Professional Documents
Culture Documents
Human Behaviour As A Long-Term Ecological Driver of Non-Human Evolution
Human Behaviour As A Long-Term Ecological Driver of Non-Human Evolution
T
he subsistence and habitat-modifying behaviours that typify archaeological record suggest that these processes likely extend con-
how humans interact with their surrounding environments siderably into prehistory 14, perhaps even to earlier than 50,000 yr bp
have broad and temporally profound effects on natural eco (years before present), and mediated by Neandertals15 in addition
systems1. At local levels, co-evolutionary relationships between to anatomically modern humans. Non-human evolutionary adap-
small-scale human societies and non-human taxa in the same com- tations to human-mediated ecological changes are also not limited
munity can be so longstanding and immersive that the removal of to direct harvesting pressure effects. For example, human land-
humans results in simplified food webs or ecosystem collapse2,3, scape modification practices, which predate the onset of agriculture
while in the context of global population growth the intensity of ~12,000 yr bp and have since then intensified1, may too be catalysts
human–environment interactions has led to numerous docu- for natural selection in affected non-human taxa6,16.
mented cases of wildlife population decline and extinction4,5. Given In this Review we describe potential mechanisms for non-human
the extent and rate that human-mediated climate change and other adaptive evolution in response to human behaviour, with human
by-products of increasing human population density are predicted behaviour defined broadly to encompass landscape modification
to impact future ecosystems6,7, these issues are of major concern to (including niche construction); the harvesting, artificial selec-
biologists, ecological anthropologists, and policymakers. tion, or translocation of non-human species; and even current and
In addition to ecosystem health and wildlife extinction risk, potential future directed practices involving genetic modification
human behaviour also influences the evolutionary biology of non- and de-extinction. These processes are illustrated by examples from
human species. Extensive morphological evolution mediated by modern evolutionary biology and enhanced with ethnographic
human behaviour has been thoroughly documented for domesti- and prehistoric perspectives that include discussion of evidence
cated plants and animals8,9, but these processes are in fact consider- for the long-term history of anthropogenic effects on the environ-
ably more widespread, extending to non-domesticated species. That ment. We focus primarily on morphological evolution, which does
is, wild populations of non-human taxa can adapt via natural selec- not encompass all adaptations (for example, genetically mediated
tion on functional genetic variation in response to human-induced behavioural or metabolic adaptations) but provides opportunities
ecological changes10. In some cases, these morphological and behav- for insights from modern biology to be translated to the archaeo-
ioural adaptations may help certain species avoid extinction. These logical and palaeontological records. The consideration of these
processes could have cascading effects on ecosystem function6,11, so data in a long-term, integrated framework will ultimately help us
they are an important factor to consider in the broader context of identify and model these evolutionary processes as they may occur
human–environment interactions. in the future.
Clear examples of anthropogenic effects on the evolutionary
biology of non-human, non-domesticated species have been docu- The antiquity of humans as trophic regulators
mented in modern studies of body size or other morphological trait Disturbance ecology quantifies the effects of natural ecosystem dis-
change in response to selective human harvesting of natural popu- ruptions that may result from abiotic forces (for example, fire, flood
lations of vertebrates, invertebrates, and plants11–13. Analyses of the and drought) or biotic variation (for example, intensive predation,
1
Department of Biology, Pennsylvania State University, University Park, Pennsylvania 16802, USA.2Department of Anthropology, Pennsylvania State
University, University Park, Pennsylvania 16802, USA. 3Huck Institutes of the Life Sciences, Pennsylvania State University, University Park, Pennsylvania
16802, USA. *e-mail: ghp3@psu.edu
disease outbreak and depletion of keystone species)17. Humans have focus on the evolution of major morphological phenotypes—that is,
major roles as biotic influencers of ecological disturbance due to traits that may be preserved in prehistoric records (unlike, for exam-
our intensive subsistence strategies that include broad-spectrum (of ple, behavioural and metabolic changes)—as one subset of a diverse
a high diversity of resources) harvesting and predation18–20, wide- spectrum of genetic-based traits in non-human species that have
spread landscape burning practices21–24, and habitat modification evolved in response to human behaviour 41–44. For the same reason
for agriculture, aquaculture, and shelter 25. In fact, the term ‘hyper- we also do not discuss bacteria, parasites, and other microorganisms
keystone species’ has recently been coined to describe humans, as whose evolutionary biology is also affected by humans45–48. However,
our subsistence behaviours in a given habitat impact multiple other to the extent that human behaviour is a long-term driver of macro-
keystone species, with extensive resultant ecosystem effects26. scale morphological change in non-human plants and animals,
Although many studies of the relationships between human we would expect these processes to have probably affected the
activity and the population health and evolution of non-human spe- evolutionary histories of other phenotypes and organisms as well.
cies have focused on recent human impacts in ‘natural’ ecosystems, Artificial selection associated with the domestication of non-
archaeological and palaeoenvironmental records demonstrate a human animals and plants is the best-known example of mor-
long-term history of major anthropogenic effects that began well phological evolution in response to human behaviour. While
prior to the origins of agriculture1. agriculture and pastoralism originated by at least ~12,000 yr bp1,
With respect to intensive harvesting, while there is fossil-record the roots of these processes likely extend considerably further into
evidence for the consumption of meat by African hominins by at least prehistory, for example to ~23,000 yr bp49 or earlier 50 for plants and
2.5 million yr bp27,28, it is unclear whether this early record reflects >30,000 yr bp for dogs51. Here, because both the scale and archaeo-
hunting or scavenging and whether this behaviour was intense or logical prehistory of morphological evolution in domesticated taxa
sporadic29. However, good evidence for repeated, systematic butch- have been reviewed extensively 8,9,52–54, we will instead focus our
ery of terrestrial vertebrate species by hominins, consistent with tar- discussion on four other mechanisms by which human behaviours
geted hunting and processing, is observed by ~780,000 yr bp30, and could be ecological drivers of morphological evolution: (i) size- or
hafted spear technology appears in the archaeological record by at trait-selective harvesting; (ii) landscape modification and urbani-
least ~500,000 yr bp31. Archaeological data from South Africa32 and zation; (iii) human-mediated ecosystem taxonomic turnover; and
Spain33 also show that broad use of coastal resources by hominin (iv) looking towards the future, climate change and direct genetic
foragers began by at least 150,000 yr bp34. modification (Fig. 1).
Analyses of pollen and charcoal records suggest the initiation
of widespread anthropogenic landscape burning between ~50,000 Response to size- or trait-selective harvesting pressures
and 40,000 yr bp in Borneo and New Guinea, coincident with the Human patterns of prey selection can vary substantially from those
respective appearances of modern humans in each region35,36. Some of other predators. Although probably not applicable to all socie-
scholars even infer that the purposeful use of fire for habitat mod- ties, at least some human populations may harvest particular spe-
ification—a practice common to both agriculturalists and hunter- cies with unparalleled intensity, and may disproportionately target
gatherers worldwide today 23,24—may have roots in the evolution of large adults of a given prey species or individuals with a specific tro-
the genus Homo close to two million yr bp21. phy feature, such as antlers or horns18. Rapid phenotypic evolution
Human–environment interactions later intensified further with may result 11.
the origins and spread of agriculture and animal husbandry begin- Morphological evolution has been especially well documented in
ning ~12,000 yr bp1. Human-mediated ecological impacts during fisheries, in which larger prey are typically of higher value and netting
this period have included extensive land clearing, terracing, water- technology may greatly enrich the catch for larger individuals based
way diversion, the construction of permanent settlements and ulti- on the sizes of the net openings13,55–57. Body length and mass reduc-
mately urban centres, translocation of non-endemic species to new tions of 25% or greater have been documented over time periods of
habitats, the potential for magnified selective hunting and fishing only one or several decades in multiple independent taxa58–61.
pressures on non-domesticated species associated with increasing In addition to marine and freshwater vertebrates, organism size
human population sizes and densities, climate change, and now changes in response to human harvesting pressures have also been
even direct genetic modification of non-human taxa. reported in natural populations of terrestrial vertebrates and even
Through these various processes humans became deeply entan- plants (Fig. 2). An excellent example, documented in a carefully con-
gled in their ecosystems, as illustrated by the cascading effects that trolled study, is that of the wild snow lotus Saussurea laniceps. This
have been observed following the removal of human activity from plant, from the eastern Himalayas, is used in traditional Tibetan and
some habitats3. For example, in desert Australia, Aboriginal predation Chinese medicine for headache and high blood pressure treatments.
of monitor lizards involves disturbing climax vegetation via burning Larger plants are considered more potent and thus collected prefer-
tracts of old-growth xeric grassland, creating a tight mosaic of veg- entially. Based on herbarium specimens (of flowering plants; when
etative succession that supports higher densities of many tradition- maximum plant height is achieved), Law and Salick62 reported a
ally hunted species37,38. A mid-twentieth-century hiatus in traditional ~45% decrease in S. laniceps heights over the past century. In contrast,
burning led to trophic collapse and coincided with the extinction of flowering heights of the sympatric and related—but substantially
at least 21 endemic marsupials2. Thus, while human predation and less desired medicinally and little harvested—snow lotus Saussurea
habitat modification can certainly affect non-human population medusa were unchanged over the same period. Moreover, modern
distributions negatively, leading to extinction and food web simpli- S. laniceps individuals subject to harvesting in non-protected areas
fication5,39,40, there is also strong evidence for significant ecological were ~70% shorter than those in protected areas62.
adaptation to the long-term keystone presence of intensive human Trophy hunters exert directional pressure on a particular
disturbance, which could include adaptive morphological evolution pheno type, potentially leading to the evolution of smaller fea-
in non-human taxa. tures across the population or higher rates of feature absence. For
example, illegal ivory hunting in Zambia resulted in an increase of
Human behaviour as a driver of non-human evolution >300% (from 10.5% to 38.2%) of tuskless female African elephants
One of our goals in this Review is to demonstrate the antiquity of (Loxodonta africana) from 1969 to 198963. Trophy hunting may also
widespread phenotypic adaptation to human behaviour by connect- have pleiotropic effects on other phenotypes that are not the direct
ing observations and insights from modern evolutionary biology targets of selection, as observed for bighorn sheep (Ovis canadensis)
to archaeological and palaeontological observations. To do so, we on Ram Mountain in Alberta, Canada. Specifically, in addition to a
Trophy feature
reduction
Harvesting
pressure
Body size
reduction
Wing shape
changes
Landscape
modification
Altered limb
proportions
Increased
seed size
Domestication
Reduced horn
and body size
Accelerated
1 year 1 year growth rate
Genetic
modification
De-extinction?
Figure 1 | Model of how human behaviour may effect non-human morphological evolution.
~30% reduction in bighown male horn length over 23 years of horn- Modern human intertidal foragers worldwide are selective and
size-focused hunting pressure (Fig. 3), body mass also decreased highly sensitive to changes in the trade-offs associated with search-
~23.5% over the same period64,65. ing for and handling different types and sizes of shellfish73–75. Because
Is trophy hunting limited or widespread among human societies, shellfish are sessile or slow on-encounter, shell size predicts the
and what is the antiquity of this behaviour? In many societies, pres- macro-nutritional return rate of molluscs, and correlates well with
tige goods do serve as costly signals of status and commitment and ethnographic measurements of the probability that foragers will
probably play an important role in the development of inequality 66,67, select an individual encountered while foraging 76. Thus, observed
and in foraging groups hunting is commonly as much a political foraging behaviour suggests selective preferences for larger mol-
activity as it is a provisioning practice68,69. Discerning archaeological luscs, generating conditions that could lead to evolutionary size
signatures of prestige hunting is difficult but has been proposed in decreases over time for species harvested intensely by humans.
several cases70. These expectations are consistently met by analyses of the archae-
Well-documented archaeological evidence for prey size changes ological record. Early evidence for routine, large-scale intertidal
in response to human harvesting pressure does exist in the record mollusc exploitation has been recorded beginning in the middle
of intertidal mollusc exploitation14 (Fig. 4). While many predators stone age (~120,000 to 60,0000 yr bp) at coastal cave sites in South
(marine, terrestrial, and avian) exploit shellfish, only humans trans- Africa71,77. These South African middle stone age shellfish assem-
port large loads to central locales for repeated processing and depo- blages have been compared to those that accumulated in midden
sition, creating substantial heaps or trash ‘middens’ of the inedible, layers dated to the later stone age (~12,000 to <1,000 yr bp), with sig-
durable mollusc shells and the remains of other taxa that come nificant observed reductions over time in the sizes of all represented
to represent temporal records of both direct human predation on species78,79, including for the Cape turban shell (Turbo sarmaticus)
these animals and their sizes71,72. Such harvesting can have substan- as depicted in Fig. 5.
tial effects, with selective pressures resulting in either morphologi- Similar evolutionary trends are evident in many coastal environ
cal change in targeted molluscs or predation resulting in shellfish ments elsewhere. For example, molluscs recovered from mid-
populations with persistently skewed age–size profiles. den layers dated to ~19,000 to 9,000 yr bp at Riparo Mochi, Italy
Figure 2 | Selected modern examples of morphological change in response to human behaviour. The presented magnitudes of phenotypic change
represent approximate percentages of difference from earliest measured value. The reported evolutionary rates (in darwins)130 reflect the magnitude of
morphological change (absolute value of the difference between the natural log of the starting trait value and the natural log of the ending trait value) per
million years, for cases with available morphological trait measurements (which excludes the elephant presence/absence example) and information on the
number of years over which the change occurred. For cases in which multiple phenotypes were analyzed, the magnitude of change depicts the first listed
phenotype. Only one of multiple modern examples of these effects in different species of salmon59,61,137 and shellfish138 were included in the figure. Photo
credits (from top to bottom): Geoff Gallice; Gabriel Pigeon; US Fish and Wildlife Service, Jon Nickles; US Fish and Wildlife Serve Division of Public Affairs,
Dan J. Pittillo; ref. 62, PNAS; US Fish and Wildlife Service, Phil Coleman; Bering Land Bridge National Preserve; Gnangarra; SEFSC Pascagoula Laboratory,
Collection of Brandi Noble, NOAA/NMFS/SEFSC/NOAA Photo Library; US NOAA; Tvabutzku1234; Jerry Kirkhart; Scott Loarie; Ken Thomas; Ken Thomas;
Gavin Schaefer; Geoff Gallice; G.-U. Tolkiehn; Ken Thomas; AnRo0002; gailhampshire; Alex Popovkin, Bahia, Brazil.
80 n = 34
n = 23
n=7
n=7 n=6
70
n=7
n = 12
60
50
Trophy hunting targeting
large male horn size
40
‘75– ‘79– ‘83– ‘87– ‘91– ‘95– ‘99– ‘03– ‘07– ‘11–
’78 ’82 ’86 ’90 ’94 ’98 ’02 ’06 ’10 ’12
Year (1975–2012)
Figure 3 | Evolutionary changes in bighorn sheep (Ovis canadensis) horn size in response to variable human trophy hunting pressures. a, Box and
kernel density plots of the horn lengths of male 4-year-old sheep from 1975–2012 at Ram Mountain, Alberta, Canada, where 2.26 male rams were sport-
harvested per year from 1973–1995 followed by a rate of only 0.27 rams per year from 1996–2011. Median values are represented by white circles; box
bottoms and tops indicate the lower and upper quartiles, respectively, whiskers approximate 95% confidence intervals of the medians, and kernel density
is shown in blue. Data are reproduced with permission from ref. 65, Wiley; see the original article for more comprehensive analyses, including with data
from sheep of different age classes and estimates of changes in horn length genetic value. b, Ram Mountain bighorns. Photo courtesy of Gabriel Pigeon.
were ~30% smaller than those in earlier layers dated to ~36,000 to been well documented for many non-human species, including as
24,000 yr bp15. Likewise, in the California Channel Islands, owl lim- changes to behaviour, distribution, and abundance2,37,38,82–85.
pets (Lottia gigantea) have experienced ~40% size reductions since Anthropogenic landscape modifications also create contexts of
the onset of human foraging ~10,000 yr bp80. novel selective pressure, providing a substrate for non-human mor-
In each of the above cases, the observed patterns of morpho- phological evolution16,86. An exceptional example comes from a long-
logical change are inconsistent with known patterns of palaeo term study of cliff swallows (Petrochelidon pyrrhonota) in Nebraska,
environmental variation, suggesting that the morphological change where roadside nesting behaviour increased in frequency starting in
probably reflects responses to anthropogenic activity. However, the 1980s with expanding bridge and culvert construction87. Overall
because growth is continuous throughout life for many of these population wing length decreased ~2% across a 30-year period,
taxa, confidently distinguishing adaptive morphological evolu- probably because shorter wings aided flight agility and reduced vehi-
tion from by-products of changes in harvest age profiles based on cle collision risk. Birds killed by cars were consistently longer-winged
archaeological midden data is also a challenge79. Still, such analy- than the overall population, and the frequency of road-killed birds
ses are possible for species with clear markers of growth cessation steadily declined despite the increasing proportion of sport-utility
(for example, for skeletal long bones with their epiphyseal fusion) vehicles over the study period87.
or at least of maturation. For example, sizes of both the largest juve- Human-mediated habitat fragmentation may also exert evolu-
niles and the smallest adult conch (Strombus pugilis) shells became tionary pressures on morphological traits—perhaps especially those
dramatically smaller over the past 7,000 years of human harvesting related to dispersal88. For example, increased wing pointedness in
pressure in Carribbean Panama—conch harvested for food today birds facilitates longer-distance mobility in the face of increased
have ~40% less meat than the conch that existed when human pre- habitat isolation89. By studying museum specimens, Desrochers90
dation began81—demonstrating a morphological evolutionary effect found that songbird relative wing pointedness increased by a mag-
rather than only a mortality profile change. nitude of 7–28% over the last century in North American forests
Without the analytical benefits afforded by extensive shell mid- heavily affected by clear-cutting. The same pattern was not observed
den deposits, archaeologists are more challenged to identify evi- for birds in early-successional and intact forests90.
dence of morphological evolution in response to human harvesting Urban environments in particular are hotbeds for behavioural
pressures on terrestrial vertebrates. Still, Stiner 15 reported a 19% specializations and population density increases among non-human
size decrease (based on humeral shaft diameter as a proxy for body taxa, even while other species have been extirpated91. These environ
size) for spur-thighed tortoises (Testudo graeca) that cannot be ments are also home to some of the strongest current examples of
explained by patterns of climatic variation across archaeological non-human adaptation in response to human landscape modifica-
layers dated to 200,000–150,000 yr bp, 100,000–70,000 yr bp, and tion92. In a recent study, Winchell et al.93 examined 319 adult male
28,000–11,000 yr bp at Nahal Meged, a Neandertal and modern Puerto Rican crested anole lizards (Anolis cristatellus) from three
human site in Israel. independent pairs of urban and nearby forested habitat populations.
Average limb lengths for urban lizards were consistently and sig-
Response to landscape modification and urbanization nificantly longer (~2% longer), and urban individuals also had sig-
Anthropogenic burning and clearing, irrigation infrastructure, nificantly greater numbers of subdigital scales (~3% greater), with
horticultural and pastoral disturbance, and structural investment both traits potentially facilitating more efficient locomotion on arti-
have widespread ecological consequences for other organisms shar- ficial surfaces. Results from a subsequent common garden experi-
ing the affected habitats. Direct and indirect effects or responses to ment suggest a genetic, and thus potentially evolutionary, basis for
human landscape modification—both positive and negative—have the phenotypic differences93. House finches in urban Arizona have
Spur–thighed tortoise
~200 kyr–11 kyr bp Humeral shaft diameter –19% 1 15
Testudo graeca
Mediterranean limpet
~36 kyr–9 kyr bp Shell diameter –33% 15 15
Patella caerulea
Conch
7 kyr bp–Present Shell height –40% 72 81
Strombus pugilis
Argenville’s limpet
~120 kyr–1 kyr bp Shell length –21% 2
5 79
Scutellastra longicosta
Barley +256%
~4 kyr bp–Present Grain weight 238 148
Hordeum vulgare
Maize +4,000%
~9 kyr bp–Present Number of kernels per ear 429 149
Zea mays
Pumpkin +6,500%
~10 kyr bp–Present Fruit diameter 420 150
Cucurbita maxima
Extinctions Domestication
≥50 25 0 25 ≥50 1 10 100 1,000 10,000+
Figure 4 | Selected archaeological examples of morphological change in response to human behaviour. The presented magnitudes of phenotypic change
represent approximate percentages of difference from earliest measured value. The reported evolutionary rates (in darwins)130 reflect the magnitude of
morphological change (absolute value of the difference between the natural log of the starting trait value and the natural log of the ending trait value) per
million years. For cases in which multiple phenotypes were analysed, the magnitude of change depicts the first listed phenotype. Photo credits (from top
to bottom): Gisella G.; Esculapio; Daniel Cavallari; Jerry Kirkhart; Jan Delsing; Jan Delsing; MerlinCharon; Justin Johnsen; Leo Michels; US Department of
Agriculture; Arnaud 25.
~2% larger beaks relative to populations in nearby desert habitat, a crisis5,39,99,103,104. In this section we consider the potential effects of
putative genetic adaptation to facilitate greater bite forces for con- human-mediated turnover in ecosystem species composition on
sumption of seeds in backyard feeders that are relatively larger and morphological evolution in non-human plants and animals: either
harder-shelled than those typically processed by the non-urban directly for translocated taxa themselves in response to their new
finches94. Adaptive beak shape changes associated with feeder- ecosystems, or indirectly for endemic species affected by invasives
based diets are also hypothesized for Central European blackcaps or by the extinction of other species in their ecosystem.
(Sylvia atricapilla)95. However, because bird song acoustics may also
adapt to urban noise environments96 and beak morphology also Translocation. Studies of potential morphological adaptation in
modulates song frequency 97, the ultimate evolutionary pressure(s) translocated species are challenged by the high levels of genetic drift
affecting urban bird beak shape need to be confirmed. potentially associated with founder events that may also result in
(non-adaptive) phenotypic change over time105,106, but evolutionary
Response to human-mediated taxonomic turnover genomic or fitness analyses may be used to evaluate hypotheses of
Human population movement, landscape modification, and har- neutral versus adaptive evolution in such cases. For example, purple
vesting practices are (and have been) associated with widespread loosestrife (Lythrum salicaria), a European wetland plant invasive
translocation and extinction of thousands of non-human species. to North America, flowers ~20 days earlier and is ~50% smaller
For example, an incredible number of plant species—more than near the northern versus southern extent of its East coast invasive
13,000, or ~4% of the known extant flora—have been translo- range107. Based on common garden experiments, these traits are
cated by humans98. While expanding global travel and trade have largely genetically controlled, and strong fitness advantages for ear-
led to an increased rate of human-mediated translocation in the lier reproduction in the north and larger plant size in the south were
past 150 years99, the purposeful or accidental introduction of non- identified in transplantation experiments107.
domesticated plants and animals to naive ecosystems has a long Introduced species can also affect the behavioural ecology
prehistory, likely back to the earliest human colonizations of at least and evolutionary biology of endemic taxa. There is mounting
some global landmasses100–102. Invasive species affect endemic taxa evidence that prehistoric translocation events induced trophic
by competitive exclusion, predation, and other processes, com- cascades with continental-scale impacts on fauna and flora108.
pounding the effects of human landscape modification and harvest- Contemporary examples of the role of invasives in shaping the
ing pressure to contribute to the ongoing anthropogenic extinction morphology of endemic taxa include South American venomous
50 n = 351 n = 1,130
n = 55 n = 893
Max operculum length (mm)
n = 396
40
30
20
10
Figure 5 | Middle stone age (MSA) to later stone age (LSA) size reductions of Cape turban shell (Turbo sarmaticus) opercula recovered from South
African archaeological refuse dumps (shell middens). a, Box and kernel density plots of opercula maximum lengths from midden layers dating to the
MSA (120–60 kyr bp) and LSA (11–0.5 kyr bp). Median values are represented by white circles; box bottoms and tops indicate the lower and upper
quartiles, respectively, whiskers approximate 95% confidence intervals of the medians, and kernel density is shown in blue. Dotted lines identify the
dated layer boundaries for each sample. Data are reproduced with permission from ref. 79, PNAS. b, Modern Turbo sarmaticus, with operculum external
surface indicated. Photo credit: MerlinCharon. c, Archaeological Turbo sarmaticus opercula that have been micromill sampled for seasonal palaeoclimate
reconstruction, from a Nelson Bay Cave (South Africa) midden layer dated to 9–7 kyr bp. Photo courtesy of Emma Loftus.
fire ants (Solenopsis invicta), introduced by the 1940s into the past century for vertebrates than the long-term background rate5.
southern United States, where they now prey on endemic fence This crisis may be connected to longer-term extinction processes
lizards (Sceloporus undulatus). Adult fence lizards at sites with the related to longstanding human landscape modification practices
longest times since ant invasion have 3.4% longer hindlimbs and and harvesting pressures.
have higher rates of body twitching and fleeing responses (which Like humans, megafauna are generally keystone trophic facili-
are more effective in longer-limbed individuals) to fire ant attack tators for other species in the same habitats111. Of ~150 terrestrial
compared with lizards at not-yet-invaded sites109. This pattern mammalian genera >44 kg living on Earth’s continents 50,000 yr bp,
is inconsistent with expected ecogeographic variation based on approximately 65% were extinct by 10,000 yr bp112. Recent analyses
morphology of museum specimens collected prior to fire ant inva- of new, high-resolution palaeoenvironmental and chronological
sion, altogether suggesting a recent history of phenotypic adapta- datasets have given some (but not all) scholars increasing con-
tion109. Meanwhile, the cane toad was introduced into Australia fidence that colonizing humans played at least indirect roles (for
in 1935. This invasive species is highly toxic upon consumption example, via landscape burning) in prehistoric megafaunal extinc-
to some but not all endemic snakes, although smaller toads are tions in multiple large island habitats1,113,114. At continental scales
less likely to deliver sufficient volumes of toxin to be fatal. Based the role of humans in this process remains less clear 115,116, with natu-
on morphological analyses of museum specimens, two snake spe- ral climate change likely a major factor in at least certain cases117.
cies (Pseudechis porphyriacus and Dendrelaphis punctulatus) have Our purpose is not to draw conclusions from these ongoing inves-
evolved smaller heads over the past 70 years that putatively preclude tigations or to review the associated debate. Rather, we suggest that
the consumption of larger toads110. if our ancestors did play roles in the extinctions of megafauna, then
those events would represent another mechanism through which
Extinction. Human-mediated extinction may have cascading human behaviour may have indirectly affected the evolutionary
effects on the evolutionary biology of surviving non-human species biology of other (non-megafaunal) non-human taxa. Regardless,
in the same ecosystem, especially when the extinctions involve key- by considering the outcomes of these past events we can raise our
stone species. We are currently experiencing a mass extinction cri- awareness of ongoing and future evolutionary responses to modern
sis, with (for example) 8–100 times higher extinction rates over the extinction processes.
Ecosystem effects from the removal of keystone megafauna likely are driving morphological evolution in non-human taxa, perhaps
included niche space changes and openings for surviving taxa, as by-products of rapid human global population growth and
resulting in opportunities for morphological evolution that might commensurate industrialization. We do not believe that smaller-
be detected in the palaeontological record. For example, in the magnitude morphological changes are not also occurring in the
~1,000 years following megafaunal extinctions in North America, present day—on the contrary, there are likely many, many ongoing
coyote (Canis latrans) femur circumferences—a proxy for body processes that are more subtle and simply below the detection reso-
size—became ~14% smaller relative to those of pre-megafaunal lution of modern evolutionary biology studies. In contrast, from the
extinction coyotes118. The observed direction of phenotypic change archaeological record we would expect considerable power to detect
is counter to expectations based on contemporary climate varia- non-human evolutionary responses to human behaviour for slower
tion, suggesting to the authors that the coyote body size change may but longer-lasting phenotypic responses to relatively less-intensive
reflect adaptation to a smaller prey base, smaller competitors, or human activities compared with the examples from modern biol-
both118. Furthermore, compared to earlier coyotes, the mandibles ogy 81. Although it is possible that faster rates of evolutionary change
of coyotes in post-megafaunal extinction North America were rela- may be less visible in lower resolution (incomplete) archaeologi-
tively less robust, and their dental morphology tended more toward cal records, a temporal increase in the intensity of relevant human
tooth cusp morphology commonly associated with grinding and impacts on the environment is the most likely explanation for
omnivory rather than shearing and carnivory 119. These changes are the observed difference in the highest observed rates between the
consistent with morphological evolution in response to megafaunal two datasets.
extinction-associated dietary shifts. Looking forward, the intensity of anthropogenic impacts on the
The vegetation consumption and trampling activities of herbivo- evolutionary biology of non-human species will likely continue to
rous megafauna helped to maintain rich, mosaic-like landscapes; increase, including via additional types of human behaviours and
following megafaunal extinction these structured habitats were impact. For example, we have entered a controversial era of human-
then replaced by denser and less-diverse landscapes111,120,121. Similar directed genetic modification of non-human taxa that includes
to the above-discussed evolutionary ecology implications of direct realized opportunities to effect morphological changes in existing
human landscape modification, such changes probably affected species131,132 and the potential to resurrect versions of extinct taxa133.
the abundance, behaviour, and perhaps evolutionary biology of Rapid climate change is also an ongoing consequence of human
surviving species. behaviour that can affect non-human evolution134,135. To date, mod-
Megafauna also had co-evolutionary relationships with many ern biology studies have been challenged to distinguish genetic-
large-seeded plants, acting as primary dispersers122. Larger seeds based morphological adaptations in response to human-induced
have higher seedling survival rates123, which probably helps to climate change from environmental-based (plastic) responses136.
explain the co-evolutionary optimization of seed size with dis- Yet based on the global scale of climate change, we predict that an
persers. Following extinction of their primary dispersers, some unprecedented number of non-human species (that are not other-
megafauna-adapted plants still benefitted from secondary dis- wise driven to extinction) could eventually exhibit signs of mor-
persers124–126, while others without extant dispersers127 have expe- phological adaptation to this particular form of human-mediated
rienced significant range contractions and even extinction in the ecological disturbance.
wild128. We predict greater magnitude evolution in fruit and seed
phenotypes facilitating secondary dispersal for the megafaunal- Conclusion
adapted plants that have best survived this transition. Once in Humans are keystone species; our pervasive habitat-modifying and
place, such morphological changes could have aided fruit acces- subsistence behaviours have widespread ecosystem effects, including
sibility for additional dispersers, likewise potentially affecting their on the evolutionary biology of non-human species. Not restricted
evolutionary ecologies. to the present, the keystone status of humans and the substantial
In fact, a major component of this process in action has been ecological impacts of human behaviour extends at least 50,000 years
documented in Brazil, where larger, seed-dispersing birds (for exam- into prehistory, with associated long-term implications for the evo-
ple, toucans, Ramphastos dicolorus) have been locally extirpated lutionary biology of many non-human taxa. Archaeological and
over the past two centuries from many but not all forest fragments. palaeontological studies that identify non-human morphological
Defaunation status accounted for ~34% of variance in seed size; the evolutionary responses to past human behaviour—the interpreta-
proportion of Euterpe edulis palm seeds with diameter > 12 mm (the tion of which can be informed by studies of these evolutionary pro-
size limit for smaller birds) was ~32% in forests with larger birds but cesses in modern systems and vice versa—are thus helping us to
near 0% in defaunated sites129, providing a mechanism for directional more comprehensively reconstruct the history and significance of
selection on seed size. anthropogenic impacts on worldwide ecosystems.
Connecting past, present, and future Received 12 September 2016; accepted 20 December 2016;
To explicitly compare the modern biology and archaeological published 21 February 2017
observations of non-human morphological evolution in response
to human behaviour we computed the ‘darwin’ statistic as an esti- References
mate of the rate of evolutionary change130 for each suitable exam- 1. Boivin, N. L. et al. Ecological consequences of human niche construction:
ple depicted in Figures 2 and 4. darwins reflect the magnitude of examining long-term anthropogenic shaping of global species distributions.
Proc. Natl Acad. Sci. USA 113, 6388–6396 (2016).
morphological change per million years. Overall, the observed rates 2. Bliege Bird, R. Disturbance, complexity, scale: New approaches to
of morphological evolution in natural (non-domestic taxa) modern the study of human–environment interactions. Annu. Rev. Anthropol.
systems are substantially higher than those from the archaeologi- 44, 241–257 (2015).
cal record. Strikingly, present-day human behaviour can apparently 3. Castilla, J. C. Coastal marine communities: trends and perspectives from
affect morphological evolution in non-human, non-domesticated human-exclusion experiments. Trends Ecol. Evol. 14, 280–283 (1999).
species at rates similar to or greater than those associated with 4. Jackson, J. B. C. Historical overfishing and the recent collapse of coastal
ecosystems. Science 293, 629–637 (2001).
longer-term domestication processes. 5. Ceballos, G. et al. Accelerated modern human-induced species losses:
The substantially higher evolutionary rates observed in the entering the sixth mass extinction. Sci. Adv. 1, e1400253 (2015).
modern versus archaeological (non-domestication) records may 6. Western, D. Human-modified ecosystems and future evolution.
reflect recently increased intensities of the human behaviours that Proc. Natl Acad. Sci. USA 98, 5458–5465 (2001).
71. Marean, C. W. The origins and significance of coastal resource use in Africa 100. Wilmshurst, J. M., Anderson, A. J., Higham, T. F. G. & Worthy, T. H. Dating
and Western Eurasia. J. Hum. Evol. 77, 17–40 (2014). the late prehistoric dispersal of Polynesians to New Zealand using the
72. Simenstad, C. A., Estes, J. A. & Kenyon, K. W. Aleuts, sea otters, and commensal Pacific rat. Proc. Natl Acad. Sci. USA 105, 7676–7680 (2008).
alternate stable-state communities. Science 200, 403–411 (1978). 101. Jones, E. P., Eager, H. M., Gabriel, S. I., Jóhannesdóttir, F. & Searle, J. B.
73. Bird, D. W. & Bliege Bird, R. L. Contemporary shellfish gathering Genetic tracking of mice and other bioproxies to infer human history.
strategies among the Meriam of the Torres Strait Islands, Australia: Trends Genet. 29, 298–308 (2013).
Testing predictions of a central place foraging model. J. Archaeol. Sci. 102. Austin, C. C. Lizards took express train to Polynesia. Nature
24, 39–63 (1997). 397, 113–114 (1999).
74. De Vynck, J. C. et al. Return rates from intertidal foraging from Blombos 103. Mooney, H. A. & Cleland, E. E. The evolutionary impact of invasive species.
Cave to Pinnacle Point: Understanding early human economies. Proc. Natl Acad. Sci. USA 98, 5446–5451 (2001).
J. Hum. Evol. 92, 101–115 (2016). 104. Clavero, M. & García-Berthou, E. Invasive species are a leading cause of
75. Meehan, B. Shell Bed to Shell Midden (Australian Institute of Aboriginal animal extinctions. Trends Ecol. Evol. 20, 110 (2005).
Studies, 1982). 105. Colautti, R. I. & Lau, J. A. Contemporary evolution during invasion:
76. Codding, B. F., O’Connell, J. F. & Bird, D. W. Shellfishing and the evidence for differentiation, natural selection, and local adaptation. Mol. Ecol.
colonization of Sahul: A multivariate model evaluating the dynamic effects 24, 1999–2017 (2015).
of prey utility, transport considerations and life-history on foraging patterns 106. Le Gros, A. et al. Invasion history and demographic processes associated with
and midden composition. J. Isl. Coast. Archaeol. 9, 238–252 (2014). rapid morphological changes in the Red-whiskered bulbul established on
77. Jerardino, A. On the origins and significance of Pleistocene coastal resource tropical islands. Mol. Ecol. 25, 5359–5376 (2016).
use in southern Africa with particular reference to shellfish gathering. 107. Colautti, R. I. & Barrett, S. C. H. Rapid adaptation to climate facilitates range
J. Anthropol. Archaeol. 41, 213–230 (2016). expansion of an invasive plant. Science 342, 364–366 (2013).
78. Avery, G. et al. The Ysterfontein 1 Middle Stone Age rock shelter and the 108. Fillios, M., Crowther, M. S. & Letnic, M. The impact of the dingo on the
evolution of coastal foraging. Goodwin Ser. 10, 66–89 (2008). thylacine in Holocene Australia. World Archaeol. 44, 118–134 (2012).
79. Klein, R. G. & Steele, T. E. Archaeological shellfish size and later human 109. Langkilde, T. Invasive fire ants alter behavior and morphology of native
evolution in Africa. Proc. Natl Acad. Sci. USA 110, 10910–10915 (2013). lizards. Ecology 90, 208–217 (2009).
80. Erlandson, J. M. et al. 10,000 years of human predation and size changes 110. Phillips, B. L. & Shine, R. Adapting to an invasive species: Toxic cane toads
in the owl limpet (Lottia gigantea) on San Miguel Island, California. induce morphological change in Australian snakes. Proc. Natl Acad. Sci. USA
J. Archaeol. Sci. 38, 1127–1134 (2011). 101, 17150–17155 (2004).
81. O’Dea, A., Shaffer, M. L., Doughty, D. R., Wake, T. A. & Rodriguez, F. A. 111. Malhi, Y. et al. Megafauna and ecosystem function from the Pleistocene to the
Evidence of size-selective evolution in the fighting conch from prehistoric Anthropocene. Proc. Natl Acad. Sci. USA 113, 838–846 (2016).
subsistence harvesting. Proc. R. Soc. B 281, 20140159 (2014). 112. Barnosky, A. D. Assessing the causes of Late Pleistocene extinctions on the
82. Smith, D. A. Garden game: shifting cultivation, indigenous hunting and continents. Science 306, 70–75 (2004).
wildlife ecology in Western Panama. Hum. Ecol. 33, 505–537 (2005). 113. Crowley, B. E. et al. Island-wide aridity did not trigger recent megafaunal
83. Herzog, N. M. et al. Fire and home range expansion: A behavioral response extinctions in Madagascar. Ecography http://dx.doi.org/10.1111/ecog.02376
to burning among savanna dwelling vervet monkeys (Chlorocebus aethiops). (2016).
Am. J. Phys. Anthropol. 154, 554–560 (2014). 114. Allentoft, M. E. et al. Extinct New Zealand megafauna were not in decline
84. Hulme-Beaman, A., Dobney, K., Cucchi, T. & Searle, J. B. An ecological and before human colonization. Proc. Natl Acad. Sci. USA 111, 4922–4927 (2014).
evolutionary framework for commensalism in anthropogenic environments. 115. Stuart, A. J. Late Quaternary megafaunal extinctions on the continents: a short
Trends Ecol. Evol. 31, 633–645 (2016). review. Geol. J. 50, 338–363 (2015).
85. Trant, A. J. et al. Intertidal resource use over millennia enhances forest 116. Wroe, S. et al. Climate change frames debate over the extinction of megafauna
productivity. Nat. Commun. 7, 12491 (2016). in Sahul (Pleistocene Australia-New Guinea). Proc. Natl Acad. Sci. USA
86. Turcotte, M., Araki, H., Karp, D., Poveda, K. & Whitehead, S. The eco- 110, 8777–8781 (2013).
evolutionary impacts of domestication and agricultural practices on wild 117. Cooper, A. et al. Abrupt warming events drove Late Pleistocene Holarctic
species. Philos. Trans. R. Soc. B 372, 20160033 (2016). megafaunal turnover. Science 349, 602–606 (2015).
87. Brown, C. R. & Bomberger Brown, M. Where has all the road kill gone? 118. Meachen, J. A. & Samuels, J. X. Evolution in coyotes (Canis latrans)
Curr. Biol. 23, R233–R234 (2013). in response to the megafaunal extinctions. Proc. Natl Acad. Sci. USA
88. Cheptou, P.‑O., Hargreaves, A. L., Bonte, D. & Jacquemyn, H. Adaptation 109, 4191–4196 (2012).
to fragmentation: evolutionary dynamics driven by human influences. 119. Meachen, J. A., Janowicz, A. C., Avery, J. E. & Sadleir, R. W. Ecological changes
Philos. Trans. R. Soc. B 372, 20160037 (2016). in coyotes (Canis latrans) in response to the Ice Age megafaunal extinctions.
89. Bowlin, M. S. & Wikelski, M. Pointed wings, low wingloading and calm air PLoS ONE 9, e116041 (2014).
reduce migratory flight costs in songbirds. PLoS ONE 3, e2154 (2008). 120. Johnson, C. N. Ecological consequences of Late Quaternary extinctions of
90. Desrochers, A. Morphological response of songbirds to 100 years of megafauna. Proc. R. Soc. B 276, 2509–2519 (2009).
landscape change in North America. Ecology 91, 1577–1582 (2010). 121. Bakker, E. S. et al. Combining paleo-data and modern exclosure experiments
91. McKinney, M. L. Effects of urbanization on species richness: A review of to assess the impact of megafauna extinctions on woody vegetation.
plants and animals. Urban Ecosyst. 11, 161–176 (2008). Proc. Natl Acad. Sci. USA 113, 847–855 (2016).
92. Alberti, M., Marzluff, J. & Hunt, V. Urban driven phenotypic changes: 122. Janzen, D. H. & Martin, P. S. Neotropical anachronisms: The fruits the
empirical observations and theoretical implications for eco-evolutionary Gomphotheres ate. Science 215, 19–27 (1982).
feedback. Philos. Trans. R. Soc. B 372, 20160029 (2016). 123. Moles, A. et al. A brief history of seed size. Science 307, 576–580 (2005).
93. Winchell, K. M., Reynolds, R. G., Prado-Irwin, S. R., Puente-Rolón, A. R. 124. Guimarães, P. R., Galetti, M. & Jordano, P. Seed dispersal anachronisms:
& Revell, L. J. Phenotypic shifts in urban areas in the tropical lizard Rethinking the fruits extinct megafauna ate. PLoS ONE 3, e1745 (2008).
Anolis cristatellus. Evolution 70, 1009–1022 (2016). 125. Pires, M. M. et al. Reconstructing past ecological networks:
94. Badyaev, A. V., Young, R. L., Oh, K. P. & Addison, C. Evolution on a local the reconfiguration of seed-dispersal interactions after megafaunal extinction.
scale: developmental, functional, and genetic bases of divergence in bill Oecologia 175, 1247–1256 (2014).
form and associated changes in song structure between adjacent habitats. 126. Jansen, P. A. et al. Thieving rodents as substitute dispersers of megafaunal
Evolution 62, 1951–1964 (2008). seeds. Proc. Natl Acad. Sci. USA 109, 12610–12615 (2012).
95. Rolshausen, G., Segelbacher, G., Hobson, K. A. & Schaefer, H. M. 127. Federman, S. et al. Implications of lemuriform extinctions for the Malagasy
Contemporary evolution of reproductive isolation and phenotypic flora. Proc. Natl Acad. Sci. USA 113, 5041–5046 (2016).
divergence in sympatry along a migratory divide. Curr. Biol. 128. Kistler, L. et al. Gourds and squashes (Cucurbita spp.) adapted to
19, 2097–2101 (2009). megafaunal extinction and ecological anachronism through domestication.
96. Slabbekoorn, H. Songs of the city: noise-dependent spectral plasticity in the Proc. Natl Acad. Sci. USA 112, 15107–15112 (2015).
acoustic phenotype of urban birds. Anim. Behav. 85, 1089–1099 (2013). 129. Galetti, M. et al. Functional extinction of birds drives rapid evolutionary
97. Giraudeau, M. et al. Song characteristics track bill morphology along changes in seed size. Science 340, 1086–1090 (2013).
a gradient of urbanization in house finches (Haemorhous mexicanus). 130. Haldane, J. B. S. Suggestions as to quantitative measurement of rates of
Front. Zool. 11, 83 (2014). evolution. Evolution 3, 51–56 (1949).
98. van Kleunen, M. et al. Global exchange and accumulation of non-native 131. Smith, M. D., Asche, F., Guttormsen, A. G. & Wiener, J. B. Genetically
plants. Nature 525, 100–103 (2015). modified salmon and full impact assessment. Science 330, 1052–1053 (2010).
99. Simberloff, D. et al. Impacts of biological invasions: what’s what and the way 132. Prado, J. R. et al. Genetically engineered crops: From idea to product.
forward. Trends Ecol. Evol. 28, 58–66 (2013). Annu. Rev. Plant Biol. 65, 769–790 (2014).