Microhabitat preferences of the freshwater prawn Macrobrachium jelskii ( Decapoda : Palaemonidae )

We analyzed the microhabitat preferences of Macrobrachium jelskii (Miers, 1877) males and females inhabiting an urban water reservoir in Cruz das Almas, Bahia, Brazil. Prawns were collected monthly, from March 2015 to February 2016, in three microhabitats, using a sieve. Each microhabitat was dominated by one macrophyte species: Eleocharis sp. (M1), Cabomba sp. (M2), and Nymphaea sp. (M3). The prawns were measured (carapace length), and categorized as juvenile males, adult males, juvenile females, non-ovigerous adult females and ovigerous adult females. An analysis of variance (ANOVA) was used to compare the number and size of individuals. The sex ratio and frequency of ovigerous females in the three microhabitats were also calculated. The sex ratio was biased towards females in M1 and did not deviate from 1:1 in M2 and M3. When prawns were separated into five categories we observed that non-ovigerous adult females were more abundant in M1, while adult males were the most abundant demographic category in M2 and M3. Juveniles of both sexes and ovigerous females showed no microhabitat preference, although M1 and M2 appeared to be more suitable for the latter. Adult females were the largest individuals in all microhabitats. Food availability, lower depth and lower predation pressure in M1 are the main factors that make M1 more suitable for M. jelskii, particularly non-ovigerous adult females and larger adult males. Intraspecific competition for shelter in M1 might also occur and adult females win this competition due to their larger body size. Therefore, adult males are found in higher abundance in M2 and M3 and the juvenile of both sexes spread evenly across all microhabitats. Our results help to understand the ecological role and the niche used by M. jelskii. Future studies on the habitat choice and predation under laboratory conditions should help to understand the behavior of this species.

Animals often search for habitats that can fulfill their biological needs (e.g.feeding, reproduction, and avoidance of predators) (Müller et al., 1999;Rocha, 2010).In freshwater environments there is a stratification of the fauna directly related to the environmental conditions and shelter availability (Thomazi et al., 2008;Giuliatti & Carvalho, 2009).This stratification occurs due to predation pressure and food resources that are exploited by different species throughout their life phases (Merrit & Cummins, 1996;Marques et al., 1999;Thomazi et al., 2008;Ferreira et al., 2010).In this context, the occurrence, abundance, and type of macrophytes forming a microhabitat strongly influence freshwater communities, as they provide nutritional resources, shelter, and breeding places for phytophilic invertebrates (Cook & Streams, 1984;Takeda et al., 2003;Lammers et al., 2009a,b;Thomaz & Cunha, 2010).
As part of the food chain of limnic environments, M. jelskii has a high ecological importance in these environments, feeding upon insect larvae, diatom and other algae, and sediment grains (Melo, 2003;Cirilo et al., 2011).In addition, like all caridean prawns, it is a prey of fishes and other animals (Bauer, 2004;Magalhães et al., 2014).From the economic point of view, M. jelskii is important in some regions of Brazil (e.g.state of Sergipe), and exploited as fishing bait, aquarium ornamentation, and food resource by the riverine population (Cirilo et al., 2011;Vera-Silva et al., 2016).
Here we investigated a population of M. jelskii in Cruz das Almas municipality, State of Bahia, Brazil.Since M. jelskii is a phytophilic species, we investigated its microhabitat preference, comparing its abundance in three types of macrophytes and considering its population structure in each microhabitat.We compared the total abundance and size (carapace length), between males and females of two classes of sexual maturity (juvenile and adult).Using information from the literature, we discuss the possible ecological factors influencing the population structure and microhabitat preference.Finally, we believe that our results give an important contribution to the knowledge of the ecology of this species, which is widely distributed throughout South American water bodies.

Study area. Specimens of Macrobrachium jelskii
were collected monthly from March 2015 to February 2016, at Doutor Braz Water Reservoir (12°40'06.6"S;39°07'11.2"W)which is located within the city limits of Cruz das Almas (Fig. 1).The reservoir is a typical lentic environment, with surrounding riparian vegetation composed mainly of ground cover and herbaceous species.Different types of aquatic macrophytes are present, wholly or partially submerged, been Eleocharis sp., Cabomba sp., and Nymphaea sp. the most abundant.Sampling protocol.The sampling sites were visually chosen according to the type and abundance of aquatic macrophytes to ensure there was no overlap between the macrophyte species.Thus, three microhabitats were selected, each one dominated by one of the most abundant macrophytes of the water reservoir: Eleocharis sp.(M1), Cabomba sp.(M2) or Nymphaea sp.(M3).It is important to note that during the pilot study, no prawns were found in sites lacking vegetation.
The three microhabitats differed both in density (see Dibble et al., 1996 for review) and distribution, which were estimated visually.M1 was a dense, marginal and emergent vegetation, found at depths shallower than 40 cm.M2 was a dense and always submerged vegetation, found at depths between 50 and 100 cm.M3 formed a less dense vegetation, with floating leaves connected to the submerged stem by long petioles, and it was dominant at higher depths (deeper than 100 cm).
In addition, numerous studies estimating spatial complexity (created by stems and leaves) and plant architecture of aquatic macrophyte species have shown that Cabomba sp. has greater spatial complexity than Eleocharis sp. and Nymphaea sp., the latter being the one with the lowest complexity (Taniguchi et al., 2003;Warfe & Barmuta, 2004;McAbendroth et al., 2005;Dibble & Thomaz, 2006, 2009;Warfe et al., 2008;Casartelli & Ferragut, 2018).Thus, we could classify the microhabitats evaluated in the present study in an increasing order of complexity, as follows: M3 (the simplest); M1 (intermediate complexity); and M2 (the most complex).Specimens of M. jelskii were captured by sweeping a sieve (diameter 50.0 cm, 0.5 mm mesh) through the vegetation of each microhabitat.The sampling effort consisted of the activity of four people during 10 minutes in each microhabitat.
In the laboratory, the prawns were identified according to Melo (2003).The sex was determined based on the presence (males) or absence (females) of the appendix masculina on the second pair of pleopods.Males were categorized as juveniles and adults based on the developmental stage of the appendix masculina observed under the microscope.Only those having a fully developed appendix masculina were considered adults, since studies indicated that this structure is important during the spermatophore transference and is only functional when fully developed (Bauer, 1976(Bauer, , 2004;;Berg & Sandifer, 1984).Females were categorized as juveniles or adults based on the average size of sexual maturity (CL = 6.67 mm) estimated by Rocha & Barbosa (2017).Therefore, prawns were separated into five demographic groups: juvenile males, adult males, juvenile females, non-ovigerous adult females and ovigerous adult females.All individuals had their carapace length (CL = from the postorbital margin to the posterior margin of the carapace) measured with a digital caliper (0.01 mm precision).Afterwards, they were fixed in 70% ethanol and kept in labeled flasks.Voucher specimens were deposited at the Museum of Zoology, Universidade Federal do Sul da Bahia, Brazil (accession number M20180358UFSB).
Statistical analyses.The Qui-square test was used to verify the differences between the overall sex ratio.A two-way ANOVA was used to check differences among the abundance of each demographic group and microhabitats.We used the abundance as a response variable, and demographic group, and microhabitat (M1, M2, and M3) as categorical variables.We calculated the ANOVA two-way in two different ways; first, the analysis was performed separating the demographic group into male and female.Second, the analysis was done subdividing males and females into five demographic groups.Significant differences would indicate microhabitat preferences.This same analysis was done using the CL as response variable to check for differences in body size among demographic groups and microhabitats.All statistical analyses were performed based on Sokal & Rohlf (1995) and Zar (1996) using Statistica 7.0 (Statsoft ® ) and Microsoft Office EXCEL ® 2003 or later, with level of significance of P < 0.05.
Finally, we tested the residuals of the ANOVAs for the assumptions of normality using a Kolmogorov-Smirnov test (Sokal & Rohlf, 1995).Homoscedasticity assumption was also tested using the statistics of Hartley (maximum value of F), Cochran (value of C), and Bartlett (Chi-squared) (Sokal & Rohlf, 1995).The assumptions were considered broken when the significance level of the tests were below the 0.05 level (Sokal & Rohlf, 1995).The normality assumption was broken for two ANOVAs: i) using as response variable the number of individuals and subdividing the demographic groups into five demographic groups (d = 0.137; P < 0.01); ii) using as response variable the CL and dividing the demographic groups only into male and female (d = 0.035; P < 0.01).The homoscedasticity assumption was broken for three ANOVAs: i) using as response variable the number of individuals and subdividing the demographic groups into five demographic groups (F max.= 740.334;C = 0.238; χ² = 177.571;P = 0); ii) using as response variable the CL and dividing the demographic groups only into males and females (F max.= 4.014; C = 0.323; χ² = 850.477;P = 0); and iii) using as response variable the CL and subdividing the demographic groups into five demographic groups (F max.= 5.976; C = 0.157; χ² = 436.356;P = 0).Therefore, we standardized both response variables (i.e.number of individuals and carapace length) for all ANOVAs.Following Sokal & Rohlf (1995) and Gottelli & Ellison (2011) recommendations we used as standardization algebra the sum of 0.5 value in each measure and then take the square root of this sum.This mathematical procedure is indicated when the response variable is counting data (e.g. the number of individuals), and when the residuals tend to have their distribution to the right (as observed in the ANOVA's residuals, using carapace length as response variable, in the normality test) (Sokal & Rohlf, 1995).

RESULTS
In total, we collected 6,428 individuals of Macrobrachium jelskii, of which 3,149 were males and 3,279 were females.Among the males captured, 114 were considered juveniles and 3,035 adults.Among the females captured, there were 1,672 juveniles and 1,607 adults, from which 220 were ovigerous.Considering the microhabitats, 2,177 (34%), 2,437 (38%), and 1,814 (28%) individuals were captured in M1, M2, and M3, respectively.There were no significant differences between the total number of individuals in each microhabitat (F = 1.529;P = 0.224).
Considering all individuals captured during the sampling period and in the three microhabitats, 51% of the population was female (without separating juveniles and adults).Therefore, the overall sex ratio was 1.04 female to 1 male (χ² = 2.59; P > 0.05).Considering only two demographic categories (males and females; Fig. 2) the number of females was significantly higher than males in M1 (F = 13.342;P < 0.001), and there was no difference between the number of males and females in M2 and M3 (F = 2.061; P = 0.135).When the data were separated into five demographic categories (Fig. 3) adult males were the most abundant demographic group in M2 and M3 (F = 5.055; P < 0.001), and there was no predominance of a demographic group in M1.Non-ovigerous adult females preferred M1 (F = 93.391;P = 0), while ovigerous adult females and juveniles of both sexes showed no microhabitat preference (F = 2.4; P = 0.093).However, in the case of ovigerous adult females, it is worth mentioning that 112 (51%) specimens were collected in M1, 88 (40%) in M2 and only 20 (9%) in M3.
When the CL was analyzed by microhabitat, females were larger than males in M1, M2, and M3 (Fig. 4).The largest individuals of both sexes were found in M1 (F = 53.6;P = 0; Fig. 4).Although individuals from M2 were the smallest, the difference was not statistically significant in comparison with individuals from M3 (see bars and confidence intervals in Fig. 4).When adult and juvenile males and adult (non-ovigerous and ovigerous) and juvenile females were analyzed separately (Fig. 5), adult females were larger in all microhabitats; ovigerous adult females were significantly larger than non-ovigerous adult females in M1 and M2 (F = 27.2.1;P = 0), but not in M3 (see bars and confidence intervals in Fig. 5).The largest and smallest juvenile females were found in M1 and M2, respectively (Fig. 5).Adult males were significantly larger in M1 than in the other microhabitats (F = 12.1; P = 0; Fig. 5) and were larger than juvenile females in all microhabitats.Juvenile males were the smallest individuals in all microhabitats (Fig. 5).The CL of non-ovigerous and ovigerous adult females and juvenile males did not differ between microhabitats (see bars and confidence intervals in Fig. 5).Finally, the results of two-way ANOVA demonstrated significant differences in the CL of males and females, both in adult and juvenile in all microhabitats (F = 2027.1;P = 0).

DISCUSSION
Increase of habitat complexity has a positive effect on biomass and structure (richness, diversity, species composition) of periphyton and the amount of particulate organic matter (Warfe & Barmuta, 2006;Casartelli & Ferragut, 2018).Compared to Nymphaea sp. the submersed structures of Cabomba sp. and Eleocharis sp.retain greater amount of particulate material colonized by periphyton, epiphyton, aquatic invertebrates, and decomposer organisms such as bacteria and fungi (Thomaz & Bini, 2003;Hogsden et al., 2007;Thomaz & Cunha, 2010;Camargo & Ferragut, 2014).Those elements are the main food items of M. jelskii diet (Melo, 2003;Cirilo et al., 2011).Therefore, although ANOVA demonstrated that there were no statistical differences between the total numbers of individuals in each microhabitat, M. jelskii were more abundant in M1 and M2 (totaling 72% of all sampled individuals).According to Williner & Collins (2002), M. jelskii prefers shallower marginal areas, which require less energy to move in the water column, in comparison with deeper and open-water areas.As observed in our study, shallower waters (less than 100 cm deep) have higher abundance of more complex macrophytes and thus provide more food resources.
Although food availability is certainly important for all demographic categories of M. jelskii, our data allows us to infer that this is particularly important for adult females.Living in food-rich habitats is especially important for females of carideans prawns that present larger size than males, which is the case for M. jelskii.In this case, females need a significant food supply to maintain their growth rate (Howard & Lowe, 1984) and the high energetic demand of ovary maturation after sexual maturity, spawning and incubation of eggs (Sastry, 1983;Pinheiro & Taddei, 2005;Hartnoll, 2006;Araújo et al., 2012;Lira et al., 2012).These biological features could explain the higher abundance of non-ovigerous adult females in M1.Moreover, although not statistically significant, the higher relative frequency of ovigerous adult females in M1 and M2 (91% of all ovigerous females sampled) is also an indicative of the importance of these food-rich microhabitats to the reproduction of M. jelskii.Similarly, ovigerous females of Palaemon pandaliformis (Stimpson, 1871) had a strong preference for E. crassipes, probably associated to the density of this vegetation, especially the roots (Paschoal et al., 2013).
The shelter availability is directly proportional to the structural complexity and density of macrophytes, where preys are protected from visual predators, especially fishes (Dibble et al., 1996;Thomaz & Cunha, 2010).Thus, in addition to the food availability, M1 and M2 could also provide a better protection against predators than M3.
However, Warfe & Barmuta (2006) conducted an experiment with artificial analogues of three macrophytes (Myriophyllum sp., Triglochin sp. and Eleocharis sp.) and observed that pygmy perch (predator fish) had their greatest impact on invertebrate grazers in the most structurally complex plant analogue, Myriophyllum sp.(macrophyte similar to Cabomba sp.).Those authors concluded that vulnerable invertebrate grazers might have been unable to perceive the risk of predation in Myriophyllum sp.compared to the other plant analogues.A similar situation might have occurred in Doutor Braz Water Reservoir, where M2, although more complex, might offer greater risk of predation.Thus, M1 although less complex than M2 would offer less risk of predation because Eleocharis sp. is a macrophyte that grows in shallower areas.This hypothesis is corroborated by the preference of the adult non-ovigerous females, the larger adult males and the larger juvenile females for M1 (see Figs. 3 and 5).In this context, Bentes et al. (2011) also observed that riparian river areas provided a better shelter to Macrobrachium amazonicum (Heller, 1862) against predators due to their lower depth and higher vegetation density.Finally, Rocha (2010) demonstrated that Macrobrachium acanthurus (Wiegmann, 1836) females prefer living in the riparian vegetation.
The higher abundance of adult males (compared to the other demographic categories) found in M2 and M3 might be related to an intraspecific competition with females (mainly adults) for shelter in M1.Our results demonstrated that the size and number of females are significantly higher than males in M1.According to Figler et al. (1999) and Lammers et al. (2009b), the body size is an important factor during the establishment of social hierarchy and shelter conquer in decapods, regardless of the sex.Therefore, based on the distribution pattern of M. jelskii in the Doutor Braz Water Reservoir, we can infer that the large females inhabiting M1 won the competition for space, displacing smaller females and males to other microhabitats.In this context, it is important to note that the largest adult males were found in M1, probably because they were large enough to won the competition for shelter from smaller juvenile females and males, which spread evenly across all microhabitats.Still, the smaller size of those individuals who could not settle in M1 would be advantageous and allow them to inhabit deeper and less dense microhabitats, since they are less conspicuous to predators.The same result was observed for Macrobrachium acanthurus and Macrobrachium olfersii (Wiegmann, 1836), in which large males were seen with groups of females in banks of E. crassipes in Salsa river, Ilhéus, Bahia (Paschoal et al., 2013).Cabomba sp. is denser than Nymphaea sp. and can grow in shallower areas near M1, where there is a higher concentration of larger adult and juvenile females.These factors would make M2 the preferred microhabitat of males (mainly adults) who could not conquer space in M1 due to their small body size, comparing with adult females.
According to Williner & Collins (2002), Macrobrachium jelskii males are more active than females.This behavioral difference might be related to the type of reproductive strategy of this species, the pure search (Rocha & Barbosa, 2017).In this strategy, males are highly mobile and their reproductive success depends on their ability to fertilize the highest number of females (Correa & Thiel, 2003;Bauer, 2004).Therefore, the search for females would stimulate the males to move between the different microhabitats in Doutor Braz Water Reservoir.During these high-activity periods, males could become more vulnerable to predation; however, their small body size is advantageous as it confers a higher mobility and low conspicuousness to predators (Correa & Thiel, 2003).
The present study demonstrated the importance of Eleocharis sp., Cabomba sp., and Nymphaea sp. in the distribution, abundance, population structure, and reproduction of M. jelskii in a lentic environment.These aquatic macrophytes were confirmed to be a very important component of the microhabitats used by the prawns.Eleocharis sp. is the most preferred microhabitat and is particularly important for adult females, while Cabomba sp. is crucial for adult males.Juveniles of both sexes dispersed equally between all microhabitats and there is an indicative that M1 and M2 are important to the reproduction of M. jelskii.The main ecological factor influencing the population of M. jelskii in these microhabitats is probably the intraspecific competition for food and shelter.Our results help to understand the ecological role and the niche used by this freshwater prawn.Future studies on the habitat choice and predation under laboratory conditions should help to understand the behavior of this species.