WEIGHT-LENGTH RELATIONSHIP OF DEMERSAL FISH OF BAIA SUL , FLORIANÓPOLIS , SANTA CATARINA STATE , BRAZIL Relação peso-comprimento dos peixes da Baía Sul , Florianópolis , Santa Catarina , Brasil

This work presents the weight-length relationship of 35 species of fish belonging to 27 families. The samples were carried out bimonthly from 2001 to 2003 using seines in Saco dos Limões, Baia Sul, Florianópolis, Santa Catarina state, Brazil. The total weight (TW) and total length (TL) data of the species with n ≥ 30 were log-transformed and plotted graphically to identify and remove outliers. The regression analysis was performed to define parameters a and b, from which the weight-length relationship of each species was verified. Growth type determination were classified according to three types: isometric (19 species), positive allometric (12) and negative allometric (4). The weight-length relationship estimate of two species is presented for the first time in Brazil, bringing new information on the life history of species, as well as important results for the evaluation of regional fishery. Key-words: demersal fish, LWR, South Brazil.


INTRODUCTION
Weight-length relationships are important in ecological studies (Vallisneri et al., 2010) because they furnish information about the life history of species, contribute to regional fishery evaluation (Nieto-Navarro et al., 2009) and permit the calculation of the production and biomass of fish populations (Giacalone et al., 2010).Parameters a and b were obtained through linear regression, and in biological terms, represent the condition factor (Mortuzza et al., 2006) and the growth type of the fish, respectively.When near 3, the allometric coefficient (b) indicates isometric growth (Froese, 2006).

MATERIAL AND METHODS
This study was carried out in Saco dos Limões, Baia Sul, Florianópolis, on the central coast of Santa Catarina state (27 o 37'31.07"S;48 o 32'12.80"W).Bimonthly samples were carried out from February 2001 to December 2003 during the day and at night at six sampling sites, for a total of 216 samples distributed equally between day and night (bottom trawling -meshes: 12 and 14 mm; length: 4.5 m; mouth width: 7.5 and 9 m in the upper and lower parts, respectively).
The specimens were preserved in ice and transported to the laboratory of Environmental Science of Vale do Itajaí (UNIVALI), where they were identified using the dichotomic keys of Figueiredo & Menezes (1978;1980;1985;2000) and Fischer (1978).Weighing was carried out using a digital scale (precision 0.01 g) and total length (TL) measured (in cm) using an ichthyometer.The adopted growth model was logtransformed (log W = log a + b*log TL), where W is the weight in grams and TL the total length in centimeters; a is the constant and b the allometric coefficient (King, 1995;Froese, 2006).These last two were estimated using a linear regression analysis of the log-transformed data (Giacalone, 2010) (Excel 2007).In order to verify if b was different para identificação e remoção de outliers.Uma análise de regressão foi realizada para definir os parâmetros a e b, a partir dos quais a relação peso-comprimento de cada espécie foi estimada.A determinação do crescimento foi classificada em: isométrica (19 espécies), alométrica-positiva (12 espécies) e alométrica-negativa (4 espécies).A relação pesocomprimento para duas espécies encontradas neste estudo é inédita no Brasil, trazendo novas informações sobre a história de vida das espécies, além de resultados importantes para a avaliação da pesca regional.

RESULTS
A total of 35.303 fish, belonging to 35 species and 27 families, was evaluated.The number of individuals varied from 37 (Mycteroperca microlepis) to 10.970 (Genidens genidens).As regards two species (Catathyridium garmani and Oligoplites saliens), as far as could be determined, this is the first work to publish length-weigth relationship data and for two species (Archosargus rhomboidalis and Mycteroperca microlepis) this relationship is presented for the first time in Brazil (Table 1).
The coefficient of determination (r 2 ) showed a strong relationship between weight and length for the fish from Baia Sul -more than 0.950 for 28 species, between 0.950 and 0.900 for six species and less than 0.900 for one species (Gobionellus oceanicus) (Table 1).The allometric coefficient (b) presented a mean value of 2.059 in Gobionellus oceanicus to 3.333 in Diapterus rhombeus (Table 1) and the t-test showed isometric (19 species), positive allometric (12) and negative allometric (4) growth (Table 1).

DISCUSSION
In the determination of the length-weight relationships, factors related to sampling (fishing gear used, number of samples, sampling site, etc.) and development phase of the fish can interfere in the results (Froese, 2006).In this way, larger samples, obtained over a long period and which comprise a larger number of size classes, make more consistent estimates of the parameters of this relationship possible.Thus, the results obtained for the Baia Sul species can be considered representative and reliable, as they were estimated based on an adequate number of individuals (n ≥ 30), which presented sizes within the ranges of variation cited for most of the species.In addition, the collections were carried out in both periods (day and night) for three years, which probably incorporated the seasonal variations of the study area.
2005 ;Frehse, 2009;Joyeux et al., 2009;Freitas et al., 2010;Haluch et al., 2011) verified the lowest values of this coefficient for eleven species (although the lowest value was 0.955), which may reflect local conditions, since there was no clear differentiation of the coefficients of determination (r 2 ) between the north-northeast (maximum of 0.993 in Joyeux et al., 2009) and south-southeast coasts (0.996 in Muto et al., 2000) and probably these differences were not significant.Among the families, some species presented higher coefficients of determination, while others presented lower coefficients in relation to the references previously cited, so that the length-weight relationship in a certain family cannot be considered more consistent than in another.Similarly, the larger or smaller lengths recorded for the species that occurred in Baia Sul were not related to the families, and may have occurred as a response of each species to the local conditions or sample effort.Giarrizzo et al. (2006) found similarity in the type of growth of the species of the families Carangidae, Mugilidae and Tetraodontidae, which, in the present study, was verified for Tetraodontidae, Sciaenidae, Ariidae, Paralichthyidae and Achiridae.
Although isometric growth has predominated in sampling (followed by positive and negative allometric growth), Froese (2006) warns that the discussion of the coefficient (b) of a single length-weigth relationship, reflects the difference of the condition of the fish from only the area and moment of collection.This same author states that consistent patterns of allometric growth are rare, and should be supplemented by an analysis of the phases of growth and discussions of the potential evolutionary benefits associated with the ontogenetic changes in the body proportions.

FINAL CONSIDERATIONS
This work presents the weight-length relationship of 35 species of fish belonging to 27 families.Growth type determination were classified according to three types: isometric (19 species), positive allometric (12) and negative allometric (4).As regards two species (Catathyridium garmani and Oligoplites saliens), as far as could be determined, this is the first work to publish length-weight relationship data and for two species (Archosargus rhomboidalis and Mycteroperca microlepis) this relationship is presented for the first time in Brazil, bringing new information on the life history of species, as well as important results for the evaluation of regional fishery.