Reproductive biology of Plagioscion magdalenae ( Teleostei : Sciaenidae ) ( Steindachner , 1878 ) in the bay of Marajo , Amazon Estuary , Brazil

Plagioscion magdalenae (pacora) is a commercially important benthopelagic sciaenid and widely distributed in the Amazon River basin. The present study describes the reproductive biology of this species in the bay of Marajo, Amazon Estuary, Brazil. The gonadal development stage, age and size at first sexual maturity (L 50 ), sex ratio, and reproductive strategy were determined. The data were collected bi-monthly from December 2005 to October 2006. A total of 251 specimens were examined, with the total length (TL) ranging between 220 and 590 mm. The weight-length relationship for females, males and grouped sexes was highly significant, showing a positive allometry. The L 50 was of 279 mm for grouped sexes, with 305 mm and 269 mm TL for females and males respectively. The sex ratio for the total number of individuals favored the males (2.02 males: 1 female). Macroscopically, the gonads were classified as immature, maturing, mature and spent. Considering the macro and microscopic evaluation of the gonads, an extended spawning period, mainly in August to February, was observed.


Introduction
The Amazon Estuary, the region where the Amazon and Tocantins Rivers flow into the Atlantic Ocean, is a complex aquatic environment where the Amazon River discharges between 1.0 and 2.8 x 10 5 m 3 /s of fresh water (Kineke & Sternberg, 1995) and an average of 3-3.5 million metric tons of sediments in the sea (Meade et al., 1985).This region is characterized by its high biological productivity, which supports a substantial fish biomass (Okada et al., 1998).
It is estimated that 24% (approximately 125,000 t) of annual Brazilian marine and estuarine catches are derived from its northern coastal and estuarine waters (IBAMA, 2008).The bay of Marajo is part of the estuary system of the Amazon coast.It is one of the most important artisanal fishing areas and represents an important feeding and reproduction area of commercially important species (Barthem, 1985;Isaac & Barthem, 1995).
Among the commercially important families in Brazil' northern region, the Sciaenidae family stands out considering their commercial and subsistence importance, whose population parameters, in particular the reproduction aspects, are largely unknown, especially for freshwater species of the genus Plagioscion.Plagioscion magdalenae is known from the Magdalena River and Amazon River basins in Colombia and Brazil, respectively (Casatti, 2005).This species has heretofore been called Plagioscion surinamensis (Bleeker 1873), as first noted by Jordan & Eigenmann (1889).The only information available on the species include general aspects of the genus (Campos, 1942;Casatti, 2003Casatti, , 2005) ) and population parameters obtained in the middle Xingu region, Amazon (Camargo & Giarrizzo, 2009).Plagioscion magdalenae is often caught in the estuarine region of the State of Para, especially in the bay of Marajo, representing 26% of this bay total fishery catch (Oliveira, 2007).
Evaluating the reproductive life history of a commercially exploited species is an important research element of fisheries to provide management options for the protection of spawning stocks (Begg, 1998).It is specifically important to estimate the size at gonadal maturation relative to capture size, in order to minimize the catch of immature individuals (Sadovy, 1996).
This paper describes the reproductive biology of P. magdalenae in the bay of Marajo Amazon Estuary, in order to contribute to the region's fishery management.

Material and Methods
The bay of Marajo, within the Amazon Estuary complex, is located south of the Island of Marajo (Fig. 1) and receives the discharge of the Tocantins River (El-Robrini, 2001).Due to the seasonal differences in the rain fall in this area, the discharge of the river greatly varies according to the period of the year and the contact river-sea may be moved in approximately 200km (Egler & Schwassmann, 1962).In the winter rainy season (January to June), this zone is moved offshore, primarily due to the strong river plume and throughout the dry summer months (July to September), the opposite pattern occurs and high salinity water flows into the inner estuary (Egler & Schwassmann, 1962;Schwassmann et al., 1989).These seasonal variations influence the biodiversity of the estuary: during the rainy season, freshwater species dominate and throughout the dry season, marine species dominate the local ichthyofauna (Barthem, 1985).
The collections were conducted bi-monthly, from December 2005 to October 2006.Sample (total of 251 specimens of P. magdalenae) was obtained from the commercial fleet operating in the bay of Marajo.The specimens were captured using drift nets of 50, 60 and 70 mm mesh size (stretched mesh size).Fish were then placed in icepacked coolers and transported to the laboratory where they were measured (total length -TL, mm) and weighed (total weight -TW, grams).Gonads were removed, individuals were hence sexed and ovary weight determined to the nearest gramme.The maturity stage was determined macroscopically, utilizing four ovarian stages: immature (A), maturing (B), mature (C) and spent (D) and microscopically considering four ovarian stages: stage I, stage II, stage III and stage IV.Both methodologies followed the maturity key of Vazzoler (1996).
The microscopic analysis was performed on 105 ovaries.Each gonad was divided into three sections and only the middle portion was fixed in Bouin's solution for 24 h.Samples were then dehydrated in alcohol, diaphanized in xylol and embedded in paraffin.Sections (5 mm thick) were stained with haematoxylin and eosin, analyzed and photographed using an Olympus (Tokyo, Japan) CH30 microscope.
The relationship between total length and total weight was established by non-linear regression represented by the mathematical expression TW = a x TL b , and adjusted by the least-squares method, with a confidence level of ±95% (p < 0.05) where a and b are parameters of the equation (Sokal & Rohlf, 1987).Coefficient b was compared between males and females by the Student's t test (Zar, 2009).The relation is said to be isometric when b = 3 (Froese, 2006) and the weight of the fish is proportional to the cube of its length, based on its dimensional equality (Harish Kumar et al., 2006).
The sex ratio was obtained for the total period, bi-monthly and by total length.The χ² test with Yates correction (Snedecor & Cochran, 1980) was applied in order to evaluate possible sex ratio differences.The reproductive period of P. magdalenae was defined by the ovaries and determined by bi-monthly frequency analysis of the maturity stages and the variation of mean values of the gonadosomatic relationship (ΔRGS) and the condition factor (ΔK).The values of ΔRGS better illustrate the physiological variation of the gonads and, the breeding season is given by the highest value of this index followed by its decrease (Vazzoler, 1996).The condition factor (ΔK) expresses the amount of reserves which is transferred to the gonads.The period immediately anterior to its peak corresponds to the breeding period (Isaac & Vazzoler, 1983).
The gonadosomatic relationship (ΔRGS) is the difference between RGS and RGS* (gonadosomatic indexes), given according to the formulae: Where GW is the ovary weight and BW (body weight) is TW -GW.
The condition factor (ΔK) is the difference between K and K*(condition factor indexes), given according to the formulae: The distribution differences of the bi-monthly values for ΔRGS and ΔK were tested using the Kruskal-Wallis nonparametric method (Kruskal & Wallis, 1952).
To analyze the size at first maturity (L 50 ) (length at which 50% of the individuals attain gonadal maturity for the first time), the percentage of adults (ovary stages B, C and D) by length was calculated and considered as dependent variable (Y), and the total length as the independent variable (X).Next, these values were adjusted by the least-squares method to a logistic curve which is given according to the formula: P = 1/(1 + exp [-r (TL -L 50 )], where P = proportion of mature individuals, r = slope curve and L 50 = mean length of sexual maturity are parameters of the equation to be estimated by the model.

Results
Of the 251 specimens of P. magdalenae, 168 were males and 83 were females.The length and weight relationship for females, males and grouped sexes was highly significant (p < 0.01) (Table 1).A positive allometry was recorded for males, females and grouped sexes.The regression coefficient (b) was significantly different between males and females (t = 1.43, p < 0.05).
The sex ratio, considering the total number of individuals, favored males (2.02 males: 1 female) (p < 0.05).Males outnumbered in most length classes and in the bi-monthly periods (p < 0.05).Significant differences (p < 0.05) for males were observed in the classes included between 360-417, 447-475 and 563-591 mm TL and in almost all months (Tables 2 and 3).
The L 50 for P. magdalenae, considering both males and females, was of 279 mm TL.Males reach maturity at sizes greater than those recorded for females, 305 and 269 mm TL, respectively (Fig. 2).
The ovaries of P. magdalenae are approximately equal size pairs and have different patterns of color, size, consistency and vascularization, depending on the ovary development stage (Table 4).
The following stages of oocyte development were identified based on the analysis of the histological crosssections and the germ cell types were described following the sequence of the oogenesis and folliculogenesis.Immature ovaries exhibit numerous oocytes in the chromatin nucleolus (stage I) and perinucleolar stage (stage II) (Fig. 3).These cells are clustered and inserted into the ovuligera lamellae.The maturing ovaries have a large amount of oocytes in the perinucleolar stage (stage II) and cortical alveoli oocytes (stage III).The mature ovaries present four types of oocytes in different development stages: chromatin nucleolus (stage I), perinucleolar (stage II), cortical alveoli (stage III) and fully vitellogenic (stage IV), which dominates (Fig. 3).The parenchyma of the spent/spawned ovary is characterized by the presence of postovulatory and atretic follicles, and oocytes in the chromatin nucleolus and perinucleolar stages, indicating occurrence of the ovary reorganization before a new reproductive cycle.
The mean values of ΔRGS and ΔK for P. magdalenae were not significantly different between the months (Kruskal-Wallis Test; p > 0.05) and an extended spawning period was reported, mainly in August to February (Figs.4a and 4b the year (Fig. 4c).This pattern was confirmed by microscopic analysis, which registered the occurrence of oocytes in various stages of development, indicating the release of gametes throughout the year, which points out an asynchronous spawning.

Discussion
In the Amazon, fishing is an extremely important socioeconomic activity, accounting for a source of food and income for much of the population.This region is renowned for its rich fishery resources.Notwithstanding, the biological information accumulated in recent decades is still limited and concentrated in the central Amazon basin (Isaac et al., 2000) and more recently in the Xingu region (Camargo & Guillard, 2009).Regarding the marine and estuarine species, information on the population dynamics is restricted to areas of the Amazon Estuary, considering a few species, such as Scomberomorus brasiliensis, Macrodon ancylodon, Cynoscion acoupa, Sciades parkeri, Cathorops spixii, Stellifer rastrifer, S. naso and Aspredo aspredo (Camargo & Isaac, 1998;Espírito Santo & Isaac, 1999;MMA, 2006).With regards to freshwater sciaenids, some studies on ecology and population dynamics are mainly available for the cogeneric species, the white hake, P. squamosissimus (Rodrigues et al., 1988;Braga, 1997Braga, , 2001;;Castro, 1999;Carnelós & Benedito-Cecilio, 2002;Stefani et al., 2005;Bennemann et al., 2006).In relation to P. magdalenae, only some information is available on the population parameters in the middle Xingu (Carmargo & Giarrizo, 2009).
The length-weight relationship has important applications in terms of fishery biology.It may be useful to calculate a population's production and biomass (Anderson & Gutreuter, 1983) and check for seasonal variations in growth (Richter et al., 2000).It is also relevant for comparing the morphology between species or populations from different habitats and/or regions (Gonçalves et al., 1997).Positive allometry (b > 3) was observed for the length weight relationship of P. magdalenae as obtained by Camargo & Giarrizo (2009) in middle Xingu.Similar results were attributed to P. squamosissimus in the surroundings of Mosqueiro Island, bay of Marajo (Viana et al., 2006).Positive allometry was also dominant for most fish species in Brazilian estuaries, including those on the Amazon Estuary (Joyeux et al., 2009).This relatively more gain in weight in relation to length, are possible under certain conditions such as fish farming and stress-free environments (Sarkar et al., 2009), but may also be attributed to a high availability of food arising from the structural complexity of estuaries (Kennish, 1990;Knox, 1986) as it was reported for most species that inhabit Guajará Bay, adjacent to Marajo Bay, which use the environment as feeding grounds (Viana et al., 2010).
In general, for fishes, males tend to mature earlier and at a smaller size, whereas females grow larger, mature later and tend to invest relatively more resources in their reproduction    (Bromley, 2003).This is recorded for most marine sciaenids (Lowerre- Barbieri et al., 1996;Hutchings et al., 2006).However, the opposite was recorded for freshwater sciaenids, as observed in this study for P. magdalenae (305 mm and 269 mm for males and females, respectively), and obtained by Castro (1999) and Marciano et al. (2005) for P. squamosissimus.This may be an specific characteristic of the genus Plagioscion or a diferential response to fishery exploitation.Lower values of length-at-first-maturity (222.5 mm TL) were obtained by Camargo & Lima (2007) in middle Xingu River, although higher estimates were also obtained in other study in Xingu River by Camargo & Giarrizzo (2009).
The sex ratio represents important information to characterize the population structure, in addition to providing specifics for studying other aspects, such as the evaluation of reproductive potential and stock size estimates (Vazzoler, 1996).According to Nikolski (1969), the sex ratio in most species is 1:1, varying according to food availability and favoring females when the food supply is adequate.For several species of the lower Amazon, a predominance of females is recorded, mainly due to the fact that males go deeper in the middle of the river, while females migrate close to the shore (Isaac et al., 2000).However, in this study, the sex ratio showed significant differences for P. magdalenae, indicating a male predominance.This difference may be related to sex segregation for feeding since it has been reported that males, in Marajo and Guajará Bays, mainly fed on crustaceans and females on teleosts (Barbosa, 2009;Luz, 2009).Wallace & Selman (1981) describe an asynchronous spawning by the concurrent presence of oocytes at all stages of development in fully-developed and partially spent ovaries.This was observed for P. magdalenae in the study area by the presence of oocytes in stages I and II, and postovulatory follicle and atretic oocytes, which indicates the elimination of mature oocytes.Macroscopically, the asynchronous spawning was suggested and an extended spawning period was reported, mainly in August to February.Asynchronous spawning were described for other sciaenids, as is the case of Micropogonias furnieri (Lowe-McConnell, 1999), Umbrina coroides (Gomes & Guzmán, 1997) and Stellifer rastrifer (Chaves & Vendel, 1997;Peres-Rios, 2001).
Overall, the histological characteristics of the ovaries of the species under study are similar to other sciaenid species (Valentim, 1998;Santos, 2006).During the spent/spawned stage, P. magdalenae presents postovulatory and atretic follicle as observed in P. squamosissimus (Valentim, 1998).In this study four gonadal stages were used according to the proposed by Vazoller (1996) and different from the classification adopted by Valentim (1998), who uses a seven-stage classification.This fact is due to a lack of standardization in the description of the ovarian development stages, which difficult the interpretation of the reproductive process in different species.
The Marajo Bay is one of the most important fishing areas of artisanal fishing in the Amazon Estuary and represents an important feeding and breeding area of commercially important species (Barthem, 1985;Isaac & Barthem, 1995).The presence of P. magdalenae individuals in all gonadal development stages indicates that the species complete their entire life cycle in the study area, which reinforces the importance of the area as breeding and nursery grounds, since immature individuals were also frequently observed.
Given that P. magdalenae is a commercially important species without any regulation and considering that the bay of Marajo is one of the most important fishing areas of the Amazon Estuary, regulation and conservation measures should be considered.The prolonged spawning cycle does not favor the implementation of a closed season for the species since protecting only a little portion of the spawning stock does not guarantee the maintenance of the population.Others management measures as controlling the minimum landing size, considering the size of first sexual maturity, is a reasonable measure to be implemented by controlling the mesh nets used, and also putting into practice an environmental education project with the fishing community, in order to promote greater awareness of this species.Protected areas and/or periods should also be taken into consideration in the case of the bay of Marajo.
the coefficient obtained by the length-weight relationship.

Fig. 4 .
Fig. 4. Mean values of ΔRGS (a) and ΔK (b) and frequency of bi-monthly percentage of the maturational stages for the females (c) of Plagioscion magdalenae in the bay of Marajo.

Table 1 .
Parameters of the length-weight relationship for males, females and groups sexes of Plagioscion magdalenae, collected in the bay of Marajo.

Table 2 .
Sex ratio by bi-monthly period for Plagioscion magdalenae in the bay of Marajo.*Significant at 5% level.

Table 3 .
Sex ratio by total length class for Plagioscion magdalenae in the bay of Marajo.*Significant at 5% level.

Table 4 .
Macroscopic description of the ovarian stages of female Plagioscion magdalenae.