The evolution of city life

Abstract

Urbanization represents a dominant and growing form of disturbance to Earth's natural ecosystems, affecting biodiversity and ecosystem services on a global scale. While decades of research have illuminated the effects of urban environmental change on the structure and function of ecological communities in cities, only recently have researchers begun exploring the effects of urbanization on the evolution of urban populations. The 15 articles in this special feature represent the leading edge of urban evolutionary biology and address existing gaps in our knowledge. These gaps include: (i) the absence of theoretical models examining how multiple evolutionary mechanisms interact to affect evolution in urban environments; (ii) a lack of data on how urbanization affects natural selection and local adaptation; (iii) poor understanding of whether urban areas consistently affect non-adaptive and adaptive evolution in similar ways across multiple cities; (iv) insufficient data on the genetic and especially genomic signatures of urban evolutionary change; and (v) limited understanding of the evolutionary processes underlying the origin of new human commensals. Using theory, observations from natural populations, common gardens, genomic data and cutting-edge population genomic and landscape genetic tools, the papers in this special feature address these gaps and highlight the power of urban evolutionary biology as a globally replicated ‘experiment’ that provides a powerful approach for understanding how human altered environments affect evolution.

1. Introduction

Urbanization is an increasingly dominant form of disturbance to Earth's ecosystems, with more than half the world's human population now living in cities and the continued rapid expansion of urban environments globally [1]. Urbanization is associated with numerous changes to the biotic and abiotic environment, which alters the structure and function of ecosystems and degrades biodiversity [2–4]. For example, cities often have increased pollution [5,6], greater habitat fragmentation [7] and warmer temperatures [8,9] than surrounding non-urban habitats. These environmental changes are associated with an increase in the prominence of non-native species and the erosion of the structure, dynamics and diversity of native species within ecological communities [10–14]. While there have been decades of research examining how urban environmental change affects ecological interactions and communities in cities, only recently have researchers begun to explore how urbanization affects the evolution of species [15,16]. A greater understanding of how species evolve in urban environments will provide insight into both fundamental and applied biological problems and facilitate the design of more sustainable cities.

The study of evolution in response to urbanization has recently emerged as an important and fast-growing area of research in evolutionary biology. A recent review summarized 192 studies of urban evolution and identified several preliminary patterns and processes from existing data [16]. Past research on the evolutionary consequences of urbanization has focused primarily on the importance of urbanization for non-adaptive (i.e. gene flow and genetic drift) rather than adaptive (i.e. natural selection) evolution. Studies of non-adaptive evolution in urban environments find that urban fragmentation often results in reduced gene flow between urban populations and stronger genetic drift, which often leads to greater genetic differentiation between urban populations and decreased genetic diversity within populations [17,18]. However, many urban populations show evidence of phenotypic change in response to urbanization [19], and a small number of studies suggest these changes may often be adaptive [20–23]. Despite advances in understanding the effects of urbanization on the evolution on natural populations, numerous gaps remain.

This special feature addresses five major gaps in our understanding of the effects of urbanization on evolution. These gaps include: (i) the absence of theoretical models examining how multiple evolutionary mechanisms interact to affect evolution in urban environments; (ii) a lack of data on how urbanization affects natural selection and local adaptation; (iii) poor understanding of whether urban areas consistently affect non-adaptive and adaptive evolution in similar ways across multiple cities; (iv) insufficient data on the genetic and especially genomic signatures of urban evolutionary change; and (v) limited understanding of the evolutionary processes by which new human commensals arise in cities. This special feature is the first standalone compilation of original research on urban evolution. The special feature includes 15 theoretical and empirical research papers that include work from a diversity of study systems. Below we expand on how each of these papers address existing gaps in our understanding of urban evolution. We emphasize the diversity of techniques used in addressing these gaps and the novel predictions made by the authors that remain to be tested. We end by considering the conceptual strides collectively made by the 15 papers in this special feature and the future outlook for the field of urban evolutionary biology.

2. Theoretical models of urban evolution

We lack explicit theoretical models that examine how non-adaptive and adaptive evolutionary processes interact to influence urban evolution. In this special feature, Santangelo et al. [24] use spatially explicit simulations to examine how gene flow, genetic drift and natural selection concurrently affect the evolution of urban–rural phenotypic clines in the expression of a discrete Mendelian-inherited trait controlled by epistatically interacting genes. The authors found that stronger drift in urban populations led to consistent changes in the distribution of the trait, resulting in clines of decreased trait expression with increasing urbanization, even in the absence of natural selection. However, clines formed by drift were consistently weaker than those formed by selection, suggesting an upper limit to the strength of evolutionary clines that can be formed by genetic drift in urban populations. This work implies that directional changes in the distribution of traits with non-additive genetic inheritance (e.g. due to epistasis) may be particularly likely in urban environments where genetic drift is often stronger due to smaller and more isolated populations caused by habitat fragmentation. Where possible, studies of urban phenotypic change should account for the genetic architecture underlying phenotypic traits to generate system-specific null models for the expected effects of gene flow, genetic drift and natural selection on evolution.

3. Natural selection and local adaptation in urban environments

Although it is well known that urbanization causes large changes to multiple biotic and abiotic environmental factors, we lack explicit comparisons of the strength of selection acting on phenotypes in urban and non-urban populations [16]. To address this gap, four papers in this special issue compare selection between urban and non-urban populations. Irwin et al. [25] estimated selection on floral and plant defensive traits of the perennial vine, yellow jasmine (Gelsemium sempervirens, figure 1a) across eight paired urban–non-urban populations in North Carolina, USA. They found stronger selection for larger floral displays in urban than non-urban populations, likely caused by greater pollen limitation in urban populations. The authors argue for the need to measure selection across multiple years to better understand how selection shapes urban phenotypes. In a 7-year study, Caizergues et al. [26] examined differences in phenotypic divergence and selection on four morphological and two life-history traits of great tits (Parus major, figure 1b) between a paired urban and forested habitat near Montpellier, France. Urban birds showed reductions in size for all morphological traits, but these reductions were explained by only minor differences in selection on these traits between habitat types. While selection on life-history traits differed between urban and forest birds, these differences could not explain patterns of life-history divergence between urban and forest great tit populations. The authors suggest that morphological and life-history divergence between urban and forest birds is driven primarily by plastic responses to the altered environmental pressures imposed by urbanization, highlighting the importance of distinguishing between genetically and plastically controlled phenotypic changes in response to urban environments. These examples highlight that differential selection between urban and non-urban populations may be common.

Figure 1.

Articles in this special feature used a diverse set of study organisms to explore the effects of urbanization on evolutionary patterns and processes. (a) Yellow jasmine (Gelsemium sempervirens, source: Wikimedia Commons). (b) Great tit (Parus major, source: Wikimedia Commons). (c) Ragweed (Ambrosia artemisiifolia, photo credit: A. Butko). (d) Water flea (Daphnia magna, photo credit: J. Mergeay). (e) Wall lizard (Podarcis muralis, photo credit: J. Beninde). (f) Burrowing owl (Athene cunicularia, photo credit: J. Tella). (g) Black widow spider (Latrodectus hesperus, photo credit: E. Tassone). (h) Brown rat (Rattus norvegicus, photo credit: L. Elliot). (i) False brome (Brachypodium sylvaticum, photo credit: K. Morse). (j) Red-tailed bumblebee (Bombus lapidaries, source: Creative Commons). (k) House sparrow (Passer domesticus, photo credit: M. Ravinet).

Divergent selection on phenotypic traits across urban and non-urban populations may lead to local adaptation in response to urbanization. Reciprocal transplants and common garden experiments using urban and non-urban populations provide a strong test of whether populations are locally adapted to urban environments. This approach was taken by Gorton et al. [27] who examined patterns of selection and local adaptation in flowering phenology across 16 urban and non-urban populations of common ragweed (Ambrosia artemisiifolia, figure 1c), planted into two urban and two non-urban common gardens in Minnesota, USA. They found consistently stronger selection on phenology among foreign plant genotypes, consistent with local adaptation to urban habitats. In addition, divergence was greater among urban populations than rural populations for both flowering time and lifetime fitness, supporting the hypothesis that cities are characterized by high levels of environmental heterogeneity within cities.

To test for evolved differences in thermal tolerance between urban and non-urban environments, Diamond et al. [28] reared urban and non-urban acorn ants (Temnothorax curvispinosus) collected from three cities in the central USA in a common garden. They found evidence for evolved increases in heat tolerance and decreases in cold tolerance among urban populations in two of the three cities they examined. Ants also had the highest fitness when reared at temperatures most closely matching their original environment (i.e. urban versus non-urban). Overall, these results are consistent with local adaptation to the contrasting thermal regimes of urban and non-urban habitats, although sampling additional cities is required to better understand the generality of this pattern.

Brans et al. [29] similarly conducted a common garden experiment with two rearing temperatures to test for evolved differences in stress physiology and life-history traits. They used 62 clonal lines of the planktonic crustacean Daphnia magna (figure 1d), collected from 13 populations along an urban to non-urban gradient in Belgium. On average, urban populations evolved improved energy storage and more efficient resistance to the increased prevalence of physiological stress in urban environments. This evolution is consistent with adaptation to an increased pace of life strategy in warmer urban environments.

Winchell et al. [30] investigated evolved differences in locomotor performance among urban and non-urban populations of the Puerto Rican crested anole (Anolis cristatellus). Previous work in this system showed that urban A. cristatellus individuals have longer limbs and more toe lamellae [22], and they frequently use the smooth, artificial vertical structures common in urban habitats [31]. Here, the authors test the hypothesis that differences in morphology and habitat use are associated with increased locomotor performance, especially on flat artificial surfaces. They test this idea by measuring locomotor performance of urban and non-urban lizards on running tracks that varied in the smoothness of natural and artificial substrates, and in inclination angles. They found that morphological traits associated with faster sprint speeds were significantly altered among urban and non-urban lizards, and that these traits improve sprint speed on both natural and man-made substrates. Given the link between locomotor performance and fitness, the authors argue that selection has shaped the increased sprint performance of urban lizards as an adaptation to increased use of smooth, flat artificial surfaces in urban environments.

The examples in this section demonstrate that urban environments frequently cause changes in selection between urban and non-urban populations, which in some instances results in local adaptation of populations to the selective pressures imposed by urbanization.

4. The repeatability of evolution across cities

Whether populations and taxa subject to similar selection pressures evolve in the same way (i.e. parallel/convergent evolution) remains an open and important question in evolutionary biology [32–34]. Owing to the spatial replication of cities, urban environments have become models for assessing the degree to which evolutionary processes and responses are convergent across replicate populations subjected to similar selective regimes, thereby shedding light on fundamental questions about the repeatability and predictability of evolution [15]. Of the 14 empirical papers in this special feature, seven studied populations across multiple cities. For example, Johnson et al. [35] examined the contributions of genetic drift, gene flow and selection on the evolution of urban–rural clines in a Mendelian-inherited antiherbivore defence trait in white clover across 20 cities in Ontario, Canada. On average, urban populations evolved lower chemical defence, and microsatellite genotyping of eight cities showed no consistent effects of urbanization on neutral genetic diversity, suggesting these patterns are not due to genetic drift. They also detected substantial gene flow among urban and rural populations. Thus, the authors argue that natural selection has driven parallel evolution in defence across multiple cities, although the agents of selection causing these evolved differences require further investigation.

Beninde et al. [36] investigated the effects of urbanization on hybridization between native and non-native lineages of the wall lizard (Podarcis muralis, figure 1e), and identified the landscape elements that influence gene flow in cities. They used microsatellite markers to genotype 826 individuals from four German cities. Hybridization was common in cities where both native and non-native lineages co-occurred. Overall, water bodies were found to be the strongest barriers to gene flow, whereas railway tracks facilitated gene flow of hybrid lineages. Given the high prevalence of non-native lineage introductions into urban habitats, the authors suggest that humans may be driving the formation of novel hybrid zones, thereby shaping the evolutionary trajectories of urban lizards in new ways.

To examine the historical and contemporary demographic consequences of urbanization, Mueller et al. [37] performed whole-genome resequencing on 137 burrowing owls (Athene cunicularia, figure 1f) from urban and non-urban populations across three cities in Argentina. They found evidence that all three cities were independently colonized from nearby non-urban populations and that little urban–non-urban gene flow has occurred post-colonization. Demographic modelling also showed that urbanization was associated with a reduction in ancestral effective population sizes, leading to the loss of rare alleles in contemporary urban populations. They argue that these results may limit their ability to detect signatures of selection across the burrowing owl genome, which represents an important avenue of future research in this system.

5. Evolutionary genomic approaches to study urban evolution

A key problem in the study of urban evolution is to understand the genetic and genomic changes underlying urban evolution. In this special feature, six research groups use genomic tools to examine both non-adaptive and adaptive evolution in response to urbanization. Miles et al. [38] used a genome-wide single-nucleotide polymorphism (i.e. SNP) dataset and a genetic network analysis to examine patterns of genetic connectivity within and between 11 urban and 10 non-urban populations of black widow spiders (Latrodectus hesperus, figure 1g) in western USA. They found greater genetic diversity and lower genetic differentiation among urban populations than among non-urban populations. In addition, many non-urban populations were only connected to one another through intermediate urban populations. The authors argue that dispersal of black widow spiders occurs primarily through human-mediated transport, highlighting the role that humans play in spreading urban pest species that pose risks to human health.

To examine the repeatability of non-adaptive evolution across urban environments, Combs et al. [39] used ddRADseq [40] to sequence greater than 15 000 SNPs from 1311 brown rats (Rattus norvegicus, figure 1h) sampled from four cities in North and South America. While levels of genetic diversity were consistent across cities, they varied among sites within cities. Rats exhibited limited dispersal ability, such that rats that were geographically closer tended to be more genetically similar and the authors identified numerous physical landscape features (e.g. roads, waterways) and socioecological factors (e.g. human food subsidies) that were important correlates of gene flow within cities. Their results suggest that knowledge of how the landscape and socioeconomic features of cities influence the movement of rats can facilitate the development of management strategies for this urban pest and minimize the transmission of rat-borne pathogens.

Arredondo et al. [41] used genomic techniques to elucidate patterns of colonization, dispersal and gene flow among populations of an invasive grass in a metropolitan area. They generated genome-wide SNP data for 176 individuals across 22 populations of the grass false brome (Brachypodium sylvaticum, figure 1i) in Portland, USA. Their data indicate that rivers act as a major conduit of gene flow, in which human recreational activity along the river is the most likely explanation for patterns of dispersal, including the expansion of this invasive species at its range margin.

Theodorou et al. [42] investigated the non-adaptive and adaptive consequences of urbanization in an important pollinator species, the red-tailed bumblebee (Bombus lapidaries, figure 1j). They analysed 110 314 SNPs from 198 Bo. lapidaries individuals collected across nine paired urban–non-urban sites in nine German cities. Overall, genetic differentiation between urban and non-urban populations was low, suggesting high levels of gene flow across habitat types. The authors identified 24 candidate loci with divergent allele frequencies between urban and non-urban populations, suggesting a genomic signal of local adaptation in response to urbanization. These data suggest numerous hypotheses that remain to be tested regarding the functional significance of the loci under selection.

6. The origin of human commensals

While urbanization is often cited as reducing local biodiversity, some species have evolved to thrive in urban environments and even depend on resources provided by humans. In this special feature, Ravinet et al. [43] use whole-genome resequencing to uncover the evolutionary history of the house sparrow (Passer domesticus, figure 1k), including the evolutionary processes that have allowed this species to thrive in the presence of humans [43]. They sequenced 120 Pa. domesticus individuals and multiple populations from each of three non-commensal European sparrow species and subspecies. Demographic modelling identified a split between the commensal and non-commensal lineages approximately 11 kya, consistent with the spread of agriculture across Europe. The authors also identified evidence of selection in two genes in the house sparrow genomes involved in adaptation to high-starch diets, and the authors argue changes in these genes may have enabled house sparrows to thrive on a high-starch diet in urban environments.

7. Conclusion and future outlook

Using theory, observations from natural populations, common gardens, genomic data, and cutting-edge population genomic and landscape genetic tools, the papers in this special feature greatly advance our understanding of urban evolution. Importantly, the special feature directly addresses some of the most pressing gaps in our understanding of urban evolutionary ecology. For example, the special feature provides the first theoretical models that predict how populations are expected to evolve in response to urbanization [24] (gap 1). Multiple papers show that urbanization is frequently associated with changes in natural selection [25–28], and that populations can locally adapt to urban environments at both phenotypic [27–30,35] and genomic scales [42,43] (gap 2). Seven studies examine the repeatability of non-adaptive and adaptive evolution across cities and they find examples of both convergent and non-convergent evolution [28,35–39,42] (gap 3). Finally, multiple studies leverage the power of genomics to understand how urbanization affects non-adaptive and adaptive evolution [37–39,41–43] (gap 4), including how new urban species [37] and well-established human commensals arise [43] (gap 5).

This special feature highlights the power of urban evolutionary biology as a globally replicated experiment that provides an unparalleled opportunity to understand how human altered environments affect evolutionary processes and patterns. In this context, there are three takeaways from the 15 papers published in this special feature: (i) cities frequently alter natural selection; (ii) altered selection in urban populations frequently drives local adaptation and parallel adaptive evolution across multiple cities; and (iii) the development of cities has given rise to new human commensal species, a process that is ongoing and deserving of additional study across a broader range of taxa and cities.

Despite the major strides taken in the special feature, large gaps and unanswered questions remain. The ecological mechanisms driving differences in selection between urban and non-urban habitats remain elusive, and we have only scratched the surface of identifying the targets and functional consequences of genome-wide patterns of selection. Future work should strive to uncover the mechanisms behind cases of convergence and non-convergence, which would enable more accurate predictions on the effects of urbanization on evolutionary processes and patterns. Finally, understanding how cities shape the evolution of urban populations can facilitate designing management strategies for urban pests and help minimize the impact of humans on the spread of invasive species. By continuing to apply a broad range of techniques in studying the responses of diverse taxa to urbanization, future work in urban environments promises to shed light on both fundamental and applied problems in evolutionary biology.

Data accessibility

This article has no additional data.

Competing interests

We declare we have no competing interests.

Funding

During the writing of this manuscripts, J.S.S. was supported by an Ontario Graduate Scholarship (OGS), L.R.R. was supported by an NSERC CGS-D and M.T.J.J. was supported by an NSERC Discovery grant and NSERC accelerator award.