Isotopic fractionation and trophic position of zooplankton species in the Upper Paraná River floodplain

This study aimed to evaluate the isotopic fractionation and trophic position of three zooplankton species ( Notodiaptomus amazonicus , Moina minuta and Bosmina hagmanni ) in the Upper Paraná River floodplain. We predict that phytoplankton is the main food resource used by these species. Three zooplankton samples and three phytoplankton samples were taken from each sampling site, with three to four samples collected for each species. The number of individuals for samples varied according to the body size: from 100 to 130 individuals for Notodiaptomus amazonicus ; 150 to 200 for Moina minuta ; and from 250 to 300 for Bosmina hagmanni . The isotopic values for δ 13 C and δ 15 N were determined using mass spectrophotometer. The isotopic fractionation of 13 C was performed according to the relationship ∆ = δ 13 C

emphasize the need to investigate the complexity of zooplankton dynamics through stable isotope analysis.This study aimed to determine the isotopic fractionation between the phytoplankton and three zooplankton species, Notodiaptomus amazonicus (Wright, 1935), Bosmina hagmanni Stingelin, 1904 and Moina minuta Hansen, 1899, and investigate the trophic position of these species in five lakes from the Upper Paraná River floodplain in the Paraná and Baía subsystems.The general hypothesis of the study predicts that phytoplankton is the main autotrophic source used by three analysed species, independent of sampled locations.
Zooplankton was sampled in two subsystems totalling five sampling stations.In the Baía subsystem, samplings were performed in Porcos, Guaraná and Maria Luiza lakes.In the Paraná subsystem, samplings were undertaken in Garças Lake and Leopoldo Backwater (Figure 1).
Porcos Lake has an elongated shape, a mean depth of 2.3 m and a length of 781.3 m, and it is directly connected to the Baía River through a 60-m wide channel.Its banks are 1 m above river level and are dominated by cyperaceans (Cyrpus) and few bushes (20%).Guaraná Lake presents a rounded shape, a mean depth of 2.1 m and a length of 386.5 m.The connection with the Baía River is 70 m long and 18 m wide.Here aquatic macrophytes predominate and the banks are covered by grasses (95%) and bushes (5%).Maria Luiza Lake is an elongated shape and connects to the Baía River through a small channel.This lake is 693.3 m long and has a mean depth of 3.3 m.This environment is located in a "várzea" field, and vegetation is composed of grasses and cyperaceans with few bushes such as azedinha, tucum and Inga uruguensis Hook et Arn.(UEM et al., 2002).
Garças Lake is situated beside the main Paraná River channel.Among the sampling stations, this lake is the longest (2,128 m) with a mean depth of 2 m.Its banks are covered by grasses and riparian vegetation.Leopoldo Backwater has a mean depth of 3.1 m and is 966.2 m long, with a 15-m wide channel connection to the Paraná River.Banks are covered by forest (100%), which consists mostly of Bambusa guadua Humb et Bonpl. and Croton, Cecropia, Inga (UEM et al., 2002).

Field samplings
Three zooplankton samples were collected from each sampling site using a plankton net (68 µm).Three phytoplankton samples were taken with a phytoplankton net (15 µm) at each locale, filtered in a Kitasato system, and retained in glass fibre filters (Whatman SeS GF 52-C) that had been sterilised in a muffle furnace for 5 hours at 450 o C.

Introduction
One main application for 13 C and 15 N stable isotopes is in understanding food chain structure (Vander Zanden et al., 1997).Isotopic studies have provided a better understanding of the energy flow along food chains.In Brazil, according to Dourado and Benedito-Cecilio (2005), such surveys have yielded important results on the relationships among aquatic organisms.
The isotopic transformations involving the relative abundances of heavy and light isotopes are called fractionations, and they occur because lighter isotopes form chemical bonds that require less energy to break.Lighter isotopes are therefore generally activated more quickly and accumulate in high concentrations in nature (Dawson and Brooks, 2001).
These isotopic signatures motivate ecologists to identify the flow of energy in the food chains composed of different trophic levels, where the number of levels represents the energy or nutrient transferences among the organisms (Lindeman, 1942).The efficiency of this energy transference describes several aspects of the structure of one ecosystem, including the number of trophic levels, the relative importance of detritivorous and herbivorous, the equilibrium values of biomass, the accumulated detritus, and the rates of material exchange (Ricklefs, 2003).
To trace this energy flow, we must understand the existing interrelations of a community, and stomach content analyses are often used to this end.However, the ingested food never represents those assimilated because items digested quickly cannot be identified through these methods (Jepsen, 1999), which are restricted to identifying food chain components.For both C and N isotopes, it is therefore essential to consider changes related to nutritional stress, reproductive activity, migratory behaviour (since migratory species may present a signal different from existing sources in the location where they were collected), and the developmental stage of the organism (Vander Zanden and Rasmussen, 2001;Benedito-Cecilio and Araújo-Lima, 2002).
In addition, it is difficult to obtain trustworthy interpretations of isotopic data because it requires understanding the isotopic fractionation that occurs between the diet and the consumer, as small organic changes affect trophic level estimations and the contributions of different producer sources (McCutchan et al., 2003).
Zooplankton is the principal link in energy transfer between other trophic levels, making it is extremely important to study the isotopic variability of zooplankton (Santana et al., 2009).Such studies can complement knowledge obtained through traditional methods, including stomach content analysis.

Isotopic fractionation
The δ 13 C fractionation between phytoplankton and zooplankton was highly variable between the sampling stations and values were considered in module.For the same species (B.hagmanni), the amplitude was around 9.5‰ between environments where species presented the lowest and the highest fractionation.These results are presented in Table 1.
Among the analysed species in the module, N. amazonicus had the highest isotopic fractionation (7.02‰) in Guaraná Lake, while the lowest value for this species (1.96‰) was in Porcos Lake.The lowest fractionation for B. hagmanni (│-6.04‰│) was observed in Leopoldo Backwater and the highest value (3.51‰) in Garças Lake.Moina minuta did not show extreme fractionation values, with the lowest (|-3.77‰|) in Guaraná Lake and the highest (3.17‰) also in Garças Lake (Table 1).
The sampled material was taken to the Laboratory of Energetic Ecology from the Universidade Estadual de Maringá, where the identification and sorting of zooplankton species were performed.From each sample, we obtained three to four samples for each one of the three zooplankton species studied.The number of individuals varied according to body size, with 100 to 130 individuals for Notodiaptomus amazonicus, 150 to 200 for Moina minuta, and from 250 to 300 for Bosmina hagmanni.
The zooplankton and phytoplankton samples were then dried at 50 o C in an oven with forced circulation for about 48 hours and sent to the Centro de Isótopos Estáveis do Instituto de Biociências at the Universidade Estadual Paulista (Unesp/Botucatu-SP) to determine 13 C/ 12 C and 15 N/ 14 N isotopic ratios with a mass spectrophotometer, calculated according to the following equation: δ 13 N = [(R sample / R standard ) -1] x 10 3 ; where: R = 15 N : 14 N (Peterson and Fry, 1987).
The values of isotopic ratios were expressed in delta (δ) and in parts per thousand (‰), relative to the international standard (Peterson and Fry, 1987) PeeDee Belemnite (PDB) (Roland et al., 2005).
The isotopic fractionation of 13 C was performed according to the relationship determined by Michener and Schell (1994): ∆ = δ 13 C zooplankton -δ 13 C phytoplankton and negative results were presented in module.To determine the possible trophic position of these species, we used the 15 N fractionation similarly estimated the phytoplankton value as -37.2‰.A similar value was found in Venezuelan lakes (Hamilton and Lewis, 1992).The present study showed that the isotopic fractionation between phytoplankton and their consumers varied from -6.04‰ to 7.02‰ for carbon.Based on this result, we suspect that zooplankton compete with components of the microbial food chain for those resources.Considering that zooplankton may explore distinctly the autochthonous and allochthonous resources (Matthews and Mazunder, 2006) as well as bacteria (Bastviken et al., 2003), which have varied isotopic ratios, the use of 1‰ fractionation, established by De Niro and Epstein (1978) and by Fry and Sherr (1984) is difficult in neotropical food web studies.
Some authors have concluded that algae are the main energy source for planktonic food chains (Forsberg et al., 1993, Vaz et al., 1999, Benedito-Cecilio et al., 2000, Finlay, 2001).Nevertheless, phytoplankton may not explain adequately the impoverishment observed in zooplankton δ 13 C values.Other studies (Jones et al. 1999, Bastviken et al., 2003) registered isotopic values ranging from -42 to -25‰.The authors suggest that methanotrophic bacteria might be significant sources of zooplankton nutrition (Hessen and Nygaard, 1992).These bacteria develop in environments rich in dissolved organic carbon during the rainy period (Calheiros and Hamilton, 1998) due to the accumulation of allochthonous organic matter and low levels of oxygen, conditions that facilitate methanogenesis (Bastviken et al., 2003).These bacterioplankton are assimilated by filter feeding organisms as reflected in their low δ 13 C values (Calheiros, 2003).
Regarding nitrogen, 3.4‰ is the fractionation proposed by Vander Zanden et al. (1997) in estimating possible trophic position for species in each sampled environment.

-(δ 15 N) -Trophic position
Determining trophic position was only possible for M. minuta and N. amazonicus.We could not obtain an isotopic signal for B. hagmanni considering the small size of this species (290 to 400 µm; Elmoor-Loureiro, 1989).Despite using 250 to 300 individuals per sample of B. hagmanni, the spectrophotometer was not sensitive enough to identify the isotopic values.For M. minuta, the trophic position was not determined in the Guaraná and Maria Luiza lakes due to low numbers of samples.The possible trophic positions for N. amazonicus showed similar values in most of these environments, whereas M. minuta showed greater variation in the different environments (Table 2).
By analysing each sampled location, we found that none of the species showed the same pattern of fractionation in relation to the other species.
The highest trophic position was verified for N. amazonicus (3.7) in Guaraná Lake, and this same species occupied a similar trophic position in all the environments.M. minuta had a higher position in Porcos Lake (2.3) and a lower position in Leopoldo Backwater (0.6) (Table 2).

Discussion
The isotopic fractionation in organismal tissue varies according to food consumed, which comes from different sources (Adams and Sterner, 2000).In the present study, the isotopic fractionation was quite distinct among species in the different sampling stations.In studies in the Amazon, Benedito-Cecilio et al. (2000) registered a δ 13 C mean value of -36.2‰ for the zooplankton.This value was considered the isotopic fractionation to be 1‰ by trophic level (De Niro andEpstein, 1978, Fry andSherr, 1984), and they Table 1.Isotopic fractionation between phytoplankton and the three zooplankton species by sampling station in the Upper Paraná River floodplain.The value was used in module (| |).LGAR = Garças Lake, LGUA = Guaraná Lake, LPOR = Porcos Lake, LMLU = Maria Luiza Lake, LLEO = Leopoldo Backwater.LGAR = Garças Lake, LGUA = Guaraná Lake, LPOR = Porcos Lake, LMLU = Maria Luiza Lake, LLEO = Leopoldo Backwater.The highest trophic level was observed for N. amazonicus (3.7) in Guaraná Lake and this species presented a similar trophic position in the rest of the environments (on average 1.67).Nevertheless, only one sample of this species was collected in this lake which may explain the high mean isotopic value.On the other hand, this species is a calanoid copepod, a selective omnivorous filter feeding organism that mainly feeds on phytoplankton and detritus (Margalef, 1983, Reynolds, 1984).

Sites/species
Moina minuta represented extreme trophic positions, with the highest was observed in Porcos Lake (2.3) and the lowest in Leopoldo Backwater (0.6), followed by Garças Lake (1.0).Here the species is possibly acting between the first and the second trophic level in Leopoldo Backwater and Garças Lake, and between the second and the third level in Porcos Lake.Such variation was also recorded in Arctic lakes by Kling et al. (1992), who reported that when the herbivorous copepod Diaptomus pribilofensis was present, the predator Heterocope septentrionalis fed preferentially on Diaptomus instead of only algae (Luecke and O'Brien, 1983).However, this observation was not recorded in all studied lakes, with Heterocope showing a preference for algae, and only rarely was Heterocope a restricted predator in these environments.The fact that M. minuta presents trophic position close to 2 may also be associated with the wide food range of cladocerans, which are filter feeding organisms that ingest particles of varied size, including nanoplankton, detritus, bacteria and phytoplankton (Margalef, 1983, Reynolds, 1984).Kling et al. (1992) emphasize that the relative enrichment between the predator zooplankton and its prey is not well known because in natural systems this relationship depends on extending the omnivorous character of these organisms.These authors observed variations in the absolute isotopic value among the systems depending on the basis of the food chain, and highlighted the difficulties inherent in estimations from mixed food sources of several origins.In this study, the absence of bacterioplankton hindered the determination of the trophic position based on the relationship proposed by Vander Zanden et al. (1997).
In the present study, we rejected the null hypothesis of the absence of interspecific and spatial variations in zooplankton community.The complexity in zooplankton dynamics examined using stable isotopes and fractionation estimations may vary depending on physiological and environmental factors.Further studies focusing on bacterioplankton are necessary for a more reliable determination of the different trophic levels of the planktonic food chain.The wide isotopic variation in zooplankton emphasizes the need to develop methods to study the trophic structure of environments subject to constant changes, namely tropical aquatic systems.The elucidation and understanding of the trophic positions of these organisms, based on traditional techniques (stomach content analysis) and stable isotopic analysis offer complementary insights into the flow of matter and energy in floodplain aquatic environment food chains as well the trophic relations involved in the ecological strategies roles distinct species.

Table 2 .
Trophic position estimate of zooplankton species by sampling station.