MATERIALS AND METHODS

The Jaffna lagoon is a shallow water body located in the northern province of Sri Lanka. It lies between approximately 79° 52’E to 80° 38’E longitude and 9⁰ 26’N to 9° 46’N latitude and has an area of about 412 km² (160 square miles) and the depth does not exceed 4 m.

Weekly random samples of Sphyraena obtusata were collected from Kakkaithevu, Navanthurai, Kurunagar and Pashaiyoor landing centres of Jaffna lagoon, Sri Lanka from November 2004 to December 2005 (Fig. 1). Thirteen field visits were made to each landing centers throughout the study period. During the field visits 70 specimens were collected from Kurunagar and Pashaioor landing centers whereas 62 specimens were collected from Kakkaithevu and Navanthurai landing centers of the Jaffna lagoon. Samples were brought to the laboratory and Total length (TL) was measured to the nearest 5 mm with a measuring board. Weight (W) was measured to the nearest 0.1 g by an electronic balance (AND, FY 300) after draining the water from the buccal cavity and wiping the moisture content on the body of fish (King, 1996). The parameters ‘a’ (proportional constant or intercept) and ‘b’ (exponent) of the Total length (TL) -Weight (W) relationship of the form W = aTL^b were estimated for males and females separately by using the logarithmic transformation log W = log a+b.log TL. Male and female S. obtusata were identified only after dissecting the specimens and this species cannot be differentiated by any special external morphological characters. External morphology of S. obtusata is shown in Fig. 2. The regression line was calculated by the method of least square regression analysis. The regression lines of male and female fishes were then analyzed further by covariance analysis using MINITAB software for significant differences.

Fig. 1:	Map showing the collection sites (a-d) of Jaffna lagoon, Sri Lanka. (a) Kakkaithevu, (b) Navanthurai, (c) Kurunagar, (d) Pasaioor, (e) Thondaimannar lagoon and (f) Jaffna lagoon

Fig. 2:	External morphology of S. obtusata

RESULTS

The size (Total length-TL) of S. obtusata ranged from 140 to 334 mm (Mean±SD = 219.42±37.00 mm). Males ranged from 149-333 mm TL (Mean±SD = 210.38±31.7 mm) while females ranged from 140 to 334 mm TL (Mean±SD = 228.73±39.7 mm). The values obtained for the mean weight by sex were simply compared by two-sample t-test (Table 1). The results showed that females were significantly (p<0.05) larger than males.

The estimates of the regression parameters of length-weight relationship for male and female, obtained by regression analysis are shown in Table 2. The equations of total length-weight relationship and their logarithmic transformation are given in Table 3. The ‘b’ values 2.898 and 2.843 obtained for male and female respectively indicate that the fish follows the cube law, its growth is proportionally three-dimensional (n = 134, r² = 0.983, 95% CL a = 0.0143 to 0.00961, 95% CL b = 2.833 - 2.964 and coefficient of variation = 0.00286 for males; n = 130, r² = 0.987, 95% CL a = 0.0116 - 0.0165, 95% CL b = 2.788 - 2.898 and coefficient of variation = 0.00245 for females). That is, with increasing age, rate of growth in terms of weight in this fish becomes slower than that of its length. Correlation coefficients(r) 0.991 for male and 0.993 for female were found to be significant (p<0.01) in both instances indicate good correlation between length and weight.

The significance of variation in the estimates of b for S. obtusata from the expected value for the ideal fish (3.0) was tested by Students’ t-test (Snedecor, 1963; Jayaprakash, 2001). Students’ t test was employed by dividing the difference between ‘b’ and ‘3’ by standard error of b (Zar, 1996). The results are as follows:

The GLMANCOVA showed the slopes (b) of males and females not to exhibit significant interaction (computed F_1,264<5.07, p>0.05). Further, comparison of regression co-efficient of males and females using GLMANCOVA for the regression of log weight on log total length of males and females showed the ‘b’ values not to show significant differences (p>0.05). The confidence intervals of ‘a’ for male and female overlap with each other and therefore there is no significant differences between the intercepts, too. Hence the following common formulae were derived for males and females.

Table 1:	Parameters obtained from two-sample t-test for mean weight of male and female Sphyraena obtusata collected from Jaffna lagoon, Sri Lanka

Table 2:	Length-weight relationship parameters of Sphyraena obtusata collected from Jaffna lagoon, Sri Lanka
R: Correlation co-efficient; N: No. of observations/sample size

Table 3:	Relationship between total length and weight of male, female and pooled sexes of Sphyraena obtusata

The regression exponent value 2.843 for females indicates a negative allometric growth in S. obtusata females while the exponent value 2.898 for males, significant from 3 (p>0.05) indicate an almost negative allometric growth in S. obtusata males, too.

DISCUSSION

A historical review on length weight relationship and recommendations for users about length weight relationships, condition factors and relative weight is provided by Froese (2006). Allen (1938) pointed out that the exponent coefficient (b) in the length-weight relationship of fishes is usually 3. Later Carlander (1969) pointed out that the b value is very close to 3.0 but varies between 2.5 and 3.5. If the ‘b’ value for a fish is 3, the fish grows isometrically; if it is greater than 3, the fish exhibits positive allometry and if it is lower than 3 the fish exhibits negative allometry (Bagenal and Tesch, 1978). Fishery biologists also stated that the ‘a’ and ‘b’ values not only differ in different species but also differ in the same species depending on sex, stage of maturity, food habits and so on (Qasim, 1973; Bal and Rao, 1984; Froese, 2006). The parameters of length-weight relationship of obtuse barracudas estimated by various authors from different parts of the world are given in Table 4.

One previous study available for length-weight relationship of S. obtusata in Bombay waters, west coast of India is that of Jaiswar et al. (2004). They reported a ‘b’ value of 2.7226, a slightly lower value than obtained in the present study.

Table 4:	The parameters of length-weight relationship of Sphyraena obtusata from different regions of the world
TL: Total length, FL: Fork length, SL: Standard length

Deviation in the growth rate ‘b’ observed during the present investigation may be the result of variations in ecology of the geographical locations or due to changes in the environmental conditions (Bagenal and Tesch, 1978). In a long-term basis, further studies on breeding biology, nutritional studies, feeding biology and age and growth will lead to a successful sustainable management of S. obtusata in Sri Lankan waters.

ACKNOWLEDGMENT

Researchers are grateful to the authorities of University of Jaffna for the financial assistance.