Non-human cooperation — candidate examples

Research long-list of animal, insect and microbial cooperation cases for the cooperative ape module, prioritized for available footage, and including the cases where cooperation is cheated.

Non-human cooperation — candidate examples

Long-list for the cooperative ape module in Session 3, and for the opening of Session 1 where the point is that cooperation is not a human invention. Over-generated deliberately; cut later.

Footage is a selection criterion here, not an afterthought. Several of these cases are chosen over more spectacular ones because three minutes of BBC film exists. Where nothing filmable exists, the entry says so.

Roughly a third of the entries are cooperation being cheated rather than cooperation working. That balance is intentional: defection in slime moulds and bacteria is the cleanest available evidence that free riding is a structural consequence of the payoff matrix rather than a human moral failing.

Two standing cautions, flagged again in the individual entries:

  • Nature documentary narration routinely upgrades coincidence into teamwork. Where the literature disputes the cooperative reading, the entry says what the dispute is.
  • Footage sourced to a BBC Earth YouTube channel clip is a cut from the parent series. The clips are 2–5 minutes and directly usable; the full sequences are longer and need the disc or the iPlayer archive.

1. Musk oxen defensive circle

The case

  • What happens: Musk oxen (Ovibos moschatus) facing wolves or bears form a line or a closed ring, adults outward with horns presented, calves inside. Individuals occasionally break out to charge and then rejoin. Against wolves the formation is close to impenetrable as long as it holds; herds that break and run lose calves.
  • Why it's a collective action case: Defense is a public good produced by joint effort, and it only exists above a participation threshold — a half-formed ring is worse than none. The individual incentive is to be on the inside. What stops defection is mostly that there is nowhere better to be: an animal that bolts from the ring is the one that gets taken, so the private and collective interest coincide once the formation exists. It is a coordination problem more than a dilemma, which makes it a good early contrast with the prisoner's dilemma cases in Session 2.
  • Could serve: Session 3, The cooperative ape. Also Session 2, The games, as a non-human stag hunt — the payoff structure is assurance, not temptation.
  • Footage: Good. Musk ox defensive formations against Arctic wolves are filmed in BBC Snow Wolf Family and Me (2014, Gordon Buchanan) and in Arctic wildlife documentary stock generally. The US National Park Service maintains a public page on muskox circle defense at Gates of the Arctic with images. Note that a full attack sequence is rare film; standing formations are common.
  • Source: Gray, D.R., The Muskoxen of Polar Bear Pass (1987). National Park Service, "Muskox: Circle Defense", nps.gov/gaar. Treat popular-web claims about precise formation times (e.g. "23–40 seconds") as unsourced — they trace to content farms, not to a paper.

2. Cape buffalo mobbing predators

The case

  • What happens: African buffalo (Syncerus caffer) herds turn on lions that have taken a member of the herd, and can drive them off and recover the victim. The best-known instance is a single amateur video shot at Kruger in 2007 in which a buffalo calf is taken by lions, dragged partly into water by a crocodile, and then recovered alive after the herd returns and mobs the pride.
  • Why it's a collective action case: Mobbing is costly to the individual and the benefit is diffuse. The free-rider option — hang back, let others charge — is available and taken by many animals in the herd. What makes it work is that the cost per participant falls sharply as numbers rise, which is the standard mobbing payoff, and a version of Olson's point about threshold effects in reverse.
  • Could serve: Session 1, Introduction to collective action, as a first case that is memorable and legible without any theory. Session 3, The cooperative ape.
  • Footage: Excellent, and unusually well-suited to teaching. "Battle at Kruger", filmed by Dave Budzinski and Jason Schlosberg, YouTube, 2007, about 8 minutes. It is amateur footage — shaky, unnarrated, no score — which is an advantage: it is a case, not a story someone has already told you how to feel about. There is also a National Geographic documentary built around it (Caught on Safari: Battle at Kruger, 2008).
  • Source: The video itself is the primary record. For the behavior generally, Prins, H.H.T., Ecology and Behaviour of the African Buffalo (1996).

3. Meerkat sentinels — and why the standard telling is wrong

The case

  • What happens: Foraging meerkats (Suricata suricatta) take turns on raised lookout, giving alarm calls that send the group to cover. It is the textbook example of altruistic vigilance, and the textbook is probably wrong.
  • Why it's a collective action case: It looks like a rota for a public good — costly personal vigilance, benefit shared. Clutton-Brock and colleagues found otherwise. Sentinels guard from safe sites, go on guard only once their own stomachs are full, take no regular turns, and are typically the first to spot and escape a predator. Over 2,000 hours of observation, no raised guard was killed. Provisioning animals with extra food increased their guarding. The conclusion is that guarding is the individually optimal activity for a fed animal when nobody else is up, which makes it not much of a collective action problem at all.
  • Could serve: Session 3, The cooperative ape — and this is the entry that earns its place by being a correction. Also Session 2, Some basic game theory, on why establishing that a behavior is genuinely costly is prior to calling it altruism. Useful as a methodological warning before the course starts making claims about human cooperation.
  • Footage: Excellent and abundant. Meerkat Manor (Animal Planet / Oxford Scientific Films, 2005–2008, four series) is built on the Kalahari Meerkat Project population. BBC Planet Earth and Africa both carry sentinel sequences. Short BBC Earth YouTube cuts exist.
  • Source: Clutton-Brock, T.H. et al., "Selfish Sentinels in Cooperative Mammals", Science 284(5420):1640–1644, 1999. https://www.science.org/doi/abs/10.1126/science.284.5420.1640

4. Fork-tailed drongos and the false alarm call

The case

  • What happens: Fork-tailed drongos (Dicrurus adsimilis) in the Kalahari follow meerkats and pied babblers and give genuine alarm calls when a raptor appears, which the mammals act on. They also give false alarm calls when a target has dug up something worth eating, and take the abandoned food. When a target stops responding to the drongo's own alarm, the drongo switches to mimicking the target species' alarm call — including the meerkat sentinel's own call, which the meerkats cannot afford to ignore.
  • Why it's a collective action case: An alarm-call system is a public good that runs on the credibility of the signal. The drongo is a free rider who has discovered that the good can be converted into private benefit by corrupting the signal, and the flexible switching between call types is a direct response to the victims' learning — an arms race in miniature. About 55% of drongo kleptoparasitism attempts use false alarms, with a 35% success rate. The system survives because drongos cannot cheat too often without destroying the credibility they depend on, which is precisely the constraint the course keeps returning to.
  • Could serve: Session 3, The cooperative ape, as the best single case for "defection is structural". Session 4, Did it ever work?, on monitoring and the cost of verification.
  • Footage: Excellent — probably the single best-filmed cheating case available. BBC Africa (2013), episode 1 "Kalahari", narrated by David Attenborough, carries the full sequence: genuine warning, then a false alarm, then the meerkats learning to ignore it, then the drongo escalating to mimicry of the meerkat's own call. Cut as "Drongo Bird Tricks Meerkats | Africa | BBC Earth" on YouTube, roughly 4 minutes. https://www.youtube.com/watch?v=tEYCjJqr21A
  • Source: Flower, T.P., "Fork-tailed drongos use deceptive mimicked alarm calls to steal food", Proc. R. Soc. B 278(1711):1548–1555, 2011. Flower, T.P., Gribble, M. & Ridley, A.R., "Deception by Flexible Alarm Mimicry in an African Bird", Science 344(6183):513–516, 2014.

5. Vervet monkey alarm calls

The case

  • What happens: Vervets (Chlorocebus pygerythrus) give acoustically distinct alarm calls for leopards, eagles and snakes, and conspecifics respond with the appropriate escape — into trees, look up, look down — even when the call is played from a hidden speaker with no predator present. Infants over-generalize (eagle alarms for any bird) and get more accurate with age.
  • Why it's a collective action case: Calling is the classic costly public good: the caller draws attention to itself and the benefit goes to everyone in earshot. The cost is real but smaller than usually assumed, and kin selection does much of the explanatory work in a group of relatives. Worth pairing with the drongo entry, since a functioning alarm system is exactly what the drongo parasitizes.
  • Could serve: Session 3, The cooperative ape. Pair directly with entry 4.
  • Footage: Moderate. The Seyfarth/Cheney playback experiments were filmed at Amboseli, and archival clips appear in primate behavior teaching material and in BBC Life of Mammals / Trials of Life material on primate communication. No single well-known documentary sequence of the calibre of the drongo film. Audio of the three call types is easy to source and works well on its own.
  • Source: Seyfarth, R.M., Cheney, D.L. & Marler, P., "Monkey Responses to Three Different Alarm Calls: Evidence of Predator Classification and Semantic Communication", Science 210(4471):801–803, 1980. https://www.science.org/doi/10.1126/science.7433999

6. Wolf pack hunting — and how little coordination it may require

The case

  • What happens: Wolf packs (Canis lupus) pursue elk, bison and moose in groups, appearing to take complementary positions — some animals holding the prey's head, others attacking flanks. On the standard telling this is role-differentiated teamwork.
  • Why it's a collective action case: If roles are real, the animal taking the head position bears more risk than the flankers, and the question is what stops the flankers from hanging back. The complication is that Muro and colleagues reproduced the main features of pack hunting — tracking, pursuit, encirclement — in simulation from two decentralized rules per wolf: move toward the prey until a minimum safe distance, and once there move away from other wolves at that distance. No communication, no roles, no leader. That does not prove wolves work this way, but it shows that apparent teamwork is not evidence of teamwork.
  • Could serve: Session 3, The cooperative ape, as a second methodological caution alongside the meerkats. Session 2, Some basic game theory, on emergent pattern versus intentional coordination.
  • Footage: Excellent. Yellowstone wolf-elk and Wood Buffalo wolf-bison sequences in BBC Frozen Planet (2011) and Planet Earth (2006, "Ice Worlds"), and extensive Yellowstone Wolf Project archive. The Frozen Planet wolf-and-bison chase in "Summer" is the most-used single sequence.
  • Source: Muro, C., Escobedo, R., Spector, L. & Coppinger, R.P., "Wolf-pack (Canis lupus) hunting strategies emerge from simple rules in computational simulations", Behavioural Processes 88(3):192–197, 2011. https://pubmed.ncbi.nlm.nih.gov/21963347/ For the field behavior, Mech, L.D. & Boitani, L. (eds), Wolves: Behavior, Ecology, and Conservation (2003).

7. Lions: leaders, laggards and the absence of punishment

The case

  • What happens: Heinsohn and Packer played recordings of intruding lions to Serengeti prides and watched which females advanced. Individuals were consistent across trials. Some always led, some participated only when least needed, some never participated at all. Leaders looked back at laggards — they evidently registered who was hanging back — and did nothing about it.
  • Why it's a collective action case: Territorial defense is a public good with real risk attached and a straightforwardly available free ride. This is the case where the standard fix fails: there is no punishment, so reciprocity is not what maintains the behavior, and the authors said plainly that existing theory does not explain it. Four stable strategy types coexisting in one group is a better picture of real cooperation than any account where everybody contributes.
  • Could serve: Session 3, The cooperative ape, and Repeated interaction and the folk theorem — this is the negative case for the punishment requirement. Runs well through the diagnostic: repeated interaction yes, observable defection yes, punishment no.
  • Footage: Good for lion territorial and hunting behavior generally — BBC Dynasties (2018) has a lion episode; Planet Earth II "Deserts" has the Namib lions; Big Cat Diary has years of Mara material. The specific playback experiments were filmed by the Serengeti Lion Project but the research footage is not packaged for teaching. Use general pride footage and narrate the experiment over it.
  • Source: Heinsohn, R. & Packer, C., "Complex Cooperative Strategies in Group-Territorial African Lions", Science 269(5228):1260–1262, 1995. https://www.science.org/doi/10.1126/science.7652573

8. Orca wave-washing

The case

  • What happens: A small population of Antarctic killer whales (Orcinus orca, pack ice ecotype B) hunt Weddell seals on ice floes by swimming in formation beneath the floe and surfacing together to generate a wave that washes the seal off. Attacks are repeated and adjusted between attempts — breaking a large floe into smaller pieces first, then washing off the remainder.
  • Why it's a collective action case: The wave cannot be produced alone; the minimum viable group is several animals swimming in synchrony, and the prey is then shared. It is a threshold public good in which the good is indivisible and the technique is culturally transmitted — Pitman's estimate is that fewer than about 100 animals worldwide do it, passed down within matrilines. The cultural-transmission angle sets up Session 5 directly.
  • Could serve: Session 3, The cooperative ape. Session 5, Multi-level selection, as non-human cultural transmission of a cooperation technology.
  • Footage: Excellent — among the best wildlife footage ever shot of a coordinated hunt. BBC Frozen Planet (2011), episode "Spring", filmed with Robert Pitman as scientific adviser. Cut as "Killer Whales Working Together to Hunt Seals on Ice | BBC Earth" on YouTube, roughly 4 minutes. https://www.youtube.com/watch?v=g1VEwsI4SlY
  • Source: Pitman, R.L. & Durban, J.W., "Cooperative hunting behavior, prey selectivity and prey handling by pack ice killer whales (Orcinus orca), type B, in Antarctic Peninsula waters", Marine Mammal Science 28(1):16–36, 2012.

9. Humpback bubble-net feeding

The case

  • What happens: Humpback whales (Megaptera novaeangliae) in Southeast Alaska and British Columbia form groups that encircle herring beneath the surface with a spiral curtain of bubbles, then lunge up through it together with mouths open. Crittercam work indicates division of labor: some animals blow the net, others dive to drive fish upward, others vocalize to herd them. Working groups are small and stable, with associations persisting over years and in some cases decades.
  • Why it's a collective action case: A public good — the net — produced jointly, benefiting all participants, and impossible alone. Stable membership over decades is the shadow of the future doing exactly what the folk theorem says it should. Fewer than 10% of the Southeast Alaska humpbacks do it, and recent work tracks the behavior spreading through the Canadian Pacific population, which makes it a second case of cultural diffusion.
  • Could serve: Session 3, The cooperative ape and Repeated interaction and the folk theorem — stable partners, repeated play. Session 5, Multi-level selection.
  • Footage: Excellent and abundant. Filmed in BBC The Blue Planet (2001) and in Alaskan whale-watch and research footage from the Alaska Whale Foundation; aerial drone footage of the full bubble spiral from above is now common and reads much better on a slide than surface-level footage.
  • Source: Wiley, D. et al., "Underwater components of humpback whale bubble-net feeding behaviour", Behaviour 148(5–6):575–602, 2011. On diffusion: "The diffusion of cooperative and solo bubble net feeding in Canadian Pacific humpback whales", Proc. R. Soc. B 293(2063), 2026.

10. Chimpanzee cooperative hunting of red colobus

The case

  • What happens: Chimpanzee (Pan troglodytes) parties hunt red colobus monkeys. At Taï in Côte d'Ivoire, Boesch describes animals taking differentiated roles — driver, blockers, ambusher — in continuous canopy where the prey cannot be cornered without them. At Gombe, hunts look much more like parallel individual pursuit in broken canopy where a lone hunter does well.
  • Why it's a collective action case: Meat is the payoff and it is divisible, so the question is whether hunters get more than non-hunters. At Taï, Boesch reports selective meat sharing that gives hunters a positive net caloric return while bystanders get little — a selective incentive in Olson's sense, attached to participation. At Gombe no such mechanism operates and cooperation is correspondingly unstable. Flag the dispute plainly: whether Taï hunting is genuinely coordinated role-taking or individual hunting that looks coordinated is contested between Boesch on one side and Gilby, Watts and others on the other, and the difference between sites may be ecological rather than social. The disagreement is itself the teaching point — the same behavior supports cooperation or not depending on whether a sharing rule rewards participants.
  • Could serve: Session 3, The cooperative ape and Olson — this is selective incentives in a non-human species. Session 5, Multi-level selection.
  • Footage: Excellent. BBC Trials of Life (1990), episode "Hunting and Escaping", has the Taï-style ambush sequence narrated by Attenborough, cut on YouTube as "Chimps Set Up an Ambush for Monkeys | Trials Of Life | BBC Earth". Netflix Chimp Empire (2023, dir. James Reed), filmed over 18 months at Ngogo in Kibale, has modern high-definition colobus hunts. Boesch's own Taï research footage exists and is partly public through the Max Planck Institute.
  • Source: Boesch, C., "Cooperative hunting roles among Taï chimpanzees", Human Nature 13:27–46, 2002. Boesch, C. & Boesch-Achermann, H., The Chimpanzees of the Taï Forest (2000). For the counter-position: Gilby, I.C. et al., "The meat-scrap hypothesis", Behav. Ecol. Sociobiol. 63:1821–1832, 2009; Gilby, I.C. et al., "'Impact hunters' catalyse cooperative hunting in two wild chimpanzee communities", Phil. Trans. R. Soc. B 370:20150005, 2015.

11. Groupers and moray eels hunting together

The case

  • What happens: Roving coral groupers (Plectropomus pessuliferus marisrubri) in the Red Sea swim to a resting giant moray eel (Gymnothorax javanicus) and perform a rapid head-shake at close range. The eel leaves its crevice and the two hunt together. The eel can enter crevices; the grouper takes fish in open water; prey has nowhere to go. Groupers also perform a headstand-and-point signal directed at hidden prey, which recruits the partner to that spot.
  • Why it's a collective action case: An interspecific mutualism with complementary rather than shared capabilities, so the joint output exceeds the sum of the parts. Neither animal shares the catch — each eats what it catches — which removes the division problem entirely and is why the partnership is stable without any enforcement. A useful counterexample: not every cooperation problem is a dilemma. Vail's analysis is also the first demonstration of referential gesture outside large-brained vertebrates.
  • Could serve: Session 3, The cooperative ape. Session 2, The games, as a case with no defection temptation to speak of — good for showing what the payoff structure has to look like for cooperation to be easy.
  • Footage: Excellent. BBC Blue Planet II (2017), episode 3 "Coral Reefs", carries the closely related coral trout and octopus partnership with the grouper's pointing display, filmed on the Great Barrier Reef. Research footage from the Bshary and Vail work in the Red Sea accompanies both papers as supplementary video. BBC Earth has short cuts of the reef sequence.
  • Source: Bshary, R. et al., "Interspecific Communicative and Coordinated Hunting between Groupers and Giant Moray Eels in the Red Sea", PLoS Biology 4(12):e431, 2006. Vail, A.L., Manica, A. & Bshary, R., "Referential gestures in fish collaborative hunting", Nature Communications 4:1765, 2013. https://www.nature.com/articles/ncomms2781

12. Cleaner wrasse and their clients

The case

  • What happens: Bluestreak cleaner wrasse (Labroides dimidiatus) hold fixed cleaning stations on reefs. Client fish — including predators that could eat them — queue, adopt a posture inviting inspection, and hold still while the wrasse removes ectoparasites, including from inside the mouth and gills. A single wrasse may handle over 2,000 client visits a day.
  • Why it's a collective action case: Mutualism with a large and continuous temptation to defect on both sides — the client could eat the cleaner, the cleaner could take the more nutritious mucus. Location is what makes it work: the station is fixed, both parties know where to find each other, interactions repeat, and clients waiting in the queue can watch. Repeated interaction, observability and the option to switch or punish — the full folk theorem ingredient list, running on a reef.
  • Could serve: Session 3, Repeated interaction and the folk theorem. This is the best non-human illustration of the shadow of the future, and it maps onto the diagnostic question by question.
  • Footage: Excellent. Cleaning stations are among the most-filmed reef behaviors. BBC The Blue Planet (2001) and Blue Planet II (2017) both carry sequences, as does BBC Trials of Life (1990, "Living Together"). Underwater cleaning-station footage is also abundant in public dive archives.
  • Source: Grutter, A.S., "Parasite removal rates by the cleaner wrasse Labroides dimidiatus", Marine Ecology Progress Series 130:61–70, 1996. Bshary, R. & Grutter, A.S., "Asymmetric cheating opportunities and partner control in a cleaner fish mutualism", Animal Behaviour 63:547–555, 2002.

13. Cleaner wrasse cheating, and what clients do about it

The case

  • What happens: Cleaners prefer client mucus to parasites. When they take mucus, the client gives a visible jolt — an abrupt body twitch — which is the observable signature of cheating and is how researchers score service quality. Clients respond according to their options: visiting clients with access to several stations switch to a different cleaner next time, while resident clients with only one local station chase and punish the cleaner instead. Cleaners also give better service when other clients are watching, and cooperating male-female cleaner pairs include males punishing females who cheat — which turns out to serve the male's own interest in the client staying.
  • Why it's a collective action case: The cleaner is a free rider on its own reputation. The mechanisms that hold it in check are exactly the ones the course names for humans: exit, punishment, and reputation among observers. The behavioral response is contingent on which of those the client actually has available, which is the diagnostic being run by a fish.
  • Could serve: Session 3, Repeated interaction and the folk theorem, and The cooperative ape. Session 4, Did it ever work? — the exit-versus-voice distinction is directly transferable.
  • Footage: Good but needs care. The cleaning interaction itself is extensively filmed (entry 12). The jolt is visible in close-up cleaning footage once you know to look for it, which makes it a strong live-annotation moment: play a cleaning sequence, point at the twitch. Bshary's laboratory and field footage from Ras Mohammed and Lizard Island accompanies the papers.
  • Source: Bshary, R. & Grutter, A.S., "Punishment and partner switching cause cooperative behaviour in a cleaning mutualism", Biology Letters 1:396–399, 2005. Bshary, R. & Grutter, A.S., "Image scoring and cooperation in a cleaner fish mutualism", Nature 441:975–978, 2006. Raihani, N.J., Grutter, A.S. & Bshary, R., "Punishers benefit from third-party punishment in fish", Science 327(5962):171, 2010.

14. Vampire bat blood sharing

The case

  • What happens: Common vampire bats (Desmodus rotundus) starve within about 60 hours of a failed feed. Bats that fed regurgitate blood to roostmates that did not. Wilkinson's 1984 field and captive work found donations predicted by an index of opportunity for reciprocation as well as by relatedness, and later work by Carter and Wilkinson found that previous help received predicted donations better than either relatedness or harassment.
  • Why it's a collective action case: The textbook case of direct reciprocity in a non-human animal, and one where the asymmetric payoff makes the trade worthwhile — the donor gives up a few hours of buffer, the recipient avoids death. Bats roost together for years, recognize individuals, and can withhold from non-reciprocators. Repetition, recognition, sanction.
  • Flag the dispute: the reciprocity reading has been contested. In the wild data, 95% of observed sharing was between close kin, and 70% of 110 observations were mothers feeding pups. Several authors argued the behavior is maintained by indirect fitness alone, with non-kin sharing explained as kin recognition error, indiscriminate altruism, or harassment. Carter and Wilkinson's controlled work is the reply, and it is a reply worth stating rather than assuming. Use the case, but do not present it as settled.
  • Could serve: Session 3, The cooperative ape and Repeated interaction and the folk theorem. This is where reciprocity gets introduced.
  • Footage: Moderate, and worse than the case deserves. Regurgitation is filmed in captive colonies — Gerald Wilkinson's and Gerald Carter's lab footage, and colony footage from the Smithsonian Tropical Research Institute in Panama. BBC Life of Mammals (2002) and various Nat Geo bat documentaries carry vampire bat feeding sequences, but sharing footage is shorter and less cinematic than the hunting footage. Expect 30 seconds of usable film, not three minutes.
  • Source: Wilkinson, G.S., "Reciprocal food sharing in the vampire bat", Nature 308:181–184, 1984. https://www.nature.com/articles/308181a0 Carter, G.G. & Wilkinson, G.S., "Food sharing in vampire bats: reciprocal help predicts donations more than relatedness or harassment", Proc. R. Soc. B 280:20122573, 2013. Carter, G.G. & Wilkinson, G.S., "Does food sharing in vampire bats demonstrate reciprocity?", Communicative & Integrative Biology 6:e25783, 2013.

15. Honeybee worker policing

The case

  • What happens: Honeybee (Apis mellifera) workers cannot mate but retain functional ovaries and can lay unfertilized male eggs. They mostly do not, and when they do, other workers eat the eggs. Ratnieks and Visscher showed strong discrimination against worker-laid male eggs in queenright colonies — over 99% removed within a day, against a much lower rate for queen-laid eggs.
  • Why it's a collective action case: The colony is not a harmonious superorganism; it is a group with a real internal conflict of interest that is suppressed by enforcement. A worker is more related to her own son than to the queen's sons, so selfish reproduction pays her — and because the queen mates with 10–20 males, each worker is more related to the queen's sons than to another worker's sons, so it pays everyone else to stop her. Policing is decentralized, mutual, and continuous. Mutual coercion, mutually agreed upon, with no agreement.
  • Could serve: Session 3, The cooperative ape — the entry that punctures the assumption that eusocial insects have solved the problem. Session 4, Coercion, as the non-human precedent for Hardin's actual prescription.
  • Footage: Moderate. Honeybee comb behavior is extremely well filmed — BBC Life in the Undergrowth (2005), Life (2009), and the observation-hive tradition going back to von Frisch. Egg-eating specifically is a close-up microscopy shot rather than a documentary sequence; research footage exists in the Ratnieks lab (Laboratory of Apiculture and Social Insects, Sussex). The waggle dance footage is the reliable crowd-pleaser if you need bees on screen.
  • Source: Ratnieks, F.L.W. & Visscher, P.K., "Worker policing in the honeybee", Nature 342:796–797, 1989. https://www.nature.com/articles/342796a0 Wenseleers, T. & Ratnieks, F.L.W., "Comparative analysis of worker reproduction and policing in eusocial Hymenoptera supports relatedness theory", American Naturalist 168:E163–E179, 2006.

16. Army ant living bridges

The case

  • What happens: Eciton burchellii army ants on the march link their own bodies into bridges spanning gaps in the trail. The bridges assemble in seconds, then move — lengthening, widening and sliding across the gap to shorten the route. They stop moving at the point where the number of workers locked into the structure costs more than the shortcut saves, and they disassemble when traffic over them falls below a threshold.
  • Why it's a collective action case: The bridge is a public good made of the contributors themselves, and every ant in it is an ant not foraging. The colony solves a cost-benefit optimization with no ant able to perceive either side of it — the trade-off is implemented entirely by local rules about traffic and contact. A useful case for showing that optimal collective outcomes do not require anyone to understand the problem.
  • Could serve: Session 3, The cooperative ape. Session 4, Polycentricity, as an argument that coordination need not be centralized.
  • Footage: Good. BBC Planet Earth (2006, "Jungles") and Life in the Undergrowth (2005) both carry Eciton raid columns and bridging. Reid's Princeton/NJIT fieldwork on Barro Colorado Island produced high-speed video published as supplementary material with the paper and widely re-cut by Princeton and National Geographic. 1–2 minutes of very strong film.
  • Source: Reid, C.R. et al., "Army ants dynamically adjust living bridges in response to a cost–benefit trade-off", PNAS 112(49):15113–15118, 2015. https://www.pnas.org/doi/10.1073/pnas.1512241112

17. Naked mole-rats

The case

  • What happens: Naked mole-rats (Heterocephalus glaber) live in subterranean colonies of up to 300 in which a single female breeds with one to three males and every other animal is reproductively suppressed, doing digging, foraging and defense instead. Jarvis identified this in 1981 as eusociality in a mammal — overlapping generations, cooperative brood care, a reproductive division of labor.
  • Why it's a collective action case: Complete reproductive self-abnegation by most of the group, which is as extreme as cooperation gets. It is held in place by a combination of high relatedness and active suppression — the queen shoves subordinates, and recent work identifies a queen-borne chemical signal (isopropyl myristate) mediating the suppression. Again: not harmony, enforcement. Pair with honeybee policing to make the point twice in different clades.
  • Could serve: Session 3, The cooperative ape. Session 5, Multi-level selection, on relatedness and group-level selection.
  • Footage: Good. Naked mole-rats are a zoo-exhibit staple and tunnel colonies are filmed in cutaway acrylic burrows — footage exists from the Smithsonian National Zoo, Pacific Science Center and the Zoological Society of London. BBC Life of Mammals (2002) and Life (2009) both include them. Visually striking in a way that holds a room.
  • Source: Jarvis, J.U.M., "Eusociality in a Mammal: Cooperative Breeding in Naked Mole-Rat Colonies", Science 212(4494):571–573, 1981. https://www.science.org/doi/10.1126/science.7209555 On suppression: "A queen odour mediates reproductive suppression in a eusocial mammal", 2025.

18. Dictyostelium: single cells building a body

The case

  • What happens: The social amoeba Dictyostelium discoideum lives as independent single cells eating bacteria in soil. On starvation, tens of thousands aggregate into a motile slug and then a fruiting body: roughly 20% of the cells die to form a rigid stalk, lifting the remaining 80% into spores that can disperse. The stalk cells get nothing. They are dead.
  • Why it's a collective action case: The cleanest case of a collective action problem with no brain, no communication in any meaningful sense, and an absolutely terminal cost to contributors. It also puts the multicellularity question in the course's own terms: a body is a group of cells that solved a collective action problem, and the solution is mostly clonality — cells in a fruiting body are usually near-identical, so the stalk cell's genes go into the spores anyway.
  • Could serve: Session 3, The cooperative ape, opening it — cooperation pre-dates animals by a long way. Session 5, Multi-level selection, since this is a major evolutionary transition in miniature.
  • Footage: Excellent, and cheap to obtain. Aggregation and fruiting-body time-lapse is a classic microscopy sequence dating to the Bonner films of the 1940s–50s; Princeton hosts John Bonner's original Dictyostelium movies publicly. Modern time-lapse is available from the Queller–Strassmann lab at Washington University and from MBL teaching collections. Roughly 60–90 seconds of time-lapse, and it is genuinely arresting on a screen.
  • Source: Bonner, J.T., The Cellular Slime Molds (1967). Strassmann, J.E., Zhu, Y. & Queller, D.C., "Altruism and social cheating in the social amoeba Dictyostelium discoideum", Nature 408:965–967, 2000.

19. Dictyostelium cheater strains

The case

  • What happens: When two genetically distinct Dictyostelium strains are mixed, they still form a single chimeric fruiting body — and in about half of such chimeras one strain contributes disproportionately few cells to the stalk and disproportionately many to the spores. Some lab-derived mutants are obligate cheaters that cannot form a stalk at all and can only reproduce by exploiting a partner.
  • Why it's a collective action case: This is the entry that does the most work in the whole file. A free rider in a system with no cognition, no intention, no culture and no morality. Whatever defection is, it is not a human character defect — it is what happens whenever the payoff to withholding exceeds the payoff to contributing, in any substrate. Obligate cheaters are also limited by their own success: they produce shortened stalks, which caps how far they can spread, which is a natural-experiment version of the point that defection is self-limiting at high frequency.
  • Could serve: Session 2, Some basic game theory — the argument that defection is rational rather than wicked, made with something that cannot be wicked. Session 3, The cooperative ape.
  • Footage: Moderate. The same time-lapse as entry 18 covers the mechanism; the cheating itself is shown through fluorescently labeled strains in chimeric fruiting bodies — still images and short microscopy clips rather than film. Available in figures and supplementary video from the Queller–Strassmann lab. This one may work better as a diagram than as footage.
  • Source: Strassmann, J.E., Zhu, Y. & Queller, D.C., Nature 408:965–967, 2000. Khare, A. et al., "Cheating by exploitation of developmental prestalk patterning in Dictyostelium discoideum", PLoS Genetics 6:e1000854, 2010. Also "In the social amoeba Dictyostelium discoideum, shortened stalks may limit obligate cheater success", PeerJ 12:e17118, 2024.

20. Pseudomonas siderophores: a public good and its cheats

The case

  • What happens: Pseudomonas aeruginosa cells secrete pyoverdine, a molecule that scavenges scarce iron from the environment. Making it is costly. The iron-pyoverdine complex is then available to any cell with the right receptor, including cells that make none. Griffin, West and Buckling showed that non-producing cheats invade producer populations, and that they do best in well-mixed conditions where the good really is shared, while spatial structure favors producers.
  • Why it's a collective action case: A textbook public good — non-excludable, costly to produce, freely consumed — in a system of bacteria. The experimental result is the one the course needs: cooperation survives when interaction is local and neighbors are relatives, and collapses when the population is well mixed. That is Nowak's spatial structure result, run in a flask, and it is the most direct available evidence for the claim that structure rather than virtue determines outcomes.
  • Could serve: Session 2, The games, on public goods and free riding. Session 3, The evolution of cooperation, as the wet-lab version of the spatial-grid result.
  • Footage: Poor. This is a plate-and-flask experiment; the visible output is fluorescent colonies and a growth curve. There is public footage of siderophore-producing colonies under UV (pyoverdine fluoresces) from microbiology teaching channels, but nothing that carries a room. Use a diagram, not film. Included because the science is load-bearing, not because it can be shown.
  • Source: Griffin, A.S., West, S.A. & Buckling, A., "Cooperation and competition in pathogenic bacteria", Nature 430:1024–1027, 2004. West, S.A. et al., "The social lives of microbes", Annual Review of Ecology, Evolution, and Systematics 38:53–77, 2007.

21. Myxococcus xanthus developmental cheats

The case

  • What happens: Myxococcus xanthus is a soil bacterium that hunts in swarms and, on starvation, aggregates into multicellular fruiting bodies in which some cells become spores. Velicer, Kroos and Lenski isolated mutants that are defective at fruiting-body development in isolation but, when mixed with wild type, over-represent themselves in the spore pool at the wild type's expense.
  • Why it's a collective action case: A second, independent instance of the same structure as entry 19, in a bacterium rather than an amoeba. Two unrelated lineages independently evolving both multicellular cooperation and the cheating that exploits it is a stronger claim than either alone: the problem is in the payoff structure, not in the organism.
  • Could serve: Session 2, Some basic game theory. Session 3, The cooperative ape. Use alongside 19 rather than instead of it; the point of having both is convergence.
  • Footage: Moderate. Myxococcus swarming and rippling time-lapse is genuinely good film — visible waves moving across an agar plate — and is available from the Velicer lab (ETH Zurich) and various microbiology labs. Fruiting-body formation time-lapse also exists. The cheating itself is not filmable, only the assay.
  • Source: Velicer, G.J., Kroos, L. & Lenski, R.E., "Developmental cheating in the social bacterium Myxococcus xanthus", Nature 404:598–601, 2000. https://www.nature.com/articles/35007066

22. Physarum network optimization

The case

  • What happens: The plasmodial slime mould Physarum polycephalum is a single enormous multinucleate cell that spreads as a network of tubes. Placed on a map of the Tokyo region with oat flakes at the positions of the surrounding cities, it grew a network comparable to the actual Tokyo rail system in efficiency, cost and fault tolerance. It also solves mazes by retracting tubes on dead-end routes.
  • Why it's a collective action case: A boundary case, and worth including for exactly that reason. There is no group here — it is one cell — so there is no free-rider problem at all, and that is the point: the infrastructure problem gets solved effortlessly precisely because the units cannot have divergent interests. Set next to entry 19, where the same organism-type in cellular form does have divergent interests and does get cheated, it isolates the variable. Also a clean illustration of distributed optimization with no planner, which Session 4's polycentricity module needs.
  • Could serve: Session 4, Polycentricity. Session 3, The cooperative ape, as the contrast case.
  • Footage: Excellent. Physarum time-lapse is widely filmed — the Tokyo rail experiment itself was filmed and published as supplementary video with the Science paper and re-cut by numerous outlets; the Paris Zoological Park's "blob" exhibit (2019) generated substantial broadcast footage; BBC Planet Earth II and Green Planet carry slime mould time-lapse. 60 seconds of time-lapse of a network forming is unusually compelling television.
  • Source: Tero, A. et al., "Rules for Biologically Inspired Adaptive Network Design", Science 327(5964):439–442, 2010. Nakagaki, T., Yamada, H. & Tóth, Á., "Maze-solving by an amoeboid organism", Nature 407:470, 2000.

23. Legume–rhizobium host sanctions

The case

  • What happens: Legumes house rhizobial bacteria in root nodules and feed them carbon; the bacteria fix atmospheric nitrogen for the plant. Kiers and colleagues forced a normally cooperative strain to stop fixing nitrogen by replacing the air around one half of a split root system with an argon-oxygen mixture containing no N₂. The plant cut the oxygen supply to those nodules and the reproductive success of the bacteria inside fell by about half.
  • Why it's a collective action case: A mutualism between organisms with no shared interest whatever, stabilized by the host detecting under-performance and sanctioning it at the level of the individual nodule. Detection and punishment, applied locally, with no central authority and no capacity on either side to negotiate. This is the closest thing in biology to a well-designed institution, and it is the answer to "why doesn't every mutualism collapse into cheating".
  • Could serve: Session 3, Ostrom and the commons — monitoring and graduated sanctions, two of the design principles, running in a root. Session 4, Coercion.
  • Footage: Poor to moderate. Nodules can be photographed and dissected — pink leghemoglobin interiors are visually distinctive — and there is agricultural extension footage of nodule inspection. The experiment itself is a gas manifold and a spreadsheet. Use a diagram. Strong case, weak film.
  • Source: Kiers, E.T., Rousseau, R.A., West, S.A. & Denison, R.F., "Host sanctions and the legume–rhizobium mutualism", Nature 425:78–81, 2003. https://www.nature.com/articles/nature01931

24. Acacia ants and the parasite of their mutualism

The case

  • What happens: Swollen-thorn acacias in Central America house Pseudomyrmex ants, providing hollow thorns to live in and nectar and protein bodies to eat; the ants attack herbivores and prune encroaching vegetation. Pseudomyrmex nigropilosa occupies the same acacias, eats the same rewards, and does not defend the tree. Janzen documented it in 1975 as a parasite of a mutualism.
  • Why it's a collective action case: A free rider in a two-species mutualism, and — more interestingly — a case where the host has counter-adaptations. The acacia's nectar is sucrose-free, which is unattractive to generalist ants; mutualist workers lose the sucrose-cleaving enzyme invertase during development, so they are locked into their host, while exploiters retain a preference for sucrose. High-reward host trees are exploited less and defended better. The host is, in effect, screening for partners with no exit option.
  • Could serve: Session 3, The cooperative ape. Session 4, Changing the payoffs — the acacia's answer to cheating is to change what the reward is, which is the subsidy-design argument in miniature.
  • Footage: Good. Acacia ants swarming out at a disturbance is a standard and very effective sequence — BBC Private Life of Plants (1995) and Life (2009, "Plants") both carry ant-acacia material, as does Planet Earth II ("Jungles"). The distinction between defenders and the non-defending parasite is not filmed as such and would need narration over generic footage.
  • Source: Janzen, D.H., "Pseudomyrmex nigropilosa: A Parasite of a Mutualism", Science 188(4191):936–937, 1975. Heil, M. et al., "Divergent investment strategies of Acacia myrmecophytes and the coexistence of mutualists and exploiters", PNAS 106:18091–18096, 2009. Kautz, S. et al., "How to prevent cheating: a digestive specialization ties mutualistic plant-ants to their ant-plant partners", Evolution 63:839–853, 2009.

25. Honeyguides and human honey hunters

The case

  • What happens: Greater honeyguides (Indicator indicator) lead human honey-hunters to wild bees' nests; the humans open the nest, take the honey, and the bird eats the wax and larvae it cannot reach alone. Spottiswoode showed the communication runs both ways: the Yao honey-hunters' specific "brrr-hm" call raised the probability of being guided from about 33% to 66% and the probability of finding a nest from 17% to 54%, against control sounds of similar volume. Hadza honey-hunters in Tanzania use a whistle instead, and the birds respond to their local human signal.
  • Why it's a collective action case: A mutualism between a wild animal and humans that is unique among living species, and which runs on a learned signal on both sides. The division is clean — humans cannot climb to what the bird finds, the bird cannot open what the humans open — so the temptation to defect is small, though honey-hunters in some cultures traditionally burn or hide the wax to keep the bird hungry, which is a documented instance of the human side cheating.
  • Could serve: Session 3, The cooperative ape, as the bridge from animal to human cooperation — literally the same case on both sides. Session 5, Henrich and the trust of strangers, on culturally transmitted signals.
  • Footage: Good. Spottiswoode's fieldwork in the Niassa Special Reserve was filmed and released with the 2016 Science paper as supplementary video, and widely re-cut by the BBC, National Geographic and NPR. The BBC Natural Histories radio programme covers it with Spottiswoode in audio. Expect 2–3 minutes of usable film showing the call, the bird leading, and the nest opened.
  • Source: Spottiswoode, C.N., Begg, K.S. & Begg, C.M., "Reciprocal signaling in honeyguide-human mutualism", Science 353(6297):387–389, 2016. https://www.science.org/doi/10.1126/science.aaf4885 Wood, B.M. et al., "Mutualism and manipulation in Hadza-honeyguide interactions", Evolution and Human Behavior 35:540–546, 2014.

Notes for selection

The three with the best footage, if only three make the cut: the drongo false alarm (entry 4), orca wave-washing (entry 8), and Battle at Kruger (entry 2). All three are 4–8 minutes, all three are available now, and between them they cover cooperation working, cooperation being cheated, and cooperation as something nobody planned.

The three that do the most argumentative work, which is not the same list: Dictyostelium cheaters (19), the selfish sentinel correction (3), and the lion laggards (7). These are the entries that establish defection as structural and prevent the module from becoming a parade of charming animals.

Cases deliberately not included. Ant "agriculture" and termite mounds — well filmed, but the collective action structure is nearly trivial given clonality, so they illustrate less than they appear to. Dolphin cooperative feeding — the footage is good, but the coordination claims in the popular literature run well ahead of what is established. Emperor penguin huddling — excellent footage, and there is a real free-rider question about position in the huddle, but the literature on whether outer positions are exploited or rotated is not settled enough to state cleanly.

Edit this page

Links to this page:

  1. Example candidates — review index
Built with LogoFlowershow