REPRODUCTIVE BIOLOGY OF Astyanax janeiroensis ( OSTEICHTHYES , CHARACIDAE ) FROM THE UBATIBA RIVER , MARICÁ , RJ , BRAZIL

Aspects of the reproductive biology of Astyanax janeiroensis in the Ubatiba river in Maricá, RJ, Brazil were evaluated in order to ascertain whether the strategy adopted by the species is causally related to environmental variables. Specimens were collected on a monthly basis from October 1994 to September 1995, through electrofishing. The Sex ratio was not significantly different from the expected frequency of 1:1 (G = 1.29; p > 0.50); nonetheless, considering three standard length classes, significantly larger numbers of females were recorded for the higher SL class (G = 11.07; p < 0.01). Size at first maturation showed no significant differences between sexes. Length-weight ratio analyses showed negative allometry for males and isometry, for females. Length structure was significantly different between sexes, suggesting that females are larger than males (D = 0.027, p < 0.01). Reproductive specimens were recorded during nine months of the annual cycle. Seasonal variations of reproductive specimens belonging to two different standard length classes displayed an asynchronous behavior, with larger specimens having a longer reproductive period and smaller ones reproducing for fewer months during the annual cycle. High values of fecundity (F), varying from 3169 to 18714 oocytes, were recorded for fish of 9.1 and 10.2 cm lengths, respectively. The correlation between weight and number of oocytes/unit of weight was positive, indicating that larger specimens produced more eggs.


INTRODUCTION
Following the heretofore narrow perspective of fish life history (Matthews, 1998), knowledge of some reproductive traits such as sexual dimorphism, egg size, fecundity and spawning time are fundamental to understand how these traits should be adapted to maximize reproductive success in a particular environment.nonetheless, despite the high plasticity of many life history traits, important constraints imposed by fundamental morphology or phylogeny are observed among different fish species.
Astyanax janeiroensis (hereinafter referred to as Astyanax) is one of the most widespread and abundant Characidae in the clear waters of the Ubatiba River.Only one coexisting Characidae (Astyanax hastatus) has been found in this stream, but its abundance is considerably lower than that of Astyanax (Mazzoni & Lobón-Cerviá, 2000).In this paper, we describe the reproductive strategy of Astyanax.The traits analyzed here include size structure, length to weight ratio, size at first reproduction, fecundity, sex ratio and reproductive season.

MATERIAL AND METHODS
The Ubatiba river (22°60'S and 42°48'W) consists of a small coastal fluvial system of the Brazilian East Basin in the state of Rio de Janeiro.This river flows through meadows deforested for farming and cattle ranching, although patches of pristine Atlantic forest are still common on the slopes of the surrounding rocky hills.Water level is regulated by rainfall (~ 1500 mm yr -1 ), but the system's main physical characteristics are the sudden spates that stress local habitats and the absence of seasonality in the hydric regime (Mazzoni & Lobón-Cerviá, 2000).
The specimens investigated in this study were collected by electrofishing (1000 W, 2-3 A) between October 1994 and September 1995 in a monthly basis (see more detailed information about the sampling methodology in Mazzoni et al., 2000b).
Sampled fish were packed in ice for transportation, processed in the laboratory and the following data obtained: i) standard length (SL -0.1mm); ii) total weight (WT -0.01 gr); iii) sex; and iv) gonadal weight (WG -0.01 gr).
Prior to preservation, the ovaries/testes were classified according to a macroscopic scale of gonadal development, taking into account the vascular irrigation and the relative occupation in the abdominal cavity, for both sexes; for females, the size and color of oocytes were also recorded, and for males, sperm release when pressing the abdomen was recorded.According to these characteristics, the following classification was considered: females -immature, mature 1, mature 2, ripe and spawned; males -immature, mature 1, mature 2, ripe and spent.The ovaries were then fixed in modified Gilson's fluid (Vazzoler, 1981) for oocyte measurements and counts.
Ovaries preserved in Gilson's fluid were stored for two weeks and shaken periodically to promote oocyte release.Oocytes were then cleaned by changing the alcohol and removing the ovarian membranes, and stored in a 70° GL alcohol solution.Fecundity (F) was determined according to Bagenal & Braum (1971) after all the vitellogenic oocytes from 25 ripe ovaries were counted and correlated to SL, WT and WT/number oocytes per unit of WT (WT/F/gr).
Standard length structure was determined by the number of individuals in different SL classes, while sexual dimorphism for length was tested by applying the Kolmogorov-Smirnov test on two samples.The SL/WG relationship was established from the Ln transformed data of SL and WT, using the Minimum Square Method; differences between sexes were tested with the angular coefficients test.The size at first maturation (SL 50 ) was determined according to the methodology presented in Vazzoler (1981).All the aforementioned analyses were made separately on males and females.
The sex ratio was determined for the whole sample and for the three SL classes, established according to SL 50 , and the bias in the proportion of individuals of each sex was tested by the log-likelihood ratio (G test).The reproductive season was tested through the temporal distribution of mean GI (Gonadosomatic Index -GI =WG / WT * 100) and ovarian/testis maturation stages.For the latter, the maturation stages were rearranged as follows: Immatureyoung fish without developed gonads and SL equal to or less than SL 50 ; non-reproductive -adult fish, SL of more than SL 50 , in mature, mature 1 and 2 and recovering gonads; and Reproductive -adult fish, SL of more than SL 50 , with ripe and spawned/spent gonads.The seasonal pattern of reproduction was tested based on the temporal variation of the mean GI of three different SL classes.

RESULTS
A total of 308 specimens were analyzed, 164 females and 144 males.The sex ratio presented no significant difference from the expected 1:1 ratio (G = 1.29; p > 0.50); however, considering the three SL classes, the higher class showed a significant difference (Fig. 1; Table 1).The standard length distribution differed between the sexes (D = 0.027, p < 0,01), indicating that males and females reached 11.4 cm (mean SL = 6.81 cm) and 10.6 cm (mean SL = 7.16 cm), respectively (Fig. 2).The lengthweight relationship for Log transformed data of SL and WT showed significant positive correlations: females -i) females: LnWT = 3.04 LnSL -3.68; r = 0.98; p < 0.01; and ii) males: LnWT = 2.73 LnSL -3.17; r = 0.98; p < 0.01.Between sexes differences for the lenght-relationship were also registered (t = 4.00; df = 285; p < 0.05) The onset of sexual maturity was determined as the SL class in which 50% of specimens were adult; males and females presented no difference for this characteristic, and 5.5 cm was found to be the size of Astyanax at first maturation (Fig. 3).
Reproductive specimens occurred throughout the year except for May to July, with a higher incidence between november and March (Fig. 4).The same results were found by monthly distribution of GI (Fig. 5).A seasonal analysis of reproduction of different SL classes specimens indicated that adult specimens longer than 9.0 cm were responsible for this pattern, since they were reproducing over an extended period.Smaller (= younger) specimens displayed markedly seasonal reproduction, since breeding specimens occurred only during some spring and summer months, from October to February (Fig. 6).
Astyanax fecundity (F) showed a positive correlation with length (SL) -F = 4543 SL -33456, r = 0.84, n = 20; p < 0.01 and weight (WT) -F = 551 WT -4117, r = 0.89, n = 20; p < 0.01, varying from 3169 to 18714 oocytes, respectively, for 9.1 and 10.2 cm long fish.A correlation analysis of WT and number of oocytes/unit of WT (WT/F/gr) showed a positive correlation (WT / F/gr = 8.1607 WT + 165.82, r = 0.57, n = 20; p < 0.01), indicating that larger specimens are more efficient in producing eggs.nikolsky (1963) referred to sexual differences in body size and length-weight relationship as the most frequent form of sexual dimorphism among fish.Among females, a large size is favored because it increases fecundity, whereas male size is affected by sexual selection, with bigger males enjoying a reproductive advantage (Greenwood & Wheeler, 1985;Parker, 1982;Shine, 1990).With   competitive behavior absent during reproductive events, Astyanax species are pelagic spawners normally reproducing in school formation during movements (Breeder & Rosen, 1966;Mazzoni & Iglesias-Rios, 2004), and the increase in potential fecundity among larger females may explain the differences in length structure observed here.

DISCUSSION
The onset of sexual maturity plays an important role, shedding light on the life-history adopted by a species during its evolution (Stearns, 1976;Lessels, 1991;Stearns & Crandal, 1984).It is an accepted fact that the optimal size for the first reproduction depends upon many factors, including the relative allocation of energy between somatic and gonadic growth (Mann et al., 1984).The benefits of larger specimens will be countered by opposing selective forces, such as survivorship and potential reproduction, for the longer maturity is delayed the lesser the chance of surviving to leave descendants (Ross, 1986).
A. janeiroensis attained sexual maturity at 5.5 cm, which is a smaller size than that found by Barbieri (1992) for Astyanax scabripinnis paranae and Agostinho et al. (1984) for Astyanax bimaculatus (Paraná basin), although these three species attain the same size.The interpretation of the selective forces acting on Astyanax suggests that reproductive precocity may be related to an adaptive behavior to compensate for the stochastic mortality imposed by the unstable hydrology of the Ubatiba River system.Such an adaptation may explain the success of small species such as Deuterodon sp.(Mazzoni & Petito, 1999) and Characidium sp.(Mazzoni et al., 2002a) when considering an unpredictable environment like that of the Ubatiba River.
The prolonged breeding season found in this study can be explained by the same rationale and was also found for other species from the Ubatiba River (Menezes & Caramaschi, 1994;Mazzoni & Petito;1999;Mazzoni et al.;2002a;Mazzoni & Iglesias-Rios, 2002).In this breeding pattern, we found that adult specimens are nonseasonal breeders, whereas younger specimens are markedly seasonal breeders, with reproduction determining an asynchronous pattern of reproduction.Hyndes et al. (1992) attributed delayed reproduction among young specimens to delayed maturation of specimens reproducing for the first time.According to nikolsky (1963), intrapopulation asynchrony may be explained as evasion for competition for breeding places.Mazzoni & Caramaschi (1995) found the same holds true for two sympatric populations of Hypostomus.Considering that Astyanax are pelagic spawners, we believe that Hyndes et al.'s (1992) proposition is probably more accurate in the case studied here.
Fecundity was shown to be positively correlated with female weight and length.nonetheless, an analysis of the correlation between weight and number of eggs per unit of weight revealed that larger females are able to produce more eggs per unit of body mass, indicating that big fish produce more offspring than small fish.
The overall reproductive strategy of A. janeiroensis suggests that a high energy budget is allocated to reproduction, mainly by bigger females,  Fig. 6 -Temporal variation of mean Gonadosomatic Index (GI) of two standard length classes.7.5 and 10.5 are mid points of classes 6.5-8.9 cm and 9.0-11.0cm. in order to compensate for the unpredictable hydrological conditions prevailing in the Ubatiba River.

Fig. 1 -
Fig. 1 -Sex ratio of Astyanax janeiroensis considering three different Standard Length classes.Grey bars correspond to females and white bars to males.

Fig. 2 -
Fig. 2 -Standard length distribution of males and females Astyanax janeiroensis from the Ubatiba fluvial system, Maricá-RJ.Grey bars correspond to females and white bars to males.

Fig. 3 -
Fig. 3 -Standard length distribution of adult specimens of Astyanax janeiroensis from the Ubatiba fluvial system, Maricá-RJ.SL 50 means standard length at sexual maturity.