Larval and adult density of the porcellanid crab Petrolisthes armatus ( Anomura : Porcellanidae ) in an Amazon estuary , northern Brazil

Petrolisthes armatus (Gibbes, 1850) is a porcellanid crab with a wide geographical distribution. In the present study we analyzed variations in the abundance of P. armatus adults and larvae over an annual cycle in the Marapanim estuary of the Amazon coastal zone, in the northeastern portion of the state of Pará, Brazil. Particularly, we focused on the presence of ovigerous females and timing of larval release, with the aim of elucidating reproductive patterns in a tropical estuarine system. The mean density of P. armatus larvae (zoea I and II) correlated positively with the salinity of the shallow waters of the estuary, whereas the abundance of adults correlated with the salinity registered in water samples collected from the benthic environment. There was also a significant positive correlation between larval density (zoea I and II) and water temperature. Ovigerous females were captured throughout the study period, from August 2006 to July 2007, but were more abundant in June and less abundant during the rainy months, between February and May. Larvae were only present during the dry season and transition months (June to January), and were absent during the rainy season (February to May). Petrolisthes armatus reproduces throughout the year in the Marapanim estuary and all developmental stages of this species (zoeal stages I and II, megalopae and adults) are found in the estuary. The results indicate that the study area is an important environment for the reproduction of this decapod.

The porcellanids are typical crustaceans, with a pelagic larval phase.Larvae disperse in the water column, where they feed and grow.This is followed by a benthic phase associated with the substrate (QUEIROGA & BLANTON 2005).Most porcellanid species pass through two zoeal stages prior to the molt that gives rise to the megalopae (OSAWA & MCLAUGHLIN 2010).
Petrolisthes armatus (Gibbes, 1850) is the most widely distributed species of the genus, being found in the western and eastern Atlantic, and eastern Pacific (MELO 1999).In Brazil, P. armatus has been recorded off the northeastern coast (Fernando de Noronha Island) and in Santa Catarina, in the extreme South (MELO 1999).On the northern coast, this species has been recorded in Maranhão and, more recently, in Pará (MELO 1999, BARROS & PIMENTEL 2001).The spatial distribution, population structure, and development patterns of P. armatus has been studied in oceanic beach and estuarine environments in the southern and southeastern Brazil (OLIVEIRA & MASUNARI 1995, MASUNARI & DUBIASKI-SILVA 1998, MIRANDA & MANTELATTO 2009, 2010a).
The larval development of P. armatus was first investigated by LEBOUR (1943), who described zoeal stages I and II, whereas GORE (1970GORE ( , 1972) ) described the complete larval development of specimens from the Atlantic and Pacific oceans, respectively.BROSSI-GARCIA & MOREIRA (1996) also described the juvenile instars of this species.STILLMAN & SOMERO (2000) and STILLMAN (2002) studied the thermal tolerance of a number of porcellanids, including P. armatus, and the consequences of this tolerance for their geographic distribution.The associations of these species with sandbanks formed by polychaetes and as a host for an isopod species have also been reported (OLIVEIRA & MASUNARI 1998, 2006, MICHELETTI-FLORES & NEGREIROS-FRANSOZO 1999, BOSA & MASUNARI 2002, MIRANDA & MANTELLATO 2010b).A number of other studies have also focused on the ecological features of the species, abundance patterns, and the morphology of the larval stages (LIMA et al. 2005, MAGRIS & LOUREIRO-FERNANDES 2005, DÍAZ-FERGUSON et al. 2008, TILBURG et al. 2010, MELO JR et al. 2012).
The only information available for P. armatus from the Amazon estuary is the first record of this species in the region (BARROS & PIMENTEL 2001) and descriptions of the stomachs of larvae and post-larvae (LIMA et al. 2005).No published data on the reproductive biology or life cycle of P. armatus are currently available.In the present study we analyzed the variations in the abundance of P. armatus larvae and adults over an annual cycle in an estuary of the Amazon coast (Marapanim).We focused on the periods when the largest numbers of ovigerous females were present, and the timing of larval release in this tropical environment.

MATERIAL AND METHODS
The northern Brazilian coast is 1,200 km long.It encompasses the mouth of the world's largest river in length and freshwater and sediment discharge, as well as the largest continuous tract of mangrove forest (SOUZA FILHO et al. 2009).This region is also characterized by macro-tides, low water transparency and a predominance of sandy and muddy bottoms (SOUZA FILHO et al. 2009, KRUMME & SAINT-PAUL 2010).
The waters of the Marapanim estuary are well mixed, being influenced by the semi-diurnal macrotides (amplitude > 5 m) (BERRÊDO et al. 2008).In the municipality of Marapanim, monthly precipitation is highest in February (760.60 mm) (ANA 2007), which is typical of the region's climate, characterized by a rainy season between February and May (MORAES et al. 2005).
Three distinct climatic phases can be identified in this region (OLIVEIRA et al. 2012): the dry season (August to Decem-ber), transition periods (January and June to July) and the rainy season (February to May).Mean temperatures range from 27°C in the rainy season to 29°C in the dry season (BERRÊDO et al. 2008).The pH becomes higher during the rainy season (from 5.74 to 6.68) because of the intense influx of organic acids produced by the mangroves (24-60 g/kg) (BERRÊDO et al. 2008).
Larvae and adults of P. armatus were collected in the estuary of the Marapanim River, which is located in the Brazilian state of Pará, part of the Amazon coastal zone, and known locally as the "Salgado paraense"."Salgado", which means salty, is a reference to the fact that this region is dominated by the Atlantic Ocean, in contrast with the sector of the coast further west, which is under the influence of the discharge of the Amazon River.
The municipality of Marapanim is located between the Mãe Grande de Curuçá (BRASIL 2002b) and the Maracanã (BRASIL 2002c) extractive reserves, which are sustainable-use protected areas, as defined by the Brazilian National Conservation System (BRASIL 2002a).Both reserves are important for the protection of local mangrove ecosystems and the subsistence and cultural identity of local communities (IBAMA 2006).
The sampling unit was a 0.5 m x 0.5 m quadrat of PVC tubing.The sampling sites were selected randomly during low tide, when the substrate was exposed.The sites were located equidistantly, following a horizontal line perpendicular to the margin of the estuary.The porcellanid crabs present in each quadrat were removed manually and the substratum was sieved (3 mm of mesh size) in order to facilitate collection of specimens.The individuals were kept on ice before being fixed in glycerol, and were subsequently stored at the Laboratório de Biologia Pesqueira e Manejo dos Recursos Aquáticos, Universidade Federal do Pará, Brazil.
During collection, samples of water were obtained from the rock pools in which the adult P. armatus were found, with a 3 mL syringe, for the determination of salinity, using an optical refractometer (Atago).Data on monthly precipitation were obtained during the study period and long-term means were ZOOLOGIA 30 (6): 592-600, December, 2013 taken from the National Water Agency (ANA 2007).The temperature was recorded only for the zooplankton samples.A multiparameter analyzer was used to measure temperatures in situ, during collecting.It was not possible to measure the temperature of the benthic samples due to the reduced volume of water.
In the laboratory, the specimens were counted and identified to species using the identification key available in MELO (1999).Males and females were identified based on OLIVEIRA & MASUNARI (1995).The abundance of benthic specimens were estimated by dividing the number of individuals collected in each quadrat by its area (0.25m 2 ), multiplied by 4 to obtain the number of individuals per m 2 .
Zooplankton samples were obtained from six sites located along the margins of the Marapanim estuary.The western margin, with the town of Marapanim, and three fishing villages (Araticum, Aracumirim, and Alegria), suffers stronger anthropogenic impact.Three sites were selected on this margin, A1, A2 and A3, at distances of 4.7 km (A1-A2) and 6.7 km (A2-A3) from one another.The other three sites, B1, B2 and B3 were established on the eastern margin of the estuary, which is virtually uninhabited.The distance between B1 and B2 was 8.2 km, and that between B2 and B3, 8.7 km (Fig. 1).
We attempted to establish sites B1, B2, and B3 exactly opposite to the corresponding points of profile A. In some cases it was not possible due to the presence of sandbanks or rocky outcrops.These sites also corresponded to the estuary's gradient of salinity, with zone I (A1 + B1) closest to the open sea, zone II (A2 + B2) intermediate, and zone III (A3 + B3) in the innermost portion of the estuary, where salinity is lowest.
The temperature and pH of the water were measured during the collection of specimens from all six sites using a YSI multiparameter analyzer.Water samples were collected in polyethylene flasks for the analysis of salinity in the laboratory, using an optical refractometer (Atago).
Zooplankton samples were collected by horizontal surface trawls using a conical-cylindrical net (200 µm of mesh size) equipped with a flowmeter at the mouth.The specimens were fixed in a formalin solution (4%).A total of 144 samples were collected, two samples from each site over the 12 months of the study period.
Each sample (1 L) was subdivided in a Folsom Plankton Splitter, following the procedure described by BOLTOVSKOY (1981).A volume of 250 mL was established as the standard sample for the analysis of the larvae of P. armatus.These larvae were separated from the other zooplankton, analyzed under a Zeiss optical microscope, dissected, and identified as zoeal stages I or II, based on the studies of GORE (1970GORE ( , 1972)). The

RESULTS
The mean and standard deviation of the environmental parameters recorded in the shallow waters of the Marapanim estuary were 28.6 ± 0.5°C for temperature; 7.8 ± 0.6 for pH; and 19 ± 9.7 for salinity.The mean salinity was 17.5 ± 10.4 in all sites where adult P. armatus were collected (benthic environment).
As salinity did not vary significantly between the shallow water and the rock-pools (H = 1.81, p = 0.18), the monthly median of pooled values are presented in Fig. 4. In this case, significantly higher values (H = 122.18,p < 0.01) were recorded during the dry season, between August (median = 28.5, range = 20-33) and December (median = 30, range = 26-31), with intermediate values being recorded during the transition month of January, and the lowest values during the rainiest months, between February (median = 8, range = 6-8) and July (median = 17.5, range = 13-24) (Fig. 4).
The highest abundance of P. armatus larvae recorded during the study period was 269.35 zoea I/100m 3 and 172.15 zoea II/100m 3 (Table I).Larval density was significantly higher (H = 31.84,p < 0.01) during the drier and transition months, in particular in October, 2006, when the highest median values were determined.Also, we registered two peaks of larval abundance, in December, 2006, andJuly, 2007 (Fig. 5).The monthly variation in density was the opposite of that found for the larvae, with significantly higher densities recorded in January (transition month) and February (rainy month) (Fig. 6).The mean density of P. armatus larvae (zoea I and II) correlated positively with the salinity of the shallow waters of the estuary, whereas that of the adults collected from the rocky outcrops correlated with the salinity of the benthos, but the density of megalopae did not (Table II).There was also a significant correlation between the density of larvae (zoea I and II) and the temperature of the estuarine waters (Table II).
Ovigerous P. armatus females were collected throughout the year, but were less common during the rainy months, from February to May.Larvae were collected only during the dry season and transition months, i.e.August-January, June, and July (Fig. 7), and were absent in the rainy season (February-May).Megalopae were collected between boulders only in August, October, January, and February, with the highest density ZOOLOGIA 30 (6): 592-600, December, 2013 Significant abundance of zoea I (H = 31.84,p < 0.01) occurred in December, 2006, with approximately 500 larvae/100 m 3 , and in July, 2007, with just over 450 larvae/100 m 3 .Peaks of zoea II abundance were recorded in July, 2007, October, 2006, and December, 2006, with approximately 297, 143, and 128 larvae/100 m 3 , respectively (Fig. 8).
In August and October all developmental stages of P. armatus were collected, including ovigerous females (Fig. 8).

DISCUSSION
The densities of P. armatus adults in tropical regions such as the Marapanim estuary in Northern Brazil tend to be higher than those recorded in colder regions.This pattern can be observed on the Brazilian coast.In the present study (tropical) maximum densities of P. armatus adults were observed in August, with 14,960 individuals/m 2 , and in February, with 9,056 individuals/m 2 .MASUNARI & DUBIASKI-SILVA (1998) and OLIVEIRA & MASUNARI (1995), by contrast, recorded maximum densities of approximately 668 and 305 individuals/m 2 , in the south coast (subtropical) and MIRANDA & MANTELATTO (2009) collected a total of 775 specimens over the course of a year on the coast of the State of São Paulo (subtropical).
The higher density of P. armatus adults recorded in tropical regions may reflect the relatively successful recruitment of this species in these environments, although the integrated knowledge on all stages of the life cycle (covering both planktonic and benthic environments) will provide a better understanding of population dynamics of this decapod species (DÍAZ-FERGUSON et al. 2008).In the present study, all the developmental stages of P. armatus (zoea I, zoea II and megalopa) were considered, allowing the identification of the period of most intensive reproductive activity and also the developmental strategies of this species.Petrolisthes armatus reproduces throughout the year in the Marapanim estuary, in the Amazon coastal zone, although ovigerous females were most common in June (at the end of the rainy season) and August (dry season).All life stages (zoea I and II, megalopa and adults) can be found in the estuary.The abundance of adults was high throughout the study period (August 2006through July 2007), which implies that the study area is important for the development of this decapod species, and therefore for its conservation.
The spawning of decapod species in tropical estuaries tends to be continuous throughout the year, in contrast with temperate estuaries (DITTEL & EPIFANIO 1990).FRANSOZO & BERTINI (2001) identified some porcellanid species that present seasonal reproduction, and others that breed throughout the year.Petrolisthes boscii (Audovim, 1826), P. rufescens (Heller, 1861), P. elongatus (H. Milne Edwards, 1837), andP. vanderhorsli Haig, 1956 are all known to reproduce seasonally in some parts of the world (LEWIS 1960, WEAR 1965, AHMED & MUSTAQUIM 1974).GEBAUER et al. (2007) recorded a distinct pattern in P. laevigatus in southern Chile, with an 11-month breeding season, starting at the end of the summer, and ending in the middle of the subsequent summer.
The Marapanim estuary offers favorable temperature conditions for reproduction throughout the year, with mean temperatures of between 27°C and 30°C.This intensive breeding activity enables P. armatus to colonize the region successfully, with densities of 41,280 individuals/m 2 .DÍAZ-FERGUSON & VARGAS-ZAMORA (2001) collected a total of 15,382 P. armatus crabs in the tropical Gulf of Nicoya in Costa Rica, between December 1997 and November 1998, with a maximum density of almost 100 individuals/m 2 .Other decapod species also reproduce throughout the year in the Marapanim estuary, such as the thalassinidean shrimps Lepidophthalmus siriboia Felder & Rodrigues, 1993and Upogebia vasquezi Ngoc-Ho, 1989(OLIVEIRA et al. 2012, SILVA & MARTINELLI-LEMOS 2012).
The reproductive period of the species of the suborder Pleocyemata is normally defined according to the abundance of ovigerous females in different periods of the year (SANT 'ANNA et al. 2009), an approach also used for P. armatus in southeastern Brazil (MIRANDA & MANTELATTO 2009).Although two peaks of more intense reproductive activity were recorded -the first in the dry season (December), when the peak in the abundance of stage I zoea was recorded (483 larvae/100 m 3 ), ovigerous females of P. armatus were observed throughout the year in the Marapanim estuary.The abundance of stage II zoea was also above 100 larvae/100 m 3 , while megalopae were recorded in the subsequent months, January (dry-rainy transition) and February (early rainy season), indicating the beginning of the recruitment period.
During the rainy months (February through May), the density of ovigerous females was lower, and no larvae were collected (excepted on April).Our hypothesis is that the absence of larvae during this period is due to the fact that they do not survive when salinity is low.We do not believe that they are dragged out from the estuary to offshore waters.Larval development from hatching to the megalopa stage does not exceed one month in either the Atlantic (GORE 1970) or the Pacific oceans (GORE 1972), which suggests that the larvae hatching between February and May are unable to survive when salinity is lower than 20.The low density of zoea I in April may represent the small number of larvae that are able to survive the prevailing conditions, reinforcing the idea that they remain in the estuary.
The larval abundance increases in the end of the rainy season, when the environmental conditions begin to favor larval development, which means more saline waters.Salinity is a key factor in the structure and distribution of decapod larvae in tropical estuarine environments (ANGER 2003, MAGRIS & LOUREIRO-FERNANDES 2011).The dispersal and recruitment of larvae of estuarine crustaceans are strongly influenced by salinity (O'CONNOR & EPIFANIO 1985).Salinity appears to be the principal factor influencing the breeding activity of P. armatus in the study area, as shown by the lower densities of ovigerous females and reduced numbers of larvae collected during the rainiest months (February-May), when salinity was significantly lower than in the dry season.
Similarly, the density of the thalassinid L. siriboia and U. vasquezi larvae in the Marapanim estuary is significantly higher during the dry season (OLIVEIRA et al. 2012).The same pattern has also been recorded in a tropical estuary in Costa Rica, where the abundance of P. armatus larvae was significantly higher during the dry season, when the salinity of the water was higher (DÍAZ-FERGUSON et al. 2008).In an estuary in the South Atlantic, TILBURG et al. (2010) also recorded significant variation in the density of P. armatus larvae in relation to salinity.
In other regions of the world, in particular the temperate zone, temperature is the principal factor influencing the reproductive patterns of decapod species, including porcellanids (HERNÁEZ-BOVÉ 2001, EMPARANZA 2007, HOLLEBONE & HAY 2007).As considerable seasonal fluctuations in salinity occur in coastal and estuarine environments, the decapod species that inhabit these areas adopt different strategies of development according to their physiological constraints (STRATHMANN & STRATHMANN 1982, ANGER 2003, 2006).Some species, such as Ucides cordatus (Linnaeus, 1763) and Uca vocator (Herbst, 1804), export their early larval stages to offshore waters, which implies a more ample dispersal strategy (DIELE & SIMITH 2006, SIMITH & DIELE 2008, SIMITH et al. 2012).Some estuarine crabs also increase their swimming ZOOLOGIA 30 (6): 592-600, December, 2013 activity at higher salinities to avoid being removed from the estuary (QUEIROGA & BLANTON 2005).In northeastern Brazil, MELO JR et al. (2012) found that P. armatus larvae were more concentrated in the midwater and surface during the flood tide, thus avoiding being removed to more internal regions.By contrast, these larvae were more concentrated at the bottom during the ebb tides, avoiding exportation, which suggested that P. armatus reproduces and spends its life cycle on the inner shelf, rather than the outer shelf (MELO JR et al. 2012).
DITTEL & EPIFANIO (1990) found all larval stages of Pinnotheres spp. in the plankton of Gulf of Nicoya and suggested that this species reproduces in this region and that larvae are retained in the system.At Marapanim, P. armatus remains in the estuary throughout its life cycle, following the same reproductive strategy described by MELO JR et al. (2012) for this species on the northeastern coast of Brazil.This hypothesis has yet to be tested experimentally, although the presence of all the developmental stages of P. armatus throughout the year at different locations within the Marapanim estuary reinforces the conclusion that this species passes through its larval phases on the inner shelf.Furthermore, P. armatus breeds throughout the year in this region and salinity is a key factor for its development.

ABSTRACT.
Petrolisthes armatus (Gibbes, 1850) is a porcellanid crab with a wide geographical distribution.In the present study we analyzed variations in the abundance of P. armatus adults and larvae over an annual cycle in the Marapanim estuary of the Amazon coastal zone, in the northeastern portion of the state of Pará, Brazil.Particularly, we focused on the presence of ovigerous females and timing of larval release, with the aim of elucidating reproductive patterns in a tropical estuarine system.The mean density of P. armatus larvae (zoea I and II) correlated positively with the salinity of the shallow waters of the estuary, whereas the abundance of adults correlated with the salinity registered in water samples collected from the benthic environment.There was also a significant positive correlation between larval density (zoea I and II) and water temperature.Ovigerous females were captured throughout the study period, from August 2006 to July 2007, but were more abundant in June and less abundant during the rainy months, between February and May.Larvae were only present during the dry season and transition months (June to January), and were absent during the rainy season (February to May).Petrolisthes armatus reproduces throughout the year in the Marapanim estuary and all developmental stages of this species (zoeal stages I and II, megalopae and adults) are found in the estuary.The results indicate that the study area is an important environment for the reproduction of this decapod.KEY WORDS.Decapoda; macrobenthos; reproduction; zooplankton.
volume of water filtered during the trawls was based on V = A*R*C, where V = the volume of water in m 3 , A = opening of the net in m 2 (for a 0.5 m diameter opening, A = 0.19625 m 2 ), R = number of rotations recorded on the flowmeter before and after each trawl (FF-FI), and C = standardization factor following calibration of the flowmeter (C = 0.32).The density of P. armatus larvae was calculated by D = n/V, where n = number of larvae collected during the sampling, and V = volume of water filtered by the net (m 3 ), expressed as larvae per 100 m 3 .The variation in the abiotic factors and the density of the larvae and adults of P. armatus during the study period -August, 2006, through July, 2007 -was evaluated using the Kruskal-Wallis nonparametric analysis of variance, given the lack of normality or homocedasticity of the data, even after the data had been transformed (logarithmic and square root).The possible relationship between the density of P. armatus and abiotic variables was evaluated using the Spearman's correlation coefficient.All analyses were run in BioEstat 5.0 ®(AYRES et al. 2007), considering ␣ = 0.05.

Figure 1 .
Figure 1.Sampling sites ( ) where the larvae and adults of Petrolisthes armatus were monthly collected at Marapanim estuary, northern Brazilian coast, between August 2006 and July 2007.

Table I .
Abundance and density of P. armatus at different stages of development recorded in the Marapanim Estuary, Pará (Brazil), between August 2006 and July 2007.

Table II .
Spearman coefficients for the correlation between the density of P. armatus (larvae and adults) and abiotic factors.Significant values in bold script (p < 0.05).