Ecological aspects of natural populations of Hyalella pampeana ( Crustacea , Amphipoda , Hyalellidae ) from the Natural Reserve Island of Martín García ( Río de La Plata , Argentina )

Freshwater amphipods of the genus Hyalella have a great importance in aquatic ecosystems due to their role in matter and energy cycling and its utilization as bioindicators of environmental health. The aim of this work was to analyze relevant population parameters of Hyalella pampeana, such as population density and structure, individual body size, sex ratio, fecundity and recruitment, and to study the possible relation of these parameters to variations in environmental variables. Samples were taken monthly during a one-year period (2006) at three small ponds in the Natural Reserve Island of Martín García, Argentina. The inland aquatic environments showed distinctive physicochemical characteristics, mainly related to pH and conductivity values. Amphipod population density varied among sites and throughout the year, with population peaks during spring, positively correlating with pH values. The operational sex ratio of the total population was almost equal to 1 male: 1 female, displaying seasonal variation. Mean fecundity and mean size of ovigerous females were higher during winter and early spring months than in summer months. The presence of ovigerous females and juveniles in all months indicated that reproduction is continuous during the year for Hyalella, in its natural environment.


INTRODUCTION
Amphipods inhabit a wide variety of aquatic environments and have a great ecological importance, since they facilitate the matter and energy transference from inferior levels of the food web to consumers (other macroinvertebrates, fish, amphibians and birds).Besides, due to their ubiquity, small body size, short life cycles and easy culturing in laboratory, amphipods are used as water quality bioindicators and as test organisms in toxicology studies (Pilgrim and Burt 1993, Ingersoll et al. 1998, Borgmann et al. 2005, lasier and Urich 2014, Javidmehr et al. 2015).
The distribution of the freshwater amphipod genus Hyalella Smith 1874 is restricted to the Western Hemisphere, with most species found in South America (Väinölä et al. 2008).Most studies in amphipod populations classically focus on Hyalella azteca Saussure 1858, a widely distributed species in aquatic ecosystems of North America, and analyze the population density and its temporal fluctuations, age and size, sex proportion, reproduction and recruitment periods (lindeman and Momot 1983, edwards and Cowel 1992, Wen 1992, 1993, Moore and Farrar 1996).The evaluation of these population parameters is of crucial importance since they allow to obtain information about the biology of a species, the stability of the population in a given habitat, its adaptability, reproductive success and persistence (probability of leaving offspring for prolonged periods), among others (Odum and Warrett 2005).
Due to the lack of ecological studies in natural populations of H. pampeana, the objective of this work was to contribute to the knowledge of the population structure and dynamics of this species in pleustonic communities of three water bodies on the Martín García Island.For this purpose, we analyzed relevant population parameters such as amphipods density, body size, recruitment, fecundity and sex ratio, and their monthly and seasonal variation.We also aim to relate some of the population parameters of H. pampeana to relevant variables on the aquatic habitat in order to better understand the natural environment where this species lives.

STUDY AREA
The Island of Martín garcía (34º 11' S and 58º 15' W) is located at the confluence of the Paraná and Uruguay rivers, Río de La Plata Superior, Argentina (Figure 1).geologically, the island is a raised and fractured block of the crystalline basement of the mass of Brasilia, later covered by Holocene and Pleistocene quaternary deposits (Ravizza 1984).In different sectors of the island, there are aquatic environments that differ in their origin and substrate (natural sandy ponds or quarries excavated on the rocky bottom), hydrological regimen (permanent or temporal) and water inputs (precipitations and/or flows from the Río de La Plata River).These environments develop carpets of floating vegetation and a wide diversity of aquatic invertebrates inhabits the pleustonic habitat.Some of them include mollusks, annelids, platyhelminthes, crustaceans and insects (Viana An Acad Bras Cienc (2019) 91(1)  e20170928 3 | 21 and placed in white trays where ovigerous females were individualized in labeled plastic vials before transportation to the laboratory.Physicochemical parameters (air and water temperature, pH, dissolved oxygen, conductivity and total dissolved solids) were registered in situ with a digital multiparametric equipment (Water Quality Meter Sper Sc.LTD).
In seasonal analysis, the sampling day date of each month was considered for selecting the corresponding season: February and March (summer); April and May, (autumn); June, July, August and September (winter); October, November and December (spring).
Sampling campaigns were conducted monthly between February and December 2006.Triplicate samples of pleuston were taken with a hand net (0.09 m 2 surface area and 150 μm mesh size) at three study sites.Site 1 ("Cantera la gata") and site 2 ("Cantera Tanque") are both abandoned quarries, and site 3 ("laguna Arenalcito") is a small natural sandy pond.While sites 1 and 3 are temporary, site 2 is a permanent habitat (Figure 1).Samples were fixed in situ with 10% v/v formalin water over a sieve (125 µm) and removed for further identification analysis.Once the vegetation was separated, the remnant material was analyzed under a stereoscopic microscope, and 20 amphipods of each sample (10 males and 10 females) were selected for dissection.The identification of the amphipods was made by observing the anatomical pieces placed in semipermanent slides under microscope and following the specialized taxonomic literature (Cavalieri 1968, grosso andPeralta 1999).
H. pampeana individuals were classified into four categories: males (individuals with enlarged second pair of gnathopods), females (individuals with visible oostegites), ovigerous females (females with eggs or juveniles inside the marsupium) and juveniles (individuals without recognizable sexual characters) (Castiglioni and Bond-Buckup 2008a).All specimens were counted, and total population density and density by categories (ind/m 2 ) were calculated.
In order to obtain the cephalothorax length (Cl, in mm), all individuals were measured from the front margin of the head to the posterior margin of the head, in lateral view (Castiglioni and Bond-Buckup 2008a).For measurements, a stereoscopic microscope complemented with a 0.1 mm precision micrometer was used.The relationship between Cl and total length (Tl, distance between the front margin of the head and the posterior margin of the telson, in lateral view) was analyzed in 25 females, 25 males and 25 juveniles of H. pampeana.Pleustonic vegetal species were identified according to Lahitte and Hurrel (1996).To obtain the plant dry weight, the material was dried during 48hs at 105 °C and weighed using a precision digital scale (Dahus, explorer).

Environmental variables
An Analysis of Variance (ANOVA) complemented with a posteriori Bonferroni test (α=0.05)was conducted to evaluate significant differences in the total mean values of the environmental variables between the three study sites.The environmental variables were also analyzed seasonally in order to account for differences along the studied period.
A Principal Component Analysis (PCA) was performed in order to analyze the relationship between the environmental variables measured, and to characterize the study sites.The variables included in the analysis were water temperature (°C), pH, dissolved oxygen (DO, mg/l), conductivity (µS/ cm), and plant dry weight (g/m 2 ).Air temperature and total dissolved solids were removed from the analysis since both variables are correlated with water temperature and conductivity, respectively.PCA was conducted using the statistic program XlSTAT 2014.

Population dynamics
Differences in total population density of H. pampeana (ind/m 2 ) between sites were analyzed by means of ANOVA and a posteriori Bonferroni test (α=0.05).Data were previously log-transformed in order to meet the requirements of the analysis.In order to evaluate the representation of males, females, ovigerous females and juveniles in the population, we also analyzed the population density by demographic categories (Wen 1992, Poi de Neiff and Carignan 1997, Galassi et al. 2006).
Minimum, maximum, mean, and standard deviation (SD) of Cl were estimated for males, females (including ovigerous females) and juveniles at each study site.Significant differences between annual mean body size of each amphipod category were compared using a one-way ANOVA (factor: sites) complemented with a posteriori Bonferroni test (α=0.05).
A t-test was conducted in order to determine body size differences between males and females at each sampling site (α=0.05;Sokal and Rohlf 1979) (Castiglioni and Bond-Buckup 2008a).The To determine temporal variations in the frequency of each category of H. pampeana during the analyzed period, the study site with higher population density (and, consequently, with higher probabilities for each size category to be represented) was selected to conduct a sizefrequency distribution analysis.H. pampeana individuals were grouped into size classes, and relative frequency histograms were constructed.The number of size classes was determined according to the value of one quarter of the Cl standard deviation measurements (Castiglioni and Bond-Buckup 2008a).
A monthly proportion analysis of juveniles and adults of H. pampeana was performed at each site.This analysis was conducted to identify reproduction and recruitment periods.Differences in the proportion 1 juvenile: 1 adult were studied by means of a Chi squared test (X 2 ; α=0.05) using the statistic software XlSTAT 2014.
Sex ratio of H. pampeana was calculated as the abundance of males/abundance of females (M: F, except ovigerous females) at each site.This ratio is called "operational" (OSR, operational sex ratio), and it considers the mean number of fertilizable females per sexually active male, in a given moment (Emlen and Oring 1977).OSR was also calculated seasonally and by size classes at each study site.A Chi squared test (X 2 ; α=0.05) was conducted in all the cases to test the null hypothesis of 1M: 1F (Sokal and Rohlf 1979).
Fecundity was determined as the number of eggs or juveniles present in the marsupium of each ovigerous female.At each study site, the minimum and maximum egg production was estimated, and the mean fecundity among the sites was compared by means of ANOVA, complemented with a Bonferroni test (α=0.05).The monthly variations in mean fecundity and size of the ovigerous females were analyzed by a Pearson correlation analysis in order to determine the relation between both parameters.Frequencies between ovigerous females and adult non-ovigerous females were analyzed seasonally in each study site using a k-proportions test (p<0.05)with the statistic software XlSTAT 2014.

Population parameters and environmental variables
In order to assess if the variation in the population parameters (population density, OSR, recruitment, fecundity) can be related to the variation in the environmental variables measured, a stepwise multiple regression analysis was performed using the SPSS v. 22 software.Variables included in the analysis were water temperature, pH, dissolved oxygen, conductivity and plant dry weight, at all sites sampled.The selection criterion adopted for environmental variables was the p-value associated with Student's t-statistic (probability of entry of the variable to the equation, p <0.01) (lepš and Šmilauer 2003).

ENVIRONMENTAL CHARACTERIZATION OF STUDY SITeS
A summary of the seasonal environmental data is shown in Table I.Mean water temperature ranged between 12.3 °C in winter and 26 °C in summer, while pH values were close to neutrality at sites 1 and 2, and slightly acidic at site 3. Dissolved oxygen values showed great variation, with a minimum of 2.17 mg/l and a maximum of 9.17 mg/l.Water conductivity and total dissolved solids (TDS) mean values oscillated between a minimum of 15 µS/cm and 10.5 ppm, and a maximum of 447 µS/cm and 295 ppm, respectively.Mean plant dry weight ranged between 27 and 85.1 g/m 2 .The most frequent vegetal species were Lemna minuta Humb.I).

Bonpl. et Kunth 1815 and Azolla filiculoides lam
According to the PCA results, components 1 and 2 explained 71.77% of the data variability (C1: 48.21% and C2: 23.56%, Figure 2).Water conductivity, pH and plant dry weight account for 85.6% of component 1 variation, while water temperature and dissolved oxygen content account for 82.9% of component 2 variation.Samples from site 2 were located at the positive sector of component 1, related to higher values of conductivity and pH, while samples from sites 1 and 3 were located at the opposite sector, related to lower values of these variables.The exception was site 1 in October, which presented a maximum value of plant dry weight of 106.3 g/m 2 .Component 2 ordered the sites according to water temperature and dissolved oxygen content.Sites located at the positive sector of the component represent the samples taken during warmer months and with higher values of dissolved oxygen.

POPUlATION DeNSITY
Higher mean population density values were registered at site 2 (5,326 ± 3,500 ind/m 2 ).The ANOVA results showed significant differences in annual mean population density (log 10 density) between the three study sites (F=21.78;p<0.01) (Figure 3).   at sites 1, 2 and 3, respectively.H. pampeana was registered throughout the whole year at the three study sites, with the exception of site 3 in May, where no amphipods were collected.A population peak occurs during spring months, with maximum values in November at sites 1 and 2 (8,163 ind/m 2 and 14,244 ind/m 2 , respectively), and in December at site 3 (1,859 ind/m 2 ).

BODY SIZE
The correlation between cephalothorax length (Cl) and total length (Tl) of males, females and juveniles of H. pampeana was positive and statistically significant in all cases (males: r=0.93, p<0.01; females: r=0.86, p<0.01 and juveniles: r=0.72, p<0.01).
Mean body size, standard deviation (SD) and range (Cl minimum-Cl maximum) of each category of H. pampeana per site are indicated in Table II.At all sites, the mean body size of males was significantly greater than that of females (site 1: F=150.8,p<0.01; site 2: F=135.6,p<0.01; site 3: F=35.6, p<0.01).The ANOVA results showed significant differences in mean size of males, females and juveniles between study sites (Table II).

POPULATION STRUCTURE AND RECRUITMENT
The study site with the highest population density was selected (site 2), and individuals of H. pampeana were grouped into 23 size classes of 0.02 mm.All the categories of H. pampeana were represented during the year (Figure 5).
The size-classes distribution analysis revealed polymodality in most of the months, with different modes in juveniles and adults.Adults predominated for most of the year.even though recruitment of juveniles occurs throughout the year, there were two periods when this event had greater intensity.During March and April, a moderate increase of recruitment was observed, with juveniles representing frequencies of 54 and 57%.Finally,  from October to December (spring), the second peak of recruitment occurs, with juvenile frequencies of 55 and 66%.
Table III shows monthly variations in the proportions of juveniles and adults at each study site.A similar trend was observed in each environment: during most of the year, the proportion was 1 adult: 1 juvenile or adults predominated.Juveniles instead, predominated during spring months.At site 3, the only site of natural origin analyzed, juveniles density increased also in April and August.No amphipods were collected during May at this site.
Sex ratio was also analyzed by size classes at the three study sites (Figure 7 a, b and c).At all sites, a greater number of females occurred in the smaller size classes in relation to males (until size class: 0.35-0.37mm at site 1, and 0.41-0.43mm at sites 2 and 3), although this ratio was not significantly different from 1M: 1F (p>0.05).However, in greater size classes (from 0.44-0.46mm at site 1, and 0.47-0.49mm at sites 2 and 3), sex ratio was skewed towards males at all sites (p<0.05).
FECUNDITY No significant differences were found in total mean fecundity of H. pampeana between sites (F=2.94;p>0.05).At site 1, fecundity was 9.97±3.7 eggs/ female (range 6-26 eggs/female); at site 2, 10.03±3.6 (range 6-27 eggs/female); and at site 3, 10.9±4.4 (range 6-27 eggs/female).Regarding monthly variation, mean fecundity and size of ovigerous  females of H. pampeana increased during winter and early spring and decreased during the summer months (Figure 8a, b and c).Both parameters correlated positive and significantly at all sites (site 1: R 2 =0.44, p<0.05; site 2: R 2 =0.53, p<0.05 and site 3: R 2 =0.59, p<0.05).Ovigerous females were collected in all seasons in the three study sites.The seasonal variation indicates a greater frequency (p<0.05) of ovigerous females during autumn and spring compared to winter in site 1, and during autumn, spring and summer in site 3.In site 2, the ovigerous females were more abundant in winter (51.2% of females were ovigerous) (Figure 9a, b  and c).

POPULATION PARAMETERS AND ENVIRONMENTAL VARIABLES
The results of the multiple stepwise regression analysis of the complete environmental dataset show that pH is the variable that better describes the variations in population density of H. pampeana at the study sites.About 61% of the variability in population density (log 10 amphipod density) can be explained by its lineal relation to pH (R 2 = 0.62; p<0.01).The resulting model corresponded to a simple regression: log 10 amphipod density = 1.23pH -4.84 (Figure 10).No significant correlations were found between other population parameters, such as  fecundity, OSR, recruitment, body size, and the environmental variables analyzed.

STUDY SITeS
The inland aquatic environments analyzed in this work differ in their biological and physicochemical characteristics.The main differences between study sites were related to water conductivity and pH values.According to Drago and Quirós (1996), one of the most important mechanisms governing Argentinean inland water chemistry is rock dominance.Sites 1 and 2 have a common origin; both are quarries drilled in the granite bottom of the island and later covered with organic matter.
Nevertheless, site 1 has a temporary hydrological regimen, with droughts in some months of the year, although it never dried during the study period.Water inputs at this site come from precipitation and floods of the Río de La Plata River (César et al. 2009)  Population peaks were registered during spring at the three sites analyzed.According to Kruschwitz (1978), increments in water temperature reduced the time required for ovarian maturation in cultures of H. azteca, shortened the intervals between ovipositions, and increased reproductive rates.This could explain the high densities in natural populations of Hyalella during warmer months, as has been reported by other authors (Casset et al. 2001, Poretti et al. 2003).In our study, H. pampeana density peaks occur together with an increment on the floating vegetal biomass.Similar results were reported by giorgi et al. ( 2005) in a Pampean stream, where amphipods increased their numbers together with the macrophyte development in spring and summer.According to Waterkeyn et al. (2008), aquatic vegetation cover is an important factor influencing invertebrate community structure, presumably because the vegetation creates structural heterogeneity and it can provide refuges and food resources.

BODY SIZE
Males of H. pampeana were larger than females.This difference in size between sexes has been reported in other Hyalella species, such as H. pleoacuta, H. castroi (Castiglioni and Bond-Buckup 2008a), H. bonariensis (Castiglioni et al. 2016), H. longistila (Bastos-Pereira and Bueno 2016), H. georginae and H. gauchensis (Ozga and Castiglioni 2017).According to low (1978), crustacean growth is similar between sexes until reproductive maturity.After that, males and females present different ecological or reproductive demands.Males invest most of their energy in reproduction, especially in searching for a female and copulation, while in females energy effort is invested in gamete production and parental care.This results in different growth rates, which are probably the main cause of size differences between sexes (Wen 1993).According to lopretto An Acad Bras Cienc (2019) 91(1) e20170928 14 | 21 (1983), molting has different functions for males and females of H. pampeana.For females, it is an event mostly related to reproduction, since its occurrence is indispensable for achieving egg fertilization.Males, on the other hand, exhibit a lower number of molt occurrences, but intervals between molts are longer, allowing considerable increments in size.
In this study, a smaller mean body size of H. pampeana was registered at site 1, when compared to the other sites.Strong (1972) found that different populations of H. azteca might have different adult body sizes, according to the selective pressure of the habitat.This includes seasonal changes in growth and reproduction and characteristics of the aquatic environment, such as substrate type, productivity, food availability and presence of predators.Although these parameters were not analyzed in the present study, the great variability in water level at site 1, the only pond studied that receives the flooding of the Río de La Plata River, could affect habitat stability and explain the smaller body size of H. pampeana registered in this environment.
Positive correlations between cephalothorax and total length like those found in this work are indicated in other Hyalella species, like H. azteca (Strong 1972, Kokkotis andMclaughlin 2002) and H. longistila (Bastos-Pereira and Bueno 2016).Measurement of cephalothorax length is an effective and easy method for obtaining information about body size, and it can be used to replace the measurement of total length.

POPULATION STRUCTURE AND RECRUITMENT
Population structure of H. pampeana showed several modes in different size categories of males, females and juveniles.Polymodality has been reported in other South American Hyalella species, such as H. castroi, H. pleoacuta (Castiglioni and Bond-Buckup 2008a), H. bonariensis (Castiglioni et al. 2016) and H. longistila (Bastos-Pereira and Bueno 2016).According to Conde and Díaz (1989), bi-or polymodality in size classes frequency distribution generally reflects recruitment pulses, differential mortality or behavioral differences, while unimodality is associated with a stable population structure with continuous recruitment and constant mortality rate.In this work, juveniles were registered in all months, indicating a continuous moderate recruitment but with more intense pulses in some periods, where this category exceeded 50% of the total population abundance.
Ovigerous females of H. pampeana were collected in most of the months, indicating that reproduction is continuous in this species throughout the year, in agreement with data reported by lopretto (1983) in ponds of the province of Buenos Aires, Argentina.Continuous reproduction has been reported in other Hyalella species that inhabit natural populations in tropical and subtropical temperature regimes (edwards and Cowell 1992, Castiglioni and Bond-Buckup 2009, Bastos-Pereira and Bueno 2016, Ozga and Castiglioni 2017, Torres et al. 2015, Castiglioni et al. 2016).According to Sainte-Marie (1991), gammaridean populations exhibit reproductive patterns that can be associated with latitude.High latitude species generally present biannual or perennial life cycles, large body size, delayed maturity, and single or few broods with many relatively large embryos, while the opposite set of characters occurs in low latitude populations.The continuous food availability in tropical and subtropical climates appears to be one of the main factors allowing all year reproduction and brooding in Hyalella populations (Steele andSteele 1991, edwards andCowell 1992).

SEX RATIO
Operational sex ratio of H. pampeana was close to 1:1 when considering the annual data set per study site.However, regarding seasonal variation, OSR was skewed towards males during autumn at sites 1 and 3.According to Moore (1981), the sex ratio of some amphipods may vary during the life cycle, with male dominance during colder months and female dominance or proportions close to 1:1 during warmer months of the year.In Amphipoda, females are available for reproduction only for a brief period during its molting cycle (Wen 1993).According to emlen and Oring (1977), the degree of spatial and temporal clumping of the limiting sex is a main factor that produces asymmetries in the sex ratio.For example, continuous long periods of sexual activity by males, coupled with brief and asynchronous periods of receptivity by females, will produce a strong skew in the OSR.When the OSR is skewed toward males, as we observed in H. pampeana during autumn in sites 1 and 3, polygyny (individual males frequently gain access to multiple females) is expected; when the skew is toward females, polyandry (individual females frequently gain access to multiple males) should occur.Seasonal fluctuation of OSR was also recorded in H. longistila (Bastos-Pereira and Bueno 2016) and H. bonariensis (Castiglioni et al. 2016).
Sex ratio analyzed by size classes showed a statistically significant predominance of males over females in larger size classes of H. pampeana.Similar results were reported for H. castroi, H. pleoacuta (Castiglioni and Bond-Buckup 2008a), H. longistila (Bastos-Pereira and Bueno 2016) and H. bonariensis (Castiglioni et al. 2016), which could be related to differences in energy consumption between sexes.Female amphipods invest most of their time and reproductive effort in egg production, while males do so in pairing.Besides, taking into account that ovigerous females do not molt, males can reach larger sizes than females.This reproductive difference may probably generate differences in adult size of both sexes (Wen 1993), and influence sex ratio in animal populations (Székely et al. 2014).H. georginae: 37.4 eggs and H. gauchensis: 25.7 eggs) (Castiglioni andBond-Buckup 2009, Ozga andCastiglioni 2017).
Both mean fecundity and size of ovigerous females increased during winter and early spring, and decreased during summer months.A reduction in fecundity together with a smaller body size of ovigerous females during warmer months of the year have also been reported in H. pleoacuta and H. castroi (Castiglioni and Bond-Buckup 2009).Temperature is considered to be the main factor in the generation of size and fecundity differences in winter and summer amphipod populations.Villarroel et al. (2000) analyzed the effect of temperature on some population parameters of H. azteca in laboratory, and reported a decrease in mean body size, fecundity, incubation time and duration of sexual amplexus with a 6 °C increase in water temperature (24 to 30 °C).Reduction in fecundity and size of amphipods during summer months could be attributed to a metabolic effect.Higher temperatures result in higher costs for females in metabolic maintenance, which could result in a lower amount of energy invested in the production of eggs during warmer months (Nelson 1980, Ward 1986).
Monthly mean size of ovigerous females of H. pampeana was positively correlated with mean fecundity.This result agrees with those reported by other authors for several Hyalella species, like H. azteca (Strong 1972, France 1992 (Ozga and Castiglioni 2017).
Although ovigerous females of H. pampeana were collected in all months of the year, their frequency varied seasonally at each site.In the permanent pond (site 2), the higher frequencies were found during colder months.According to data reported for populations of H. pleoacuta and H. carstica (Castiglioni et al. 2008a) and H. bonariensis (Castiglioni et al. 2016), this greatest occurrence of ovigerous females could be associated with greater food availability during winter (macrophytes).On the other hand, ovigerous females were more frequent during warmer months in the two temporary ponds (sites 1 and 3).Although these ponds never dried completely during the study period, they are more exposed to water level fluctuations than site 2 (site 1 receives water during floods of the Rio de La Plata River, and site 3 is a small pond in the base of a sand bank that only receives water for precipitation).Unlike permanent ponds, temporary habitats provide less stable physicochemical conditions and, consequently, less time is available for invertebrates to complete their life cycles, colonization and community development (Waterkeyn et al. 2008).
We suggest that the seasonal changes in occurrence of ovigerous females among the different ponds could be related to the environmental stochasticity to which they are exposed.Therefore, future studies about the influence of hydroperiod and other related parameters can be important for better understanding the dynamics of amphipods populations in these habitats.

POPULATION PARAMETERS AND ENVIRONMENTAL VARIABLES
In this study, significant correlations between monthly amphipod population density and water pH were found.There are a few studies on the Ricciardi (2010), water conductivity influences the growth and survival of gammaridean amphipods.When exposed to conductivity or calcium levels at the lower limit of their tolerance range, crustaceans may suffer reduced growth, lower reproduction, and increased mortality by osmotic stress (Zehmer et al. 2002).This hard water requirement could explain the preference of Hyalella species for waters with elevated conductivity, as has been reported in other studies (Miserendino 2001, Miserendino and Pizzolón 2000, galassi et al. 2006).

CONCLUSIONS
The results of this work suggest that although Hyalella is a ubiquitous genus, the environmental characteristics of each aquatic microhabitat are important for the establishment and persistence of its populations.Hyalella pampeana was registered at the three study sites, showing variable abundance throughout the year.Among the variables analyzed in the aquatic environment, pH was an important factor affecting amphipod density.Population peaks occur during spring, as in other Hyalella species.
Males were significantly larger than females, and positive correlations between cephalothorax and total length were found in all the demographic categories (males, females and juveniles).Ovigerous females and juveniles were registered in all months, indicating continuous reproduction and recruitment; both events have more intense pulses in some periods of the year.Sex ratio of the total population was close to equality, but, when analyzed seasonally, males predominated in autumn.Fecundity was lower compared to another Hyalella species and it fluctuated seasonally, smaller ovigerous females with fewer eggs were found during summer months.Future studies might increase the knowledge about the relations between natural populations of H. pampeana with their environment, as well as their possible use as test organisms in ecotoxicological tests.

Figure 1 -
Figure 1 -Location of the Natural Reserve Island of Martín García (Argentina) in (a) South America; (b) the estuary of Rio de La Plata River.(c) Map of the Island showing study sites 1, 2 and 3.

Figure 3 -
Figure 3 -Total mean population density (ind/m 2 ) of H. pampeana at each study site (lines indicate standard deviation).Values with different letters differ statistically (p<0.05).

Figure 5 -
Figure 5 -Monthly size-frequency distribution of H. pampeana at site 2 (year 2006).The numbers in the right corner of each graph represent the number of juveniles (J), males (M), females (F) and ovigerous females (OF).

FECUNDITY
Total mean fecundity values of H. pampeana at each site were similar to those reported by lopretto (1983) in laboratory cultures (14 and 12 eggs/ female at thermal regimes of 10-22 °C and 23-26 °C, respectively).Fecundity of H. pampeana was also comparable to mean fecundity values estimated for H. carstica; 12.6 eggs/female(Torres et al. 2015).However, when compared with other south American Hyalella species, females of H. pampeana produce a lower mean number of eggs (H.pleoacuta: 36.1 eggs; H. castroi: 31.4 eggs;

Figure 9 -
Figure 9 -Seasonal variation in the frequency of ovigerous females (%) of Hyalella pampeana in (a) site 1, (b) site 2 and (c) site 3. Values with at least one letter in common did not differ statistically (p>0.05).

Figure 10 -
Figure 10 -Simple regression model.Variables: pH, and log 10 amphipod density.Regression equation and value of the coefficient of determination (R 2 , p<0.01) in the text.Samples references as in Figure 2.

TABLE I Main environmental parameters (seasonal mean ± standard deviation) of three inland aquatic sites on the Island of Martín García (year 2006). The total mean values (N=11) that showed significant differences between sites are indicated with letters of the same type (capital, small, capital bold and small bold). Values with at least one letter in common did not differ statistically (p >0.05).
An Acad BrasCienc (2019) 91(1) e20170928 8 | 21

TABLE II Summary of body size data of males, females and juveniles of H. pampeana and results of the Analysis of Variance (ANOVA) and a posteriori Bonferroni test (α=0.05) for the CL mean data recorded at sites 1 to 3. Annual mean CL (mm), standard deviation (SD), number of individuals (N) and CL range (minimum-maximum, in mm). Capital, small and bold letters show differences between mean CL of males and females at sites 1, 2 and 3, respectively.
POPULATION DYNAMICS OF Hyalella pampeanaAn Acad Bras Cienc(2019) 91(1) e20170928 10 | 21

TABLE III Monthly variation in the proportion of juveniles and adults of H. pampeana at each site with respective results of Chi squared test for goodness of fit (X 2 ). Asterisks indicate significant differences in the proportion 1 adult: 1 juvenile (p<0.01). Density values (ind/m 2 ) of juveniles and adults are indicated in parenthesis.
, Othman and Pascoe 2001) H. castroi, H. pleoacuta (Castiglioni and Bond-Buckup 2009), H. carstica (Torres et al. 2015), H. bonariensis (Castiglioni et al. 2016), H. gauchensis and H. georginae