484 



Fishery Bulletin 98(3) 



that the 1993 zones were completed later in the calendar 

 year than those formed in 1994. The hypothesis that the 

 frequency of zone completion, as the slope of the regression, 

 was equal to one cycle/yr was therefore retained for the 

 ventral axis for all species pooled. 



Timing of zone formation 



FourL. eiythropterus had otoliths with two OTC marks that 

 bracketed a fully formed opaque zone (Fig. 5), indicating 

 that these annuli were formed in the austral spring in 

 September-November 1993. A Hotelling paired sample test 

 for equality of means in circular distributions showed that 

 there were significant differences in estimates of calendar 

 closing dates iMCCD) from the direct method between 

 reading axes (n=60 ndf=2 ddf=58 F=3.41 P>F=0.04). 

 There was also variability amongst species, with the most 

 numerous species completing an annulus along the ventral 

 axis in mid-August by L. johnii, in early September by L. 

 eryfhropterus, and in early November by L. sebae. 



By pooling data from the ventral and sulcal axes by 

 taking the mean of each pair we were able to overcome 

 problems associated with anomalous otolith growth, resolve 

 differences in the estimates, and not significantly change the 

 degrees of freedom in any subsequent tests for significance 

 of timing of zone formation. This procedure tightened the 

 estimates of closing date for L. sebae and generally gave a 

 best estimate of spring-early summer for formation of the 

 annulus in the most numerous species (Table 5). Results 

 from the "method of best fit" were inside the bounds of the 

 confidence limits of the direct method for L. johnii and L. 

 sebae and were similar for L. malabaricus. L. monostigma, 

 and L. rivulatus but were significantly later in the year 

 for L. erythropteriis (Table 5). The discrepancies between 

 methods fori. argentimaculatus,L. bohar.L. carponotatus. 

 and L. vifta reflected both the small number of fish and 

 difficulties in measurement and interpretation of their 

 smaller otoliths. The estimates for the less numerous 

 species must be considered to be tentative. 



There was a marked coincidence between the August 

 2nd minimum in water temperature and a peak in August 

 for frequency of 1994 closing dates (Fig. 11). Extra opaque 

 zones counted or deemed to be false annuli may have 

 formed after storms and salinity fluctuations at Cape 

 Ferguson. To investigate this possibility, we estimated 

 closing dates using Equation 7 from the radii of anomalous 

 "extra" opaque zones visible on both reading axes in 

 otoliths of L. johnii (sl947, sl949, sl998. Fig. 6), L. sebae 

 (sl979, sl982. Fig. 4), and L. eryfhropterus (sl991. Fig. 

 5) identified as outliers on Figure 8. All but one of these 

 fish (sl998) were in the group of captives. Compared with 

 salinity and sea state during the months of captivity, there 

 appeared to be a relationship between a sharp fall and rise 

 in salinity of at least 1.4 ppt for March-May 1994 and the 

 completion of these "extra" zones in the following month, 

 although periods of rough and very rough sea conditions 

 also occurred most commonly at this time. Salinity dropped 



