Cappo et al.: A new approach to validation of periodicity and timing of opaque zone formation 



485 



from 36.0 ppt on 3 February to 34.8 ppt on 1 April and 

 rose again to 36.2 ppt on 1 May. The closing dates for these 

 false annuli were distributed in March (n=l), April (/!=2), 

 and May («=3). 



Discussion 



The use of increment widths on otolith sections to examine 

 timing of opaque zone formation underpins marginal 

 increment analysis (Fowler and Short, 1998) and informal 

 inferences made from the relative position of OTC marks 

 (Cheat and Axe, 1996; Ferreira and Russ, 1992, 1994; 

 Francis et al., 1992). The direct method developed in our 

 study has essentially formalized and extended this use to 

 allow for robust parametric tests and has three advantages 

 over previous approaches to direct validation. First, the 

 use effractions of otolith growth allows linear regression of 

 the number of whole or partially completed (or whole and 

 partially completed) increment cycles on time at liberty, 

 and both measures are expressed correctly as continuous 

 variables. This method allows periodicity to be estimated 

 directly from all recaptures with at least one opaque zone 

 outside the OTC mark, including those fish at liberty for 

 less than one year. Second, there is no need for subjective 

 choice of a fixed date of completion of opaque zones to allow 

 for comparisons between expected and observed counts 

 of opaque zones. Third, the calendar date of completion 

 of outer opaque zones can be estimated independently of 

 the dates of marking or sacrifice for all recaptures with 

 at least one opaque zone outside the OTC mark. Previous 

 inferences about timing have been limited to the opaque 

 zones that are immediately adjacent to OTC marks. Such 

 contrasts are not always readily available, especially when 

 tagging and recaptures cannot be spread throughout the 



J'A^S^'n'D J F'l^'A'M'j J^A S ON DiJ'F'M'A 

 1993 1994 1995 



Figure 11 



Monthly sea surface temperature ( 'C ) in the cages during captivity 

 of marked fish, plotted with the monthly frequency of closing dates 

 (CD) in 1994 estimated with the "direct method" with Equation 

 6 for nine species from Figure 9. The estimates have been pooled 

 and averaged for the ventral and sulcal axes for each of 61 fish. 



year because of seasonal differences in fishing effort or 

 vulnerability offish to capture. 



Periodicity and timing of opaque zones 



We tested, and retained, the hypothesis that there was 

 an annual periodicity of formation of opaque zones for 

 the pooled study species, around a mean of 0.96 ±0.32 

 increment cycles/yr. Anomalous sections and "false" annuli 

 were identified objectively and could be related to periods of 

 environmental stress in six cases. The model also identified 

 interannual difference in timing of zone completion by a 

 group of L. erythropterus , which clearly inflated overall 

 estimates of periodicity. Despite this difference, estimates 

 of periodicity in the range of 0.78-1.21 cycles/yr for both 

 reading axes, accompanied by confidence intervals, were 

 robust for L. erythropterus, L.johnii, L. malabaricus, and 

 L. sebae. This result supported the use of thin sections 

 of otoliths for accurate age estimates of lutjanids in the 

 central GBR and supported estimates of longevity proposed 

 for some of our study species (eg Sheaves, 1995; Newman 

 et al., in press; Loubens-). The results for the other species 

 showed similar trends, and no interspecific differences in 

 periodicity were detected, but the numbers of marked fish 

 were low for L. argentimaculatus,L. bohar,L. carponotalus, 

 L.gibbus,L. monostigma,L. rivulatus, andL. vitta. Twofold 

 differences in otolith growth rate between captive and 

 tagged L.johnii and L. sebae did not affect the periodicity 

 of annulus formation, further demonstrating the capacity 

 of these otoliths to accurately reflect age under varying 

 conditions of somatic growth. 



Substantial variability in the timing of opaque zone 

 formation has been noted by Beckman and Wilson ( 1995), 

 at the individual, population, and species levels. Our best 

 estimates of closing dates indicated that, on average, 

 opaque zones recognized as annuli in lutjanid otoliths 

 were completed within months of the austral spring 

 and early summer, around early September for L. 

 erythropterus and L. johnii, late September for L. 

 malabaricus, and late October for L. sebae. These 

 differences were significant and indicated a timing 

 of formation that is somewhat earlier for L. 

 erythropterus andL. johnii than the spring-summer 

 peak of opaque zone formation in tropical fishes 

 identified by Fowler ( 1995 ) and Beckman and Wilson 

 (19951. Other studies have concluded that lutjanids 

 and serranids on the GBR form opaque zones in 

 winter-spring (Ferreira and Russ, 1992, 1994; New- 

 man etal., 1996). 



Lack of definition of the term "formation" makes 

 it difficult to compare our results with those of 

 previous studies. Our use of single OTC marks 

 and measurements of only the outside edges of 

 opaque zones allowed inferences about the timing of 

 completion, not commencement, of annuli. We could 



Loubens, G. 1980. Biologie de quelques de poissons du 

 lagon Neo-Caledonian. III. Croissance. Cahiers de flndo- 

 pacifique2l2): 101-1.53. Foundation Singer-Polignac, 43, 

 avenue Georges-Mandel, 75016 Paris. ISSN 0180-9954. 



