Due to our intensive subsistence and habitat-modification strategies—including broad-spectrum harvesting and predation, widespread landscape burning, settlement construction, and translocation of other species—humans have major roles as ecological actors who influence fundamental trophic interactions. Here we review how the long-term history of human–environment interaction has shaped the evolutionary biology of diverse non-human, non-domesticated species. Clear examples of anthropogenic effects on non-human morphological evolution have been documented in modern studies of substantial changes to body size or other major traits in terrestrial and aquatic vertebrates, invertebrates, and plants in response to selective human harvesting, urbanized habitats, and human-mediated translocation. Meanwhile, archaeological records of harvested marine invertebrates and terrestrial vertebrates suggest that similar processes extend considerably into prehistory, perhaps to 50,000 yr BP or earlier. These results are consistent with palaeoenvironmental and other records that demonstrate long-term human habitat modification and intensive harvesting practices. Thus, while considerable attention has been focused on recent human impacts on ‘natural’ habitats, integrated evidence from modern biology and archaeology suggests a deep history of human entanglement with our ecosystems including substantial effects on the evolutionary biology of non-human taxa. The number and magnitude of such effects will probably increase given the continued intensification of anthropogenic activities and ecosystem impacts, including climate change and direct genetic modification.
The subsistence and habitat-modifying behaviours that typify how humans interact with their surrounding environments have broad and temporally profound effects on natural ecosystems1. At local levels, co-evolutionary relationships between small-scale human societies and non-human taxa in the same community can be so longstanding and immersive that the removal of humans results in simplified food webs or ecosystem collapse2,3, while in the context of global population growth the intensity of human–environment interactions has led to numerous documented cases of wildlife population decline and extinction4,5. Given the extent and rate that human-mediated climate change and other by-products of increasing human population density are predicted to impact future ecosystems6,7, these issues are of major concern to biologists, ecological anthropologists, and policymakers.
In addition to ecosystem health and wildlife extinction risk, human behaviour also influences the evolutionary biology of non-human species. Extensive morphological evolution mediated by human behaviour has been thoroughly documented for domesticated plants and animals8,9, but these processes are in fact considerably more widespread, extending to non-domesticated species. That is, wild populations of non-human taxa can adapt via natural selection on functional genetic variation in response to human-induced ecological changes10. In some cases, these morphological and behavioural adaptations may help certain species avoid extinction. These processes could have cascading effects on ecosystem function6,11, so they are an important factor to consider in the broader context of human–environment interactions.
Clear examples of anthropogenic effects on the evolutionary biology of non-human, non-domesticated species have been documented in modern studies of body size or other morphological trait change in response to selective human harvesting of natural populations of vertebrates, invertebrates, and plants11,
In this Review we describe potential mechanisms for non-human adaptive evolution in response to human behaviour, with human behaviour defined broadly to encompass landscape modification (including niche construction); the harvesting, artificial selection, or translocation of non-human species; and even current and potential future directed practices involving genetic modification and de-extinction. These processes are illustrated by examples from modern evolutionary biology and enhanced with ethnographic and prehistoric perspectives that include discussion of evidence for the long-term history of anthropogenic effects on the environment. We focus primarily on morphological evolution, which does not encompass all adaptations (for example, genetically mediated behavioural or metabolic adaptations) but provides opportunities for insights from modern biology to be translated to the archaeological and palaeontological records. The consideration of these data in a long-term, integrated framework will ultimately help us identify and model these evolutionary processes as they may occur in the future.
The antiquity of humans as trophic regulators
Disturbance ecology quantifies the effects of natural ecosystem disruptions that may result from abiotic forces (for example, fire, flood and drought) or biotic variation (for example, intensive predation, disease outbreak and depletion of keystone species)17. Humans have major roles as biotic influencers of ecological disturbance due to our intensive subsistence strategies that include broad-spectrum (of a high diversity of resources) harvesting and predation18,
Although many studies of the relationships between human activity and the population health and evolution of non-human species have focused on recent human impacts in ‘natural’ ecosystems, archaeological and palaeoenvironmental records demonstrate a long-term history of major anthropogenic effects that began well prior to the origins of agriculture1.
With respect to intensive harvesting, while there is fossil-record evidence for the consumption of meat by African hominins by at least 2.5 million yr BP27,28, it is unclear whether this early record reflects hunting or scavenging and whether this behaviour was intense or sporadic29. However, good evidence for repeated, systematic butchery of terrestrial vertebrate species by hominins, consistent with targeted hunting and processing, is observed by ∼780,000 yr BP30, and hafted spear technology appears in the archaeological record by at least ∼500,000 yr BP31. Archaeological data from South Africa32 and Spain33 also show that broad use of coastal resources by hominin foragers began by at least 150,000 yr BP34.
Analyses of pollen and charcoal records suggest the initiation of widespread anthropogenic landscape burning between ∼50,000 and 40,000 yr BP in Borneo and New Guinea, coincident with the respective appearances of modern humans in each region35,36. Some scholars even infer that the purposeful use of fire for habitat modification—a practice common to both agriculturalists and hunter-gatherers worldwide today23,24—may have roots in the evolution of the genus Homo close to two million yr BP21.
Human–environment interactions later intensified further with the origins and spread of agriculture and animal husbandry beginning ∼12,000 yr BP1. Human-mediated ecological impacts during this period have included extensive land clearing, terracing, waterway diversion, the construction of permanent settlements and ultimately urban centres, translocation of non-endemic species to new habitats, the potential for magnified selective hunting and fishing pressures on non-domesticated species associated with increasing human population sizes and densities, climate change, and now even direct genetic modification of non-human taxa.
Through these various processes humans became deeply entangled in their ecosystems, as illustrated by the cascading effects that have been observed following the removal of human activity from some habitats3. For example, in desert Australia, Aboriginal predation of monitor lizards involves disturbing climax vegetation via burning tracts of old-growth xeric grassland, creating a tight mosaic of vegetative succession that supports higher densities of many traditionally hunted species37,38. A mid-twentieth-century hiatus in traditional burning led to trophic collapse and coincided with the extinction of at least 21 endemic marsupials2. Thus, while human predation and habitat modification can certainly affect non-human population distributions negatively, leading to extinction and food web simplification5,39,40, there is also strong evidence for significant ecological adaptation to the long-term keystone presence of intensive human disturbance, which could include adaptive morphological evolution in non-human taxa.
Human behaviour as a driver of non-human evolution
One of our goals in this Review is to demonstrate the antiquity of widespread phenotypic adaptation to human behaviour by connecting observations and insights from modern evolutionary biology to archaeological and palaeontological observations. To do so, we focus on the evolution of major morphological phenotypes—that is, traits that may be preserved in prehistoric records (unlike, for example, behavioural and metabolic changes)—as one subset of a diverse spectrum of genetic-based traits in non-human species that have evolved in response to human behaviour41,
Artificial selection associated with the domestication of non-human animals and plants is the best-known example of morphological evolution in response to human behaviour. While agriculture and pastoralism originated by at least ∼12,000 yr BP1, the roots of these processes likely extend considerably further into prehistory, for example to ∼23,000 yr BP49 or earlier50 for plants and >30,000 yr BP for dogs51. Here, because both the scale and archaeological prehistory of morphological evolution in domesticated taxa have been reviewed extensively8,9,52,
Response to size- or trait-selective harvesting pressures
Human patterns of prey selection can vary substantially from those of other predators. Although probably not applicable to all societies, at least some human populations may harvest particular species with unparalleled intensity, and may disproportionately target large adults of a given prey species or individuals with a specific trophy feature, such as antlers or horns18. Rapid phenotypic evolution may result11.
Morphological evolution has been especially well documented in fisheries, in which larger prey are typically of higher value and netting technology may greatly enrich the catch for larger individuals based on the sizes of the net openings13,55,
In addition to marine and freshwater vertebrates, organism size changes in response to human harvesting pressures have also been reported in natural populations of terrestrial vertebrates and even plants (Fig. 2). An excellent example, documented in a carefully controlled study, is that of the wild snow lotus Saussurea laniceps. This plant, from the eastern Himalayas, is used in traditional Tibetan and Chinese medicine for headache and high blood pressure treatments. Larger plants are considered more potent and thus collected preferentially. Based on herbarium specimens (of flowering plants; when maximum plant height is achieved), Law and Salick62 reported a ∼45% decrease in S. laniceps heights over the past century. In contrast, flowering heights of the sympatric and related—but substantially less desired medicinally and little harvested—snow lotus Saussurea medusa were unchanged over the same period. Moreover, modern S. laniceps individuals subject to harvesting in non-protected areas were ∼70% shorter than those in protected areas62.
Trophy hunters exert directional pressure on a particular phenotype, potentially leading to the evolution of smaller features across the population or higher rates of feature absence. For example, illegal ivory hunting in Zambia resulted in an increase of >300% (from 10.5% to 38.2%) of tuskless female African elephants (Loxodonta africana) from 1969 to 198963. Trophy hunting may also have pleiotropic effects on other phenotypes that are not the direct targets of selection, as observed for bighorn sheep (Ovis canadensis) on Ram Mountain in Alberta, Canada. Specifically, in addition to a ∼30% reduction in bighown male horn length over 23 years of horn-size-focused hunting pressure (Fig. 3), body mass also decreased ∼23.5% over the same period64,65.
Is trophy hunting limited or widespread among human societies, and what is the antiquity of this behaviour? In many societies, prestige goods do serve as costly signals of status and commitment and probably play an important role in the development of inequality66,67, and in foraging groups hunting is commonly as much a political activity as it is a provisioning practice68,69. Discerning archaeological signatures of prestige hunting is difficult but has been proposed in several cases70.
Well-documented archaeological evidence for prey size changes in response to human harvesting pressure does exist in the record of intertidal mollusc exploitation14 (Fig. 4). While many predators (marine, terrestrial, and avian) exploit shellfish, only humans transport large loads to central locales for repeated processing and deposition, creating substantial heaps or trash ‘middens’ of the inedible, durable mollusc shells and the remains of other taxa that come to represent temporal records of both direct human predation on these animals and their sizes71,72. Such harvesting can have substantial effects, with selective pressures resulting in either morphological change in targeted molluscs or predation resulting in shellfish populations with persistently skewed age–size profiles.
Modern human intertidal foragers worldwide are selective and highly sensitive to changes in the trade-offs associated with searching for and handling different types and sizes of shellfish73,
These expectations are consistently met by analyses of the archaeological record. Early evidence for routine, large-scale intertidal mollusc exploitation has been recorded beginning in the middle stone age (∼120,000 to 60,0000 yr BP) at coastal cave sites in South Africa71,77. These South African middle stone age shellfish assemblages have been compared to those that accumulated in midden layers dated to the later stone age (∼12,000 to <1,000 yr BP), with significant observed reductions over time in the sizes of all represented species78,79, including for the Cape turban shell (Turbo sarmaticus) as depicted in Fig. 5.
Similar evolutionary trends are evident in many coastal environments elsewhere. For example, molluscs recovered from midden layers dated to ∼19,000 to 9,000 yr BP at Riparo Mochi, Italy were ∼30% smaller than those in earlier layers dated to ∼36,000 to 24,000 yr BP15. Likewise, in the California Channel Islands, owl limpets (Lottia gigantea) have experienced ∼40% size reductions since the onset of human foraging ∼10,000 yr BP80.
In each of the above cases, the observed patterns of morphological change are inconsistent with known patterns of palaeoenvironmental variation, suggesting that the morphological change probably reflects responses to anthropogenic activity. However, because growth is continuous throughout life for many of these taxa, confidently distinguishing adaptive morphological evolution from by-products of changes in harvest age profiles based on archaeological midden data is also a challenge79. Still, such analyses are possible for species with clear markers of growth cessation (for example, for skeletal long bones with their epiphyseal fusion) or at least of maturation. For example, sizes of both the largest juveniles and the smallest adult conch (Strombus pugilis) shells became dramatically smaller over the past 7,000 years of human harvesting pressure in Carribbean Panama—conch harvested for food today have ∼40% less meat than the conch that existed when human predation began81—demonstrating a morphological evolutionary effect rather than only a mortality profile change.
Without the analytical benefits afforded by extensive shell midden deposits, archaeologists are more challenged to identify evidence of morphological evolution in response to human harvesting pressures on terrestrial vertebrates. Still, Stiner15 reported a 19% size decrease (based on humeral shaft diameter as a proxy for body size) for spur-thighed tortoises (Testudo graeca) that cannot be explained by patterns of climatic variation across archaeological layers dated to 200,000–150,000 yr BP, 100,000–70,000 yr BP, and 28,000–11,000 yr BP at Nahal Meged, a Neandertal and modern human site in Israel.
Response to landscape modification and urbanization
Anthropogenic burning and clearing, irrigation infrastructure, horticultural and pastoral disturbance, and structural investment have widespread ecological consequences for other organisms sharing the affected habitats. Direct and indirect effects or responses to human landscape modification—both positive and negative—have been well documented for many non-human species, including as changes to behaviour, distribution, and abundance2,37,38,82,
Anthropogenic landscape modifications also create contexts of novel selective pressure, providing a substrate for non-human morphological evolution16,86. An exceptional example comes from a long-term study of cliff swallows (Petrochelidon pyrrhonota) in Nebraska, where roadside nesting behaviour increased in frequency starting in the 1980s with expanding bridge and culvert construction87. Overall population wing length decreased ∼2% across a 30-year period, probably because shorter wings aided flight agility and reduced vehicle collision risk. Birds killed by cars were consistently longer-winged than the overall population, and the frequency of road-killed birds steadily declined despite the increasing proportion of sport-utility vehicles over the study period87.
Human-mediated habitat fragmentation may also exert evolutionary pressures on morphological traits—perhaps especially those related to dispersal88. For example, increased wing pointedness in birds facilitates longer-distance mobility in the face of increased habitat isolation89. By studying museum specimens, Desrochers90 found that songbird relative wing pointedness increased by a magnitude of 7–28% over the last century in North American forests heavily affected by clear-cutting. The same pattern was not observed for birds in early-successional and intact forests90.
Urban environments in particular are hotbeds for behavioural specializations and population density increases among non-human taxa, even while other species have been extirpated91. These environments are also home to some of the strongest current examples of non-human adaptation in response to human landscape modification92. In a recent study, Winchell et al.93 examined 319 adult male Puerto Rican crested anole lizards (Anolis cristatellus) from three independent pairs of urban and nearby forested habitat populations. Average limb lengths for urban lizards were consistently and significantly longer (∼2% longer), and urban individuals also had significantly greater numbers of subdigital scales (∼3% greater), with both traits potentially facilitating more efficient locomotion on artificial surfaces. Results from a subsequent common garden experiment suggest a genetic, and thus potentially evolutionary, basis for the phenotypic differences93. House finches in urban Arizona have ∼2% larger beaks relative to populations in nearby desert habitat, a putative genetic adaptation to facilitate greater bite forces for consumption of seeds in backyard feeders that are relatively larger and harder-shelled than those typically processed by the non-urban finches94. Adaptive beak shape changes associated with feeder-based diets are also hypothesized for Central European blackcaps (Sylvia atricapilla)95. However, because bird song acoustics may also adapt to urban noise environments96 and beak morphology also modulates song frequency97, the ultimate evolutionary pressure(s) affecting urban bird beak shape need to be confirmed.
Response to human-mediated taxonomic turnover
Human population movement, landscape modification, and harvesting practices are (and have been) associated with widespread translocation and extinction of thousands of non-human species. For example, an incredible number of plant species—more than 13,000, or ∼4% of the known extant flora—have been translocated by humans98. While expanding global travel and trade have led to an increased rate of human-mediated translocation in the past 150 years99, the purposeful or accidental introduction of non-domesticated plants and animals to naive ecosystems has a long prehistory, likely back to the earliest human colonizations of at least some global landmasses100,
Studies of potential morphological adaptation in translocated species are challenged by the high levels of genetic drift potentially associated with founder events that may also result in (non-adaptive) phenotypic change over time105,106, but evolutionary genomic or fitness analyses may be used to evaluate hypotheses of neutral versus adaptive evolution in such cases. For example, purple loosestrife (Lythrum salicaria), a European wetland plant invasive to North America, flowers ∼20 days earlier and is ∼50% smaller near the northern versus southern extent of its East coast invasive range107. Based on common garden experiments, these traits are largely genetically controlled, and strong fitness advantages for earlier reproduction in the north and larger plant size in the south were identified in transplantation experiments107.
Introduced species can also affect the behavioural ecology and evolutionary biology of endemic taxa. There is mounting evidence that prehistoric translocation events induced trophic cascades with continental-scale impacts on fauna and flora108. Contemporary examples of the role of invasives in shaping the morphology of endemic taxa include South American venomous fire ants (Solenopsis invicta), introduced by the 1940s into the southern United States, where they now prey on endemic fence lizards (Sceloporus undulatus). Adult fence lizards at sites with the longest times since ant invasion have 3.4% longer hindlimbs and have higher rates of body twitching and fleeing responses (which are more effective in longer-limbed individuals) to fire ant attack compared with lizards at not-yet-invaded sites109. This pattern is inconsistent with expected ecogeographic variation based on morphology of museum specimens collected prior to fire ant invasion, altogether suggesting a recent history of phenotypic adaptation109. Meanwhile, the cane toad was introduced into Australia in 1935. This invasive species is highly toxic upon consumption to some but not all endemic snakes, although smaller toads are less likely to deliver sufficient volumes of toxin to be fatal. Based on morphological analyses of museum specimens, two snake species (Pseudechis porphyriacus and Dendrelaphis punctulatus) have evolved smaller heads over the past 70 years that putatively preclude the consumption of larger toads110.
Human-mediated extinction may have cascading effects on the evolutionary biology of surviving non-human species in the same ecosystem, especially when the extinctions involve keystone species. We are currently experiencing a mass extinction crisis, with (for example) 8–100 times higher extinction rates over the past century for vertebrates than the long-term background rate5. This crisis may be connected to longer-term extinction processes related to longstanding human landscape modification practices and harvesting pressures.
Like humans, megafauna are generally keystone trophic facilitators for other species in the same habitats111. Of ∼150 terrestrial mammalian genera >44 kg living on Earth's continents 50,000 yr BP, approximately 65% were extinct by 10,000 yr BP112. Recent analyses of new, high-resolution palaeoenvironmental and chronological datasets have given some (but not all) scholars increasing confidence that colonizing humans played at least indirect roles (for example, via landscape burning) in prehistoric megafaunal extinctions in multiple large island habitats1,113,114. At continental scales the role of humans in this process remains less clear115,116, with natural climate change likely a major factor in at least certain cases117. Our purpose is not to draw conclusions from these ongoing investigations or to review the associated debate. Rather, we suggest that if our ancestors did play roles in the extinctions of megafauna, then those events would represent another mechanism through which human behaviour may have indirectly affected the evolutionary biology of other (non-megafaunal) non-human taxa. Regardless, by considering the outcomes of these past events we can raise our awareness of ongoing and future evolutionary responses to modern extinction processes.
Ecosystem effects from the removal of keystone megafauna likely included niche space changes and openings for surviving taxa, resulting in opportunities for morphological evolution that might be detected in the palaeontological record. For example, in the ∼1,000 years following megafaunal extinctions in North America, coyote (Canis latrans) femur circumferences—a proxy for body size—became ∼14% smaller relative to those of pre-megafaunal extinction coyotes118. The observed direction of phenotypic change is counter to expectations based on contemporary climate variation, suggesting to the authors that the coyote body size change may reflect adaptation to a smaller prey base, smaller competitors, or both118. Furthermore, compared to earlier coyotes, the mandibles of coyotes in post-megafaunal extinction North America were relatively less robust, and their dental morphology tended more toward tooth cusp morphology commonly associated with grinding and omnivory rather than shearing and carnivory119. These changes are consistent with morphological evolution in response to megafaunal extinction-associated dietary shifts.
The vegetation consumption and trampling activities of herbivorous megafauna helped to maintain rich, mosaic-like landscapes; following megafaunal extinction these structured habitats were then replaced by denser and less-diverse landscapes111,120,121. Similar to the above-discussed evolutionary ecology implications of direct human landscape modification, such changes probably affected the abundance, behaviour, and perhaps evolutionary biology of surviving species.
Megafauna also had co-evolutionary relationships with many large-seeded plants, acting as primary dispersers122. Larger seeds have higher seedling survival rates123, which probably helps to explain the co-evolutionary optimization of seed size with dispersers. Following extinction of their primary dispersers, some megafauna-adapted plants still benefitted from secondary dispersers124,
In fact, a major component of this process in action has been documented in Brazil, where larger, seed-dispersing birds (for example, toucans, Ramphastos dicolorus) have been locally extirpated over the past two centuries from many but not all forest fragments. Defaunation status accounted for ∼34% of variance in seed size; the proportion of Euterpe edulis palm seeds with diameter > 12 mm (the size limit for smaller birds) was ∼32% in forests with larger birds but near 0% in defaunated sites129, providing a mechanism for directional selection on seed size.
Connecting past, present, and future
To explicitly compare the modern biology and archaeological observations of non-human morphological evolution in response to human behaviour we computed the ‘darwin’ statistic as an estimate of the rate of evolutionary change130 for each suitable example depicted in Figures 2 and 4. darwins reflect the magnitude of morphological change per million years. Overall, the observed rates of morphological evolution in natural (non-domestic taxa) modern systems are substantially higher than those from the archaeological record. Strikingly, present-day human behaviour can apparently affect morphological evolution in non-human, non-domesticated species at rates similar to or greater than those associated with longer-term domestication processes.
The substantially higher evolutionary rates observed in the modern versus archaeological (non-domestication) records may reflect recently increased intensities of the human behaviours that are driving morphological evolution in non-human taxa, perhaps as by-products of rapid human global population growth and commensurate industrialization. We do not believe that smaller-magnitude morphological changes are not also occurring in the present day—on the contrary, there are likely many, many ongoing processes that are more subtle and simply below the detection resolution of modern evolutionary biology studies. In contrast, from the archaeological record we would expect considerable power to detect non-human evolutionary responses to human behaviour for slower but longer-lasting phenotypic responses to relatively less-intensive human activities compared with the examples from modern biology81. Although it is possible that faster rates of evolutionary change may be less visible in lower resolution (incomplete) archaeological records, a temporal increase in the intensity of relevant human impacts on the environment is the most likely explanation for the observed difference in the highest observed rates between the two datasets.
Looking forward, the intensity of anthropogenic impacts on the evolutionary biology of non-human species will likely continue to increase, including via additional types of human behaviours and impact. For example, we have entered a controversial era of human-directed genetic modification of non-human taxa that includes realized opportunities to effect morphological changes in existing species131,132 and the potential to resurrect versions of extinct taxa133. Rapid climate change is also an ongoing consequence of human behaviour that can affect non-human evolution134,135. To date, modern biology studies have been challenged to distinguish genetic-based morphological adaptations in response to human-induced climate change from environmental-based (plastic) responses136. Yet based on the global scale of climate change, we predict that an unprecedented number of non-human species (that are not otherwise driven to extinction) could eventually exhibit signs of morphological adaptation to this particular form of human-mediated ecological disturbance.
Humans are keystone species; our pervasive habitat-modifying and subsistence behaviours have widespread ecosystem effects, including on the evolutionary biology of non-human species. Not restricted to the present, the keystone status of humans and the substantial ecological impacts of human behaviour extends at least 50,000 years into prehistory, with associated long-term implications for the evolutionary biology of many non-human taxa. Archaeological and palaeontological studies that identify non-human morphological evolutionary responses to past human behaviour—the interpretation of which can be informed by studies of these evolutionary processes in modern systems and vice versa—are thus helping us to more comprehensively reconstruct the history and significance of anthropogenic impacts on worldwide ecosystems.
How to cite this article: Sullivan, A. P., Bird, D. W. & Perry, G. H. Human behaviour as a long-term ecological driver of non-human evolution. Nat. Ecol. Evol. 1, 0065 (2017).
We thank G. Pigeon, R. Klein and T. Steele for providing the data from their studies used in Figs 3 and 5; G. Pigeon and E. Loftus for providing images; and M. Aylward, C. Bergey, R. Bliege Bird, B. Codding, E. Davenport, R. Klein, and D. Schussheim for helpful comments on earlier drafts of the manuscript. This material is based on work supported by grants from the National Science Foundation (BCS-1554834 to G.H.P.; BCS-1459880 to D.W.B.) and by the National Science Foundation Graduate Research Fellowship Program (DGE1255832 to A.P.S.). Any opinions, findings, and conclusions or recommendations expressed in this material are those of the authors and do not necessarily reflect the views of the National Science Foundation.