Extinction Is Not Random: Which Species Are Most at Risk

Table of Contents

Why do some animal species face a particularly high risk of extinction, while others survive even major changes to their environment? Steller’s sea cow was hunted intensively, the crescent nail-tailed wallaby was driven back by introduced foxes and cats, and the Bramble Cay melomys ultimately lost what little habitat it had.

Every extinct species has its own story, and many extinction events cannot be reduced to a single cause. Yet comparisons across many such cases show that certain biological traits repeatedly coincide with greater vulnerability. These include body size, reproductive strategy, island endemism, and the ability to evade threats.

But these traits do not cause extinction. A species does not disappear because it is large, flightless, or restricted to a small island. The direct causes are specific threats—nowadays mostly human-induced—such as hunting and fishing, habitat loss, introduced species, pollution, or climate change. Diseases can also severely reduce populations.

Instead, biological traits help determine how vulnerable a species is to such threats and how well it can absorb losses. Conservation therefore depends not only on the dangers a species faces, but also on the question: Which traits make it especially sensitive to those dangers?

Large animals are often in a bad position

One risk factor that repeatedly emerges across different animal groups is body size. For example, a study published in 2026 by Simon Schowanek and his team found that large tropical mammals had poorer long-term survival prospects than smaller species.

This pattern also appears in birds. Amir Fromm and Shai Meiri examined 469 known bird extinctions since the late Pleistocene for a 2021 study. The extinct species were markedly larger overall than today’s avifauna: Their median body mass was 288 grams, compared with 36 grams among living species. The difference remained even when only birds capable of flight were compared.

Similar relationships have been found in reptiles and marine fishes. A global reptile study by Alex Slavenko and colleagues showed that extinct lizards and turtles were larger on average than extant species. Among marine fishes, large species are also more often threatened and more heavily affected by fishing. However, the effect of size does not apply equally to every animal group: No such relationship with size was found in snakes, for example; findings for freshwater fishes are inconsistent, and the pattern is also far less clear in amphibians—some studies even found a higher risk among very small species. In crocodilians, the reptile study found the opposite pattern.

Several biological relationships help explain why large species are often more vulnerable. They frequently need more food and larger habitats, usually occur at lower population densities, and often have a slow pace of life. Their populations may therefore be less able to compensate for losses.

In addition, large animals have often been—and remain—especially attractive to humans: They provide large quantities of meat or have coveted body parts such as horns, antlers, tusks, or fur. Large animals have consequently often been exposed to intense hunting pressure. Fromm and Meiri, for example, cite archaeological sites where the bones of moas, elephant birds, and other large birds show evidence of human use.

Large body size is therefore not a universal death sentence. But when it coincides with slow reproduction and intense human exploitation, populations can be especially vulnerable.

Slow reproduction as a risk factor

Large body size and slow reproduction are closely linked. Schowanek and his team were nevertheless able to separate their effects statistically. Their results showed that even among mammals of similar size, species with longer generation times had a higher risk of extinction.

The relationship is easy to understand. Species that reach sexual maturity late and reproduce slowly are less able to compensate for population losses. If many adult animals are killed, their population may take years or decades to recover.

A 2024 study of 2,813 vertebrate species shows that a slow pace of life can also be problematic in other animal groups. According to the study, island birds and island mammals often have longer generation times than their mainland relatives—a trait that can increase their vulnerability to human threats.

Similar relationships are also found among amphibians and fishes. A global study of more than 2,000 amphibian species showed that species producing few offspring per reproductive event were more threatened. In marine fishes, late sexual maturity in particular indicates that severely depleted populations may recover only slowly.

Slow reproduction therefore becomes a risk primarily when additional losses occur. Intensive hunting, fishing, or other pressures can reduce populations faster than reproduction can replenish them.

Island species are especially vulnerable

A heightened risk of extinction is particularly apparent among island species. Of the extinct bird species studied by Fromm and Meiri, around 90% were island endemics. Among the living bird species in their dataset, the corresponding figure was only 24%.

The pattern is not limited to birds. Of 268 known mammal extinctions in the Holocene, 225 occurred on islands. Among reptiles, 73 of 82 known extinction events over the past 50,000 years occurred on islands. The ratio is even more pronounced for the past 3,000 years: 64 of 65 known reptile extinctions occurred on islands. Across several animal groups, insularity is thus one of the most conspicuous risk factors.

There are several reasons for this. Island species usually have small populations and limited geographic ranges. Many islands have also experienced long periods of evolutionary isolation. Remote islands often originally lacked terrestrial mammalian predators, so some species evolved without effective defenses against such enemies.

With the arrival of humans, these conditions often changed within a short time. Introduced rats, cats, and other animals ate eggs and young, hunted adults, or altered habitats and food webs. Island birds were affected particularly severely. One well-known example is Lyall’s wren, whose disappearance was significantly influenced by introduced cats.

Traits that had been successful for long periods under an island’s original conditions could suddenly create exceptional vulnerability after humans and new animal species arrived.

Flightlessness—an especially high risk

Among birds, flightlessness is a particularly conspicuous risk factor. Of the 469 extinct species studied by Fromm and Meiri, 116 were flightless—roughly one in four. According to their data, 68% of the known flightless bird species still in existence at the end of the Pleistocene have since become extinct. Of these losses, 58 affected rails alone.

Flightlessness was therefore far more widespread in the recent past than it is today. Entire groups such as moas and elephant birds disappeared completely, and several bird orders now have no flightless representatives even though they once did.

Yet flightlessness is not an evolutionary mistake. Flight consumes a great deal of energy, and on islands without terrestrial predators, the ability to fly can become less important. This adaptation becomes problematic when humans or introduced predators appear: Flightless birds cannot escape threats by taking flight, while the eggs and young of ground-nesting species are also easy to reach.

The dodo, the moas, and numerous extinct island rails show how dangerous the combination of flightlessness, island life, and new predators or human persecution can be.

Exposure and refuges

Flightlessness is a trait specific to birds. A similar general principle is found in other animal groups: How easily can a threat reach a species—and what options does it have for evading that threat?

In Schowanek’s study, strictly terrestrial mammals disappeared more frequently over the long term than species that can climb at least some of the time. One possible explanation is that animals that climb are harder for humans or terrestrial predators to reach. Yet living in trees does not offer general protection from extinction: If the forest is lost, that dependence can become a disadvantage.

A similar pattern occurs among sharks and rays. Species that live in shallow waters or use only a narrow depth range are at greater risk, partly because they are more exposed to fishing. Species that occupy a broader range of depths can, by contrast, retreat at least partly into areas less affected by certain fishing methods. What matters, then, is not the lifestyle alone, but how it interacts with a specific threat.

Why carnivores may be at risk

Among the mammals studied, a higher proportion of animal matter in the diet was associated with a higher long-term risk of extinction. Carnivores often occupy high positions in the food chain, frequently require large home ranges, and depend on sufficient prey. If prey populations decline, predators can therefore also come under pressure.

For example, Haast’s eagle may have become extinct on New Zealand’s South Island partly because its most important prey, the moas, disappeared. Large predators also frequently come into conflict with humans, for example when they kill livestock or are perceived as a danger.

However, the relationship is less clear-cut than those involving body size or generation length. Closely related mammals often resemble one another not only in their diets, but also in other traits. When Schowanek and colleagues accounted for this evolutionary relatedness, the relationship became much weaker. It therefore does not follow that carnivores are generally at greater risk than herbivores.

Can a larger brain aid survival?

There is evidence that behavioral flexibility may also influence extinction risk in mammals. In Schowanek’s study, species with a relatively larger brain had better long-term survival prospects. This does not refer to absolute brain size, but to a brain larger than would be expected for the animal’s body size.

A 2022 study of Late Quaternary mammal extinctions had already reached a similar conclusion. One possible explanation is that a larger brain may be associated with greater behavioral flexibility—for example, the ability to use new food sources or respond to changing environmental conditions.

The relationship is not unequivocal, however. Large brains require a great deal of energy and can be associated with a slower pace of life. Some studies of living mammals have therefore found the opposite relationship. Whether a relatively large brain offers an advantage probably depends on the environmental changes and threats a species faces.

Small geographic ranges—high risk

Another fundamental risk factor is geographic range size. A widespread species can often absorb local losses because populations survive in other regions. For an endemic species with a tiny range, by contrast, a single major change can affect a large proportion—or even all—of its global population. The Bahama nuthatch, for example, may have disappeared for good in 2019 after several hurricanes struck its home island of Grand Bahama.

The link between small geographic ranges and greater risk appears across different animal groups. A global analysis of 5,246 anuran species, for example, found that species with small ranges were especially likely to have declining populations. Small geographic ranges are also among the traits associated with greater threat levels in marine fishes.

This relationship also partly explains the exceptional vulnerability of island species. It is not the island itself that matters, but, among other things, the fact that many island endemics occur only across very small areas. The same problem affects mainland species that live, for example, only in a single mountain range, river system, lake, or forest.

Specialization can become a trap

Ecological specialization can also make a species more vulnerable. Generalists can use different food sources or inhabit a range of habitats. Specialists, by contrast, may depend on a particular food, habitat, or type of nesting site. If precisely that resource disappears, they have few alternatives available.

This is clearly illustrated by coral reef fishes. Some species are closely tied to particular corals or reef habitats. Studies show that such specialists are especially sensitive to the loss of their preferred corals—for example, through coral bleaching or other changes to the reef ecosystem.

Specialization is not inherently disadvantageous. It can be highly successful under stable conditions. It becomes a risk primarily when the environment changes rapidly and the required resource disappears. As with other risk factors, the principle is: The fewer alternatives a species has, the more difficult it may find it to respond to change.

Combinations of factors increase extinction risk

None of the factors discussed so far inevitably causes extinction on its own. Large, slowly reproducing species can thrive over very long periods, as can flightless birds or island species with stable populations. What often matters is the combination of several traits with a specific threat.

A large, slowly reproducing island endemic with a small geographic range, for example, may be far less able to withstand intensive hunting than a small, fast-reproducing, widespread species. Additional pressures such as introduced predators can heighten its vulnerability further.

The dodo combined several such traits: It was large and flightless, occurred only on Mauritius, and had evolved there in the absence of terrestrial mammalian predators. With the arrival of humans and introduced animals, these conditions changed within a short time.

A study of 241 coral reef fish species in the Indo-Pacific shows that such combined effects also occur in other animal groups. Particularly large species with small geographic ranges were especially vulnerable to intense human pressure.

The key point, then, is not to view a species as “destined for extinction” based on individual traits. Instead, such traits indicate how much scope a population has to respond when new pressures arise.

Extinction changes our idea of “normal”

The fact that species do not disappear at random has another consequence: When species with particular traits are lost disproportionately over long periods, the composition of the animal world changes. The species we observe today are therefore not a random sample of the diversity that once existed.

This is especially apparent in birds. Today, fewer than 0.5% of bird species are flightless; flightlessness was far more widespread in the recent past. The size distribution of the world’s birds has also changed because a disproportionate number of large species disappeared. Fromm and Meiri therefore caution against inferring how common particular traits used to be based solely on living species.

This problem is reminiscent of the concept of shifting baselines: The current state can easily become the benchmark for what is considered “normal.” Yet if numerous species have already disappeared, an impoverished animal world may mistakenly appear to be the natural starting point.

The study by Schowanek and his team illustrates a similar problem. The shorter the period considered, the more some relationships between biological traits and extinction risk changed. Traits associated with greater losses over the long term could even appear to be linked to better survival prospects in present-day local data.

One possible explanation is clear: Many of the most sensitive species are already gone. Anyone looking only at today’s survivors may therefore underestimate which traits made species vulnerable in the past.

Many losses remain unknown

Our knowledge of past extinction events is incomplete. Fromm and Meiri assume that the 469 extinct bird species in their study represent only a proportion of the actual losses. Many remote islands have received insufficient study, numerous fossils have yet to be described scientifically, and the most recent known record of a species need not coincide with the actual date of its extinction.

There is also potential bias in the fossil record: The bones of large birds are more likely to be preserved and are easier to discover than those of smaller species. Their share of known extinction events could therefore be overestimated. Historical accounts and archaeological finds nevertheless suggest that large birds were indeed particularly severely affected by humans.

A study published in 2026 by Søren Faurby and colleagues shows how large the gap may be. In their preferred model, the researchers estimated a median of 881 undocumented bird extinctions. Many of these unknown species probably lived on islands, where small geographic ranges and a patchy fossil record make extinct species harder to discover.

Known extinction events therefore probably represent only a proportion of the true loss. On islands in particular, the scale of past losses may still be significantly underestimated.

What does this mean for conservation?

Biological traits can help identify particularly vulnerable species at an early stage. No individual trait means that a species will become extinct. But the presence of several known risk factors can be an important warning sign.

What always matters is the specific threat these traits encounter. A small geographic range, slow reproduction, or limited options for evasion become problems primarily when hunting, fishing, habitat loss, introduced species, or other pressures are simultaneously affecting a population.

Such traits are not evolutionary mistakes. They could remain successful for long periods under the conditions in which they arose. The danger emerges when the environment changes faster than a species can respond. Humans can alter habitats within a few decades, severely deplete populations, or introduce new predators—evolutionary adaptations, by contrast, require many generations.

Especially for species whose population trends are still poorly known, such risk profiles can indicate which species should be studied more closely or afforded greater protection as a precaution. They do not, however, replace data on populations and specific local threats.

Schowanek’s study also shows that long-term risk patterns cannot simply be applied to every present-day population. Whether traits become advantageous or disadvantageous also depends on the period considered and local conditions. In some current local datasets, for example, large, slowly reproducing mammals appeared comparatively resilient. This could be due in part to the fact that particularly sensitive species have already disappeared and many of the areas studied were at least partially protected.

For conservation, this means that biological traits are warning signs, not predictions. Their interaction with the actual threats to a population remains crucial.

Extinction does not strike species at random

The various studies do not reveal a universal “extinction profile.” Not every risk factor applies to every animal group, and the same trait can have very different consequences under different conditions.

Nevertheless, certain patterns recur: large body size, slow reproduction, small geographic ranges, island endemism, strong specialization, and limited options for evading threats. Flightlessness can play an additional important role among birds; among mammals, there is evidence of other influences, such as diet and relative brain size.

What usually matters is not a single trait, but the interaction between biological vulnerability and a specific threat. A trait becomes a risk primarily when it leaves a species with little scope to respond to hunting, fishing, introduced species, or rapid changes to its habitat.

The stories of extinct species therefore show more than why individual species disappeared. Only by comparing many such stories do recurring patterns become visible. They also remind us that the animal world we observe today has already been shaped by these unevenly distributed losses. Today’s natural world is therefore not the starting point; it has been shaped by millennia of nonrandom extinction events.

---

Sources

• Bak, T. M., Camp, R. J., Heim, N. A., et al. (2023). A global ecological signal of extinction risk in marine ray-finned fishes (class Actinopterygii). Cambridge Prisms: Extinction, 1, e25. https://doi.org/10.1017/ext.2023.23
• Dembitzer, J., Castiglione, S., Raia, P., & Meiri, S. (2022). Small brains predisposed Late Quaternary mammals to extinction. Scientific Reports, 12(1), 3453. https://doi.org/10.1038/s41598-022-07327-9
• Faurby, S., Matthews, T. J., Triantis, K. A., & Sayol, F. (2026). Quantifying the unrecorded loss of avian phylogenetic diversity. Ecography, 2026(2), e08267. https://doi.org/10.1002/ecog.08267
• Fromm, A., & Meiri, S. (2021). Big, flightless, insular and dead: Characterising the extinct birds of the Quaternary. Journal of Biogeography, 48(9), 2350–2359. https://doi.org/10.1111/jbi.14206
• Graham, N. A. J., Chabanet, P., Evans, R. D., et al. (2011). Extinction vulnerability of coral reef fishes. Ecology Letters, 14(4), 341–348. https://doi.org/10.1111/j.1461-0248.2011.01592.x
• Loiseau, N., Mouillot, D., Velez, L., et al. (2024). Inferring the extinction risk of marine fish to inform global conservation priorities. PLOS Biology, 22(8), e3002773. https://doi.org/10.1371/journal.pbio.3002773
• Mellin, C., Mouillot, D., Kulbicki, M., et al. (2016). Humans and seasonal climate variability threaten large-bodied coral reef fish with small ranges. Nature Communications, 7, 10491. https://doi.org/10.1038/ncomms10491
• Pincheira-Donoso, D., Harvey, L. P., Cotter, S. C., et al. (2021). The global macroecology of brood size in amphibians reveals a predisposition of low-fecundity species to extinction. Global Ecology and Biogeography, 30(6), 1299–1310. https://doi.org/10.1111/geb.13287
• Sandretti-Silva, G., Schwarz Meyer, A., Caron, F. S., et al. (2026). Phenotypic and ecological correlates of population decline in the world’s anurans. Ecology and Evolution, 16(3), e73168. https://doi.org/10.1002/ece3.73168.
• Schowanek, S. D., Sheil, D., Dupont, P., et al. (2026). Predictors of extinction risk in large tropical forest mammals: From global to local. Science Advances, 12(29), eaeb7543. https://www.science.org/doi/10.1126/sciadv.aeb7543
• Slavenko, A., Tallowin, O. J. S., Itescu, Y., et al. (2016). Late Quaternary reptile extinctions: Size matters, insularity dominates. Global Ecology and Biogeography, 25(11), 1308–1320. https://doi.org/10.1111/geb.12491
• Smith, R. (2026, August 20). How well do past extinctions predict extinction risk today? Phys.org. https://phys.org/news/2026-08-extinctions-extinction-today.html
• Xiong, Y., Rozzi, R., Zhang, Y., et al. (2024). Convergent evolution toward a slow pace of life predisposes insular endotherms to anthropogenic extinctions. Science Advances, 10(28), eadm8240. https://doi.org/10.1126/sciadv.adm8240