884 



Fishery Bulletin 92(4), 1994 



Swan Estuary 



Males: 

 Females: 



y = 6.152 + 0.95 lx - 0.875* 2 - 0.400* 3 (R 2 = 0.931, P < 0.001, n = 499) 

 y = 6.174 + 1.171* - 0.479* 2 - 0.230* 3 (R 2 = 0.931, P < 0.001, n = 568) 



Examination of the otoliths suggests that the curvi- 

 linearity at the upper end of these relationships (Fig. 

 3) is due to the otoliths of larger fish tending to 

 thicken rather than lengthen. 



The lengths offish of a given age class were highly 

 variable (Figs 4 and 5). For example, the lengths of 

 female fish that were about four years old in Wilson 

 Inlet ranged from 478 to 631 mm and those that were 

 about three years old in the Swan Estuary ranged 

 from 351 to 591 mm. The predicted lengths offish, 

 derived from generalized growth curves, were greater 

 when lengths at age rather than back-calculated 

 lengths were used for fish of ages 1 and 2 (Table 1). 

 The high values for R 2 for the generalized growth 

 curves, derived from both lengths at age and back- 

 calculated lengths, show that these curves fit the data 

 well (Table 2). The oldest male and female C. 

 macrocephalus caught in Wilson Inlet were 12 3 /4 

 years old (718 mm, 1885 g) and 9 3 A years old (670 

 mm, 1738 g), respectively. The corresponding values 



for fish from the Swan Estuary were 5 years (582 mm, 

 1142 g) and 6 3 A years (683 mm, 1880 g), respectively. 



The use of common curves in the cases of both 

 lengths at age and back-calculated lengths for each 

 of the two sexes in each system accounted for 89 to 

 94% of the observed variance. By assuming that a 

 difference exists between the growth curves of the 

 two sexes in each system and with each of the two 

 methods, the fit was improved by only 0.003% for 

 back calculated data for the Swan Estuary and 0.3% 

 for length at age data for Wilson Inlet. 



Applying likelihood ratio tests, the length-at-age 

 growth curves for males and females differed signifi- 

 cantly in both the Wilson Inlet (P<0.001) and Swan 

 Estuary populations (P<0.05). Back-calculated 

 growth curves calculated for the two sexes also dif- 

 fered significantly (P<0.001) in Wilson Inlet but not 

 in Swan Estuary. 



The use of a common curve for each sex by using 

 both lengths at age and back-calculated lengths for 



Wilson Inlet 

 Males 



Swan Estuary 

 Males 



0.5 (III (IS 



In Otolith Radius (mm) 



2.0 1.5 III (I 5 (1 5 

 In Otolith Radius (mmi 



Figure 3 



Relationships between the natural logarithms of total length (x) and lapilli ra- 

 dius (y) of the eel tailed catfish Cnidoglanis macrocephalus from Wilson Inlet 

 and the Swan Estuary. Broken lines represent the best fit for a linear regres- 

 sion, the solid lines the best fit for the cubic polynomial equation. 



