Introduction
Studies comparing islands and mainland communities have been mostly conducted using organisms such as plants and animals (e.g. mammals, birds and reptiles) for understanding insular patterns of biodiversity (Brown, Reference Brown1978; Mueller-Dombois, Reference Mueller-Dombois2001; Helmus et al., Reference Helmus, Mahler and Losos2014; Medina et al., Reference Medina, Vásquez, Moreno and Torres-González2015). These studies provide insights into the biogeography process where an island can act as a natural laboratory due to its isolation and dynamic (Lomolino, Reference Lomolino2000a).
According to the biogeography of islands theory, the insular dynamic is characterized by its reduced area, geographic isolation and age, which select species and diversity (MacArthur & Wilson, Reference MacArthur and Wilson1967). The equilibrium of these populations is adapted and efficient in their environment, but they are very sensitive due to isolation and reduced size, which can induce some extinctions (MacArthur & Wilson, Reference MacArthur and Wilson1967; Vitousek et al., Reference Vitousek, Loope and Andersen1995). In light of this theory, some studies (e.g. Lomolino, Reference Lomolino2000b; Losos & Schluter, Reference Losos and Schluter2000) have corroborated that small areas (such as islands) harbour low richness when compared to large areas (such as mainland), while other studies have found the opposite or no substantial effect in the relation of size to richness (e.g. Dunn & Loehle, Reference Dunn and Loehle1988; Lomolino, Reference Lomolino2000b).
Although insular research that use attributes of parasites are still scarce, some studies have demonstrated that prevalence and intensity of infection might be higher on islands, with the highest infection values on the smallest island (Casanova et al., Reference Casanova, Miquel and Fons1996; Miquel et al., Reference Miquel, Casanova, Fons, Feliu, Marchand, Torres and Clara1996) when compared to those of the mainland (Lewis, Reference Lewis1968; Gregory & Munday, Reference Gregory and Munday1976). Thus, more studies are needed on the ecology of parasites, which investigate the response between the effects of environmental and host features (Poulin, Reference Poulin2004). The parasites represent a hidden diversity (Poulin & Morand, Reference Poulin and Morand2000), and some research seeks to understand how this diversity and other parasitological parameters can be influenced by environmental conditions (Poulin, Reference Poulin2004). For example, host characters such as feeding, body size, sex, age, evolutionary history and the type of environment where the host lives can influence the parasite communities (Aho, Reference Aho, Esch, Bush and Aho1990; Barton, Reference Barton1999; Muzzall et al., Reference Muzzall, Gillilland, Summer and Mehene2001; Bolek & Coggins, Reference Bolek and Coggins2003; Brooks et al., Reference Brooks, León-Règagnon, McLennan and Zelmer2006; Hamann et al., Reference Hamann, Gonzalez and Kehr2006; Campião et al., Reference Campião, Ribas, Morais, Silva and Tavares2015).
In this sense, amphibians are considered as good models for studying parasite communities due to their lifestyle, inhabiting both aquatic and terrestrial ecosystems, which allows contact with different parasite species. Numerous reproductive modes can be highlighted by the great ability of these vertebrates to exploit several types of microhabitats, especially those from the Atlantic Rain Forest (Myers et al., Reference Myers, Mittermeier, Mittermeier, Fonseca and Kent2000; Haddad & Prado, Reference Haddad and Prado2005). Haddadus binotatus ((Spix, 1824() (Anura, Craugastoridae) is an endemic frog from the Brazilian Atlantic Forest biome. It lives on the forest floor and its distribution is from southern Bahia to Rio Grande do Sul States, Brazil (Frost, Reference Frost2020). This anuran species has an average size, with the female (46–63 mm) generally larger than the male (32–44 mm) (Heyer et al., Reference Heyer, Rand, Cruz, Peixoto and Nelson1990; Canedo & Rickli, Reference Canedo and Rickli2006). In the breeding season, the eggs are released as foam on land, and there is no tadpole stage in the development of these frogs (Hedges et al., Reference Hedges, Duellman and Heinicke2008; Frost, Reference Frost2020). Furthermore, they may prey on a variety of items of different sizes, representing an important component of the local food web as well as getting parasites by trophic transmission (Coco et al., Reference Coco, Júnior, Fusinatto, Kiefer, Oliveira, Araujo, Costa, Sluys and Rocha2014). Concerning the records of parasites, there are only two studies reporting the helminth fauna associated with H. binotatus (Travassos, Reference Travassos1925; Aguiar et al., Reference Aguiar, Morais, Cicchi and Silva2014), but no study has compared parasite communities between host populations from mainland and islands.
Several Brazilian islands of the Atlantic Forest, such as those from São Paulo State, can present similar characteristics to the mainland due to their continental origin, which occurred in the last glaciation during the Pleistocene (Martin et al., Reference Martin, Mörner, Flexor and Suguio1986). These characteristics of the continental island can result in similar fauna and phytophysiognomies as those of the continent, allowing comparisons between patterns of abundance and biodiversity from each locality.
Here, considering the ecology of insular environments and the features of parasites and the host species, we predicted that the type of environment (insular or continental) can influence the richness and abundance of parasites associated with the frog H. binotatus. To test this effect, we predicted that: (1) insular environments will be more similar to each other in respect to parasite community, as well as continental forest fragments sharing more similarities; (2) the fragments from the mainland will have higher species richness and diversity of parasites due to their wide area compared to the limited area of islands; and (3) the islands will present a higher abundance of parasites due to their limited area, which can promote more aggregation of the insular hosts and, thus, more reinfection and higher intensity of infection. Based on these predictions, we aimed to describe an effect of the environment (i.e. insular and continental) in the richness, diversity and abundance of the parasite community associated with populations of H. binotatus collected in continental and insular regions from the Atlantic Forest, south-eastern Brazil.
Materials and methods
Study areas and collection of amphibian hosts
Eighty-six specimens of H. binotatus were collected using pitfall traps or visual encounter surveys in four localities from the Atlantic Forest during the breeding seasons of the years 2004, 2005, 2006, 2009 and 2010: Núcleo Santa Virginia (NSV), Serra do Mar State Park (n = 23); municipality of São Luis do Paraitinga (SLP) (n = 30); Anchieta Island (ILA) (n = 9); and Moela Island (ILM) (n = 24). The localities from the continent are NSV (23°24′S, 45°03′W), which is located within a preserved area, and SLP (23°13′S, 45°18′W), which is located in a non-preserved area of a mountainous region of the Serra do Mar, São Paulo State. ILA (23°45′S, 45°033′W), belonging to the municipality of Ubatuba, is an insular protected area (Anchieta Island State Park), despite the intense influence of tourists and its short distance from the north coast of São Paulo State (approximately 800 m) (Guillaoumon et al., Reference Guillaoumon, Marcondes and Negreiros1989; Cicchi et al., Reference Cicchi, Serafim, Sena, Centeno and Jim2009). ILM (24°03′S, 46°16′W) is an insular concession of the Brazilian Navy due to the maintenance of a lighthouse there, and the island is 2.5 km from the southern coast of São Paulo State, in the municipality of Guarujá (fig. 1). Currently, both islands are considered protected areas, despite their past environmental impacts (Secretaria de Estado de Insfraestrutura e Meio Ambiente, 2019).
Fig. 1. Map of the São Paulo coast, indicating the four localities where the anurans Haddadus binotatus were collected. Insular: (1) Moela Island and (2) Anchieta Island. Continental: (3) Núcleo Santa Virginia and (4) São Luis do Paraitinga.
The collected anurans were killed with thiopental sodic and then, after the necropsy and helminth survey were carried out, they were fixed with formaldehyde (10%), preserved in alcohol 70% and deposited at the Museu de Zoologia da Universidade Estadual de Campinas (ZUEC 24874 and 24875) and Coleção Herpetológica da Universidade Regional do Cariri (URCA-H 10688–10700, 10946–10968).
Collection of helminths and procedures in the laboratory
All organs and body cavity were examined for helminths with the aid of a stereomicroscope. Recovered parasites were counted and site of infection registered. Nematodes were fixed using a hot solution with 93 parts of alcohol 70° GL, five parts of formalin, and two parts of acetic acid (AFA), while acanthocephalans were first kept on cold water for the exposition of proboscis and then fixed with AFA solution. Subsequently, all helminths were maintained in labelled bottles with alcohol 70%. In the laboratory, helminths were mounted in temporary slides for observation of taxonomic structures using a computerized system of image analysis (LAS DIC, Leica Microsystems, Wetzlar, Germany). The nematodes were cleared with lactic acid or lactophenol, and acanthocephalans were stained with chloridric carmine, dehydrated with alcoholic series and then cleared with eugenol (Amato et al., Reference Amato, Boeger and Amato1991). Voucher specimens were deposited in the Coleção Helmintológica do Instituto de Biociências de Botucatu (CHIBB 8861–8887) of the Universidade Estadual Paulista/UNESP. The identification of parasite taxa was performed using Yamaguti (Reference Yamaguti1961, Reference Yamaguti1963), Vicente et al. (Reference Vicente, Rodrigues, Gomes and Pinto1991), Anderson et al. (Reference Anderson, Chabaud and Willmott2009), Gibbons (Reference Gibbons2010) and papers with species descriptions. In the morphological analyses, the presence of juveniles (larvae of Nematoda) and the absence of males (e.g. nematodes of Cosmocercidae) did not enable us to make any specific identifications since these helminths present conservative characters. Thus, we separated the cosmocercids into morphospecies based on the main morphological differences.
Data analysis
To characterize the helminth communities of each anuran population among the continents and islands localities, we analysed the prevalence, mean abundance, mean intensity of infection and mean richness according to Bush et al. (Reference Bush, Lafferty, Lotz and Shostak1997). To describe the diversity for each infracommunity, we used the Brillouin diversity index, and then we provided the mean diversity for the component community (bootstrap 95%). We performed a cluster analysis with the Bray–Curtis index (B) to explore potential similarities among helminth communities from each host population, taking into account the mean intensity of infection of each helminth taxa. These descriptive analyses were performed using Sigma-Stat version 3.1 (SigmaStat, 2005) and PAST (Hammer et al., Reference Hammer, Harper and Ryan2001), adopting a significance level at the value of P ≤ 0.05.
Accumulation curves for helminth species were generated to evaluate sampling effectiveness by randomizing samples 1000 times using EstimateS 9.1.0 (Colwell, Reference Colwell2006). According to Poulin (Reference Poulin1998), this estimator is the most invariable and it present reduced susceptibility for overestimation in the real richness, independently of how frequent species appear in the observed community. For constructing curves of observed and estimated richness, we used Statistica software version 7.1 (Statsoft, 2005).
We used generalized linear mixed models to investigate whether richness and abundance of parasites (response variables) are affected by the predictor variables ‘environment’ (insular or continental) and ‘locality’ (NSV, SLP, ILA and ILM) for each individual frog. These predictor variables were also used in linear mixed models to investigate the diversity as response variable. In each model set, we considered the ‘environment’ as a fixed effect and the ‘locality’ as random effect. For richness and abundance data, our models ran with negative binomial distribution using the ‘glmer.nb’ function in the lme4 package (Bates et al., Reference Bates, Maechler, Bolker and Walker2014) in R (R Development Core Team 2020, version 4.0.0). For diversity, a Gaussian distribution was most appropriate for the model, and we used the ‘lmer’ function in the lmerTest package (Kuznetsova et al., Reference Kuznetsova, Brockhoff and Christensen2017) in R (R Development Core Team 2020, version 4.0.0).
Results
A total of 86 specimens of H. binotatus were analysed from the following localities: NSV (n = 23), SLP (n = 30), ILA (n = 9) and ILM (n = 24). The sampling efforts were representative according to richness estimators (fig. 2). The overall richness of helminths (number of taxa) in each locality was as follows: eight (NSV), four (SLP), seven (ILA) and eight (ILM). Helminths identified up to the genus level (e.g. Ochoterenella sp., Oxyascaris sp., Rhabdias sp. and Physaloptera sp.) were considered the same morphospecies in the host populations, whereas Cosmocercidae sp. 1, Cosmocercidae sp. 2, Cosmocercidae sp. 3 and Cosmocercidae sp. 4 were treated as distinct species.
Fig. 2. Sampling efforts in observed (solid line) and expected (dashed line) richness within each community of helminths associated with Haddadus binotatus from ILA (Anchieta Island), ILM (Moela Island), NSV (Núcleo Santa Virginia) and SLP (São Luis do Paraitinga).
In NSV, we recorded 258 helminths associated with 22 of 23 sampled frogs (table 1); the mean diversity for the component community was 0.17 ± 0.05 (0–0.68) and the mean richness was 1.54 ± 0.16 (1–3). In SLP, 170 helminths were recorded in 24 of 30 sampled frogs (table 1). The mean diversity and the mean richness for the component community from SLP were 0.16 ± 0.04 (0–0.49) and 1.42 ± 0.10 (1–2), respectively. In this way, the samples from continental localities (NSV and SLP) numbered ten helminth taxa and shared two species (Oxyascaris sp. and Rhabdias sp.) (table 1).
Table 1. Helminths associated with populations of Haddadus binotatus from Núcleo Santa Virginia (NSV) (n = 23), São Luis do Paraitinga (SLP) (n = 30), Anchieta Island (ILA) (n = 9) and Moela Island (ILM) (n = 24), São Paulo State, Brazil.
NH, number of helminths recorded; P%, prevalence; MA, mean abundance; MII, mean intensity of infection; min–max, range; SE, standard error; SI, sites of infection; Cav, cavity; Lin, large intestine; Lu, lungs; Sin, small intestine; St, stomach.
In ILA, we recorded 695 helminths associated with eight of nine sampled frogs (table 1); the mean diversity for the component community was 0.21 ± 0.07 (0–0.46) and the mean richness was 1.87 ± 0.35 (1–4). In ILM, 435 helminths were recorded in all 24 sampled frogs (table 1). The mean diversity and the mean richness for the component community from ILM were 0.45 ± 0.08 (0–1.27) and 2.29 ± 0.22 (1–5), respectively. Thus, the samples from islands (ILA and ILM) numbered 11 helminth taxa and shared four species (Aplectana pintoi Travassos, Reference Travassos1925, Cosmocerca brasiliensis Travassos, Reference Travassos1925, Physaloptera sp. and Rhabdias sp.) (table 1).
The four populations of anurans shared only Rhabdias sp., whereas the anurans from islands shared four taxa (A. pintoi, C. brasiliensis, Physaloptera sp. and Rhabdias sp.), and the continental ones shared two (Oxyascaris sp. and Rhabdias sp.). In addition to Rhabdias sp., the comparisons NSV–ILM and SLP–ILA both presented Ochoterenella sp. and Oswaldocruzia subauricularis (Rudolphi, 1819), respectively.
A similarity cluster analysis (Bray–Curtis index, co-phenetic coefficient of 0.93), using the mean intensity of infection of each helminth taxa, showed two principal groups regarding the four anuran populations: (1) continental populations (NSV and SLP) (B = 0.06) and (2) the insular population (ILA and ILM) (B = 0.19).
Our models showed a significant influence of the environment on richness and abundance of helminths associated with H. binotatus (table 2), where the islands presented high richness and abundance of helminths. On the other hand, the effects of environment and locality were not significant for helminth diversity (table 2).
Table 2. Effects of environment and study site on response variables (richness, diversity and abundance of helminths associated with Haddadus binotatus from south-eastern Brazil).
Predictor variables included environment with two levels (continent and insular) and the studied sites with four levels (NSV, SLP, ILA and ILM). The study site was considered as random effects in the three models. Akaike information criterion (AIC) values are presented in the models with Negative Binomial distribution.
Discussion
This is the first study to compare helminth fauna of H. binotatus from island and continent localities and to verify a significative influence of environment on richness and abundance of helminth communities. We found high richness and abundance of helminths in hosts from the insular environments studied, and the findings suggested a greater similarity between islands than that found between continental fragments.
We detected eight new records of helminth associated with H. binotatus (Pseudoacanthocephalus lutzi (Hamann, 1891), C. brasiliensis, Cosmocerca parva, A. pintoi, Oxyascaris sp., Ochoterenella sp., O. subauricularis and Spiroxys sp.). As expected, no one of these four anuran populations presented trematodes in the component community, probably due to the direct reproductive mode of this frog, as they do not enter water bodies to spawn (Haddad & Sazima, Reference Haddad, Sazima and Morellato1992; Canedo & Rickli, Reference Canedo and Rickli2006). Consequently, the adult frogs and tadpoles are not infected with aquatic cercariae. Other studies (e.g. Bolek & Coggins, Reference Bolek and Coggins2003; Kehr & Hamann, Reference Kehr and Hamann2003; González & Hamann, Reference González and Hamann2006) have correlated the composition of the parasite community with the habit of hosts, suggesting the predominance of monoxenic nematodes in terrestrial frogs and trematodes in aquatic or semiaquatic amphibians.
For each component community, a dominant species that presented the greatest prevalence and mean intensity of infection was observed (table 1). Haddadus binotatus from ILM presented a dominance by cysthacanths (Acanthocephala), which usually reach frogs by the trophic transmission of the first intermediate host (e.g. an arthropod) (Santos & Amato, Reference Santos and Amato2010). Besides the availability of an arthropod as prey for frogs, acanthocephalans demand a bird as final hosts, as reported in other studies that point out the important role of birds in parasite exchange among ecosystems (e.g. Karvonen & Valtonen, Reference Karvonen and Valtonen2004; Poulin & Leung, Reference Poulin and Leung2011). These parasites could finish their development in sea birds that usually nest or rest on coastal islands (Neves et al., Reference Neves, Vooren, Bugoni, Olmos, Nascimento, Neves, Bugoni and Rossi-Wongtschowski2006). In the other component communities, the nematodes C. parva, Cosmocercidae sp. 3 and C. brasiliensis were dominant in NSV, SLP and ILA, respectively. These nematodes are expected as dominant in this frog species, which occupies a terrestrial habit, thus enabling the contact between infective larvae from the soil and skin and eyes of the host (Anderson, Reference Anderson2000).
Despite the differences in the composition of parasite communities, some parasites were found in more than one host population (e.g. Rhabdias sp., C. brasiliensis and O. subauricularis), indicating a historic relationship between these hosts and parasites, which would have adapted to insular conditions since they were separated from the mainland by the ocean. When the islands were created, during the oscillations in the sea level approximately 11,000 years ago (Pleistocene) (Martin et al., Reference Martin, Mörner, Flexor and Suguio1986; Souza et al., Reference Souza, Suguio, Oliveira and Oliveira2005), insular hosts could have then retained these parasites.
In this context, the similarity in helminth fauna between ILM and ILA (B = 0.19) could also be explained by their continental origin, when they were considered as a continuous mainland. Moreover, comparable conditions imposed by this type of environment can contribute to the likeness on communities from these islands, even though they have a different history of human occupation and differences in the distance from the continent; it is possible that these recent factors were not sufficient to modify completely these two insular communities. In other words, ILM and ILA can share similar factors that modulate parasite communities. The proximity between ILA and the coast does not seem to contribute to the similarity with the communities from continent fragments since only Rhabdias sp. was verified in ILA and NSV, and comparing ILA and SLP only two taxa were shared (Rhabdias sp. and O. subauricularis). These findings corroborate the two groups formed (ILA–ILM and NSV–SLP) by cluster similarity analysis.
ILA and ILM presented the highest mean richness and mean abundance compared to fragments from the continent. We confirmed these findings by generalized linear mixed models, which showed the influence of insularity on the increase of richness and abundance of parasite of H. binotatus. The positive influence of these islands on richness was different from the others studies, which found that insularity affects decreasing parasite richness (e.g. Fromont et al., Reference Fromont, Morvilliers, Artois and Pontier2001; Bellocq et al., Reference Bellocq, Morand and Feliu2002). Some studies assume that high abundance could be an effect of the small competition caused by low richness in insular environments (Dobson, Reference Dobson1985; Fromont et al., Reference Fromont, Morvilliers, Artois and Pontier2001); however, our results do not corroborate with such studies, because in addition to the high richness we found high abundance in islands. In this context, two frameworks of determinants of parasite communities have been considered: the ‘founder effect’ and the ‘epidemiological argument’ (Poulin, Reference Poulin2004). The first derives from biogeography theory, and it assumes that insular hosts possess a subset of parasites, which was the result of a small number of migrants with few parasites species causing species loss during colonization (Miquel et al., Reference Miquel, Casanova, Fons, Feliu, Marchand, Torres and Clara1996; Morand & Guégan, Reference Morand and Guégan2000). In other words, insular hosts should harbour lower parasites richness than those on the mainland. Unlike the ‘founder effect’, our results showed a high richness and abundance in islands, which can probably be explained by the ‘epidemiological argument’. This argument considers high population density due to insularity and other features of the host as factors that increase parasite transmission success among individuals (Morand & Guégan, Reference Morand and Guégan2000; Poulin, Reference Poulin2004). Also, we cannot disregard the possible impact levels of human interference in these areas from the Atlantic Rain Forest as factors which can influence the parasite communities. In the diversity model, we did not verify a significative effect of insular or continental environment. However, considering the mean diversity of helminths, the islands presented the highest numbers, and, in the same sense of richness, this finding does not corroborate with the previous studies (e.g. Fromont et al., Reference Fromont, Morvilliers, Artois and Pontier2001; Bellocq et al., Reference Bellocq, Morand and Feliu2002).
The presence of a helminth species in a locality can depend on the probability of colonization from nearby localities and the habitat's general suitability for parasite establishment (Poulin & Leung, Reference Poulin and Leung2011). In other words, speciation, extinction and dispersion are the three main processes that could influence biological richness observed presently (Ricklefs, Reference Ricklefs1987; Poulin, Reference Poulin1995). This also explains the differences in the richness of helminth communities. According to Poulin (Reference Poulin2004), there is no general rule as a key factor promoting rich parasite faunas; a prediction can be important in one study and unimportant in another. There are several factors related to parasite colonization and extinction, and, for this reason, multivariate approaches are essential for understanding at least part of the contributory determinants of parasites richness (Poulin, Reference Poulin2019).
According to our proposed hypotheses, we confirmed that insular environments are more similar to each other in respect to parasite community, and continental forest fragments can also share more similarities; furthermore, the similarity index was higher between islands than fragments of the mainland. Our findings did not corroborate with the second hypothesis, since we found a significative increase of insular environment in parasite richness and no significant influence of environment on diversity, although the islands showed the highest numbers of mean diversity. On the other hand, the third hypothesis was confirmed, with a significative influence of insularity in abundance.
Thus, our findings fill some knowledge gaps concerning parasite communities of insular host populations with the conclusion that richness and abundance of helminth parasites of the frog H. binotatus are positively affected by the island environment.
Acknowledgements
We would like to acknowledge our colleagues from the laboratory who helped us in our fieldwork. We are grateful to the two anonymous reviewers that provided constructive and useful suggestions. Our special thanks to A.C. Wunderlich and M.R. Lima for their help in the mixed models and interpretations, and R.A. Kalwashita-Ribeiro for map production.
Financial support
Financial support for this study was provided by the Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq) (A.A., MSc scholarship 130583/2011-1). We acknowledge the Fundação de Amparo à Pesquisa do Estado de São Paulo (FAPESP) for the scholarships granted to F.H. Yamada (2011/22603-3, 2014/14298-4), L.A. Anjos (FAPESP 12/20978-2) and for financial support (2008/50417-7, 2008/58180-6, 09/15104-0). L.A.F. da Silva is grateful for Coordenação de Aperfeiçoamento de Pessoal de Nível Superior/CAPES for her scholarship. R.J. da Silva (309125/2017-0) and D. H. Morais (313241/2018-0) are supported by Conselho Nacional de Desenvolvimento Científico e Tecnológico/CNPq.
Conflicts of interest
None.
Ethical standards
The authors assert that all activities contributing to this work comply with the ethical standards in this type of study (SISBIO collection permission number 31716-2; IBAMA permission number 567/05).
