Species composition , richness and nestedness of lizard assemblages from Restinga habitats along the brazilian coast

Habitat fragmentation is well known to adversely affect species living in the remaining, relatively isolated, habitat patches, especially for those having small range size and low density. This negative effect has been critical in coastal resting habitats. We analysed the lizard composition and richness of restinga habitats in 16 restinga habitats encompassing three Brazilian states (Rio de Janeiro, Espírito Santo and Bahia) and more than 1500km of the Brazilian coast in order to evaluate if the loss of lizard species following habitat reduction occur in a nested pattern or at random, using the “Nestedness Temperature Calculator” to analyse the distribution pattern of lizard species among the restingas studied. We also estimated the potential capacity that each restinga has to maintain lizard species. Eleven lizard species were recorded in the restingas, although not all species occurred in all areas. The restinga with the richest lizard fauna was Guriri (eight species) whereas the restinga with the lowest richness was Praia do Sul (located at Ilha Grande, a large coastal island). Among the restingas analysed, Jurubatiba, Guriri, Maricá and Praia das Neves, were the most hospitable for lizards. The matrix community temperature of the lizard assemblages was 20.49° (= P <0.00001; 5000 randomisations; randomisation temperature = 51.45° ± 7.18° SD), indicating that lizard assemblages in the coastal restingas exhibited a considerable nested structure. The degree in which an area is hospitable for different assemblages could be used to suggest those with greater value of conservation. We concluded that lizard assemblages in coastal restingas occur at a considerable level of ordination in restinga habitats and that some restinga areas such as Jurubatiba, Guriri, Maricá and Praia das Neves are quite important to preserve lizard diversity of restinga environments.


Introduction
Habitat fragmentation is well known to adversely affect species living in the remaining, relatively isolated, habitat patches, especially for those having small range size and low density (Foster, 1980;Terborgh and Winter, 1983;Fahrig, 2003).Because habitat fragmentation may differentially affect the species within an assemblage, the loss of species does not necessarily occur at random, but may occur in a nested pattern (Atmar andPatterson, 1993, 1995).Assemblages are nested when species present at species-poor sites are non-random subsets of the species present in species-rich areas (Cutler, 1991;Atmar and Patterson, 1993;Maron et al., 2004;Fischer and Lindenmayer, 2005a, b).Different factors including extinction, colonisation, habitat, and niche structure may cause these nested patterns, and some degree of nestedness appears to be the rule in nature (Wright et al., 1998;Patterson and Atmar, 2000).
The eastern coast of Brazil has been intensively occupied and modified by humans in the last five centuries and is currently the most densely populated region of the country (Rocha et al., 2003).Most coastal areas in Brazil are dominated by sand dune habitats covered with herbaceous and shrubby vegetation, known as "restingas" (Suguio and Tessler, 1984).The restingas are part of the Atlantic rainforest biome (Eiten, 1992), which is one of the so-called biodiversity hotspots (Mittermeier et al., 2005).Many restingas were naturally isolated from each other as a result of topography (Suguio and Tessler, 1984).Additionally, as a consequence of the disturbances continuously imposed to restinga habitats, most of their original area has been degraded and a large portion of it has been lost (Rocha et al., 2007).Presently, most of the eastern Brazilian coastal is composed of remnants of restinga habitats separated by human-dominated landscapes (Rocha et al., 2007).Different lizards assemblages are found in the remaining restingas (Rocha, 2000) and some of such fragments contain an important set of endemic and threatened vertebrate species (Rocha et al., 2005).
In this study we analysed the lizard composition and richness of restinga habitats in order to evaluate if the loss of lizard species following habitat reduction occur in a nested pattern or at random.We also estimated the potential capacity that each restinga has to maintain lizard species (that is, the hospitality of each area, sensu Patterson and Atmar , 2000).

Study Areas and Methods
We intensively sampled and recorded lizards in 16 restinga habitats encompassing three Brazilian states (Rio de Janeiro, Espírito Santo and Bahia) and more than 1500km of the Brazilian coast (Figure 1, Table 1).Restingas are Quaternary sand dune habitats characterised by sparse, mostly xerophilous shrubby and herbaceous vegetation (Suguio and Tessler, 1984).These habitats vary both floristically (Araújo and Henriques, 1984) and structurally (Rocha and Bergallo, 1997).Estimates in Nova Viçosa, Itaúnas, Linhares, Pontal do Ipiranga, Marataízes, Neves (Espírito Santo state), and Grumari (Rio de Janeiro state) were made from December to January 1996 whereas Trancoso, Prado (Bahia state), Guriri, Setiba (Espírito Santo state), Grussaí, Macaé, Maricá, Massambaba and Praia do Sul (Rio de Janeiro state) were sampled from November 1999 to February 2000.The general geographical location of each area was recorded using a Garmin II GPS set to WGS-84 datum.
We intensively sampled each restinga during four consecutive days through Time-constrained Visual Encounter Surveys (Crump and Scott Junior, 1994), keeping sampling effort equivalent (30 min transectation within each one hour interval from dusk in the morning to dusk in the night, totaling about six and half hour of search/day).Searches were undertaken from 06:00 to 19:00h which includes the peaks of activity of most lizards of restingas (Araújo, 1984;Rocha, 1988;Bergallo and Rocha, 1993;Teixeira-Filho et al., 1995;Vrcibradic and Rocha, 1996;Zaluar and Rocha, 2000;Hatano et al., 2001) and the start of activity of the only nocturnal species in the restingasthe invasive Hemidactylus mabouia (Moreaú de Jonnés, 1818) (Rocha et al., 2002).Weather conditions during sampling were as similar as possible.During sampling we carefully searched for lizards in all different potential microhabitats, such as the border and interior of bushes, on trees and shrubs, on the leaf litter and on bromeliads.In order to reduce differences due to vegetation structure, we searched for lizards in areas similar in physiognomy, characterised by sandy soil covered predominantly with thickets of shrubs and herbaceous vegetation at similar distances from the sea (approximately 200 m).Linhares was an exception because it is an isolated patch of restinga located about 12 km inland from the coast (Rocha et al., 1997).
We used the "Nestedness Temperature Calculator" of Atmar andPatterson (1993, 1995) to analyse the distribution pattern of lizard species among the restingas.This method analyses the matrix of presence and absence of species and calculates its "temperature", which indicates the level of order or disorder of the matrix.The lower the temperature of the matrix (level of nestedness of the system), the lower is its disorder and less stochastic is the distribution of the species in the studied assemblage.For this analysis we used 5000 randomisations.We also evaluated the capacity that each restinga has to maintain lizard species ("hospitality"), and potential areas to be preserved (Atmar andPatterson, 1993, 1995).
The matrix community temperature of the lizard assemblages was 20.49° (= P <0.00001; 5000 randomisations; randomisation temperature = 51.45 ± 7.18° SD), indicating that lizard assemblages in the coastal restingas exhibited a considerable nested structure.In general, nested assemblages occur as a result of processes such as selective immigration or extinction, habitat nestedness and fragmentation (Fischer and Lindenmayer, 2005a, b).In the case of coastal restingas, the nested pattern of lizard assemblages may result in part from the degradation and fragmentation of restingas along the coast which led to the increasing isolation of these habitats (Rocha et al., 2007) and in part from the distinct geological histories of each area.However, our dataset does not allow distinguishing between these two putative causes for the nested pattern found.Nevertheless, the frogs assemblages in these restinga habitats were also nested (Rocha et al., 2008).Fragmented systems tend to exhibit distinctive patterns of species richness and species composition (Atmar and Patterson, 1995;Patterson and Atmar, 2000) and this was the tendency we found with the lizard assemblages.
Among the restingas analysed, Jurubatiba, Guriri, Maricá and Praia das Neves, were the most hospitable for lizards (Table 2).For frogs, Praia das Neves was also considered the most hospitable restinga in the eastern Brazilian coast (Rocha et al., 2008).The degree in which an area is hospitable for different assemblages could be used to suggest those with greater value of conservation.In the case of restinga lizards, the three endemic species were present in the restingas identified as most hospitable.However, at the same time, these areas are still not protected as conservation units, with the exception of Jurubatiba which is the only restinga already protected as a National Park in Brazil.
We concluded that lizard assemblages in coastal restingas occur at a considerable level of ordination in restinga habitats and that some restinga areas such as Jurubatiba, Guriri, Maricá and Praia das Neves are quite important to preserve lizard diversity of Brazilian restinga environments.

Figure 1 .
Figure 1.Map showing the sampling localities at 16 restinga habitats in Rio de Janeiro, Espírito Santo and Bahia states, Brazil.

Table 1 .
Restingas sampled for lizard species in 16 restinga areas in the states of Rio de Janeiro, Espirito Santo and Bahia, eastern Brazilian coast.Coordinates based on datum WGS-84.