**5. Local adaptation in tree species**

An adaptation is a characteristic that enhances the survival or reproduction of organisms that bear it, relative to alternative character states, especially ancestral condition. The only way to an adaptation to evolve is by means of natural selection, so it can be also stated that adaptation is a characteristic that evolved by natural selection (Futuyma, 2005).

Local adaptation is the set of patterns and processes observed across local populations of the same species connected, at least potentially, by dispersal and gene flow. It is generally the case when resident genotypes in a deme have on average a higher relative fitness in their local habitat than genotypes originated from other habitats (Kawecki and Ebert, 2004 and references therein).

Local adaptation can be observed in a continuous population, in which sampling units are arbitrary, but is more commonly observed in fragmented populations, since they are discrete units of perennial populations in well delimited habitat patches (Kawecki and Ebert, 2004). Considering that the most existing neotropical tree populations have been fragmented by land use and urban occupation, local adaptation is a relevant area of study for neotropical tree species. Even though, most studies on local adaptation in plants are only available for herbaceous plants in temperate regions (Leimu and Fischer, 2008).

We aim to gather information on local adaptation in plants that can be used to evaluate its strength and outcomes for neotropical tree populations, considering their present fragmented character. Our understanding of this topic is mainly based on the papers of Kawecki and Ebert (2004), Savolainen and colleagues (2007), and Leimu and Fischer (2008).

The ability to adapt may be compromised in small populations because of reduced genetic diversity, caused by genetic bottlenecks or founder effects, which have as consequences an increase in genetic drift and inbreeding. In addition to reduced genetic variation and genetic drift, local adaptation can also be constrained by variation in natural selection. Temporal environmental variability may involve opposing selection pressures and thus constrain adaptation. In contrast, spatial heterogeneity of the habitats of plant origin favours selection for reduced dispersal and increases habitat fidelity, which may in turn favour the evolution of local adaptation (reviewed in Leimu and Fischer, 2008).

Reproductive traits have a role in determining the extent of local adaptation in natural populations. Gene flow, for example, can hinder local adaptation. This is true because protected polymorphism in a heterogeneous environment may be maintained even if dispersal results in complete mixing of the gene pool. In such a case demes will not differentiate genetically, i.e. there will be no local adaptation. Nevertheless, the existence of a pattern of local adaptation despite gene flow certifies to the strength of natural selection imposed by particular environmental factors (Kawecki and Ebert, 2004). Also, if local adaptation is constrained by lack of genetic variation, dispersal and gene flow between populations can enhance local adaptation by increasing genetic variation within populations and potential to respond to selection (Leimu and Fischer, 2008).

Spatial environmental heterogeneity favours reduced dispersal and habitat fidelity, which make conditions for local adaptation more favourable. It should be noted that environmental heterogeneity favours the evolution of adaptive phenotypic plasticity. In the absence of costs of and constraints on plasticity, a genotype that in each habitat produces the locally optimal phenotype would become fixed in all demes. Adaptive phenotypic plasticity would thus lead to adaptive phenotypic differentiation, but without underlying genetic differentiation. The failure of the metapopulation to evolve such ideal plasticity is thus a pre-requisite for local adaptation (Kawecki and Ebert, 2004 and reference therein).

Shortlived and self-compatible species tend to be more strongly differentiated at a smaller scale than long-lived and outcrossing species, and so the former are expected to show stronger adaptation to local conditions. Therefore, neotropical tree species, which are longlived, outcrossing and as a group includes several examples of self incompatibility, are expected to have weak local adaptation (Leimu and Fisher 2008).

Leimu and Fischer (2008) conducted the first quantitative review on local adaptation in plants, assembling papers that reported comparisons of the performance of plants from local and foreign populations. Among these studies, local genotypes performed on average better than foreign genotypes at their site of origin. However, divergent selection favoured locally adapted plants only in less than half of the pair-wise site comparisons. This suggested to them that local adaptation is less widespread than commonly believed. In this study, they also found that local adaptation appeared to be independent of some plant life-history traits, the degree of spatial and temporal habitat heterogeneity, and of the geographic distance between study populations, but was strongly affected by population size. This clear role of population size for the evolution of local adaptation raises considerable doubt on the ability of small plant populations to cope with changing environments. Thus, in the context of fragmentation process in the Neotropic small fragments, featured by low genetic variability as a consequence of genetic drift and endogamy, might not be able to respond to different selective pressures of changing environments and develop local adaptation.

#### **6. Brazilian Atlantic forest fragments: Case studies of species from Tibagi River Basin**

#### **6.1 The Tibagi project**

In this section of the chapter we intend to describe our experience in population genetic studies of neotropical trees remnants in Brazil. All species that we have studied have a high importance value index or IVI (the sum of relative dominance, relative frequency and

Reproductive traits have a role in determining the extent of local adaptation in natural populations. Gene flow, for example, can hinder local adaptation. This is true because protected polymorphism in a heterogeneous environment may be maintained even if dispersal results in complete mixing of the gene pool. In such a case demes will not differentiate genetically, i.e. there will be no local adaptation. Nevertheless, the existence of a pattern of local adaptation despite gene flow certifies to the strength of natural selection imposed by particular environmental factors (Kawecki and Ebert, 2004). Also, if local adaptation is constrained by lack of genetic variation, dispersal and gene flow between populations can enhance local adaptation by increasing genetic variation within populations

Spatial environmental heterogeneity favours reduced dispersal and habitat fidelity, which make conditions for local adaptation more favourable. It should be noted that environmental heterogeneity favours the evolution of adaptive phenotypic plasticity. In the absence of costs of and constraints on plasticity, a genotype that in each habitat produces the locally optimal phenotype would become fixed in all demes. Adaptive phenotypic plasticity would thus lead to adaptive phenotypic differentiation, but without underlying genetic differentiation. The failure of the metapopulation to evolve such ideal plasticity is thus a

Shortlived and self-compatible species tend to be more strongly differentiated at a smaller scale than long-lived and outcrossing species, and so the former are expected to show stronger adaptation to local conditions. Therefore, neotropical tree species, which are longlived, outcrossing and as a group includes several examples of self incompatibility, are

Leimu and Fischer (2008) conducted the first quantitative review on local adaptation in plants, assembling papers that reported comparisons of the performance of plants from local and foreign populations. Among these studies, local genotypes performed on average better than foreign genotypes at their site of origin. However, divergent selection favoured locally adapted plants only in less than half of the pair-wise site comparisons. This suggested to them that local adaptation is less widespread than commonly believed. In this study, they also found that local adaptation appeared to be independent of some plant life-history traits, the degree of spatial and temporal habitat heterogeneity, and of the geographic distance between study populations, but was strongly affected by population size. This clear role of population size for the evolution of local adaptation raises considerable doubt on the ability of small plant populations to cope with changing environments. Thus, in the context of fragmentation process in the Neotropic small fragments, featured by low genetic variability as a consequence of genetic drift and endogamy, might not be able to respond to different

pre-requisite for local adaptation (Kawecki and Ebert, 2004 and reference therein).

and potential to respond to selection (Leimu and Fischer, 2008).

expected to have weak local adaptation (Leimu and Fisher 2008).

selective pressures of changing environments and develop local adaptation.

**River Basin** 

**6.1 The Tibagi project** 

**6. Brazilian Atlantic forest fragments: Case studies of species from Tibagi** 

In this section of the chapter we intend to describe our experience in population genetic studies of neotropical trees remnants in Brazil. All species that we have studied have a high importance value index or IVI (the sum of relative dominance, relative frequency and relative density) in the Tibagi River basin, Paraná, Brazil and, due to the rapid degradation of ecosystems associated with this river basin, its present occurrence was limited to highly impacted forest remnants. The Tibagi River Basin has a great importance in the economic and social development of one of the Brazilian states of greatest economic impact in the country, however, the degree of landscape devastation that has taken place since the beginning of the last century has been threatening the biodiversity of the ecosystem and even physical and chemical characteristics of its rivers. For these reasons, a group composed of around 51 researchers from the University of Londrina and other research centers, in partnership with COPAT (Consortium for Environmental Protection of the Tibagi River basin) and funded by Klabin S/A and Araucaria Foundation for the Support of Scientific and Technological Development of Paraná, developed the project "Aspects of Fauna and Flora of the Tibagi River basin" or Tibagi Project. The Tibagi Project produced a wealth of valuable information aimed at the recovery and conservation of the river basin as a whole. We limit ourselves here to present the knowledge gained about the influence of two factors on the genetic diversity of populations of neotropical tree species remnants: habitat fragmentation and local adaptation to seasonally flooded river banks.

The Tibagi River basin is composed of 65 direct tributaries and hundreds of sub-tributaries in an area of approximately 25,000 Km2, covering 54 counties in the Paraná State, Brazil. Its landscape, belonging to the Atlantic Forest biome, presents important climate and soil variations in the north-south axis that allows it to be divided into Upper, Middle, and Lower Tibagi. On the Upper Tibagi, the predominant vegetation is steppe grassy-woody also known as general fields, with patches of Araucaria forest. In the Middle Tibagi, there is a transition zone between rain forest and mixed semideciduous forest with some patches of fields and scrubs. In the Lower Tibagi, before fragmentation, the dominant vegetation was the semi-deciduous forest, which formed a continuum with the dense rain forest of the Brazilian coast. With massive deforestation due to lumber extraction, urbanization and expansion of the agricultural front, only small forest fragments remain in this area adding up to only 2.4% of the original forest cover in the lower Tibagi and 12.7% in the Tibagi River basin as a whole (Ribeiro, 2009).

The riparian forests are plant formations that surround bodies of water and for that reason, are deeply influenced by them. On the other hand, water quality and other physical characteristics of rivers are also heavily influenced by the presence and degree of conservation of the riparian forests. In river banks, where the vegetation is scarce or not present at all, events such as erosions can be up to 30 times more significant than in areas where the riparian vegetation is present. Moreover, it is estimated that 80-90% of sediments and pollutants generated in agricultural areas are filtered by riparian vegetation before reaching the bodies of water (Naiman & Decamps, 1997). From the ecological point of view, these forests work as ecological corridors, linking different ecosystems and thereby promoting gene flow, both by land, and across the river. These are environments in which the diversity is determined by the occurrence of flood events with consequent flooding of the river margins. These environments can be very heterogeneous when flooding events are short, sporadic and with little intensity; however, they can be slightly heterogeneous when these flooding events are constant or very prevalent and intense.

Flooding is the main limiting factor that acts on the riparian forest, since it changes dramatically the physical and chemical characteristics of the soil, which is the main substrate for plant growth. In flooded soil, the diffusion of gases is severely reduced, with a sharp drop in oxygen levels, reduction of gas exchanges, accumulation of toxins, changes in pH and nutrient availability (Drew 1992). When flooding extends to total or partial submergence of the plants, the negative effects on vegetation are even greater due to reduced availability of light and carbon dioxide (Crawford, 1993). Once flood period is over, the soil becomes compacted and a series of injuries from re-aeration process arises due to the accumulation of reactive oxygen species in plant tissues during the flooding period (Crawford, 1993).

However, most of the riparian plant species survive the flooding stress through the development of morphologic, anatomic and physiological changes (collectively referred to as the low oxygen escape syndrome - LOES; Bailey-Serres & Voesenek, 2008) that reestablished, at least in part, minimum energy levels for plant survival. Other plant species survive with avoidance strategies, completing their life cycles between periods of flooding. There are also those species that have better relative growth rates in flooded soil. The ability to tolerate periods of flooding gives plants a competitive advantage over those that do not tolerate this stress, which is evidenced by the example of the tolerant species *Cecropia pachystachya*, *Sebastiana commersoniana* and species of the genus *Inga*, which are most often found in areas affected by periodic flooding than in other plant formations, where inter-specific competition limits its distribution. In general, local adaptation in challenging environments may either favour the reduction of genetic variation through natural selection or lead to the expression of phenotypic plasticity of identical genotypes (Stöcklin et al., 2009).

In this light, it seems clear that the recovery strategies and/or conservation of genetic diversity of tree species of the riparian forests on the Tibagi River basin should consider: (i) knowledge about the history of fragmentation and its influence on the genetic structure of the affected populations, (ii) knowledge about the diversity of mechanisms of tolerance to flooding and their possible genetic determination, (iii) and the influence of local adaptation to flooding events on the distribution of genetic variation within and among populations. In order to answer these questions, many common species from Tibagi river basin were studied, regarding flood tolerance and occurrence of morphological, anatomical and physiological changes in response to it. For eight of these species, we used DNA-based molecular markers (RAPD, AFLP and SSR) to obtain estimates of population genetic parameters such as polymorphic loci, total genetic diversity and genetic differentiation within and among populations. The eight sampled species have high importance value index (IVI) in Tibagi River basin (five are among the ten species with highest IVI by region) and cross-pollination system, but vary in regard to gender distribution and dispersal mechanisms for pollen and seeds (Table 1).

In Tibagi River basin, the analysis of the pluviometrical events from 1932 to 1990 showed that there were ten significant flood events in the region in this period, while the fragmentation of the vegetation was more pronounced between 1920 and 1960 (Medri et al., 2002). Considering that the first study published by the group comprised a minimum of 76 years between the collection of plant material and the first pronounced flooding event registered in the region and 83 years in relation to the beginning of the fragmentation process; in the meantime, the considered species advanced at least three generations (Table 1). It was expected at first that: (i) the genetic diversity of populations of degraded areas might have been reduced compared to preserved areas, (ii) the pioneer species have been less affected by the effects of fragmentation than the species of late successional stages, (iii) there has been significant genetic differentiation between fragments and (iv) local adaptation to flooding has caused genetic structure, probably by reducing the diversity in populations that regularly experience stress.


1 Pi: Pioneer; Si: Secundária inicial; St: Secundária tardia; Cl: Clímax. 2 Ou: Outcrossing; Vp: Vegetatively propagated; Mo: Monoecious; Mo-Si: Monoecious, Self

incompatible; Di: Dioecious; He: Hermaphrodite; Un: unknown.

3 Bi: Biotic; Ab: Abiotic.

4 In press.

402 Genetic Diversity in Plants

substrate for plant growth. In flooded soil, the diffusion of gases is severely reduced, with a sharp drop in oxygen levels, reduction of gas exchanges, accumulation of toxins, changes in pH and nutrient availability (Drew 1992). When flooding extends to total or partial submergence of the plants, the negative effects on vegetation are even greater due to reduced availability of light and carbon dioxide (Crawford, 1993). Once flood period is over, the soil becomes compacted and a series of injuries from re-aeration process arises due to the accumulation of reactive oxygen species in plant tissues during the flooding period

However, most of the riparian plant species survive the flooding stress through the development of morphologic, anatomic and physiological changes (collectively referred to as the low oxygen escape syndrome - LOES; Bailey-Serres & Voesenek, 2008) that reestablished, at least in part, minimum energy levels for plant survival. Other plant species survive with avoidance strategies, completing their life cycles between periods of flooding. There are also those species that have better relative growth rates in flooded soil. The ability to tolerate periods of flooding gives plants a competitive advantage over those that do not tolerate this stress, which is evidenced by the example of the tolerant species *Cecropia pachystachya*, *Sebastiana commersoniana* and species of the genus *Inga*, which are most often found in areas affected by periodic flooding than in other plant formations, where inter-specific competition limits its distribution. In general, local adaptation in challenging environments may either favour the reduction of genetic variation through natural selection or lead to the expression of phenotypic plasticity of identical genotypes

In this light, it seems clear that the recovery strategies and/or conservation of genetic diversity of tree species of the riparian forests on the Tibagi River basin should consider: (i) knowledge about the history of fragmentation and its influence on the genetic structure of the affected populations, (ii) knowledge about the diversity of mechanisms of tolerance to flooding and their possible genetic determination, (iii) and the influence of local adaptation to flooding events on the distribution of genetic variation within and among populations. In order to answer these questions, many common species from Tibagi river basin were studied, regarding flood tolerance and occurrence of morphological, anatomical and physiological changes in response to it. For eight of these species, we used DNA-based molecular markers (RAPD, AFLP and SSR) to obtain estimates of population genetic parameters such as polymorphic loci, total genetic diversity and genetic differentiation within and among populations. The eight sampled species have high importance value index (IVI) in Tibagi River basin (five are among the ten species with highest IVI by region) and cross-pollination system, but vary in regard to gender distribution and dispersal

In Tibagi River basin, the analysis of the pluviometrical events from 1932 to 1990 showed that there were ten significant flood events in the region in this period, while the fragmentation of the vegetation was more pronounced between 1920 and 1960 (Medri et al., 2002). Considering that the first study published by the group comprised a minimum of 76 years between the collection of plant material and the first pronounced flooding event registered in the region and 83 years in relation to the beginning of the fragmentation process; in the meantime, the considered species advanced at least three generations (Table 1). It was expected at first that: (i) the genetic diversity of populations of degraded areas

(Crawford, 1993).

(Stöcklin et al., 2009).

mechanisms for pollen and seeds (Table 1).

Table 1. Biological features and genetic parameters of the tree species studied in the Tibagi Project initiative. Information on biological traits was obtained in the literature and by personal communications of experts. When more than two populations of a given species were studied, the Fst values are represented as an average of those populations. Ss: Sucessional stages; Ra: Reprodutive age, in years; MS (SD): Mating system (Sex distribution); Pd: Polen dispersal; Sd: Seed dispersal; Long: Longevity, in years; Mm: Molecular marker; Np: Number of populations; Pp: Percentage of polymorphic loci; Ht: heterozigosity or genic diversity, when either codominant or dominant markers are respectively applied; Ref: references.

#### **6.2 Local adaptation of tree species to flooding**

In the Amazonian floodplain forests there are some regions where the water column reaches up to 7m with submersion of trees for long periods, and regions where this stress is less intense. Comparing the different areas of flooding, it is possible to observe a significant variation in the occupation pattern of these areas ranging from monospecific to over 150 species ha-1 (Wittmann et al., 2007 and references therein). It is clear that the stress intensity and duration of flooding periods determine the diversity of species occupying these areas. In the Tibagi River basin, the most critical flooding period is the three wettest summer months (December-January-February), with areas that may remain flooded for a few days to several months depending on the intensity of the phenomenon. Torezan & Silveira (2002), conducted several phytosociological studies in forest fragments along the Tibagi River Basin comparing areas under different intensities of occasional floodings. It was observed that the higher intensity, duration and size of the flooded area, the lower species richness was found. In a fragment with 100% of flooded area, there were 42 species/ha, while in fragments located in non-flooded areas, or with an insignificant area subject to flooding, the number of tree species was often greater than 100/ha, reaching up to 127 in one single area. Thus, we consider as a starting point the hypothesis that the stress of flooding, typically observed in the region, should determine the occupancy of the periodically flooded banks, favoring the establishment of species tolerant to this stress.

Several studies with species that are found in the Tibagi River basin were conducted in a greenhouse to answer this hypothesis. Plant species not found in the wetlands and the species most often found in the wetlands have been challenged in artificial flooding treatments with different intensities and durations. Briefly, the results showed that some tree species naturally distributed in this river basin but that do not occur in flooded areas, did not tolerate the stress of artificial flooding; also, there was a great variation in the response of the species from wetlands. In addition to this information, the provenance trials, conducted to compare the performance of populations of the same species from areas periodically flooded and non flooded, showed that for the studied species, *Luehea divaricata* and *Parapiptadenia rigida*, the plants that originated from populations naturally flooded were more efficient in responding to waterlooging stress (De Carvalho et al., 2008) or tolerated higher levels of stress, which was not tolerated by the plants of the other provenance (Silva et al., 2010). In a similar study with the tree species *Aegiphila sellowiana*, Medri et al. (2011) used plants regenerated from seeds collected in four regions that presented occasional flooding, showing a variation in the response of individual tolerance to waterlogging. In this experiment, 46.7% of the plants died, while the remaining individuals developed morphoanatomical alterations common to tolerant species, supporting up to 80 days flooding (Medri et al., 2011).

When these results were obtained (2000 - 2002) we believed that there should be a genetic background related to flood tolerance, but the variation in responses observed between the tolerant species pointed to the existence of a large number of genes and/or a strong effect of genotype-environment interaction in determining the tolerance. With these in mind, we used Random Amplified Polymorphic DNA (RAPD) to see if there was a reduction in genetic variation in populations tolerant to flooding, and whether it was possible to detect variation between populations that respond differently to the stress. In our experiments of performance comparison, the percentage of polymorphism observed *in situ* among adults of *Luehea divaricata* and *Parapiptadenia rigida* was greater (approximately 6%) in the areas subjected to periodic flooding than in the highest regions, which is never affected by stress. The studied populations of *L. divaricata* and *P. rigida*, showed genetic differentiation of 10.48 % for the first species and of 6.00% for *P. rigida* (while about 90% of the observed variation was attributed to the variation found within populations). These results suggested that perhaps the expected reduction in genetic diversity of riparian communities has been masked by the balance between different evolutionary forces that could be occurring *in situ*. Considering the proximity of the sampled areas and the fact that both are crossbreeding species, there certainly is gene flow between the flooded and non flooded populations, which in turn must ensure the re-establishment of genetic diversity in each generation. Possibly, the lower inter-specific competition and greater light availability on the river banks may represent, for the pioneer species studied, a more important factor than natural

comparing areas under different intensities of occasional floodings. It was observed that the higher intensity, duration and size of the flooded area, the lower species richness was found. In a fragment with 100% of flooded area, there were 42 species/ha, while in fragments located in non-flooded areas, or with an insignificant area subject to flooding, the number of tree species was often greater than 100/ha, reaching up to 127 in one single area. Thus, we consider as a starting point the hypothesis that the stress of flooding, typically observed in the region, should determine the occupancy of the periodically flooded banks, favoring the

Several studies with species that are found in the Tibagi River basin were conducted in a greenhouse to answer this hypothesis. Plant species not found in the wetlands and the species most often found in the wetlands have been challenged in artificial flooding treatments with different intensities and durations. Briefly, the results showed that some tree species naturally distributed in this river basin but that do not occur in flooded areas, did not tolerate the stress of artificial flooding; also, there was a great variation in the response of the species from wetlands. In addition to this information, the provenance trials, conducted to compare the performance of populations of the same species from areas periodically flooded and non flooded, showed that for the studied species, *Luehea divaricata* and *Parapiptadenia rigida*, the plants that originated from populations naturally flooded were more efficient in responding to waterlooging stress (De Carvalho et al., 2008) or tolerated higher levels of stress, which was not tolerated by the plants of the other provenance (Silva et al., 2010). In a similar study with the tree species *Aegiphila sellowiana*, Medri et al. (2011) used plants regenerated from seeds collected in four regions that presented occasional flooding, showing a variation in the response of individual tolerance to waterlogging. In this experiment, 46.7% of the plants died, while the remaining individuals developed morphoanatomical alterations common to tolerant species, supporting up to 80 days flooding

When these results were obtained (2000 - 2002) we believed that there should be a genetic background related to flood tolerance, but the variation in responses observed between the tolerant species pointed to the existence of a large number of genes and/or a strong effect of genotype-environment interaction in determining the tolerance. With these in mind, we used Random Amplified Polymorphic DNA (RAPD) to see if there was a reduction in genetic variation in populations tolerant to flooding, and whether it was possible to detect variation between populations that respond differently to the stress. In our experiments of performance comparison, the percentage of polymorphism observed *in situ* among adults of *Luehea divaricata* and *Parapiptadenia rigida* was greater (approximately 6%) in the areas subjected to periodic flooding than in the highest regions, which is never affected by stress. The studied populations of *L. divaricata* and *P. rigida*, showed genetic differentiation of 10.48 % for the first species and of 6.00% for *P. rigida* (while about 90% of the observed variation was attributed to the variation found within populations). These results suggested that perhaps the expected reduction in genetic diversity of riparian communities has been masked by the balance between different evolutionary forces that could be occurring *in situ*. Considering the proximity of the sampled areas and the fact that both are crossbreeding species, there certainly is gene flow between the flooded and non flooded populations, which in turn must ensure the re-establishment of genetic diversity in each generation. Possibly, the lower inter-specific competition and greater light availability on the river banks may represent, for the pioneer species studied, a more important factor than natural

establishment of species tolerant to this stress.

(Medri et al., 2011).

selection caused by flooding, enabling the maintenance of higher levels of diversity in wetlands.

We needed a model that would allow us to access the isolated effect of flooding on the genetic diversity of populations of plants tolerant and intolerant to stress. In this experimental model we use the tree species *A. sellowiana*, which displays a gradation of responses to flooding, ranging from death to survival periods of up to 80 days of stress. When comparing plants of *A. sellowiana* which survived the stress of artificial flooding (tolerant), with plants which died (intolerant), the genetic differentiation around 6% was detected between the two groups and it was not possible to detect important variation in the percentage of polymorphism between both of them (Medri et al., 2011). Given these results, it became clear that the vast phenotypic variation between tolerant and intolerant plants of *A. sellowiana* can be, at least in part, related to the genetic patterns observed when using the DNA-based marker RAPD. A similar situation was found by Bekessy et al. (2003) who used RAPD markers to study genetic variation in populations of the South American tree *Araucaria araucana*. Even though we have yet to acquire the knowledge of which genes determine the stress tolerance of flooding on tree species it is possible that this character is directly influenced by many genes as suggested by Sairam et al. (2008) in their review entitled "Physiology and biochemistry of waterlogging tolerance in plants".

Interestingly, considering the results obtained by the group, especially with the local adaptation experiment, carried out *ex situ* with *A. sellowiana*, it becomes clear that the genetic diversity within and between populations cannot be considered the only factor in the election of sources for the recovery of degraded riparian forests. In short, the RAPD markers used in these studies allowed us to detect consistent results for the three species studied since more than 94% of genetic variation detected can be attributed to variation found within populations, following the pattern usually observed for pioneer species, with preferably cross-fertilization. In the absence of knowledge on the mechanisms of inter-specific diversity of tropical trees on response to flooding, we may conclude, erroneously, based only on estimates of distribution of the genetic diversity, that both the populations that occupy the margins of the flooded rivers and those never flooded could be used as seed sources for the recovery of degraded riparian areas. However, the difference in performance between plants, as measured by morphological parameters, indicate that populations locally adapted to the stress would provide a better material for the recovery of degraded riparian areas.

#### **6.3 Effects of fragmentation in natural populations from Tibagi River Basin**

In several years of studies observing the effects of fragmentation and local adaptation on the genetic diversity of the remaining tree populations of the Tibagi River Basin, eight species were evaluated mostly with dominant molecular markers (Table 1). The information gathered here reflect observations for post-fragmented populations, and the number of generations elapsed since the beginning of the fragmentation process is equal to one or two for centenary species, and no more than ten generations for the species with longevity from 20 to 25 years.

As seen among tropical species (Matallana et al., 2005), all species present higher frequencies of cross-fertilization and animals play an important role in pollination and/or seed dispersal. These characteristics can be readily related to the high levels of genetic diversity (or heterozygosity) observed in the sampled populations (Table 1). In addition, pollen and seed dispersal by animals, often observed in tropical species, should provide high gene flow between nearby populations in a continuous gradient, while favouring the genetic differentiation over long distances (Givinish, 1999 and references there in). In such cases, the observed genetic distance between populations was positively correlated with their geographic distances. This hypothesis could be verified from the combined analysis of two studies conducted with *P. rigida*, where two nearby populations (Silva et al., 2010) and eight populations distributed over a long geographic scale (Souza et al., 2011) were evaluated. For nearby populations, the Fst value obtained was only 6.2%, whilst the average FST between the eight populations was 22.8%, and the correlation between genetic and geographic distance was positive and significant. Similarly, the observed values of FST between geographically close and distant populations of *A. sellowiana* ranged from 9.56% to 50.26%, respectively (Medri et al. 2011; Oliveira, 2010). For this species in particular, the genetic differentiation between distant populations may be even more pronounced than the one seen for *P. rigida* due to the occurrence of vegetative propagation.

When the genetic diversity among populations is compared between the studied outcrossing species, one can see that there are significant variations among the observed values of FST (6% - 30%, Table 1), which is suggested to be related to successional stage that each species occupies. Fst values for the initial or secondary pioneer species vary between 6% - 19% and the values for the late secondary or climax species vary between 21% - 30%. In both cases, genetic differentiation among populations assume moderate to high values, which is, in principle, expected to occur among tropical species. Tropical species are often pollinated by animals or have their seeds dispersed by them, often have mixed breeding system in which selfing rates can be changed depending on the environment and are represented by low-density populations (Dick et al. , 2008). These characteristics, together, make the tropical species more dependent on the quality of the ecosystem where they are inserted then the temperate species, and therefore more susceptible to the effects of fragmentation. Under the effect of fragmentation, tropical species suffer a reduction in the availability of pollinators and/or seed dispersers with the reduction in the number of individuals, experiencing a higher proportion of inbreed crosses and genetic drift, with a resulting increase in genetic differentiation between populations.

It is suggested that the observed tendency of lower values of Fst between the pioneer species than among the secondary is a reflection of the life history of these species. The pioneer species represent the first successional stage of the forest, being able to invade areas not yet occupied, including harsh environments and forest edges. Thus, given the ecological role they play, the pioneers are heliophyte, experience rapid growth, lower inter-specific competition, increased investment in reproduction (r selection), a higher number of generations per unit time and aggregate distribution. Fragmentation, therefore, does not cause too negative effects on the pioneer species. In fact, while fragmentation reduces the number of individuals of the species of later successional stages and completely alters the environment they occupy, the pioneer species can be instead favored by this process. Also, the pioneer species from the fragments of the Tibagi River Basin have advanced several generations (10 -15 generations) since the beginning of the fragmentation process. So, after several cycles of cross-fertilization, gene flow between populations may have indeed come relatively high to overcome the effects of fragmentation, allowing lower levels of genetic differentiation among populations.

seed dispersal by animals, often observed in tropical species, should provide high gene flow between nearby populations in a continuous gradient, while favouring the genetic differentiation over long distances (Givinish, 1999 and references there in). In such cases, the observed genetic distance between populations was positively correlated with their geographic distances. This hypothesis could be verified from the combined analysis of two studies conducted with *P. rigida*, where two nearby populations (Silva et al., 2010) and eight populations distributed over a long geographic scale (Souza et al., 2011) were evaluated. For nearby populations, the Fst value obtained was only 6.2%, whilst the average FST between the eight populations was 22.8%, and the correlation between genetic and geographic distance was positive and significant. Similarly, the observed values of FST between geographically close and distant populations of *A. sellowiana* ranged from 9.56% to 50.26%, respectively (Medri et al. 2011; Oliveira, 2010). For this species in particular, the genetic differentiation between distant populations may be even more pronounced than the one

When the genetic diversity among populations is compared between the studied outcrossing species, one can see that there are significant variations among the observed values of FST (6% - 30%, Table 1), which is suggested to be related to successional stage that each species occupies. Fst values for the initial or secondary pioneer species vary between 6% - 19% and the values for the late secondary or climax species vary between 21% - 30%. In both cases, genetic differentiation among populations assume moderate to high values, which is, in principle, expected to occur among tropical species. Tropical species are often pollinated by animals or have their seeds dispersed by them, often have mixed breeding system in which selfing rates can be changed depending on the environment and are represented by low-density populations (Dick et al. , 2008). These characteristics, together, make the tropical species more dependent on the quality of the ecosystem where they are inserted then the temperate species, and therefore more susceptible to the effects of fragmentation. Under the effect of fragmentation, tropical species suffer a reduction in the availability of pollinators and/or seed dispersers with the reduction in the number of individuals, experiencing a higher proportion of inbreed crosses and genetic drift, with a

It is suggested that the observed tendency of lower values of Fst between the pioneer species than among the secondary is a reflection of the life history of these species. The pioneer species represent the first successional stage of the forest, being able to invade areas not yet occupied, including harsh environments and forest edges. Thus, given the ecological role they play, the pioneers are heliophyte, experience rapid growth, lower inter-specific competition, increased investment in reproduction (r selection), a higher number of generations per unit time and aggregate distribution. Fragmentation, therefore, does not cause too negative effects on the pioneer species. In fact, while fragmentation reduces the number of individuals of the species of later successional stages and completely alters the environment they occupy, the pioneer species can be instead favored by this process. Also, the pioneer species from the fragments of the Tibagi River Basin have advanced several generations (10 -15 generations) since the beginning of the fragmentation process. So, after several cycles of cross-fertilization, gene flow between populations may have indeed come relatively high to overcome the effects of fragmentation, allowing lower levels of genetic

seen for *P. rigida* due to the occurrence of vegetative propagation.

resulting increase in genetic differentiation between populations.

differentiation among populations.

Unlike the pioneers, the late secondary and climax species tend to have less aggregated spatial distribution, to be ombrophilous and to occupy more specialized niches. These species spend their energy more in the inter-specific competition than in reproduction (k selection) and have slow growth and far fewer generations per unit time. The species of late successional stages, considered here, represent the first or second generation postfragmentation and thus are believed to reflect the immediate consequences of the fragmentation process. Interestingly, one of the secondary species studied, *Aspidosperma polyneuron*, was found in a continuous distribution, in a plateau followed by a high declivity (Damasceno et al., 2011), and the other, *Maythenus aquifolium*, was found in fragments separated by up to 30 km (Sayhun et al., 2010); however, the FST values observed in the two situations were similarly high, suggesting that part of the genetic differences found between the populations of these species also linked to adaptive characteristics (Sayun et al. 2010; Damasceno et al., 2011).

Another important information that we could extract from the obtained results for the sampled populations in these regions of lower, middle and upper Tibagi (Ruas et al. 2011 In press; Ruas et al. 2011; Medri et al. 2011 In press; Oliveira et al, 2011) is that the intense fragmentation towards middle-lower Tibagi has provided a significant increase in inbreeding coefficient (FIS), loss of alleles and reduced genetic diversity compared to populations from the upper Tibagi (Table 1). These factors also influenced the genetic differentiation between populations of the upper Tibagi (where fragmentation is less evident) and the others. The only exception to this result was observed in *A. sellowiana*. When analyzing the species *A. sellowiana* with microsatellite markers we also observed the formation of two distinct groups of populations (Dendrogram using Nei´s genetic distance, 1978), one comprised of populations from the middle Tibagi region and the other pertaining to lower Tibagi (Oliveira, 2011 submitted). Even though *A. sellowiana* is able to perform vegetative reproduction and also its propagules can reach as far as 10m distance from mother tree, we evidenced highly significant negative values of FIS indicating excess of heterozygosity in tree of the five populations. When considering the effect of recent genetic bottlenecks in these populations, the values obtained by the software bottleneck indicated that four of the five populations showed significant values of heterozygosity excess, when considering the Infinite Allele Model, and only one population presented significant levels of heterozygosity excess for all tree models (Infinite Allele Model, Two-phase Model and Stepwise Mutation Model). Migration rates calculated for these populations demonstrated that exchange of 30% of migrants between two populations from the middle Tibagi region. Such results demonstrate that some of these populations are suffering more than others, the impact of fragmentation and also founder effect and gene flow are playing a key role in the shaping of the genetic variability within these populations of *A. sellowiana* promoting a weak balance between the evolutionary forces of genetic drift and migration in this post fragmentation period landscape.

Lately, this research group has been engaged in the development and inter-specific transference of microsatellite primers (SSR) for the species *L. divaricata* (Ruas et al., 2009), *A. sellowiana* (Ruas et al., 2010), *P. rigida*, *A. polyneuron*, *C. xanthocarpa*, *C. gonocarpum* and *A. polyneuron*. The use of SSR markers will allow more accurate estimates about the distribution of genetic variation and the effects of fragmentation and local adaptation between the species of the Tibagi River basin. Such is possible due to the codominant nature of these markers, allowing to access allelic information and thus to estimate the number of alleles, heterozygosity, inbreeding and gene flow among natural populations. Although much work is yet to be done with these and other species of the Tibagi River basin, the present studies certainly represent a great step for the comprehension of the present availability of genetic resources and of its relation to the life history of species from forests fragments pertaining to the Tibagi River basin.
