Length-weight relationships and biomass of the main microcrustacean species of two large tropical reservoirs in Brazil

Length-weight equations were determined for the main microcrustacean species of Três Marias and Furnas reservoirs (state of Minas Gerais, Brazil), and evaluated the fluctuations of their biomass. We examined the following species: Thermocyclops minutus, Bosminopsis deitersi, Bosmina hagmanni, Ceriodaphnia cornuta, and Moina minuta (Três Marias Reservoir), and Notodiaptomus henseni, Daphnia ambigua, Ceriodaphnia silvestrii, Diaphanosoma spinulosum, D. fluviatile, and Bosmina freyi (Furnas Reservoir). Dry weight was obtained in a microbalance for each size class (Cladocera) or developmental stage (Copepoda). Microcrustacean mean biomass varied from 5.76 mg DW.m and 20.36 mg DW.m (Furnas Reservoir) and from 3.75 mg DW.m and 18.14 mg DW.m (Três Marias Reservoir). Significant differences (p < 0.000) between seasons were registered with higher biomass during the rainy seasons. Thermocyclops minutus was the most important species in Três Marias, whereas in Furnas, N. henseni contributed in equal proportion. In Furnas, the higher cladoceran biomass was explained by the contribution of larger-sized species, such as D. ambigua, C. silvestrii, D. spinulosum, and D. fluviatile. Even though both reservoirs are considered oligotrophic, there were significant differences in the estimated microcrustacean biomass, which reinforces the importance of this parameter for the description of the real contribution of each species in the community.


Introduction
The majority of studies on the zooplankton community in Brazilian aquatic ecosystems involve only the taxonomic list or the numerical densities of these organisms.However, because of the wide range in body size (from mm to cm), this community is better characterized by the biomass (McCauley, 1984;Rocha et al., 1995;Bozelli and Huszar, 2003).Some taxa that contribute significantly in numbers, such as rotifers and protozoans, have less importance for biomass, whereas the larger-sized copepods and cladocerans may be more important for this parameter (Pace, 1986;Matsumura-Tundisi et al., 1989;Melão et al., 2005;Bonecker et al., 2007).Moreover, because of their larger body size compared to rotifers, microcrustaceans are preferential prey of fish, especially visual predators (Sá Júnior, 1994).
The need for high-precision equipment is one of the main reasons for the relatively few studies on zooplankton biomass, and explains why many studies use lengthweight regressions described in the literature (Esteves and Sendacz, 1988;Bezerra-Neto and Pinto-Coelho, 2002;Pinto-Coelho et al., 2005a;Corgosinho and Pinto-Coelho, 2006;Panarelli et al., 2010;Santos et al., 2010).Most often, these regressions are generalizations for higher taxonomic levels and are based on species of temperate climates.Moreover, species-specific effects (lifecycle duration, assimilation capacity, and nutritional state), environmental variables (temperature, pollution, climate and geographical factors, nutrients, food quantity and quality, predation) and methodological particularities (preservation or not in formalin, time of preservation, measurement criteria) may influence the determination of the length-weight regression of each species (Bottrell et al., 1976;McCauley, 1984;Blettler and Bonecker, 2006).
Recently, with the development of limnological research in Brazil and wider acquisition of microbalances (an instrument capable of making precise measurements of weight of objects of relatively small mass: of the order of a million parts of a gram), studies on length-weight relationships (Maia-Barbosa and Bozelli, 2005;Santos et al., 2006;Castilho-Noll and Arcifa, 2007) and dry weight and biomass (Melão andRocha, 2004 and2006;Sendacz et al., 2006;Magalhães et al., 2006;Blettler and Bonecker, 2007;Bonecker et al., 2007;González et al., 2008;Rosa, 2008) have increased.
Studies that consider biomass values are important because they allow comparison of different environments, providing a common unit to evaluate zooplankton groups.Generally, eutrophic environments support higher biomass of smaller individuals than more oligotrophic ones (Pace, 1986;Esteves and Sendacz, 1988).Biomass studies are also important to determine secondary production, another parameter that is relatively seldom evaluated in Brazil.
Aiming to increase fish production in Três Marias and Furnas reservoirs, the project "Technical-Scientific Study to Delimit Aquaculture Areas in the Lakes of Furnas and Três Marias Reservoirs -MG, Brazil" (SEAP/SECTES/FUNDEP-UFMG), of which this study is a part, was proposed to select areas with oligotrophic waters within these reservoirs to install cage systems for tilapia farming.In this project, abiotic and biotic variables were measured to estimate the support capacity of these areas, with the objective of minimizing the impact on water quality.The knowledge generated about the zooplankton will help, in the future, to use these organisms as a food source for fish as well as possible bioindicators for monitoring water quality, mitigating the impacts of fish production on aquatic communities.
In this context, the aim of this study was to determine length-weight relationships of the main microcrustacean species of two tropical reservoirs, Três Marias and Furnas (state of Minas Gerais), as well as to evaluate the biomass fluctuations of these species in two arms of these reservoirs, during the dry and rainy seasons.In this context, we also made a complete comparison with all papers available that provide values of microcrustaceans size, dry weight and biomass in many lakes and reservoirs of Brazil considering their trophic state.

Study area
Três Marias (18°12' S and 45°15' W) and Furnas (46°19' W and 20°40' S) are two large reservoirs in the state of Minas Gerais (Brazil).Both were constructed ca.40 years ago, mainly for power generation, and are also used for recreation, professional and sport fishing, irrigation, and water supply.Três Marias Reservoir is located on the Upper São Francisco River in the central-western part of Minas Gerais.The reservoir was completed in 1960, and silting is the main impact that compromises its power generation and water quality, followed by the replacement of the vegetation along the shore by Eucalyptus plantations and large cattle ranches (Sampaio and López, 2003).Furnas Reservoir is located in the Grande River basin in southern Minas Gerais; its north arm is represented by the Grande River, and the south arm by the Sapucaí River.This reservoir was completed in 1962, and the main impacts are monoculture crops and cattle ranching, sewage discharge, solid residues, and loading of agricultural chemicals (Nogueira et al., 2008).Some morphometric characteristics of the two reservoirs are presented in Table 1.
In each reservoir, we selected two arms with different surrounding land uses (see Figure 1).In Três Marias Reservoir, we selected the Barrão arm (Z max = 21.8 m, with preserved Cerrado, the Brazilian savanna) and Extrema arm (Z max = 20.9m, with Eucalyptus monoculture), located between Morada Nova de Minas and Três Marias municipalities.In Furnas Reservoir, we selected the Varjão arm (Z max = 17.5 m, with headwaters in Paredão Municipal Park and coffee monoculture in the surroundings) and the Mendonça arm (Z max = 25.5 m, with cattle ranching on native grasslands), located between Guapé and Capitólio municipalities.These arms were selected in order to reflect the different land uses around the reservoirs.

Procedures
Samples were collected every two days, during four weeks, in two dry seasons (July/August 2006 and July/August 2007) and two rainy seasons (January/March 2007 and January/March 2008), considered the most representative of the variations in temperature and precipitation.In the project "Technical-Scientific Study to Delimit Aquaculture Areas in the Lakes of Furnas and Três Marias Reservoirs -Minas Gerais, Brazil", the cages will be located within the arms to avoid the effect of reservoir management.Thus, samples were collected at the midpoint of each arm, in the limnetic region, far from the main channel of São Francisco and Grande Rivers.
For the qualitative and quantitative zooplankton samples, vertical hauls were made with a plankton net of 68 mm mesh size.Because of the presence of drowned original vegetation ("paliteiros") at the sampling stations, hauls were made in the euphotic zone as determined by Secchi disk.Organisms were narcotized with gasified water, stained with Rose Bengal, and preserved with 4% buffered formalin.Subsamples of 1.0 mL were counted in an Olympus (CBA) optical microscope, in a Sedgewick-Rafter chamber.The data are presented as organisms per m 3 .During the counts, nauplii, copepodids, and adults of each species of Copepoda were enumerated.The literature used to identify species could be find in Brito et al. (2011).
For this work, we considered the most abundant microcrustacean species (those that comprised at least 5% of the zooplankton community during the sampling To determine the length-weight relationships, we used specimens from the dry season of 2006, when 30 individuals from each size class (determined for Cladocera) or stages of development (Copepoda) (Bird and Praire, 1985) were measured with an ocular micrometer.Nauplii of T. minutus were the exception once 200 individuals were measured and weighed due to smaller weight.For Copepoda, the measurements were made from the cephalothorax to abdomen, excluding the furca

Reservoir Três Marias Furnas
Flooded area 1 100 km 2 1 440 km 2  López andSampaio, 2003 andTundisi et al., 1993.or caudal ramus; for Cladocera, measurements were made from the head to the base of the carapace, excluding the spine.For Daphnia ambigua, the measurements were made from the base of the helmet to the base of the carapace, excluding the spine.In this way, we determined the size classes and the percent contribution of neonates, juveniles, and adults of cladoceran species.The specimens were rinsed three times in distilled water, and eggs and embryos were removed from adult females, dried at 60 °C for 24 h (Bottrell et al., 1976;McCauley, 1984), and weighed in a Sartorius SE2 microanalytical balance (accuracy of 0.1 mg).The data are presented as dry weight (DW).Although the literature contains reports of weight losses of 10 to 40% for preserved organisms (Dumont et al., 1975;McCauley, 1984;Blettler and Bonecker, 2006), we made no correction for losses due to preservation.
Mean dry weights for each size class (Cladocera) or stages of development (Copepoda) were used to determine length-weight regression after logarithmic transformation of weight (W in mg DW) and length (L in mm), according to the formula of Bottrell et al. (1976): where a = estimate of intercept and b = estimate of slope.
Using length-weight regressions and lengths obtained for other sampling seasons, it was possible to estimate biomass from the product of density (org.m -3 ) and mean dry weight (mgDW) (Winberg and Duncan, 1971) for Três Marias and Furnas reservoirs.Biomass data are presented as mg DW. m -3 .
Differences in values of length and dry weight were assessed through Student's T test, while differences in biomass values between arms, reservoirs, and seasons (dry and rainy) were tested through two-way Analysis of Variance, followed by a Tukey's post-hoc comparison test (Sampaio, 2002).All statistical analyses and regressions were performed with Statistica 7.0 (StatSoft).
Mean length (mm) and respective mean dry weight (mgDW) of developmental stages of the main microcrustacean species in Três Marias and Furnas reservoirs are presented in Table 2.For the species that occur in both reservoirs, generally shorter lengths were observed in Três Marias (t-values for C. cornuta = 7.776; M. minuta = 6.891 and T. minutus = 4.343; p < 0.0001), with the exceptions of B. deitersi (t = 0.965; p = 0.341) and B. hagmanni (t = 0.824; p = 0.414), where differences were not significant.Lower dry weights were also obtained in Três Marias, but differences were not significant for bosminids (t = 1.561; p = 0.127 and t = 1.347; p = 0.249 respectively).
Length-weight regressions obtained in this study (Table 3) were considered adequate for dry weight estimates of copepod and cladoceran species from Três Marias and Furnas reservoirs, since the differences between the observed and estimated weights were not significant (p ³ 0.928).Comparing common species, regressions obtained for T. minutus, B. hagmanni, and M. minuta from Furnas were characterized by a steeper slope (b value), indicating a greater increase in dry weight for an individual of the same length in Três Marias Reservoir.Steeper slopes in the regressions of B. deitersi and C. cornuta were obtained in Três Marias.
As observed for the density, microcrustacean biomass (see Figure 2) showed seasonal variations, with higher values recorded in the rainy season (F = 131.81;p = 0.000) (mean values in dry and rainy seasons of 3.75 mg DW.m -3 and 18.14 mg DW.m -3 in Três Marias and 5.76 mg DW.m -3 and 20.36 mg DW.m -3 in Furnas, respectively).However, differently from density, higher values of microcrustacean biomass were observed in Furnas (F = 2.808; p = 0.04).Higher values of biomass in Três Marias were obtained in the Barrão arm, and in Furnas, in the Mendonça arm, but the differences were not significant (p > 0.06).Copepods always comprised the majority of microcrustacean biomass in both reservoirs, between 68 and 80% in Três Marias and between 53 and 75% in Furnas.

Discussion
Body-size and dry-weight values obtained for the microcrustacean species in Três Marias and Furnas reser-voirs were within the range of those obtained by several investigators in other Brazilian aquatic ecosystems.Considering differences in methodology, the following studies did not mentioned or did not make corrections for losses due to preservation: Maia-Barbosa ( 2000  (2005), individuals in Três Marias and Furnas showed higher values than those estimated for cladocerans from the "natural" (unimpacted) station of that lake, however lower values than observed at the station in the area impacted by bauxite tailings.According to these authors, the ingestion of tailing particles and the retention of this material within the valves (imposible to remove during preparation for weighing) is the reason for the higher weight of individuals from the impacted station.
Daphnia ambigua showed a larger mean body size, however a lower mean dry weight than observed by Castilho-Noll and Arcifa (2007) respectively).Lake Monte Alegre is considered an eutrophic environment, and therefore individuals of this species may have a better nutritional condition and consequently higher dry weight.However, dry weight values near those obtained in this study were observed by Sendacz et al. (2006) in both Ponte Nova (oligotrophic) and Guarapiranga reservoirs (eutrophic) (São Paulo).
Bosminopsis deitersi showed a smaller mean body size and lower mean dry weight compared to adults in Broa (0.61 mg) (Matsumura- Tundisi et al., 1989) and Lagoa Dourada reservoirs (Melão and Rocha, 2006) (dry weights of neonates, juveniles, and adults: 0.23, 0.55, and 0.70 mg, respectively).However, comparisons with these environments must be made with care, because nonpreserved individuals were used.
Moina minuta, Ceriodaphnia cornuta, and C. silvestrii from Três Marias and Furnas reservoirs showed lower mean dry weights compared to Ponte Nova and Guarapiranga reservoirs (Sendacz et al., 2006); however, the last species showed higher values than those obtained by González et al. (2008) for Bariri Reservoir (SP) (mean dry weight of adults: 1.44 mg).Length and dry weight of C. silvestrii found by Santos et al. (2006) (dry weights for neonates, juveniles, and adults of 1.337, 2.113, and 5.335 mg, respectively) were higher than those obtained in Três Marias and Furnas.However, the indivuduals of this study were raised in the laboratory, and their weights were estimated from non-preserved individuals.In Bariri Reservoir, Diaphanosoma spinulosum and Bosmina hagmanni showed higher mean dry weights for adults (1.897 and 2.233 mg, respectively) (González et al., 2008).In Broa Reservoir (Matsumura-Tundisi et al., 1989), copepodids (I-V) (420 mm and 0.430 mg) and males and females (600 mm and 0.780 mg) of Thermocyclops minutus showed larger body sizes and higher mean dry weights than observed in Três Marias and Furnas reservoirs.For this species also, non-preserved individuals were measured.
These differences among the estimates of body size and dry weight for microcrustaceans reflect the responses of species to the varied conditions in aquatic ecosystems (Dumont et al., 1975;Boersma and Vijverberg, 1996;Blettler and Bonecker, 2006).
Length-weight relationships for microcrustacean species were confirmed in the present study.Some studies have failed to find linearity in this relationship (Magalhães et al., 2006;González et al., 2008).However, these authors did not use logarithmic transformation, a procedure that strengthens length-weight relationships due to linearization of the data (Blettler and Bonecker, 2006).
Biomass fluctuations in Três Marias and Furnas reservoirs were more pronounced in the rainy season, as also observed by Rosa (2008) in five reservoirs of the Furnas Centrais Elétricas S.A. System and by Melão et al. (2005) in Lagoa Dourada Reservoir.In some studies, higher values were obtained in the dry season (Rocha and Matsu-mura-Tundisi, 1984;Sendacz et al., 2006, Santos-Wisniewski andRocha, 2007); while in others, this pattern most often was not evident, due to different contributions of zooplankton groups in the two seasons (Rietzler et al., 2004;Corgosinho and Pinto-Coelho, 2006).However, in all studies, the highest values of biomass were consistently related to sites and seasons with higher food availability (nutrients, chlorophyll-a, suspended solids).In fact, in Três Marias and Furnas reservoirs, higher values of chlorophyll-a and suspended solids were observed in the rainy season (Brito et al., 2011).However, significant correlations with microcrustacean biomass were obtained only for total and organic suspended solids (r = 0.560 and 0.547; p < 0.0001; respectively).In these reservoirs, this could be an important non-algal food source for the species studied.According to Marzolf (1990), the phytoplankton alone does not satisfy all the nutritional requirements of zooplankton.Pinto-Coelho et al. (2005a and2005b) suggested that when the correlations between chlorophyll-a and zooplankton biomass are weak, detritus is more important than primary production.
The dominance of T. minutus in density and biomass in Três Marias can be explained by the predominant food available in this reservoir, the cyanobacteria.This copepod species can capture large cyanobacteria colonies such as Microcystis, Botryococcus, and Aphanocapsa (Matsumura- Tundisi et al., 1997).However, because cyanobacteria have low nutritional value (Rietzler and Espíndola, 1998), the greater consumption of this group by T. minutus in Três Marias can affect its development, resulting in smaller-sized individuals and lower dry weight than for the same species in Furnas Reservoir.
Higher biomass values in Furnas may be related to larger-bodied species such as calanoids (N.henseni), daphniids (D. ambigua and C. silvestrii), and sidids (D. spinulosum and D. fluviatile), which have higher dry weights than the smaller bosminids that predominate in Três Marias.Castilho-Noll and Arcifa (2007) observed that, despite its smaller body size, D. ambigua weighed more than D. gessneri, probably due to its rounded body shape, reflecting a higher biomass.Similarly, for B. hagmanni in Três Marias, the higher dry weight of this species resulted in a higher biomass contribution than B. deitersi, which dominated in numbers.Therefore, we can conclude that for microcrustaceans, biomass better represents the community structure than does numerical density alone.
The microcrustacean biomass values estimated in Três Marias and Furnas reservoirs are close to those observed in other reservoirs considered oligotrophic such as Segredo and Mourão (Blettler and Bonecker, 2007) in state of Paraná; Ponte Nova (Sendacz et al., 2006) and Lagoa Dourada (Melão et al., 2005) in São Paulo.In the natural lakes Dom Helvécio and Carioca (Minas Gerais) these values also were close to these oligotrophic reservoirs (Table 5).
In Lake Batata, lower cladoceran biomass values occured at the natural station, while at the station im-pacted by bauxite tailings, higher biomass values -close to mesotrophic environments -were recorded.In Broa Reservoir, microcrustacean biomass showed peaks from 60 to 80 mg DW.m -3 (Matsumura- Tundisi et al., 1989)when the study was carried out, this reservoir was considered mesotrophic by authors.
In naturally eutrophic environments such as Lake Amarela (Z max = 2 m, in the final stage of succession, and densely colonized by macrophytes), higher biomass values were also obtained, with an annual mean of 173.85 mg DW.m -3 for microcrustaceans (Matsumura- Tundisi and Tundisi, 1986).
Several authors have described a positive relationship between zooplankton biomass and the trophic state of tropical and temperate aquatic environments (Pederson et al., 1976;Pace, 1986;Esteves and Sendacz, 1988;Rocha et al., 1995;Rocha et al., 1997;Pinto-Coelho et al., 2005c).These authors considered differences in phytoplankton quantity and quality as the food source, and in zooplankton community composition and density among environments, as well as the nutritional state of individuals and other factors that influence biomass.
Although they are both considered oligotrophic, Três Marias and Furnas reservoirs showed remarkable differences in microcrustacean biomass.In Três Marias, Thermocyclops minutus, a common species in Brazilian aquatic environments (Reid, 1989) was the most important in terms of density, and comprised up to 80% of the biomass.Bosminopsis deitersi, B. hagmanni, C. cornuta, and M. minuta were not important in this reservoir, in contrast to other Brazilian lakes and reservoirs (Maia-Barbosa, unpublished data; Melão and Rocha, 2006;Sendacz et al., 2006).In Furnas, copepods were also the dominant group in density and biomass; however, N. henseni contributed in the same proportion as T. minutus.The higher Cladocera biomass in this reservoir was composed mainly of the larger-sized species D. ambigua, C. silvestrii, D. spinulosum, and D. fluviatile, especially during the rainy season.Because of their greater availability in terms of biomass and higher nutritional quality (Santeiro and Pinto-Coelho, 2000;Pinto-Coelho et al., 2005a), larger-sized microcrustaceans may be considered a more appropriate food source for juveniles as well as adults of zooplanktivorous fishes in Furnas Reservoir.

Table 2 -
Mean length (mm), mean dry weight (mg DW), and number of replicates for the main microcrustacean species of Três Marias and Furnas reservoirs (Minas Gerais).