Growth Pattern and Length-weight Relationship of Scomberoides lysan (Pisces: Carangidae) from the Northern Waters of Sri Lanka

MATERIALS AND METHODS

Sample collection: Northern part of Sri Lanka is situated between 79°E to 80°E longitude and 9°N to 10°N latitudes in the Indian ocean. Samples were collected weekly from the commercial catches at Passaiyoor, Delft, Karainagar and Point Pedro landing centers during January 2010 to December 2010, using a specific drift gill net for Genus Scomberoides with the mesh size of 7", 21 ply (Katta valai). Beach seines and trap net (Kalankatti valai) fixed in shallow waters are also operated to caught S. lysan in waters around Jaffna peninsula.

Morphometric measurements: Total Length (TL) of each fish was measured from the anterior most edge of the lower lip (tip of snout) to the distal part of the caudal fin to the nearest mm with a measuring board and Total Weight (TW) was measured to the nearest 0.1 g by an electronic balance (AND, HF-1200G). All specimens were sorted by sex according to the morphological analysis of gonads as male, female or unsexed.

Regression analysis: The relationship between the length and weight of a fish is usually expressed by the equation W = a L ^b (Ricker, 1973); where, W is the weight of the fish, L is the total length, “a” is the intercept and “b” is the slope. If a fish is growing isometrically (increasing in all dimensions at the same rate) and doubles in length, its weight will increase in relation to the increase in volume; that is by 8 (or 2³) times (King, 2007). Thus there is a cubic relationship between length (L) and weight (W) and ‘b’ is close to 3 in isometric growths and ‘a’ is a constant determined empirically.

Statistical analysis: Weights and lengths of male, female and unsexed were log transformed and the resulting linear relationship fitted by the computer based linear regression analysis. Significance of the regression was assessed by General Linear Model Analysis of Covariance (GLMANCOVA). The “b” values obtained for male, female and unsexed were tested by Student’s t-test to see whether the ‘b’ values differ significantly from 3 or not (Zar, 1999). Two sample t-test was performed to compare the mean weight data of male and females to distinguish the significant difference between male and female. All statistical analysis was done by MINITAB (Version 14) statistical software and the significance level was tested at 95% confidence interval.

RESULTS AND DISCUSSION

A total number of 892 fish (299 males, 354 females and 239 unsexed) were analyzed. Length range of fish collected is 22.2-62.5 cm for males, 22.0-71.0 cm for females and 14.6-36.1 cm for unsexed. The estimates of the regression parameters for male, female and unsexed are given in Table 1. The r² values obtained while calculating the regression parameters are 0.9485 for males, 0.9490 for females and 0.9435 for unsexed. For male, female and unsexed S. lysan obtained residual degrees of freedom are 297, 352 and 237 whereas residual sum of squares are 1.505861, 1.348065 and 0.590727, respectively. Linear regression relationships between total length and weight of male, female and unsexed S. lysan are given in Table 2. Curvilinear relationships obtained for male, female and unsexed are W = 0.0037 L ^3.1319, W = 0.0058 L ^3.0048 and W = 0.0179 L ^2.6496, respectively.

The results of student’s t-test to analyze the significance of variation in the estimates of ‘b’ for S. lysan from the expected value for the ideal fish (3.0) are as follows:

Student t-test showed that the ‘b’ values obtained for male and unsexed were significantly different (p<0.05) from 3 indicating positive allometric growth for male and negative allometric growth for unsexed whereas the ‘b’ value obtained for female very close to 3, did not deviate significantly (p>0.05) from 3 indicating isometric growth.

Table 1:	Length-weight relationship parameters of S. lysan
N: No. of observations, df: Degrees of freedom, ‘b’: Regression exponent, ‘a’: Constant, SS: Sum of squares, r: Correlation coefficient

Table 2:	Linear regression relationship between total length and weight of male, female and unsexed S. lysan

Table 3:	Parameters obtained from two-sample t-test for mean weight of males and female of S. lysan

Fig. 1:	Percentage frequency of total length distribution of male, female and unsexed S. lysan

Correlation coefficient (r²) 0.9485 for males, 0.9490 for females and 0.9435 for unsexed found to be significant (p<0.001) in all instances indicating good correlation between length and weight of S. lysan.

Comparison of regression co-efficient of male, female and unsexed using GLMANCOVA for the regression of log weight on log total length for S. lysan male, female and unsexed revealed the b values show significant differences (p<0.05) between each other. The confidence intervals of ‘a’ value for male, female and unsexed do not overlap with each other and therefore there is a significant difference between the intercepts, too.

Results of the two sample t-test are given in Table 3. It showed that males were significantly (p>0.05) heavier than females. The frequency distribution of total length for male, female and unsexed S. lysan individuals is shown in Fig. 1. Highest percentage of frequency (50.4%) was observed for S. lysan individuals of 21-30 cm total length class interval.

Length-weight relationship of fish and fish related organisms have been studied widely in all parts of the world by several authors. Different values for the exponent coefficient (b) for different fish have been recorded in different parts of the world. Earlier, Allen (1938) pointed out that the exponent coefficient (b) computed from 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. This statement could be supported by the previous work carried out for various fish species for example 3.171 for Amblygaster sirm, 3.24 for Carcharhinus albimarginatus (Kulbicki et al., 1993); 2.459 for Sepioteuthis lessoniana (Sivashanthini et al., 2009b); 3.359 for Leuciscus cephalus Karatas and Faith CAN (2005); 2.880 for Tenualosa ilisha (Ahmed et al., 2008) and 2.929 for Atherina sp. (Bouriga et al., 2011). These ‘b’ values cannot be simply compared for consistency as these are different species with different metabolic rate.

Widely accepted concept is that if the exponent value is 3, the fish grows isometrically, if it is greater or less than 3 fish grows allometrically (Tesch, 1968). The ‘b’ value depends on several factors such as age, body shape and amount of fat present, sex, maturity stage, season, temperature, salinity and available nutrient food (Moutopoulos and Stergiou, 2002). Taskavak and Bilecenoglu (2001) and Ozaydin and Taskavak (2006) stated that the parameter ‘b’ in the length-weight relationship vary seasonally, even daily and between habitats.

Length-weight relationship of S. lysan have not been studied in Sri Lanka, so far and this is the first study to compute such parameters. Some of the earlier results on length-weight relationship of Scomberoides species from other parts of the world are shown in Table 4. In New Caledonia and South Africa ‘b’ value of 2.896 and 2.685 was obtained, respectively for S. lysan (Letourneur et al., 1998). But ‘b’ value of 2.937 was obtained for S. tol from Karnataka waters, India (Abdurahiman et al., 2004). The ‘b’ value obtained in the present study for male (3.1319) and female (3.0048) is slightly higher than the previously recorded values. It may be due to the variations in ecology of the geographical locations, food availability and different environmental conditions. However the ‘a’ and ‘b’ values obtained for male and female S. lysan were superimposed in the plot of log ‘a’ versus ‘b’ available for S. lysan, Carangidae and 1300 miscellaneous species in FishBase 2011 (Froese and Pauly, 2011) and it is shown in Fig. 2a and b. The parameters obtained from the study are useful fundamental factors applicable in future culture trials as well as in population dynamics studies.

Table 4:	The parameters of length-weight relationships of genus Scomberoides from different regions of the world ( SL = standard length, FL = Fork length, TL = Total length)

Fig. 2a:	Plot of log ‘a’ versus ‘b’ for male S. lysan (Black dot indicates the current estimate; X axis denotes exponent coefficient ‘b’ and Y axis denotes constant ‘log a’). Source: FishBase 2011

Fig. 2b:	Plot of log ‘a’ versus ‘b’ for female S. lysan (Black dot indicates the current estimate; X axis denotes exponent coefficient ‘b’ and Y axis denotes constant ‘log a’). Source: FishBase 2011

CONCLUSION

Length-weight relationships for male, female and unsexed S. lysan were W = 0.0037*L^3.1319, W = 0.0058*L^3.0048 and W = 0.0179*L^2.6496, respectively. Comparison of regression co-efficient of male, female and unsexed using GLMANCOVA revealed that the ‘b’ values show significant differences (p<0.05) between each other. The statistical analysis confirmed that males exhibit positive allometric growth, females exhibit isometric growth and unsexed exhibit negative allometric growth. The parameters obtained can be applied in future culture trials and in population dynamics studies.

ACKNOWLEDGMENT

Authors are grateful to the authorities of National Research Council, Colombo, Sri Lanka for the financial assistance in the form of research grant (Grant No. 07-19).