INTRODUCTION

Urbanization represents one of the major processes which have had profound effects on biodiversity and its distribution at a global scale (^{Aronson 2014}). The need to maintain the urban habitat in a homeostatic condition, directly affects the habitat and the conservation of the regional biota (McKinney 2006). Considering that recent sources (^{ONU 2014}) estimate a continuing population growth and urbanization across the globe, the threats to biodiversity are likely to increase. Although cities currently represent about 3% of the world’s land usage, their effects on biodiversity extend far beyond their municipal borders (^{Grimm et al. 2008}). The regions with the highest percentages of people living in urbanized areas are Europe, Latin America and the Caribbean, and Northern America with 73%, 80% and 82% respectively. Chile is the third most urbanized country (89%) in South America, only preceded by Argentina (92%) and Uruguay (95%) (^{ONU 2014}).

As urbanization continues to expand, with obvious consequences on abundance and species richness, efforts directed towards the conservation within urban landscapes could support regional and global biodiversity conservation, restoration and education as well as improve human well-being (^{Dearborn & Kark 2010}, ^{Faeth et al. 2011}). The expansion of cities not only alter the habitat of native species but also generates habitat for exotic species that are adapted to urban conditions (^{McKinney 2006}), promoting the replacement of local biotas with cosmopolitan species in a process known as biotic homogenization (^{McKinney & Lockwood 1999}, ^{Sax & Gaines 2003}). Furthermore, the introduction of exotic species has the potential to cause biological invasions (^{Richardson et al. 2000}), making urban areas pools of potentially spreading taxa. Hence, the balance between native and exotic species could be considered as a first indicator of the effects of urbanization on biodiversity.

So far, relatively few cities have been studied in terms of plant species composition with the majority of them being in Europe (e.g. ^{Grapow et al. 1996}, ²⁰⁰⁶, ^{Kent et al. 1999}, ^{Maurer et al. 2000}, ^{Leporatti et al. 2001}, ^{Van der Veken et al. 2004}, ^{Altay et al. 2010}, ^{Carretero 2010}, ^{Ricotta et al. 2010}, ^{Milovic & Mitic 2012}, ^{Stešević et al. 2014}). In these analyses, emerges that the native species fraction is commonly higher than the exotic one. Two studies encompassing a total of 117 European and 25 non-European cities have shown that < 50% of the urban flora were native species (^{Lososova et al. 2012}, ^{La Sorte et al. 2014}). Yet, the knowledge about urban biodiversity is considerably inferior in other continents. As for South-America, the attempts to describe urban floras are localized and incipient (^{Pauchard et al. 2006}). However, they show a variable representation of native species ranging between 19 and 31%, whereas exotic species ranging between 69 and 81% (see ^{Méndez 2005}, ^{Córdova 2013}, ^{Moro & Castro 2015}). Not surprisingly, given the lack of information, these cities have not been included in such large-scale studies as the aforementioned ones. These floristic studies on South American cities show that the number of exotic species is higher than the number of native species. Nevertheless, a generalization on the base of these results is difficult because the sampling was not systematic, and it would be premature to affirm whether the greater number of exotic species reflects a general trend contrasting the one of European cities.

The biogeographic region of central Chile (30-36o S), displays a native flora of global importance because of its high endemism (45.8% according to ^{Marticorena 1990}). The region mainly encompasses the Mediterranean-type climate flora of Chile and comprises 70% of the cities of the country, and 62% of the population (^{INE 2005}). Although the human impact on the vegetation structure of the extra-urban landscape has been investigated (^{Fuentes et al. 1989}, ^{Figueroa et al. 2011}), studies analysing the composition of the floras within the cities are currently scarce. Particularly in central Chile, studies on urban flora have been carried out only in Santiago (^{Figueroa et al. 2016}, ²⁰¹⁸, ^{Fischer et al. 2016}, ^{Hernández & Villaseñor 2018}), Temuco (^{Romero-Mieres et al. 2009}), and Curicó (^{Lozano-Diéguez & Teillier 2014}). Although these studies have analysed different floristic components (i.e. trees, shrubs, and/or herbs; planted or spontaneous species) and habitats (streets, abandoned sites, public and private parks), they have found that exotism (i.e., the proportion of exotic species respect to the total species) ranges between 73 and 92%. Thus, up to date there are no systematic studies that have described and compared the floristic composition of the cities of central Chile and showed how the native flora of the Mediterranean region is represented within them.

The present study investigates the composition of the ornamental flora of five cities of central Chile: La Serena, Valparaíso, Santiago, Rancagua, and Talca. These cities lie on a latitudinal gradient and represent the main urban centres within their respective administrative regions. We realized a systematic sampling of streets and squares aimed at describing and comparing the composition of planted woody species (trees and shrubs). Because of their representation and ornamental importance, we also included palms and bamboos species as well as semi-woody shrubs and succulent. Two main questions leaded our investigation: What is the proportion of native and exotic species in these cities? And does this proportion differs in all the five cities or follows a generalized pattern? Considering the latitudinal disposition of the cities studied, we expect to find a gradient in the representation of native and exotic species associated with environmental conditions observed along the latitude (i.e. temperature and precipitations, see ^{Luebert & Pliscoff 2006}). Because the richness of naturalized species increases with the latitude in extra-urban habitats (^{Fuentes et al. 2013}), it was hypothesized a similar trend for exotic species

in cities. Additionally, we characterize the composition of the species found in this sampling describing their biogeographical origin, their incidence and their distribution in streets and squares. With this information, we want to draw attention to the scarce representation of the native flora within the cities of central Chile.

METHODS

ANALYSES

Its was calculated the overall representation of exotic and native species (total pool of species) and compared this result with the proportion of exotic and native species found in each city to assess for statistical difference among cities. To statistically assess the significance of the differences, we used the Pearson’s chi-squared, where the observed frequencies in each city were compared with the expected frequencies generated from the total pool of species. By similar procedure, its was also compared the origin of the taxa between different habitat types (streets and squares). Furthermore, its was evaluated whether there was an overall significant difference in the origin of the species and whether the origins of the species were the same among cities.

TABLE 1 The five cities of central Chile object of this study with information about foundation year, area, density and number of plots sampled. / Cinco ciudades estudiadas con información acerca de año de fundación, área, densidad de habitantes y número total de plots empleados en el muestreo. 

FIGURE 1 Political map of Chile showing regional division, and close up to the area of central Chile including the cities sampled in this study. / Mapa político de Chile mostrando la división regional administrativa y la ubicación de las cinco ciudades estudiadas. 

Additionally, we wanted to assess the incidence of the native and exotic species in the studied cities. We calculated incidence as the number of plots in which the species occurred divided by the total number of plots. This was studied as a function of their biogeographical origin, i.e. native and exotic species. A Kolmogorov-Smirnov test was performed in order to assess the significance of the differences in the incidence between native and exotic species.

RESULTS

A total of 302 species were recorded for the five cities (Annex 1); among these, 275 (91.1%) species were angiosperms and only 27 (8.9%) were gymnosperms (Annex 1). Among the angiosperms, we found 71 families and 177 genera (Annex 1). The most represented families in terms of number of species were Rosaceae (35 species) and Fabaceae (30 species), followed by Malvaceae and

Oleaceae (10 species each). The rest of the families included between 1 and 9 species (Annex 1). At generic level, Prunus (12 species) and Acacia (9 species) were the most diverse (Annex 1). The rest of the genera were represented by less than 6 species. As for gymnosperms, we recorded 6 families (Annex 1); Cupressaceae (11 species), Pinaceae (9 species) and Araucariaceae (4 species) were the most diverse families (Annex 1). At generic level, 16 genera were recorded among the gymnosperms, and Cupressus and Araucaria were the most represented, with 4 species each (Annex 1).

Of the 302 species found in this study, its was recorded 42 (13.9%) native and 260 (86.1%) exotic species (Table 2). These frequencies differ of an equitable distribution (50 and 50%; χ2= 157; f.d.= 1; P < 0.01), and show that each city had a similar proportion of exotic and native species with no significant difference (Table 2). Similar results were obtained when comparing streets and squares: systematically a higher richness of exotic than native plant, following proportions as 88.6% and 11.3% for exotic and native species in streets, and 85.3% and 14.7% for exotic and native species in squares (Table 3).

TABLE 2 Numbers of exotic and native species found and results from Pearson’s chi-squared test comparing the observed frequencies of native and exotic species of the cities to the expected frequencies from the Total pool species. / Número total de especies nativas y exóticas registradas en el muestreo de cinco ciudades y los resultados de la prueba chi-cuadrado de Pearson establecidos al comparar las frecuencias de nativas y exóticas de cada ciudad respecto de la frecuencia esperada a partir del pool de especies totales. 

TABLE 3 Distribution of the native and exotic species in streets and squares in each of the studied cities and results from Pearson’s chi-squared test. This test compares the frequencies observed in each city with the expected one from the total pool of species. / Distribución de especies nativas y exóticas en calles y plazas de las cinco ciudades estudiadas y resultados de la prueba de chicuadrado de Pearson. Esta prueba compara las frecuencias observadas en cada ciudad con la esperada a partir del pool total de especies. 

TABLE 4 Species frequency according to their geographical origin (Geographical region) in each studied city. At the bottom of the table are results from Pearson’s chi square test; here the observed frequencies were compared with expected frequencies, which were obtained from the total pool species. / Número de especies según su región geográfica de origen (Geographical region), presente en cada ciudad estudiada. En la parte inferior de la tabla se encuentran los resultados de la prueba de chi cuadrado de Pearson; aquí las frecuencias observadas se compararon con las frecuencias esperadas, que se obtuvieron a partir del pool total de especies. 

Of the 42 native species, 6 were extra-limital natives: Araucaria araucana (Molina) K. Koch, Tara spinosa (Molina) Britton & Rose, Erythrostemon gilliesii (Hook.) D. Klotzsch, Cylindropuntia tunicata (Lehm.) F.M.Knuth, Hebe salicifolia G. Forst., and Schinus areira L. On the other hand, the 260 exotic species recorded fell in 18 biogeographical categories (Table 4); 39.6% of the exotic species, had their original distribution in different regions of Asia, including Temperate Asia, Tropical Asia and Australasia, with Temperate Asia alone counting for 24.2% (Table 4). North American species added up 14.2%, South America account for 9.6% of the exotic species, while African and European species 8% and 4.2% respectively (Table 4). The rest of the taxa showed original distributions which combines different regions and continents (Table 4).

No statistical differences were obtained when comparing the relative proportion of species falling in each category among the five cities (Table 4).

The 302 species recorded, ranged incidence values between 0.001 and 0.430. In 92.1% of the cases (278 species), incidence was < 0.1, while the remaining 24 species showed higher values between 0.1 and 0.430 (Fig. 2). The most frequent exotic species were Nerium oleander L., Melia azedarach L., Liquidambar styraciflua L., Ligustrum lucidum W.T. Aiton, Ligustrum ovalifolium Hassk., Prunus cerasifera Erhr., Robinia pseudoacacia L., and Acer negundo L., ranging between 53.9 and 85.8%; meanwhile, the most widely distributed native species were Cestrum parqui (Lam.) L’Hér., Cryptocarya alba (Mol.) Looser, Acacia caven (Mol.) Mol., Maytenus boaria Mol., Quillaja saponaria Molina, and Schinus areira L., ranging between 12.7% and 54.1%. When comparing incidence values between exotic and native species for each city and for all the cities together, no statistical difference were observed (see Fig. 2). Of the 302 species found in our sampling, 41 species (13.6%) were shared by all cities while 119 species (39.4%) were unique to one city. Of the former, 37 were exotic among which the aforementioned Acer negundo, Robinia pseudoacacia and Prunus cerasifera, and 4 were the native Quillaja saponaria, Acacia caven, Cryptocarya alba and Schinus areira. These findings show an ornamental trend that results in a few species widely distributed and many species narrowly distributed.

FIGURE 2 Percentage of native and exotic species plotted against their incidence classes, compared among the five cities and overall. All the cities show a similar pattern of species incidence, with a few taxa well represented across the plots of their boundaries and many taxa present in only a few plots (rare species); comparison were performed by a Kolmogorov-Smirnov test (La Serena D= 0.1, P= 0.925; Valparaíso D= 0.2, P= 0.700; Santiago D= 0.1 P= 0.304; Rancagua D= 0.1, P= 0.294; Talca D= 0.1, P= 0.9; Total D= 0.1, P= 0.171). / Distribución porcentual de la incidencia de especies exóticas y nativas en cinco ciudades estudiadas. Las distribuciones fueron comparadas usando pruebas de Kolmogorov-Smirnov (La Serena D= 0.1, P= 0.925; Valparaíso D= 0.2, P= 0.700; Santiago D= 0.1 P= 0.304; Rancagua D= 0.1, P= 0.294; Talca D= 0.1, P= 0.9; Total D= 0.1, P= 0.171). 

DISCUSSION

Previous studies on urban floras have shown that cities may support strikingly high biotic diversity (e.g. ^{Godefroid & Koedam 2007}) and possibly represent the regional flora to which the cities belong (e.g. ^{La Sorte et al. 2014}). However, our study shows that native species represent approximately 14% of the urban flora, nearly 25% of the Chilean native trees (≈ 120 species; see ^{Rodríguez et al. 1983}) and only 0.81% of the total species described for Chile (≈ 5,000 species; see ^{Marticorena & Quezada 1985}). Thus, the representation of the regional flora within the urban context appears to be low for central Chile. With regards to the exotic flora (which includes naturalized and non-naturalized species), it is difficult to establish quantitative comparisons because the study of their diversity has focused mainly upon the naturalized plants (e.g. ^{Arroyo et al. 2000}, ^{Figueroa et al. 2004}, ^{Pauchard et al. 2004}, ^{Castro et al. 2005}, ^{Fuentes et al. 2008}). In a floristic guide, ^{Hoffmann (1998}) summarized 94 exotic woody species for the total Chilean cities, of which 98% were found in this study. In a more exhaustive study, Rodríguez et al. (2005) recorded 158 exotic species (trees and arborescent species) for Chile, of which 64% were found in this study. These values indicate that within Chilean cities, the exotic trees and arborescent species are more diverse (i.e. greater species richness) and better represented with respect to the total floras (native and exotic sets) than the native ones.

These results are in line with the results of other studies carried out on planted and spontaneous flora of other cities of central Chile. ^{Lozano-Diéguez & Teillier (2014)} and ^{Romero-Mieres et al. (2009)} found that only 8% of the ornamental flora of Curicó and 27% of the ornamental flora of Temuco, respectively, was native. Recently, ^{Hernández & Villaseñor (2018}) reported an increase in the representation of native trees within Santiago over the last 12 years. Nevertheless, ^{Figueroa et al. (2016}) highlighted that the composition of the vascular spontaneous flora of Santiago was represented by 15% of native species and 85% of exotic species. ^{Fischer et al. (2016}) analysed the composition of spontaneous weeds in grassland and wooded areas of 15 parks of Santiago and found that exotic species contributed for > 90% to the total diversity. ^{Gärtner et al. (2015}) conducted a study on the ruderal herbs spontaneously growing in public spaces of Santiago and found that 16% of the taxa were native while 84% were exotic; additionally, ^{Figueroa et al. (2018)} found that 84% of the taxa present in private and public parks in Santiago were exotic species. Interestingly, by using a systematic sampling for the five studied cities we recorded 302 species of which approx. 86% were exotic and 14% were native; these figures were numerically consistent among the five cities studied not showing evidence of a gradient trend as we initially hypothesized by comparison with naturalized plants in extra-urban habitats (^{Fuentes et al. 2013}). Additionally, the proportion of native and exotic urban flora encountered in central Chile, differ from the European trend where better representation of the regional flora has been found (^{Aronson et al. 2014}, ^{Celesti-Grapow et al. 2013}, La Sorte et al. 2014). The representation of exotic over native species when comparing cities from European and central Chile also extends to the habitats within the cities, particularly between squares and streets (see ^{Lososová et al. 2012}). In fact, in the analysed cities of Chile, there is a greater representation of exotic species than native ones in both streets and squares whereas native species dominate urban squares and streets in European cities (^{Lososová et al. 2012}, ²⁰¹⁶). Probably, these differences can be attributed to the use of a greater diversity of native species as ornamentals in European cities (Lososová et al. 2012, ^{Kowarik et al. 2013}), a fact that does not seem to be the case in central Chile (^{Figueroa et al. 2016}, ²⁰¹⁸).

On the other hand, and despite the efforts to promote a change (e.g. ^{Riedemann & Aldunate 2001}, ²⁰⁰³, ^{Riedemann et al. 2006}, ²⁰⁰⁸) the scarce number of native species cultivated for ornamental programme purposes, could be related to the lack of knowledge around their possible use as ornamental plants, as well as to the misconception that Chilean native plants grow slowly. The only native species contrasting this general trend were Schinus areira, Quillaja saponaria, Maytenus boaria, Acacia caven and Cryptocarya alba whose incidence values were the highest among the natives and whose ornamental value, is well known. Given that some species such as Cryptocarya alba are considered vulnerable in the Metropolitan Region of the country (^{Benoit 1989}, ^{Riedemann & Aldunate 2001}), their use in urban settings could be a valuable mean for conservation. In Chile, the gymnosperms include various conifers among which there are native species of Araucariaceae, Cupressaceae, Podocarpaceae and Ephedraceae. In this study, we only found three species of gymnosperms reported for Chile: the native Araucaria araucana and the exotic Pinaceae Pinus radiata D. Don and Pseudotsuga menziesii (Mirb.) Franco. That means that the 94.1% of the Chilean gymnosperm flora (17 species), according to ^{Marticorena & Rodríguez (1995}), would not be represented within the cities of this study. Similarly, this study showed that the representation of the angiosperms was not higher.

With regards to the origin of the ornamental exotic flora, the findings were not consistent with ^{Matthei (1995}), ^{Arroyo et al. (2000}), ^{Figueroa et al. (2004}), ^{Castro et al. (2005}), who report that most of the naturalised flora in Chile comes from Mediterranean Eurasia. On the other hand, ^{Lozano-Diéguez & Teillier (2014)} and ^{Romero et al. (2009)} found that 29% and 19% of the ornamental flora of the Chilean cities of Curicó and Temuco respectively, were of Asian origin. The prominent number of exotic species, especially from Asia, can be related with the large number of plants that European gardeners introduced first to Europe, and then to Chile in the nineteenth century, reflecting the English and French influence over the urban space’s ornamentation style in Chile (^{Lozano-Diéguez & Teillier 2014} and references therein).

Consistent with other studies, the analysis on species incidence has shown a general tendency to use many rare species and a few common species in streets and squares (^{Lososová et al. 2012}, ^{Kowarik et al. 2013}). That is to say, only a minority of the species found were encountered in a consistent number of plots. Examples are Nerium oleander, Melia azedarach, Liquidambar styraciflua, Ligustrum lucidum, Ligustrum ovalifolium, Prunus cerasifera, Robinia pseudoacacia, and Acer negundo, all exotic to Chile. If making use of many rare species tends to increase diversity, the use of the same species in different cities decreases variability when comparing them.

It shows that exotic species strongly contribute to the homogenization of the floras of the cities studied (^{Aronson et al. 2014}, La Sorte et al. 2014, ^{Lososová et al. 2016}). In fact, 90.2% of the species (37 out of 41) shared by all the five cities were exotic; however, the urban homogenization for Chilean cities needs further investigation. The other issue with the predominance of the exotic flora within the urban context, is that much of the species naturalised in Chile such as Rubus (Rosaceae), Rosa rubiginosa L. (Rosaceae), Acacia dealbata Link (Fabaceae) among the others, have turned out to be invasive and have replaced the native flora in many areas (^{Teillier 2008}). Currently, we ignore if the urban populations of these species maintain reproductive relationships with the rural populations.

The use of native flora in cities could benefit in different ways. Not only it can provides support to the regional plant diversity; it can also create stepping stones or corridors for natural populations, as well as provide environmental education and ecosystem services (^{Dearborn & Kark 2010}). Theoretically, species that are well adapted to the local environmental conditions require less intervention for their maintenance (e.g. water requirements), which implicates the possibility of amortising the costs of maintenance. Another aspect, less obvious but realistically important for conservation, is the awareness of the public of native biodiversity. Some studies (^{Rozzi et al. 2003}, ^{Ballouard et al. 2011}) have shown that when people are asked to recall species of their cities, they mostly named exotic ones and they are keener to prioritize virtual exotic iconic biodiversity over local biodiversity. Additionally, it has been pointed out that native plants species may be preferred by native birds, thus a higher representation of native flora it is expected to attract native avifauna to urban areas (^{Díaz & Armesto 2003}, ^{White et al. 2005}).

Conservation within urban landscapes could support regional and global biodiversity (^{Dearborn & Kark 2010}). Thus, a better monitoring of the urban biota in areas of high regional biodiversity is certainly needed (^{Aronson 2014}). Our study shows that, at least in central Chile, the native flora is poorly represented and scarcely taken into consideration in urban planning ornamentation programmes. Cities are points of entry for numerous ornamental plants that could potentially naturalize, and spread their distribution toward extra-urban habitats, exerting an important impact upon the native biodiversity (^{Kowarik et al. 2013}). The massive use of exotic flora in street and squares, especially from Asia, can be linked to historical as well as to socio-cultural drivers. Taking into consideration the importance of these drivers is fundamental to improve the sustainability of our cities and their relationship with the regional environment.