% reduction in dengue incidence and an 86% reduction in hospitalizations, according to the World Mosquito Program. Results in other settings are still being gathered. The same bacterium is a parasite in one context and a public health tool in another.

Aphids and Buchnera: partners that cannot split

Pea aphids feed on plant sap, which is rich in sugar and poor in some amino acids. A bacterium called Buchnera lives inside special aphid cells and supplies the missing nutrients. The association is ancient, and the bacterium's genome has shrunk to a small fraction of the size of the genome of its free-living relative E. coli. The partners can no longer live apart. Biologists call such bacteria endosymbionts, from the Greek for "living inside."

Symbiogenesis: when partners merge

The most radical symbioses are the ones that become a single organism. The theory of symbiogenesis, advanced by Lynn Margulis, holds that mitochondria and chloroplasts began as free-living bacteria that took up long-term residence inside other cells. The idea was resisted for years, and it gained wide acceptance as molecular evidence showed that mitochondrial DNA resembles bacterial DNA far more than nuclear DNA.

The evidence is still the best argument for the theory. Both organelles keep their own small circular genomes, divide on their own schedule inside the cell, and use molecular machinery related to that of bacteria. Mitochondria trace to an alphaproteobacterium, and chloroplasts to cyanobacteria. For the full story of the first, see the article on how mitochondria became the cell's powerhouses, and for the scientist, the profile of Lynn Margulis.

Recent work refines the picture. A 2026 analysis in Nature found that archaea from the Asgard group made a dominant contribution to most conserved eukaryotic systems, while alphaproteobacteria, the mitochondrial ancestor, contributed mainly to energy transformation, iron-sulfur cluster assembly, and mitochondrial translation. The authors describe mitochondrial acquisition as a relatively late event, and they do not give absolute dates. Estimates in the literature place the original partnership roughly 2 billion years ago, but the timing is uncertain and debated.

An analogy helps. A merger of two companies can leave behind two sets of legal records, even after the staff have fused into one workforce. Mitochondria still carry a small genome of their own, the leftover paperwork of a partnership that began as a separate business. The photosynthetic side of the story is told in the article on how photosynthesis works.

Symbiosis is not the only route for genes to cross species lines. Microbes also swap DNA directly, a process described in the article on horizontal gene transfer. Together these routes show that the tidy tree of life has many tangled branches. The broader theory they fit within is explained in the article on evolution by natural selection.

How biologists decide which label fits

Because the labels depend on measurement, researchers use specific tests:

1. Removal experiments. Take the partner away and measure growth, survival, or reproduction in each species. The anemone work above is an example.
2. Cost-benefit accounting. Measure what each partner pays and gains, for example carbon traded for phosphorus.
3. Context testing. Repeat the measurement in different conditions, such as nutrient-rich versus nutrient-poor soil, or healthy versus stressed hosts.
4. Cheater tests. Check whether hosts can detect and punish partners that give little, as in the soybean experiment.
5. Time scale. Look at the whole life cycle, not a single season.

The table shows how this plays out across the examples above.

Partnership	Usual label	Evidence	Caveat
Legume and rhizobia	Mutualism	Plants cut oxygen to cheating nodules; cheaters' reproduction fell about 50%	Cost to the plant varies with soil nitrogen
Coral and zooxanthellae	Mutualism	Algae supply a large share of coral energy	Heat stress can break the partnership
Trees and shared fungal networks	Mutualism (individual trees and fungi), unproven (between trees)	Mycorrhizal fungi aid nutrient and water uptake	Tree-to-tree transfer evidence is thin
Remoras and sharks	Commensalism or mutualism	Fish gain transport and scraps	Effect on the host is poorly measured
Wolbachia and insects	Parasite or mutualist	77% drop in dengue incidence in a trial	Depends on host species and environment
Lichen fungus and alga	Mutualism or controlled parasitism	Alga can often live alone, fungus cannot	Labels differ by species

What people get wrong about symbiosis

"Symbiosis always means cooperation." It does not. The original definition includes parasitism, and parasitism is common.

"Mutualists are being kind." Evidence suggests that cooperation often persists because hosts can punish cheaters and reward good partners. Biologists describe mutualism as trade, not charity.

"Trees share food through a forest-wide internet." The fungi are real and important. The claim that mature trees routinely feed seedlings and send warnings through shared networks is not well supported, according to the 2023 perspective by Karst, Jones, and Hoeksema.

"Humans are mostly bacteria." The old figure of ten bacterial cells per human cell was overstated. Current estimates put the ratio near 1.3 to 1 in a reference man, and bacterial cells are far smaller than human cells.

"A parasite wants to kill its host." Usually it does not. A parasite that kills its host too quickly may lose its own home, and evolution often favors parasites that spread without killing. Plenty of long-running partnerships, including some mutualisms, may have started as parasitism.

Frequently asked questions

Q: Is the human microbiome a mutualism or a parasitism? A: For most resident microbes under normal conditions, it is closer to mutualism: they get a stable habitat and food, and the host gets help with digestion, immune training, and protection against pathogens. The same organisms can cause disease if they enter the wrong place, such as the abdominal cavity. Some researchers prefer the term "commensal" for many gut species, because the benefit to the host is hard to quantify for each one.

Q: Can a symbiosis change from one type to another? A: Yes. Wolbachia is a reproductive parasite in many insects but protects some mosquitoes from dengue virus. Mitochondria may have begun as parasites or prey, then became essential partners. Context can shift a relationship in the short term, for example when gut bacteria enter the bloodstream.

Q: How common is parasitism? A: Very. One 2008 analysis in PNAS estimated that approximately 75% of the links in food webs involve a parasitic species.

Q: Are viruses symbionts? A: Viruses are obligate parasites of cells, since they cannot replicate on their own. Some endogenous viral sequences have been integrated into genomes over evolutionary time and may now serve useful functions, but most active infections are parasitic. How viruses use host cells is explained in the article on how viruses work from infection to replication.

Q: What is the difference between symbiosis and mutualism? A: Symbiosis is the broad category: any long-term, intimate association between different species. Mutualism is one kind of symbiosis, in which both partners benefit. Commensalism and parasitism are the others. Some ecologists use "symbiosis" to mean mutualism only, so it is worth checking how an author is using the word.

Q: Why does coral bleaching happen? A: Heat stress disrupts the partnership between the coral and its algae. The algae leave or die, and the coral loses color and a large share of its energy supply. If the stress passes, the coral can sometimes recover its algae. If it persists, the coral can starve.

The takeaway

Symbiosis turns out to be less a catalog of friendly pairings than a way of looking at life as a set of negotiated deals. Partners trade, police each other, cheat, and sometimes merge, and the same pair can move from harm to help when the environment changes. The strongest evidence is also the oldest: the organelles in your cells began as free-living bacteria, so the line between "a partner" and "a part of you" was drawn by evolution, not by biology textbooks. When you read a claim that a relationship is purely cooperative or purely harmful, the right follow-up question is what was measured, and under what conditions.

Sources

• Karst J, Jones MD, Hoeksema JD. "Positive citation bias and overinterpreted results lead to misinformation on common mycorrhizal networks in forests." Nature Ecology & Evolution, 2023. University of Alberta summary.
• Our World in Data. How many people does synthetic fertilizer feed? Cites Erisman et al., Nature Geoscience, 2008, and Rosa and Gabrielli, Environmental Research Letters, 2023.
• "Dominant contribution of Asgard archaea to eukaryogenesis." Nature, 2026. Full text on PubMed Central.
• Brundrett MC, Tedersoo L. "Evolutionary history of mycorrhizal symbioses and global host plant diversity." New Phytologist, 2018;220:1108-1115.
• Kiers ET, et al. "Host sanctions and the legume-rhizobium mutualism." Nature, 2003;425:78-81.
• Sender R, Fuchs S, Milo R. "Revised estimates for the number of human and bacteria cells in the body." PLOS Biology, 2016;14(8):e1002533.
• Dobson A, et al. "Homage to Linnaeus: How many parasites? How many hosts?" PNAS, 2008;105(Suppl 1):11482-11489.
• Zug R, Hammerstein P. "Still a host of hosts for Wolbachia: analysis of recent data suggests that 40% of terrestrial arthropod species are infected." PLoS ONE, 2012.
• Vyas A, et al. "Behavioral changes induced by Toxoplasma infection of rodents are highly specific to aversion of cat odors." PNAS, 2007.
• Margulis L. Symbiosis in Cell Evolution. W.H. Freeman, 1981; Douglas AE. The Symbiotic Habit. Princeton University Press, 2010.

---

Read next:

• The Skin Microbiome: Your Body's Outer Ecosystem
• Dengue Virus Explained: Why It's Spreading and How It Works