Diet and food resource partition among benthic tadpoles of three anuran species in Atlantic Forest tropical streams

Abstract

The study of the diet of tadpoles can provide important information about the maintenance and dynamics of populations and communities of tropical amphibians. Tadpoles are in general herbivores and consume mostly algae. Despite this, little is known about how specific guilds use resources that are available in the same space and at the same time. It is thought that species of tadpoles segregate based on food use, especially in lotic environments where food resources may be a limiting factor. Conversely, it is believed that the use of food resources is similar among tadpoles with the same ecomorphotype. This study examined the diet of Aplastodiscus cavicola, Aplastodiscus sibilatus and Bokermannohyla capra syntopic tadpoles, as similarities may exist between these benthic species, all of which are adapted to the flow of water. This study was undertaken in the Michelin Ecological Reserve, municipality of Igrapiúna, state of Bahia, Brazil. To analyze the diet of the tadpoles, nonmetric multidimensional scaling (NMDS) was performed to evaluate the similarity between the diets of the species. Analysis of similarity (ANOSIM) was carried out to detect whether diet differed among tadpole species. The diet of these tadpoles consists mainly of algae, especially diatoms, plants, protozoa, fungi and microscopic animals. The NMDS identified three clusters representing the three species studied. ANOSIM detected significant differences among the diets of the species, which indicates partition of food resources.

Keywords:

• feeding guilds
• amphibian
• Anura
• lotic environment
• coexistence
• niche overlap

Introduction

The coevolution of sympatric species and the adaptation of these species to the environment can reduce overlap in the use of available resources, thereby determining species abundance and diversity in a community. Resource allocation, therefore, is important for the coexistence of two or more species and spatial and temporal dimensions, and food availability, are considered the most important factors for specific niche delineation (Pianka 1974; Schoener 1974). For tropical frogs, spatial and temporal dimensions have traditionally been considered the most important factors, particularly the choice of breeding sites (Crump 1971; Toft 1985), although some studies have identified food resource partition among adults of sympatric and phylogenetically closely related species (e.g., Menin et al. 2005).

The temporal dimension can be decisive, especially for tadpoles, since bodies of water are often subject to periodic droughts. Additionally, competition for food can be fierce when many species use the same body of water, leading to resource allocation. Tadpoles are considered specialized filter feeder herbivores (Duellman & Trueb 1994), as they consume a large variety and quantity of algae, especially in the early stages of their development (Wickramasinghe et al. 2007). However, tadpoles can also consume protozoa, fungi and microscopic animals available in their habitat, which characterizes them additionally as omnivores (Altig & Johnston 1989).

This scenario suggests that in a given community of tadpoles, food resource allocation may be detected indirectly through the different ecomorphotypes (sensu Altig & Johnston 1989) of these organisms. Each ecomorphotype is associated with a specific microhabitat and utilizes different food resources, which are also associated, therefore, with the respective microhabitat (Rossa-Feres et al. 2004).

The benthic tadpole is one of the most representative ecomorphotypes of tropical environments and is found in both lentic and lotic ecosystems (McDiarmid & Altig 1999). However, lotic environments can undergo more drastic natural perturbations than lentic environments (e.g., flood pulses, Resh et al. 1988), resulting in lower algae diversity and primary production (Rodrigues & Bicudo 2001; Cardinale et al. 2005) and consequently affecting food availability for tadpoles. Therefore, at least two challenges are faced by tadpoles in lotic environments: low availability of food resources and water flow.

With some exceptions (e.g., Kupferberg 1997; Dutra & Callisto 2005; Sousa Filho et al. 2007), there are relatively few studies of the diet of anuran larvae in lotic environments. Studies of the composition of the diet of tadpoles are especially important in the tropics, where despite taxonomic diversity being severely threatened by factors that can cause irreparable declines; little is known of the trophic importance of tadpoles in many aquatic environments (Altig et al. 2007). This study investigated the allocation of food resources among benthic tadpoles of three hylid species (Aplastodiscus cavicola, Aplastodiscus sibilatus and Bokermannohyla capra) from lotic communities based on the hypotheses that similar ecomorphotypes use similar food resources and that species that coexist in the same micro-habitat must share food sources.

Methods

Tadpole collection

Tadpoles were collected in December 2008, February, April, July and September 2009, and January 2010 from the Michelin Ecological Reserve (MER) (13°50′ S, 39° 10′ W), located in the municipality of Igrapiúna, in south-eastern Bahia, Brazil. The air temperature in the region varies between 21°C and 30°C and annual rainfall is approximately 2000 mm (Camurugi et al. 2010).

The MER includes two fragments of Atlantic rainforest and a variety of natural ecosystems. Both fragments are surrounded by agroforestry systems, including areas of rubber silviculture governed by two different understory management systems: overall management, in which the understory is totally removed, and partial management, in which only the understory of the main access route to the rubber trails is removed. The understory of the stream margins has been maintained. Several springs, rivers and streams run through the reserve.

Tadpoles of 12 anuran species (Allobates olfersioides, A. cavicola, Aplastodiscus ibirapitanga, A. sibilatus, B. capra, Dendropsopus haddadi, Phasmahyla timbo, Proceratophrys renalis, Proceratophrys schirchi, Rhinella hoogmoedi, Scinax strigilatus and Vitreorana spp.) were found in the streams. While some of these species are benthic, only A. cavicola, A. sibilatus and B. capra are of similar size and possess similar morphological characteristics, some of which are closely related to swimming in flowing water (visible lateral line, narrow fins, long tail and developed caudal musculature) (Mercês & Juncá 2010; Mercês et al. 2011). These three species are widely distributed among the streams of MER and present for large parts of the year. The analyzed tadpoles were taken from four streams of one of the Atlantic rainforest fragments [forest stream, FS: FS1 (13° 50′ 29.8″ S 39° 15′ 2.2″ W), FS2 (13° 50′ 24.8″ S 39° 15′ 0.8″ W), FS3 (13° 50′ 28″ S 39° 14′ 28.3″ W) and FS4 (13° 51′ 3.2″ S 39° 13′ 47.4″ W)] and three streams located in an area of partially managed rubber silviculture adjacent to the forest fragment [rubber plantation stream, RPS: RPS1 (13° 52′ 12.5″ S 39° 14′ 3.2″ W), RPS2 (13° 52′ 36.4″ S 39° 14′ 10.8″ W) and RPS3 (13° 53′ 0.3″ S 39° 14′ 31.9″ W)], where the understory had been partially removed. Various types of nets with different mesh sizes and thicknesses (1, 2 and 3 mm²) were used to capture the individuals. After collection, the tadpoles were anesthetized, immediately killed in 10% alcohol, and preserved in 6% formaldehyde in order to conserve food items.

A. cavicola and B. capra tadpoles were found in both FS and RPS streams, while A. sibilatus tadpoles were found only in FS streams. The diets of 20 tadpoles of A. cavicola (FS = 9, RPS = 11), 10 tadpoles of A. sibilatus (FS) and 17 tadpoles of B. capra (FS = 11, RPS = 6) were analyzed. All tadpoles were from developmental stages 26 to 28 (Gosner 1960). The occurrence of tadpoles from two or three of the species studied was recorded for all streams sampled. However, tadpoles of more than one species from development stages 26 to 28 occurred in only four streams: tadpoles of three species were found in FS2, B. capra and A. sibilatus tadpoles were found in FS1 and FS4, and B. capra and A. cavicola tadpoles were found in FS3 and RPS1. In addition to these streams, tadpoles of A. cavicola in development stages of 26–28 were found in RP2 and RP3.

Diet analysis

The intestine of each tadpole was removed through an incision from the cloaca to the oral disk. Each intestine was then opened in a Petri plate and its contents were removed by longitudinal cuts. After the intestinal contents were removed, 1.5 ml of 4% formaldehyde was added and the mixture was placed into a microtubule. Approximately 0.05 ml of this solution was analyzed on 75 × 24 mm glass slides and 50 × 20 mm coverslips under a light microscope (Leica DMLS2) at 10× magnification and an Optovar 2×. When necessary, 40× lens magnification was used. The food items were identified to the lowest possible taxonomic level and quantified. The identification of the items was based on examination of specialized bibliographies. The items ingested by tadpoles were quantified based on numerical frequency (%NF) and frequency of occurrence (%FO). The importance index was obtained by calculating the %NF plus the %FO divided by two, in accordance with Colli et al. (2003). The collected specimens were deposited in the Museu de Zoologia da Universidade Estadual de Feira de Santana.

The nonmetric multidimensional scaling method (NMDS) was performed to evaluate the similarity between the diets of tadpoles of the three species under analysis, using a matrix of relative abundance of the food items taken from each individual (relative abundance of each matrix cell = number of items of food category y consumed by individual a divided by total number of items consumed by individual a). The Sorensen index was used as a distance measure and two dimensions were selected. In accordance with Xavier and Napoli (2011), it was performed with 50 starting configurations, the instability value of 0.0005 for the stability criterion, 100 iterations to evaluate the stability of the solution and 500 as the maximum number of iterations. The Monte Carlo test was used to evaluate whether NMDS extracted a stronger axis than expected by chance. Analysis of similarity (ANOSIM) was then performed in order to detect whether the diet of tadpoles differed among species, using 10,000 permutations and the Bray-Curtis similarity index. The significance level was 0.05, adjusted according to the Bonferroni method (Tabachnick & Fidell 2007). NMDS and ANOSIM were performed using the PCord 4.1 (McCune & Mefford, 1999) and Past 2.17C (Hammer et al. 2001) softwares, respectively.

Results

The diet of the three species was primarily based on algae, protozoa and plant fragments. A total of 898 sample items of the diet of A. sibilatus tadpoles were collected, distributed in 19 food categories. Samples of the diet of A. cavicola tadpoles totaled 1958 items distributed into 20 food categories, while samples of the diet of B. capra tadpoles totaled 2815 items distributed among 24 categories (Table 1). There were 16 food categories that occurred in the diet of the three species. Based on the importance index scores of over 50, Bacillariophyceae, Trinema and Spermatophyta fragments were important food items for the three species. Other food categories were important for tadpoles of only two species, although consumed by tadpoles of the three species. Oscillatoria was important for the tadpoles of two species of Aplastodiscus. Plancktonlyngbya and Arcela were important categories for tadpoles of A. sibilatus, although in tadpoles of A. cavicola, they scored 47.8 and 46.6, respectively. Nebela was important for the diet of A. cavicola tadpoles and fungi were important for the diets of A. sibilatus and B. capra.

Table 1. Items found in the diet of Aplastodiscus cavicola, Aplastodiscus sibilatus and Bokermannohyla capra tadpoles collected from Atlantic Forest streams of the MER, and rubber silviculture streams. %NF – numerical frequency, %FO – frequency of occurrence, I – importance index, UND – undetermined.

	A. cavicola			A. sibilatus			B. capra
Food items	%NF	%FO	I	%NF	%FO	I	%NF	%FO	I
Bacillariophyceae	24.4	100	62.2	17.3	100	58.6	15.77	100	57.89
Cianophyceae
Oscillatoria	4.8	100	52.4	15.1	100	57.6	0.32	5.88	3.10
Plancktonlyngbya	20.7	75	47.8	15.4	100	57.7	5.19	64.71	34.95
Geitlerinema	–	–	–	–	–	–	0.21	5.88	3.05
Limnothrix	–	–	–	–	–	–	0.64	23.53	12.08
Spirulina	1.6	55	28.3	0.7	50	25.3	0.78	35.29	18.04
Filament UND	–	–	–	–	–	–	17.90	88.13	53.07
Euglenophyceae
Euglena	1.3	50	25.7	–	–	–	0.82	52.94	26.88
Trachelomonas	1.9	79	3.8	0.7	40	20.3	2.45	88.23	45.34
Chlorophyceae
Filament UND	0.4	20	10.2	0.1	20	10.1	2.98	94.12	48.55
Oedogonium	0.4	15	7.7	0.9	40	20.4	–	–	–
Spermatophyta (Fragments)	13.1	100	56.6	19.2	100	59.6	17.83	100	58.92
Cyclea
Arcela	3.1	90	46.6	6.1	100	53.3	0.53	70.59	35.56
Assulina	0.6	35	17.8	0.1	10	10.1	0.14	17.65	8.89
Centropix	0.6	30	15.3	0.1	20	10.1	0.32	29.41	14.87
Diffugia	1.2	40	20.6	1.8	40	20.9	3.97	74.47	39.83
Didinium	1.1	50	25.5	1.6	60	30.8	–	–	–
Nebela	4.9	100	52.4	4.6	80	42.3	1.38	58.82	30.10
Quadrullela	0.7	40	20.3	0.6	40	20.3	0.25	23.53	11.89
Filoreticulosea
Euglypha	3.8	75	39.4	5	60	32.5	3.73	82.35	43.04
Trinema	11.4	100	55.7	4.7	100	52.3	7.57	100	53.78
Fungi	2.1	40	21.0	6	100	53.3	15.13	94.12	54.62
Acari	–	–	–	–	–	–	0.11	17.65	8.88
Rotifera	0.1	5.0	2.5	0.2	10	5.1	0.25	17.65	8.95
Crustacea	–	–	–	–	–	–	0.07	11.76	5.92
Animal UND	–	–	–	–	–	–	0.78	47.06	23.92

Based on the tadpole diets, the NMDS arranged the sample into three clusters representing the three studied species (Figure 1). However, there was an overlap between the two clusters that represented the diets of the two Aplastodiscus species. The NMDS stress was 26.0 and the Monte Carlo test indicated that the NMDS result was stronger than expected by chance (p = 0.002).

Figure 1. Clusters formed from NMDS analysis of the diet of tadpoles studied in the MER. Triangle – Bokermannohyla capra, circle – Aplastodiscus sibilatus, square – Aplastodiscus cavicola.

ANOSIM confirmed that there were significant differences among the diets of the three species of tadpoles (r = 0.536, p < 0.0001).

Discussion

Food items found in the diet of tadpoles of A. cavicola, A. sibilatus and B. capra were similar to those of other tadpoles, with their diets consisting primarily of algae, plants, protozoa (e.g., Inger 1986; Sekar 1992), fungi and microscopic animals (e.g. Rossa-Feres et al. 2004; Sousa Filho et al. 2007).

A high diversity and abundance of algae is usual in tadpoles of different anuran species (Dutra & Callisto 2005; Sousa Filho et al. 2007). However, the ingestion of a large amount of algae by anuran larvae will not necessarily result in nutritional gain (Pryor 2003; Akers et al. 2008). In contrast, at least one experimental study has found that tadpoles that ingested algae associated with diatoms showed increased growth and rapid development (Kupferberg et al. 1994). Diatoms are one of the items of major importance in the tadpole diet of the three species in this study, and in the diet of all benthic tadpoles (e.g., Rossa-Feres et al. 2004).

While protozoa diversity was relevant in the diet of the three species of tadpoles in this study (nine genera of protozoa were found), some of which had a high importance index (e.g., Trinema); in general, the importance of these organisms was low compared with other taxa (e.g., Bacillariophyceae, Fungi). Studies that address the importance of protozoa in the diet of tadpoles are rare (e.g., Nathan & James 1972) and few taxa of the group have been documented in the diet of amphibian tadpoles (Arias et al. 2002; Rossa-Feres et al. 2004).

Fungi were an important item in the diet of B. capra and A. sibilatus tadpoles. The presence of fungi in the diet of tadpoles is also rarely reported in literature even though it is an important component in oligotrophic habitats (Altig et al. 2007). Studies that address the importance of fungi in the diet of tadpoles were not found, however, Rossa-Feres et al. (2004) identified mycophagous eating habits in Leptodactylus fuscus.

Food items such as rotifers, crustaceans and mites found in the diet of tadpoles in this study have been reported for other species (e.g., Candioti 2005; Dutra & Callisto 2005; Heinen & Abdella 2005; Pfennig et al. 2006; Sousa Filho et al. 2007; Wickramasinghe et al. 2007), and were usually associated with the diet of tadpoles in the later stages of development, unlike the three species of this study, which were in the early stages of development. The high nutritional value of invertebrates in terms of protein and energy is worthy of note (Bowen et al. 1995).

A. sibilatus, A. cavicola (Mercês & Juncá 2010) and B. capra (Mercês et al. 2011) can be considered benthic tadpoles (Altig & Johnston 1989), a guild characterized by Wells (2007) as having general eating habits. The results of this study confirmed similarity in eating habits but also illustrated differences according to NMDS analysis, which could be the result of resource allocation. This study confirmed the hypothesis that there was allocation of available food sources by these species, which coexist in the same micro-habitat. One must consider, however, that differences in food supply may emerge. Besides, in some streams, tadpoles from developmental stages 26 to 28 (sensu Gosner 1960) from only one species were found. Nevertheless, considering the occurrence in streams of tadpoles in stages 26–28, there was more overlap between B. capra and A. sibilatus (two streams) and B. capra and A. cavicola (two streams), than A. sibilatus and A. cavicola (one stream). Despite this, there was considerable similarity between the diets of the two Aplastodicus species, while the greatest difference in diets was between these species and B. capra tadpoles, indicating the possible importance of the phylogenetic evolution of food habits.

Acknowledgements

The authors would like to thank the Michelin Center for Biodiversity and the UEFS Post-graduate Zoology Program for logistical support, Ivania Batista for help in identifying food items, Ednei de Almeida for help in the field and laboratory work, Arielson Protázio for his review of the manuscript.

Disclosure statement

No potential conflict of interest was reported by the authors.

Funding

This work was supported by the Michelin Center for Biodiversity; the Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (CAPES) for the scholarship awarded to ASP e FJMS; the Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq) for the scholarship awarded to FAJ [305457/2009-8]; Fundação de Amparo a Pesquisa do Estado de São Paulo [2010/52321-7]; Conselho Nacional de Desenvolvimento Científico e Tecnológico [563075/2010-4].