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



475 



and reporting of the periodicity and 

 timing of annulus formation outside 

 OTC marks. 



The most common analysis of per- 

 iodicity has been that of informally 

 comparing the observed number and 

 position of annuli distal to the OTC 

 mark with an expected number of 

 annuli chosen from the whole number 

 of years or seasons elapsed between 

 marking and recapture. The most 

 extensive studies, comprising large 

 numbers of long-term recoveries of 

 marked fish, reported the results as 

 either percentage agreement (Mac- 

 Lellan and Fargo, 1995) or as a ^test 

 (McFarlane and Beamish, 1995). Both 

 approaches could not include fish 

 recovered at liberty for less than one 

 year and did not allow exploration 

 of the sources of error evident in 

 the comparisons. Low recovery rates 

 of fish at liberty for more than 2-3 

 years are a feature of such tagging 

 programs ( Campana and Jones, 1998 ) 

 and sample sizes are usually small. 

 A median of only 21 fish was reported by Francis et al. 

 (1992) from 14 earlier OTC mark-recapture studies. Such 

 characteristics prevent the use of linear regressions, such 

 as those used in validating daily microincrements (e.g. 

 Foreman, 1996). The timing of opaque zone formation is 

 most commonly inferred by informally noting the relative 

 position of opaque zones, OTC marks, and the otolith 

 margins on diagrams and photomicrogi'aphs (Choat and 

 Axe, 1996; Ferreira and Russ. 1992, 1994; Fowler and 

 Doherty, 1992; Murphy et al., 1998). 



We propose that validation can be formalized into models 

 to develop separate estimators of both the periodicity and 

 timing of formation of opaque zones based on ratios of 

 increment measurements around OTC marks. Our primary 

 goals were to determine directly if the distinct opaque 

 zones in sectioned otoliths of 11 species of Lutjanus from 

 the central GBR were formed once per year and to estimate 

 the dates of completion of these zones. Long-term recovery 

 of useful sample sizes of GBR lutjanids marked with OTC 

 is difficult because of ontogenetic changes in habitat of 

 many species, the low levels of fishing effort, and tag 

 shedding. Therefore we held captive fish to complement 

 recoveries from a field-tagging program. 



Our objectives were to develop a model for otolith 

 growth past OTC marks that would allow tests of the null 

 hypothesis of annual periodicity of opaque zone formation 

 and that would provide mean estimates and measures 

 of variation of the date of completion of these zones for 

 comparison with environmental variables. These objectives 

 required that we compare interpretations of otoliths and 

 otolith growth between captive and wild fish, that we 

 compare different reading axes and readers, and that 

 we compare model estimates with the earlier validation 

 approaches. 



Materials and methods 

 Field techniques 



Fish were collected with baited lines or traps ( see Newman 

 and Williams, 1995) in shallow water (<10 m) around the 

 wharf at Cape Ferguson and Pandora Reef, and in deeper 

 water (<37 m) in the GBR lagoon off Cape Bowling Green 

 and the bases of Rib and Myrmidon Reefs (Fig. 1). There 

 were two separate groups of fish used in the study. One 

 group ( n = 1 70 ) of seven species were tagged and released in 

 the shallow locations, where commercial and recreational 

 fishing was prohibited, and repeated visits were made 

 under research permits to recapture tagged fish. A second 

 gi'oup (n=251) of 12 species were transported from the 

 deeper locations and housed in two sea cages at Cape 

 Ferguson wharf The cages had a volume of 17.7 m^ (cage 

 dimensions: 2.95 m long x 2.45 m deep x 2.45 m wide) 

 and were constructed from 25-mm galvanized steel tube 

 covered in square, galvanized weldmesh (50 mm or 25 

 mm). Fish in each cage were fed with 2.5 kg of chopped 

 pilchards iSardinops sagax) each night. 



Tetracycline was administered to both groups offish in the 

 form of Ten-amycin/MA injectable solution (oxytetracycline 

 hydrochloride at 100 mg/mL) into the coelomic cavity at a 

 dosage rate of 50 millilitres of Terramycin/MA per kilogram 

 of estimated fish weight. In the field-tagging program, the 

 fork length (FL mm) of all fish was measured and they were 

 double-tagged below the spinous dorsal fin with numbered 

 "T-bar" anchor tags or nylon-tipped dart tags. Tag legends 

 advised retaining the whole fish upon recapture, but some 

 fish recaptured after short times at liberty were injected and 

 released again. Batches of captive fish were measured and 

 injected once with OTC. SomeL. erythropterus were injected 



