Habitats and trophic relationships of Chironomidae insect larvae from the Sepotuba River basin , Pantanal of Mato Grosso , Brazil

Benthic habitats are linked by physical processes and are essential elements in assessing of the distribution dynamics of Chironomidae dipteran insects and their role in aquatic ecosystems. This work presents results of distribution patterns of chironomids larvae in 38 sites that are abundant in the study site, inhabiting the substrate of the main river channel, rapids, tributary brook, floodplain lakes and reservoir along the Sepotuba River from its mouth at the Paraguay River to the headwater region. A total of 1,247 larvae was registered. The most abundant taxa were Polypedilum (Tripodura) sp. (25.2%), Cricotopus sp.3 (23.0%) and Tanytarsus sp. (15.0%). Fissimentum desiccatum were found only in the reservoir; Fissimentum sp.2 and Tanytarsus cf. T. obiriciae sp.2 in floodplain lakes, and Goeldichironomus sp. in the main channel. The low diversity of the sites S06 and S35 is caused by the near-exclusive presence of the species Cricotopus sp.3, alone or together with one or another taxon (Tanytarsus sp., Djalmabatista sp.3). Collectors-filterers represent 16%, collectors-gatherers 15%, predators 11% and scrapers only 1%. The predators dominated in the secondary channel (±88 ind/m2), corresponding to 40% of the total of this group. Cryptochironomus sp.2 (34%) and Ablasbemyia gr. annulata (26%) were the most abundant among the predators. The differences along the river course are decisive for the formation of distinct or discontinuous communities and the limits become obvious though the interrelations between the populations in the community, as for instance, competition for food and habitats.


Introduction
Continuous loss of habitat in recent decades has enhanced studies on the survey of species richness, in order to prioritise the conservation of areas with high diversity.A definition of biodiversity includes all terrestrial and freshwater organisms -including plants, animals, and microbes -at scales ranging from genetic diversity within populations, to species diversity, to community diversity across landscapes (Sala et al., 2000).In this context, invertebrates are recognised as important components of biodiversity (Kim, 1993;Kremen et al., 1993;Oliver and Beattie, 1996;Yen and Butcher, 1997), because they are important in all ecosystems in terms of species numbers and biomass.
The large extension of Brazilian territory causes that much of its diversity is still unknown.Thus, this work constitutes the first register of the Chironomidae community in the Sepotuba River, an important tributary of the Paraguay River north of the Pantanal.Its discharge contributes to the wave-like inundation patterns so famously transversing the Pantanal from north to south when the rainy season starts and which gives rise to the wetlands ecology.
If the studies about species density and repartition of the diet between members of numerous Chironomidae larvae produce important information about distribution and their function (Trivinho-Strixino and Strixino, 1998;Nessimian et al., 1999), to study a new area will increase the knowledge about the distribution patterns of this group.
We studied the structure of the Chironomidae community in different habitats (main river channel, rapids, tributary brook, backwaters and reservoir) along the Sepotuba River and their trophic relationships, allowing us to detect the main factors that control the distribution patterns of this community.

Study Area
The catchment of the Sepotuba River is situated between 14°20' to 16°00'S and 57°00' to 58°40'W, north of the city of Cáceres.It is part of the Upper Paraguay River basin, which is part of the transitional area between the realm of the Cerrado and of the Amazon forest (Ab'Saber, 1983).In the Pantanal floodplains there are many shallow floodplain lakes and lentic aquatic systems (Da Silva et al., 2001).
The Sepotuba River has as tributaries several blackwater streams, and floodplain-lakes with clear water, which support large quantities of aquatic macrophytes, including Eichhornia azurea, Pistia stratiotes, Salvinia spp.The hydrologic characteristics of the river are similar to those of the Paraguay River.Floods are caused by heavy rains in the upper parts of the basin, and propagate to the region of the Pantanal.
The study area comprises different aquatic systems, defined here as "main river channel", "rapids", "tributary brook", "floodplain lakes", and "reservoir" along the Sepotuba River from its mouth at the Paraguay River to the headwater region.38 sampling sites were chosen along the river (Table 1; Figure 1).

Material and Methods
Collections were carried out with a Petersen grab (0.0189 m 2 ) during February and March 2002 at 38 sampling sites, with the exception of sampling site S32, with three samples taken with a Surber sampler with an area of 0.09 m 2 .At each site four samples were taken, three for biological and one for sediment granulometry and organic matter content analyses.All material was fixed in the field in 70% alcohol.
Concurrently with the collections of biological material, water transparency, pH, electric conductivity (mS/cm), dissolved oxygen (%) and water temperature (°C) were measured.The granulometric texture was determined according to the Wentworth Scale (Wentworth, 1922).Organic matter content was determined from 10 g of sediment sample by incineration at 560°C.
The samples for biological analysis were washed through a series of sieves with smallest mesh size 0.2 mm.In the laboratory, samples were sorted under a stereoscopic microscope.
The Chironomidae larvae were fixed on slides with Euparal, separating the head capsule from the rest of the body.The taxa were identified using the keys of Epler (2001), Coffman andFerrington Junior (2008), andTrivinho-Strixino (2011).The structure of the community was analysed based on the systematic composition, Shannon-Wiener diversity index (H') and Margalef's index of richness (Magurran, 2004).
The mounted slides were used to analyse the food items of the digestive tract, using the keys Bicudo and Bicudo (1970) and Parra and Bicudo (1995) and the larvae were fixed on slides with Euparal.
According to Baselga (2010), we calculated three measures of beta diversity, the Sorensen dissimilarity index (b SOR ), the Simpson dissimilarity index (β SIM ) and the complementary of b SIM was called residuals, by a routine that we wrote in R (R Foundation for Statistical Computing, 2011).The total beta diversity is represented by b SOR and b SIM is the species turnover part of beta diversity.The formulae that we used are as follows: where a is the number of species in both cells, b is the number of species exclusive to the focal cell and c is the number of species exclusive to the adjacent cell.
In order to evaluate the groups based on biological composition we calculated a cluster analysis by PC-Ord software (McCune & Mefford, 1999).

Abiotic variables
The water depth of the main channel varied from 0.2 m to 10.0 m, while the dissolved oxygen varied between 19.5% and 115.2%, and the pH between 5.00 and 7.47 (Table 4).In sampling sites with low water depth (0.3 m) the oxygen saturation reached its maximum.At the site with the greatest water depth (10 m), the saturation was 72.1% and the pH = 6.06.
Abiotic factors of 17 sampling sites in the main channel of the Sepotuba River showed some significant  Shredder-herbivore, Collector-gatherer correlations among each other (correlation coefficient of Spearman).There was a negative correlation between dissolved oxygen and water depth (r=-0.506,p=0.038) and between dissolved oxygen and pH (r=0.619,p=0.008).
Another correlation in the main channel was recorded between the transparency and conductivity of the water (r=-0.573,p=0.016).However, at a few sites of this habitat the lowest values of transparency (0.55 to 1.60 m) coincided with greater values of the conductivity (45.3 to 25.5mS.cm - ).Not surprisingly, in the substrate higher organic matter content was significantly correlated with higher amount of mud (r=0.686,p=0.002).
At six sampling sites of the stream, the depth was positively correlated with the water transparency (r =0.882, p=0.019), and negatively with the presence of pebbles (r= -0.882, p=0.019) and of very coarse sand (r= -0.819, p= 0.020).A greater number of correlations also appeared between the sediment types, as gravel and very fine sand (r= -0.941, p=0.005), gravel and medium sand (r=0.880,p=0.020), and medium sand and very fine sand (r=-0.885,p= 0.018).

Diversity index and species richness
Among the sites the diversity had a characteristic variation (Figure 3B), with the highest diversity at the following sites: main channel S05 (0.86), floodplain lake S09 (0.85), streams S22 (0.83) and S26 (0.80).P. (Tripodura), Cryptochironomus sp.2, Tanytarsus sp., and A. gr.Annulata were recorded in sites with higher diversity.Sites with zero diversity were S06 and S35, where only the species Djalmabatista sp.3 and Cricotopus sp.3, were recorded.The indices of diversity showed major values for sites in the main channel (0.98 and 0.71), and floodplain lakes (0.89 and 0.74) (Table 5).However, the mean values of Margalef's richness index and the taxa richness were higher in the lotic systems (main channel and streams).

Food items, density, diversity and correlations of functional feeding groups
A great part of the larvae food was constituted by algae, including in the guts content of predators, like A. gr.annulata and Djalmabatista sp.3.The algae present in the guts content belong in their great majority to the class Bacillariophyceae and few to the class Cyanophyceae.The mean density values of functional feeding groups are shown in Figure 3C.Shredders-herbivores predominate in the major part of the systems, reaching 57% of the total.Collectors-filterers represent 16%, collectors-gatherers 15%, predators 11% and scrapers only 1%.In the streams, the population density of shredders-herbivores reached 294 ind.m -2 (±450.68); in the reservoir reached 51 ind.m -2 (± 42.93) and, in the rapids 580 ind.m -2 (±642.10),where 72% of the larvae belonged to Cricotopus sp.3.
Collectors-filterers represented 35% in the main channel (109 ind.m -2 ).Larvae of Tanytarsus sp.(93 ind.m -2 ) made up the major portion among this group (40%).Collectorsgatherers represented 45% (139 ind.m -2 ±237.24) at the sites of the rapids and reached a population density of 49 ind.m -2 (±110.43) in the main channel.The predators dominated in the secondary channel (±88 ind.m -2 ), corresponding to 40% of the total of this group.Cryptochironomus sp.2 (34%) and A. gr.annulata (26%) were the most abundant among the predators.Scrapers (Phaenopsectra) occurred only in the sampling site S20 in the main channel.
Comparisons of the structure of communities of Chironomidae larvae (density and composition) and abiotic factors in the main channel revealed few significant correlations, mainly with the type of sediment.The density of the species Caladomyia sp.showed a negative correlation with water temperature (r= -0.562, p=0.018) and a positive one with the coarseness of the sediment in the main channel, with the closest correlation with the sediment fraction "coarse sand" (r=0.643,p=0.005), followed by "very coarse sand" (r=0.605,p=0.010) and "medium sand" (r=0.583,p=0.014), while the correlation with "fine sand" was negative (r= -0.512, p=0.035).Another negative correlation to sediment parameters was found for Tanytarsus sp. and muddy sediment (r= -0.671, p=0.003).
Among the functional groups, the diversity of the collectors-gatherers was highest in the habitats "stream" (0.83) and "main channel" (0.63) (see Table 3).Characteristic taxon in this habitats were mainly Goeldichironomus sp., Pelomus cf.P. psamophilus and Caladomyia sp.The greatest diversity of predators was found in the main channel (0.64), where A. gr.annulata, Cryptochironomus cf. C. reshchikov, Djalmabatista sp.3 and Labrundinia sp.2 were the characteristic taxon for this habitat.
Collectors-filterers had low diversity, but were present in most of the habitats, e.g.Tanytarsus sp. and Rheotanytarsus sp.1, and sporadically Tanytarsus cf.T. obiriciae (lake) and Stempellina sp.1 (main channel) (Table 3).The main channel was the only habitat where scrapers were recorded (Figure 2B), represented by Phaenopsectra sp.1, thus the diversity of this site was zero.
The detritus included rests of macrophytes, algae, fungi spores, exuviae, and clay.Exuviae of head capsules were recorded only in larvae of the predator A. gr.annulata.Besides algae, spores of fungi were recorded sporadically in the predator Procladius sp., and commonly in the collectors-gatherers Fissimentum cf.F. desiccatum and Cricotopus sp.3 and the collector-filterer Tanytarsus sp.
No significant correlations were found between biotic and abiotic variables at the sampling sites of the lake, rapids, reservoir, and secondary channel.
Comparing the diversity of the functional feeding groups (n=5) in the six sampled habitats, similarities were obvious only between the main channel and the stream (correlation: r=0.900, p=0.037) despite different diversity values.

Beta diversity and cluster analysis
The Beta diversity values between the sampling sites and dietary sources were high, and most of this variation was assigned to spatial turnover (Table 6).Overall, the poorer sampling sites were grouped by the cluster analysis, where there was the occurrence of rarer taxa (Figure 4).Many sampling sites which had high richness did not form clusters and in these places there was the occurrence of more common taxa.

Discussion
Connectivity to floodplains increase allow an exchange of organic matter and biota between stream and riparian wetlands (Wantzen and Junk, 2000;Junk and Wantzen, 2004) and several components of hydrological connectivity can be distinguished in riverine systems (Wantzen and Junk, 2005;Coops et al., 2008).Within the Sepotuba River, the channels and streams show a lateral connectivity gradient of decreasing exchange between lake water and river water.
Though characteristic for lotic systems (Trivinho-Strixino and Strixino, 2005), Rheotanytarsus was recorded in lakes with muddy ground of the Sepotuba basin.Similarly Butakka (2000) encountered larvae of this genus in floodplain lake close to a channel connecting to the main river (Cuiabá River).Both the systems of communication between the water bodies (floodplain lakes, river channels, permanent connections locally called "corixos") and the degree of connectivity in floodplain systems of the Pantanal can increase the drift of organisms from the channels to the floodplain lakes, generally during inundation periods.
The low diversity of the sites S06 and S35 is caused by the near-exclusive presence of the species Cricotopus sp.3, alone or together with one or another taxa (e.g., Tanytarsus sp. and Djalmabatista sp.3).According to Strixino and Trivinho-Strixino (1998) and Henriques-Oliveira et al. (2000), habitats with rocky substrate and littoral with sand and pebbles in zones of erosion are favourable for the colonisation by larvae of Cricotopus and Corynoneura.These genera contain typical organisms of lotic environments that are rarely encountered in lacustrine systems.
The correlations between the environmental variables and the density of Chironomidae larvae are in accordance with the study of Nessimian et al., (1999), Henriques-Oliveira et al. (2003) and Amorim et al. ( 2004), who found that the sediment type and the water depth were the main factors that influence the species richness and distribution of invertebrates.
The greater presence of Tanytarsus and the lesser presence of Chironomus cf. C. strenzkei in the sampling sites in lakes may indicate an environment with characteristic water quality of high oxygen saturation, low productivity and low eutrophication (Strixino and Trivinho-Strixino, 1998;Marques et al., 1999).Some species of Chironomus prefer or are restricted to water bodies with humic and fulvic substances.Larvae of this genus were encountered in lakes with organic matter and muddy bottom.
From the point of view of the trophic determination one can argue that the associated species showed preference for a certain type of sediment due primarily to the food resources, reflected by the similarity of the diet between the different functional feeding groups.Nessimian et al. (1999) suggested that correlations between two taxa can prove the common use of biotopes and food resources.The analysis of the digestive content in this study confirms this.These authors found that generalist species and opportunists do not exercise pressure on only one food item (Sanseverino and Nessimian, 2008), but rather have an impact on various food items, confirming that the majority of chironomids are generalists.
Observations of the digestive content from the present study confirm the concept of these authors, as several species swallowed the same algae species, e.g., F. desiccatum and Cricotopus sp.3 (Eunotia, Navicula and Encyonema), Chironomus cf. C. strenzkei and A. gr.annulata (Gomphonema and Pinnularia), and Tanytarsus and Cricotopus sp.3 (Cymbella and Eunotia).These results are valid not only for food resources, but can also reveal spatial relationships, as several species explore the same microhabitats, suggesting that the larva is a non-selective feeder and ingests food items in the proportions they occur in the surrounding water.
Collectors-filterers that constitute a considerable part of the functional groups (Tanytarsus, Rheotanytarsus, Tanytarsus cf.T. obiriciae), consume the dissolved organic matter originating from decomposing plants with high microbial activity, or filter living cells of phytoplankton.According to Trivinho- Strixino and Strixino (1991), organic matter is one of the major food sources in lakes and reservoirs that contribute to the balance of the systems in terms of diversity and richness of faunistic groups.
Marques et al. (1999) confirmed that in genera of Tanypodinae the distribution is limited by the presence of vegetation fragments and algae, by the type of sediment, or by the concentration of nutrients.A. gr.annulata, recorded in channels with sandy bottom with mud feed predominantly on algae.According to Pinder (1986), these larvae have an extremely varied diet of animal and algae items.
The greater diversity in the streams and the main channel lend support to the hypothesis that these habitats play an essential role in the maintenance of diversity of the functional groups of these sites.Heino and Paasivirta (2008) suggested an increase of chironomids diversity from head waters to mid-sized rivers and the patterns shown by the diversity indices that take into account the number of individuals also matched these predicted diversity patterns along the river relatively well.However, species density varied primarily with a major water chemistry gradient and not with a stream.The groups formed from cluster analysis agree with these results, because there was a tendency for grouping of sites in relation to species richness, concurrent with the clustering of taxa in relation to its rarity in the sampling sites.This fact indicates that the morphotypes Chironomidae showed no increase in the richness by the river continuum, concluding that the group responded better to some physical or chemical change of water.
Different habitat characterisations show that some groups indicate preference either for the substrate type or for the habitat type.e.g., larvae of Endotribelos, Goeldichironomus (collectors-gatherers) and Phaenopsectra (scraper) inhabited the sandy bottom of lotic systems.However, Sanseverino and Nessimian (2001) encountered larvae of these genera living preferentially in pockets of litter in areas of deposition and erosion.
In addition, the values of the beta diversity analysis indicate a high variation in the composition of morphotypes between environments.This variation was largely attributed to spatial turnover, which allows us to assume that the high dispersal ability of the group has made the species reach the whole region studied.Beta diversity is a wellestablished and widely used concept in ecology and has been measured in different ways and for various purposes (for a review see Tuomisto, 2010), including the turnover, i.e. changes in species composition among local assemblages (Baselga, 2010).
The differences along the river are decisive for the formation of distinct or discontinuous communities, as the changes are more likely in the physical environment, and the limits become obvious though the interrelations between the populations in the community, as for instance, competition for food and habitats, and predation by small fishes.
Thus, variations in the composition and richness among environments can be attributed to local factors governing the occurrence of certain taxa, contributing to compositional dissimilarity between the sampling sites.Moreover our results reveal that these changes may be different for various organisms in accordance with their traits, such as their dispersal abilities.

Figure 1 .
Figure 1.Map of the study area of the Sepotuba River with the sampling sites.

Figure 3 .
Figure 3. a-Mean density and standard error of Chironomidae larvae at 38 sampling sites in the Sepotuba River basin.b-Shannon-Wiener index of diversity (H') of Chironomidae larvae at the sampling sites of the Sepotuba River.c -Density according to functional feeding groups in different habitats of the Sepotuba River.BA= Floodplain lake, MA= Main channel, ST= Stream, RE= Reservoir, RA= Rapids, SC= Secondary channel.

Figure 4 .
Figure 4. Cluster Dendrogram for Chironomidae occurrence in the collection sites of two-way cluster analysis.

Table 1 .
Study area with 38 sampling sites in different aquatic systems in the Sepotuba River basin.

Table 2 .
Habitats and functional feeding groups of Chironomidae of the Sepotuba River Basin.Stream -bottom sandy with gravel.Collector-gatherer Beardius sp.3 Secondary channel and Stream -littoral and bottom sandy or muddy; floodplain lake -bottom gravely-sandy Collector-gatherer Chironomus cf. C. strenzkei Shallow floodplain lake -littoral with organic matter; Channel, Stream and Reservoir -bottom rocky.

Table 5 .
Mean values of the Shannon-Wiener diversity index for functional feeding groups and Evenness, Margalef's richness index, and Taxa richness in the different habitats in different habitats.BA= Floodplain lake, MA= Main channel, ST= Stream, RE= Reservoir, RA= Rapids, SC= Secondary channel.

Table 6 .
Beta diversity analysis with the data of Chironomidae taxa occurrence in the collection sites and the occurrence of food items in the different Chironomidae taxa.