- Top of page
- Materials and methods
1. Understanding the mechanisms that affect invasion success of alien species is a major issue in current ecological research. Although many studies have searched for either functional or habitat attributes that drive invasion mechanisms, few researchers have addressed the role of phylogenetic diversity of alien species.
2. Here, using data from 21 urban floras located in Europe and eight in the USA, we show that the phylogenetic diversity of alien species is significantly lower than that of native species, both at the continental scale and at the scale of single cities.
3. Second, we show that if archaeophytes and neophytes (non-native species introduced into Europe before and after AD 1500, respectively) are analysed separately, archaeophytes show lower phylogenetic diversity than neophytes, while the phylogenetic structure of neophytes is indistinguishable from a random sample of species from the entire species pool.
4. Our results suggest that urban aliens are subject to environmental filters that constrain their phylogenetic diversity, although these filters act more strongly upon archaeophytes than neophytes.
5. Synthesis. Despite the huge taxonomic diversity of plants imported into European and American cities, the strong environmental filters imposed by cities constrain the functional diversity of urban floras, which is reflected in their generally low phylogenetic diversity. Urban alien floras are mainly composed of phylogenetically related species that are well adapted to anthropogenic habitats, although these filters are stronger for species groups with longer residence times.
- Top of page
- Materials and methods
Through the influence of human-related activities, the Earth’s biota have experienced the persistent weakening of biogeographical barriers to dispersal. This has resulted in the establishment and spread at increasingly broader scales of an increasing number of alien species (Vitousek et al. 1997; Lockwood 2004; McNeely 2005; Lambdon et al. 2008). For vascular plants, there have been substantial increases in species richness at local and regional scales as a consequence of elevated levels of biotic interchange (Sax & Gaines 2003; Sax et al. 2005). Therefore, it could be argued that vascular plants have become one of the primary beneficiaries of human-influenced biotic interchange. In addition, urban areas contain the greatest proportion of alien plants and act as hubs for onward dispersal of these species (Sukopp & Werner 1983; Kowarik 1990; Pyšek 1998; Roy, Hill & Rothery 1999; Wittig 2004; Chytrýet al. 2005, Chytrý et al. 2008; Tait, Daniels & Hill 2005; Celesti-Grapow et al. 2006). Thus, when documenting ecological consequences of biological invasions, urban vascular floras are an informative focal group (La Sorte, McKinney & Pyšek 2007).
A number of studies have shown that human settlements provide distinctive ‘niche opportunities’ (sensuShea & Chesson 2002) that have allowed many alien species to become established. For instance, alien species with higher temperature requirements and tolerance for arid environments tend to occur in city centres where the ‘urban heat island effect’ is more pronounced (Godefroid 2001; McKinney 2006).
From an evolutionary perspective, functionally related species that coexist in the same habitat often share a common origin and phylogenetic history, such that what is now called phylogenetic diversity and functional diversity are usually interrelated (Darwin 1859). When traits that render a species capable of colonizing a given habitat are phylogenetically conserved, phenotypic attraction (habitat filtering) promotes a taxonomically clumped flora in which co-occurring species that are adapted to similar niches are more related than expected by chance. Conversely, when distantly related taxa are phenotypically attracted and have converged on similar niche use, phenotypic attraction generates phylogenetically overdispersed communities (Cavender-Bares & Wilczek 2003; Kraft et al. 2007). Since, in both cases, the environment affects the functional and phylogenetic organization of a species assemblage (Knapp et al. 2008), we expect that urbanization will affect the phylogenetic structure of alien species assemblages.
While the influence of urbanization on plant functional traits has been confirmed by several authors (e.g. Kleyer 2002; Chocholoušková & Pyšek 2003; Williams et al. 2005; Lososováet al. 2006), little is known about the effects of urbanization on the phylogenetic diversity of alien species. It has been suggested that phenotypic and phylogenetic relatedness between native and alien species reduces the success of invasion (Darwin’s naturalization hypothesis; see e.g. Daehler 2001; Duncan & Williams 2002). The implication is that, because of limiting similarity due to overlap in resource use, native species can hinder the invasion of close relatives (see Procheşet al. 2008). In support of the proposed pattern, Strauss, Webb & Salamin (2006) found that highly invasive grass species are, on average, significantly less related to native grasses than expected from a random sampling of the phylogenetic supertree of all grass species of California. This confirmed previous work of Rejmánek (1996), who found that European grasses from alien genera were over-represented in California’s naturalized flora.
An alternative perspective to Darwin’s naturalization hypothesis suggests that, as native species possess functional traits that render them compatible with local environmental conditions, alien species with high phylogenetic relatedness to natives are more likely to share those well-suited traits, which enable them to succeed (Procheşet al. 2008). The idea that phylogenetic similarity between native and alien species may favour invasion processes is supported by work showing that taxonomic clustering is a major driver of community assembly (Webb et al. 2002; Cavender-Bares et al. 2004). This effect is particularly important at coarse spatial scales where plant-to-plant competitive interactions become irrelevant. Ricotta et al. (2008) tested the importance of taxonomic similarity in regulating species’ co-occurrence using data from 15 local species assemblages from portions of the urban flora of Rome, Italy. Their results indicate that in most cases the local species assemblages have a higher degree of taxonomic similarity than species assemblages randomly put together from the entire flora of Rome. Knapp et al. (2008) compared the phylogenetic diversity of urbanized areas in Germany with those of rural areas. They found that phylogenetic diversity of urban areas does not reflect the high species richness found there. Hence, high urban species richness is mainly due to closely related species that are functionally similar and adapted to disturbances associated with urbanization.
In principle, due to their very diverse origin, the phylogenetic structure of urban aliens could be expected to be significantly overdispersed when contrasted with the entire urban species pool. In this study we (i) test this assumption by analysing the phylogenetic diversity of alien species assemblages from a number of urban floras located on two continents, Europe and North America, and (ii) explore whether there is a difference in patterns shown by two groups of European alien species, archaeophytes and neophytes. The two groups differ in their residence times, with archaeophytes present in European landscapes for several millenia and neophytes present for several centuries (Pyšek, Richardson & Williamson 2004; Pyšek & Jarošík 2005). Another principle difference is the region of origin, which is more diverse for neophytes (Lambdon et al. 2008). Further, the majority of archaeophytes is confined to arable land and urban wasteland, while neophytes occur in a wider range of habitats (Pyšek, Richardson & Williamson 2004; Pyšek et al. 2005).
- Top of page
- Materials and methods
At the continental scale, the MPD of alien species was significantly lower than that of native species (t = 11.768; P < 0.001 for US cities and t = 6.035; P < 0.001 for European cities). At the local scale, for all US floras and for 17 out of 21 European floras, the actual MPD of urban aliens was significantly lower than the corresponding null values (P < 0.05; two-tailed test; Table 2). That is, on average, the phylogenetic structure of alien species assemblages was more clumped than that of assemblages randomly compiled from the entire urban species pool, rejecting the null hypothesis that urban aliens are just a random sample of the urban species pool. Also, for six US floras and for 11 European floras, the actual MPD of native species assemblages was significantly higher than the null expectation, meaning that native species are phylogenetically less clumped than expected from the null model.
Table 2. Mean phylogenetic distances (MPD) of alien and native species in 8 US and 21 European cities and the average (mean of 999 randomisations) MPD values of an equal-sized random sample of the entire urban species pool
|City||Alien species||Native species|
|Actual MPD||Average random MPD||Actual MPD||Average random MPD|
|United States of America|
| New York||260.782**||271.532||273.294||271.552|
| Saint Louis||252.708**||262.909||264.695**||262.939|
| Washington DC||249.734**||262.991||266.113**||262.967|
| Berlin, West (Germany)||258.514||260.276||261.003||260.155|
| Birmingham (UK)||246.908**||261.771||264.991||261.944|
| Brighton (UK)||238.520**||251.863||256.872**||251.804|
| Brno (Czech Republic)||247.051||249.070||250.806||249.183|
| Brussels (Belgium)||250.626*||256.614||258.735*||256.383|
| Chemnitz (Germany)||251.654**||260.622||264.711||260.679|
| Dublin (Ireland)||246.044**||256.857||260.922*||256.861|
| Exeter (UK)||244.370**||258.996||263.012||258.896|
| Halle an der Saale (Germany)||252.182**||259.100||263.289||258.943|
| Hannover (Germany)||247.907**||260.531||263.612**||260.569|
| Kingston upon Hull (UK)||243.642**||257.453||264.097**||257.568|
| Leeds (UK)||243.812**||261.106||265.256**||261.176|
| Leicester (UK)||247.157**||259.952||264.666*||259.941|
| Leipzig (Germany)||248.921**||254.912||260.481**||254.842|
| London (UK)||252.210**||259.651||264.151**||259.603|
| Plymouth (UK)||250.281**||259.143||262.464*||259.067|
| Plzeň (Czech Republic)||250.631*||259.751||262.794||259.806|
| Prague (Czech Republic)||254.070**||259.933||261.565||259.975|
| Rome (Italy)||267.930||260.352||258.386||260.182|
| Sheffield (UK)||259.193||261.877||262.807||261.954|
| Warsaw (Poland)||244.514**||260.688||266.274**||260.619|
When archaeophytes and neophytes of the European floras are analysed separately, the results become more complex. For MPD, the null hypothesis that the phylogenetic structure of alien species is indistinguishable from the structure of the entire urban species pool is rejected 19 times out of 21 for archaeophytes (Table 3). For 18 urban floras, the phylogenetic structure of the archaeophytes is more clumped than in random assemblages. But for the flora of Rome the archaeophytes show a phylogenetic structure that is significantly overdispersed (with a larger MPD for the archaeophytes than the entire flora) as compared to the corresponding null values. On the other hand, for MPD, the null hypothesis is accepted 14 times out of 21 for neophyte assemblages (Table 3). This means that in most cases, the phylogenetic structure of the neophytes is indistinguishable from a random sample of species from the entire species pool. For the remaining seven cities, the phylogenetic structure of neophytes was more clumped than in random assemblages.
Table 3. Mean phylogenetic distances (MPD) of archaeophyte and neophyte alien species in 21 European cities with the average MPD values of an equal-sized random sample of the entire urban species pool (mean of 999 randomisations)
|Actual MPD||Average random MPD||Actual MPD||Average random MPD|
|Berlin, West (Germany)||251.397||260.332||259.507||260.163|
|Brno (Czech Republic)||243.496*||249.080||248.059||249.135|
|Halle an der Saale (Germany)||242.165**||259.058||261.001||259.027|
|Kingston upon Hull (UK)||238.823**||257.269||244.516**||257.565|
|Plzeň (Czech Republic)||242.557**||259.889||257.118||259.879|
|Prague (Czech Republic)||240.930**||260.202||261.179||259.966|
For NMPD the phylogenetic structure of the archaeophytes was significantly clumped within particular clades for 18 out of 21 urban floras (Table 4), whereas for neophytes the general tendency was towards actual NMPD values that were slightly, though non-significantly, higher than the corresponding null values.
Table 4. Mean phylogenetic distances to the nearest relative (NMPD) of archaeophyte and neophyte alien species in 21 European cities with the average NMPD values of an equal-sized random sample of the entire urban species pool (mean of 999 randomisations)
|Actual NMPD||Average random NMPD||Actual NMPD||Average random NMPD|
|Berlin, West (Germany)||47.807**||68.700||48.813||47.836|
|Brno (Czech Republic)||38.324||43.426||36.814||38.575|
|Halle an der Saale (Germany)||38.514**||52.801||53.314||51.918|
|Kingston upon Hull (UK)||54.010**||70.817||58.751||56.136|
|Plzeň (Czech Republic)||43.112**||60.323||65.833||58.769|
|Prague (Czech Republic)||38.702**||52.551||47.689||46.208|
At the European scale, these local discrepancies between the phylogenetic structure of archaeophytes and neophytes resulted in a significant difference in the MPD and NMPD values of both species groups (t = 4.011, P < 0.001 for MPD, and t = 2.865, P = 0.009 for NMPD; Fig. 1).
Figure 1. Box plots of MPD and NMPD values of archaeophyte (Archaeo) and neophyte (Neo) alien species in 21 European cities.
Download figure to PowerPoint
- Top of page
- Materials and methods
Phylogenetic diversity summarizes the degree of evolutionary relationships within species assemblages, thus providing valuable information about mechanisms of community organization (Webb et al. 2002; Knapp et al. 2008). The availability of detailed phylogenies, along with methods for the construction of supertrees, now allows for the integration of phylogenetic information into studies of species assembly (Webb et al. 2002). Phylogenies constructed from supertrees usually contain pervasive polytomies below the family and genus level. Due to this lack of resolution, information on the phylogenetic organization of species assemblages is inevitably lost. However, given the robustness of the method used, we can be quite confident that the lack of resolution in the original supertree does not influence the results obtained (see Webb 2000).
Overall, our results suggest that alien species play a significant role in determining plant diversity within urban floras both in Europe and the USA. However, that diversity is not distributed at random within floras. Under the assumption of evolutionary trait conservatism (Donoghue 2008), phylogenetic niche conservatism together with the presence of selective environmental filters is an important mechanism in shaping the phylogenetic structure of urban alien plant assemblages. On the one hand, urbanization is closely associated with increasing opportunities for the introduction of alien species; on the other hand, cities are richly endowed with favourable habitats for the establishment of alien plants (McKinney 2006). Human disturbance creates physical conditions allowing the establishment of alien species outside their natural habitat. A straightforward example is the tendency for urban areas to have higher air temperatures compared to their rural surroundings. This ‘urban heat island effect’ promotes the establishment of species whose distribution is limited by cooler temperatures (Sukopp & Werner 1983; Godefroid & Koedam 2007). Other examples are the high proportion of surface runoff and of hard surfaces that increase the aridity of some urban habitats, and the high alkalinity of many urban soils (affected by adjacent concrete and other lime-based materials), which promotes the growth of plants that are adapted to soils with higher pH values (Sukopp 2004; Godefroid, Monbaliu & Koedam 2007; Thompson & McCarthy 2008).
Accordingly, we found that for all US and for most European floras, alien species have a higher phylogenetic aggregation than a random sample of the entire species pool. Using the general approach proposed by Cadotte & Lovett-Doust (2001), we found that alien species in the USA are significantly overrepresented by six families: Boraginaceae, Brassicaceae, Caryophyllaceae, Chenopodiaceae, Fabaceae and Solanaceae, while alien species in Central Europe are mainly overrepresented by the families Asteraceae, Brassicaceae, Chenopodiaceae, Poaceae and Solanaceae (see Pyšek, Sádlo & Mandák 2002).
Nonetheless, looking separately at archaeophytes and neophytes in European floras, substantial differences in MPD and NMPD across these two classes of residence time were found. That is, archaeophytes displayed the highest and neophytes the lowest level of phylogenetic aggregation. The lower MPD and NMPD of archaeophytes are probably related to the archaeophytes’ more restricted origin in comparison to that of neophytes, and to their strong habitat specificity, resulting from their adaptation to anthropogenic habitats having taken place mainly in agricultural areas. Accordingly, Apiaceae, Caryophyllaceae, Chenopodiaceae and Scrophulariaceae are typical archaeophyte families.
Given their habitat specificity, demonstrated by their high local phylogenetic aggregation (high NMPD values), archaeophytes possess a number of ecological, evolutionary and biogeographical characteristics that have promoted their successful colonization of warm and dry urban environments within Europe, including large distributional ranges, long-term associations with anthropogenic environments and human-mediated biotic interchange (Lososováet al. 2004; Pyšek et al. 2005; Sádlo, Chytrý & Pyšek 2007; La Sorte et al. 2008).
A notable exception is the flora of Rome in which archaeophytes have a phylogenetic structure that is significantly overdispersed. This may be ascribed to the fact that the composition of archaeophytes in Rome is quite different from that in cities in Central or Northern Europe (Celesti-Grapow 1995). Besides a group of species from the steppes of Central Asia (such as cereal weeds), which are common in Central and Northern Europe, archaeophytes in Central Europe also include species of Mediterranean origin (e.g. Pyšek, Sádlo & Mandák 2002; Preston, Pearman & Hall 2004) that persist thanks to the ‘heat island effect’. The origin of the archaeophytes in Italy is more diverse (Celesti-Grapow et al. 2009). Most Southern European species belong to the local flora, with the majority of Roman archaeophytes having been introduced through trade occurring in the Mediterranean among civilizations that established in ancient times in peninsular Italy and on surrounding islands. First, there were the Phoenicians and the Greeks, whose colonies occurred along the coasts of the Mediterranean Basin. These cultures were followed by the Etruscans and the Romans, whose trade extended to Central Italy, Northern Africa, Southwest Asia and Southern Europe. Furthermore, the city of Rome is much older than the other cities in our analysis such that archaeophytes have had more time to adapt to human land use than species in other parts of Europe. Finally, the generally warmer climate of the Mediterranean compared to Northern and Central Europe also contributes to the high phylodiversity of achaeophytes in Rome: even with the urban heat island effect promoting archaeophytes from warmer climates, their diversity in cities might depend, at least partially, on migration of species from source populations in rural areas. The rural areas can support species from warmer climates in Italy, but they only do so in a restricted way in the cooler areas of Northern and Central Europe.
In contrast, while, by definition, the species pool of archaeophytes is restricted, neophytes are still being introduced and represent a continually expanding species pool with a much broader geographical origin (Pyšek, Jarošík & Kučera 2003; Lambdon et al. 2008). Whereas neophytes tend to be better represented within the families Amaranthaceae, Fabaceae, Onagraceae, Polygonaceae and Solanaceae, they have maintained a high level of phylogenetic diversity in European cities, in terms of both MPD and NMPD, that is not distinguishable in most cases from the entire urban species pool. Nonetheless, in spite of their diverse origin, except for Sheffield (where the effect is non-significant), none of the urban neophytic floras shows a significantly overdispersed phylogenetic structure. This means that urban neophytes are still subject to environmental filters that constrain their phylogenetic structure, although these filters are weaker than for archaeophytes.
As already noted by Thompson, Hodgson & Rich (1995), ecological attributes of successful aliens are strongly habitat-dependent, such that relatedness of invaders to the native biota may be one useful criterion for predicting the key ecological characteristics of invasive species and potentially invasible ecosystems. For instance, while in the essentially closed communities of cool, damp climates, clonal growth and competitive ability seem to be important attributes of invasiveness, r-selected characteristics assume greater significance in drier, more open habitats (Thompson, Hodgson & Rich 1995). As ecologically important traits are usually conserved through evolutionary history (Donoghue 2008), it follows that phylogenetic community structure has important consequences for understanding invasiveness.
In light of our results, we can estimate how alien species will impact upon the (phylogenetic) diversity of urban areas based on their time of introduction. Our findings suggest that while the phylogenetic aggregation of urban archaeophytes reflects their long-term and broad-scale association with anthropogenic activities, neophytes, which are more recent invaders, are likely not to be as well-adapted to the environmental, ecological and anthropogenic conditions of urban habitats (Pyšek, Richardson & Williamson 2004). In addition, this temporal approach to urban invasibility automatically considers the ecological, evolutionary and geographical dissimilarity between introduced and native species: the farther back in time the introduction occurred, the shorter the geographical distance to the native flora; and the less dissimilar the environments in the native and introduced regions, the more likely the species will be adapted to biotic and abiotic conditions in the new region. Accordingly, phylogenetic relatedness of invaders to native communities may be one useful parameter for identifying threats to local native species and for prioritizing management efforts regarding alien species.