Reproductive pattern of the South American endemic shrimp

.

An accurate description of the spawning stock and its reproductive dynamics are central for the understanding the recruitment processes and the development of spawning stock/recruitment relationships in penaeid species (Crocos and van der Velde 1995). Spawning season for penaeid shrimps have been determined in previous studies by changes in the percentage of ripe females in the catch (O'Connor 1979, Crocos and Kerr 1983, García 1985, Crocos 1987, Bauer and Vega 1992, Bauer and Lin 1994, Crocos and Coman 1997, Minagawa et al. 2000, Crocos et al. 2001. Shrimp fisheries in southeastern Brazil target the most profitable species, such as the pink shrimp Farfantepenaeus brasiliensis (Latreille, 1817) and F. paulensis (Pérez-Farfante, 1967), the white shrimp Litopenaeus schmitti (Burkenroad, 1938), and the seabob shrimp Xiphopenaeus kroyeri (Heller, 1862). The increase of the fishing fleet and the decrease of landings of commonly exploited species have contributed to the expansion of the Artemesia longinaris Bate, 1888 fishery (D´Incao et al. 2002, Costa et al. 2005. The geographical distribution of this species is restricted to the western Atlantic, from Rio de Janeiro (23° S), Brazil, to Rawson (43°S ), Argentina. This shrimp remains in the marine environment throughout its life cycle (Boschi 1969).
One of the main objectives concerning the study of the reproductive ecology of benthic invertebrates is to test latitudinal trends in the timing of reproductive activity and recruitment of juveniles. This approach can be used to test the effect of a given environmental stimulus (proximal factors) and to predict selective pressure (ultimate factors) accounting for a particular seasonal pattern of spawning and recruitment of juveniles. This methodology may also provide a basis to predict possible changes resulting from natural or anthropogenic disturbances (Bauer and Vega 1992).
In organisms with a short life cycle that are subjected to high fishing pressure, biological processes (environmental factors and food resources) that have occurred in one year define the stock's abundance in that same year. Interannual variability of temperature can cause changes in the reproductive pattern of a species in the same zone (Leal-Gaxiola et al. 2001).
São Paulo State is located between 23º20' S and 25º15' S latitude, and has a subtropicaltropical climate. There are only a few accounts of the reproduction of dendrobranchiate shrimps in this region. Almost all the published information regards populations of the species X. kroyeri (Rodrigues et al. 1993, Nakagaki and Negreiros-Fransozo 1998, Fransozo et al. 2000, Castro et al. 2005 and Rimapenaeus constrictus (Stimpson, 1874) . Little is known about the reproductive periodicity of A. longinaris off the Brazilian coast. , Castilho (2004), and Costa et al. (2005) studied aspects of its spatial and temporal distribution. Studies have been conducted in Argentinean waters, where this species is important not only for commercial fisheries, but also as a major component of the marine food web (Boschi 1969, Boschi and Scelzo 1977, Gavio and Boschi 2004.
The objective of this study was to determine the size at morphological and physiological sexual maturity of A. longinaris, to analyze the interannual and seasonal variations in spawning intensity and recruitment of juveniles during five and a half years off the coast of São Paulo, and to determine the relationship between its breeding season and the seabottom water temperature.

MATERIALS AND METHODS
Shrimps were collected monthly from January 1998 to June 2003 along the northern coast of São Paulo in the Ubatuba (23º30'S, 45º09'W), Caraguatatuba (23º37'S, 45º25'W) and São Sebastião regions (23º48'S, 45º23'W). In each month, fourteen 2-km transects were trawled for 30 min, at depths between 5 and 45 m. A shrimp boat equipped with two doublerig nets (mesh size 20 mm; cod end 15 mm) was used for trawling.
On each transect, salinity and temperature (bottom and surface water), depth, organic matter content, and grain size of sediments were monitored. Detailed descriptions of the sampling methods and analysis of environmental factors for the period are available elsewhere (Costa et al. , 2005. Because this species inhabits the bottom, we used only the bottom temperature measurements for analysis. All captured individuals were sexed and measured (to nearest 0.1 mm). Carapace length (CL) was chosen as the size dimension, corresponding to the distance from the orbital angle to the posterior margin of the carapace. Size-frequency distributions were constructed separately to estimate the seasonality for males and females using 1.0-mm CL size intervals.
The relative frequency (%) of adults in each size class was plotted, and the logistic function ) ( 50 1 1 C L C L r e y    was fitted to the data. CL 50% corresponds to the size at which 50% of the individuals are considered mature/ adults, and r is the slope of the curve. Fitting was done by the least-squares method (Aguillar et al.1995, Vazzoler 1996, requiring a size-range overlap of adults and youngs of at least two size classes. The shrimps were therefore arranged in 1.0-mm size intervals. The reproductive condition of females was determined by macroscopic observation of the degree of ovarian development (color and volume occupied by the gonads) according to Castilho (2004) and . Ovaries categorized as immature varied from thin, transparent strands to thicker strands. Ovaries of adult females were much larger and thicker, and colored white (spent), light green (developing), or green to olive green (ripe).
Breeding intensity of the population was estimated as the percentage of mature females for each month and season. We chose the smallest size class which contained a female with a mature ovary as the lowest size limit for adult females as recommended by Bauer and Lin (1994). The reproductive status of males was assessed by examining the shape of the petasma, which is fused in adult individuals (Pérez-Farfante 1969, Boschi andScelzo 1977).
Recruitment was defined as the percentage of juveniles of the total number of individuals in each month and season. Juvenile size classes were defined separately for males and females. Spearman's correlation coefficient was used to test the null hypothesis of no association between temperature and (a) the frequency of recruits and (b) the frequency of breeding females. In this analysis, monthly temperature and abundance values were plotted. Student's T-test was used to test size differences between male and females. The data were log-transformed prior to the analysis to test for normality (Zar 1999).

RESULTS
During this study, a total of 15 839 shrimps (10 288 females and 5 551 males) were captured. Female mean size was 14.4 ± 3.0 mm CL, ranging from 5.8 to 27.3 mm. The mean size of the males was 12.1 ± 3.0 mm CL, ranging from 5.5 to 21.9 mm CL. A twotailed Student's T-test indicated significant size differences between sexes (p< 0.05).
Size at sexual maturity was estimated at CL 50% = 11.0 mm for males and CL 50% = 13.4 mm for females. The largest female with immature gonads measured 10.9 mm CL; the smallest female bearing developed gonads was 9.9 mm. The largest immature male measured 10.5 mm; the smallest male with fused petasma was 8.9 mm.
The reproductive condition of females and males is illustrated in size-frequency distributions from seasonal collections (Fig.  1). Annual cycles showed an increment of carapace size with modals of smaller sizes in summer and larger sizes in spring. For immature individuals, probable growth was estimated from summer until winter, when juveniles were few or absent. New recruitment was observed in spring. Larger specimens were obtained in the seasons with lower water temperatures, such as spring and winter 2001 ( Fig. 1, 2).
Mature females occurred in every season, with higher percentages in summer, principally in 2000 and 2003. Peaks of females with spent ovaries were obtained in autumn (2001,2002,2003) and winter (1998,1999,2000) (Fig. 3). The abundance of mature females generally followed the same pattern as the bottom-water temperature (Fig. 4). In each year of the study, high percentages of females with mature gonads were recorded in months with temperatures near 22ºC. The bottom temperature varied between 17ºC to 28ºC and correlated significantly with the relative frequency of breeding females during January 1998 and June 2003 (Spearman, p< 0.001).
Temperature had a predominant influence in the months of warmer water (> 23ºC), when lower percentages of mature females were recorded. The smallest percentage of females with mature ovaries was observed in 1998, with maximum water temperature of 27ºC and a minimum of 22ºC. The temperature then decreased in November 1998, and during  December 1998 and January 1999 (spring and summer) a peak of spawning females occurred. Elevated percentages of mature females were observed in January 2000 (summer, 22.8ºC), and this proportion was maintained in the subsequent three months (temperature = 20.6, 20.7 and 21.8ºC). The temperature increased in May 2000 (22.9ºC) when the number of mature females declined. In the other seasons, there were similar fluctuations in the percentages of spawning females and in bottom temperature (Fig. 2, 3, 4).
Recruitment estimates for A. longinaris based on the percentage of juvenile shrimps showed no evident seasonal patterns. The juveniles were found in low numbers during all months over the entire study period, except in November 2002 when immature individuals comprised 55% of the catch (Fig. 5).

DISCUSSION
Sexual dimorphism according to size occurred in A. longinaris: females grew larger than males, indicating differential growth rates between sexes. According to Boschi (1969), sex-related body-length differences between males and females are a general rule among penaeids. Gab-Alla et al. (1990) and  reported slower growth rates in males and suggested that reproductive processes are related to this difference.
In Argentina, Boschi (1997) found shrimps with different carapace lengths in each latitudinal region: for females 37 mm in Chubut (43ºS), and 29 mm in Mar del Plata (37ºS); for males 27 mm in Chubut and 24 mm in Mar del Plata. In the present study (23ºS), mean lengths were smaller: 27 mm for females and 21 mm for males. In Brazil the same latitudinal effect was observed in the sexual maturity of A. longinaris: in Rio Grande do Sul, females reached sexual maturity at 17 mm carapace length (Dumont 2003), 4.6 mm larger than the maturity size observed in the present study (CL 50% = 13.4 mm). Bauer (1992) compared the longevity and size of females of Sicyonia spp. in tropical, subtropical and cool-temperate regions. Females of the larger species of Sicyonia (cool-temperate region) live at least two years, so that individual females have the opportunity to breed during a period   70  60  50  40  30  20  10  0   28  27  26  25  24  23  22  21  20  19  18  17 of the year which might be most favorable for larval development and settlement. Thus, there is selection for seasonality in breeding patterns in these species. The small tropical sicyoniids live less than one year, and may settle and grow to sexual maturity at any time during the year. The same pattern may occur in A. longinaris when comparing growth and reproduction of populations in a tropical region (São Paulo state) and a more southern cool-temperate region (Mar del Plata). The presence of juveniles and mature females throughout the year, as observed in January 1998 and June 2003, suggests that this species breeds continuously. However, there is evidence for more intense reproductive activity in summer than in spring. Otherwise, there were no seasonal trends in recruitment of juveniles. The seasonal variation in the frequency of females bearing mature ovaries was similar as reported for R. constrictus and X. kroyeri (Costa et al. 2004, Nakagaki andNegreiros-Fransozo 1998, respectively). Vega-Pérez (1993) reported that during spring and summer the ocean chlorophyll content (phytoplankton production) was usually higher when the SACW (South Atlantic Central Water) intruded into the region during upwelling events. The phytoplankton production favored a subsequent production of herbivorous zooplankton resulting in the highest density of plankton organisms during summer (Pires-Vanin and Matsuura 1993). This variation coincides with our observations of the changes in frequencies of females with mature gonads, suggesting that food availability for larval protozoea (indicated by phytoplankton production) may be an important selective factor shaping the seasonal breeding pattern in this species.
In cool-temperate regions such as the Mar del Plata (37ºS), Christiansen and Scelzo (1971) observed highly seasonal breeding and spawning in A. longinaris with high percentages of mature females from October to January, whereas breeding apparently ceases during the rest of the year. For this species, the classical paradigm of continuous reproduction in the tropics and seasonal reproduction at higher latitudes, increasingly restricted in time with an increase in latitude, is supported by the results of the present study.
A. longinaris can be considered as a typical cold-temperate species. Boschi (1969), Ruffino and Castello (1992),  and Costa et al. (2005) stated that the species occurs in a temperature range from 15 to 21ºC. Off southeastern Brazil, the upwelling current of the SACW is responsible, among other changes in bottomwater characteristics, for the cooling of coastal water during summer (Castro-Filho et al. 1987). The intrusion of the SACW was detected in this study during spring and summer, when mean water temperatures decreased (Negreiros-Fransozo et al. 1991, Mantelatto and Fransozo 1999, Bertini et al. 2001. This may explain the lack of a significant correlation between the relative frequency of breeding females and bottom-water temperature. The shrimps A. longinaris and P. muelleri both migrate to the northern coast of São Paulo during the intrusions of cold SACW (Costa et al. , 2005. We assume that the largest individuals captured in spring had migrated from other stocks located at higher latitudes. Temperature was a specific environmental stimulus for the ovary development cycle, with a significant correlation between water temperature and the frequency of mature females. Petriella and Bridi (1992) observed the same relation in Mar del Plata, where high percentages of mature females of A. longinaris were present in months with higher temperatures. Christiansen and Scelzo (1971) and Petriella and Bridi (1992) demonstrated that spawning in A. longinaris in Mar del Plata was usually centered in the warmer months with sea-water temperatures higher than 20ºC. However, off a tropical coast such as São Paulo state, a smaller percentage of spawning females was found in the warmer months (March) when the mean bottom water temperature of 28ºC is 8ºC higher than at Mar del Plata.
Insemination in species whose females have a closed thelycum occurs immediately after ecdysis (spent females), and the ovarian development cycle is closely linked with the molt cycle (Dall et al. 1990). The intrusion of the SACW was responsible for the decrease in the water temperature and the increase in plankton production, suggesting that the end of spring and beginning of summer were the principal reproductive months and, with the higher water temperatures at the end of summer, spawning occurred, followed by molting.
Recruitment was interpreted as episodic, because there was no apparent correlation between the percentages of mature females and immature individuals. A similar phenomenon has been reported for R. constrictus (Bauer andLin 1994, Costa and. Probably, the complex and variable stockrecruitment relationships in penaeid shrimps depend on the food supply (primary and secondary productivity), predators and physical conditions of each region.
The present results may be used to develop a more appropriate fishery policy in the study region. Local fishing grounds should be more wisely exploited to reduce the ongoing decline of stocks. Trawl fishing may severely disturb the benthic environment in the Ubatuba region, and future studies of latitudinal variations in populations of A. longinaris could answer many questions about the degree of their reproductive plasticity. A molecular and morphological study could measure the differentiation of the populations in each region.

ACKNOWLEDGMENTS
The authors are grateful to the Fundação de Amparo à Pesquisa do Estado de São Paulo-FAPESP (#94/4878-8, #97/12108-6, #97/12106-3, #97/12107/0, and #98/3134-6) and the Conselho Nacional de Desenvolvimento Científico e Tecnológico-CNPq for financial support during collections and analyses. We thank many colleagues from the NEBECC group who helped with sampling and laboratory analyses, to Janet Reid for her constructive comments on early drafts of the manuscript and great help with English language, to María Andrea Gavio for her constructive comments of the resumen and great help with Spanish language, and the Instituto Brasileiro do Meio Ambiente e dos Recursos Naturais Renováveis (IBAMA) for granting permission to collect the shrimps.