A critical appraisal of marginal increment analysis for assessing temporal periodicity in band formation among tropical sharks.

Marginal increment ratio (MIR) analyses were conducted as part of age and growth studies on three coastal/semi-oceanic species, the smalltail shark, Carcharhinus porosus, dagger-nose shark, Isogomphodon oxyrhynchus and the night shark C. signatus, and two ubiquitous oceanic species, blue shark, Prionace glauca, and whitetip shark, C. longimanus, collected in equatorial areas off Brazil with the aim of establishing the interspecific temporal nature of vertebral band formation. Monthly variations in marginal bands were analyzed using mean MIR on the entire sample as the standard method for all species. Reasons for the inconclusive results regarding these species are critically appraised with respect to three main sources of bias that are associated with marginal increment analysis (MIA). Bias due to insufficient sample sizes may have hampered the analysis for I. oxyrhynchus and C. longimanus due to movements from shallow waters to seamounts for the former species and to extensive migrations for the latter. Bias due to data collection over too long a period is thought to have influenced monthly mean MIR for C. porosus and P. glauca. For the latter, individuals from different age groups lay down rings at different times, making band deposition inconsistent between individuals. Finally, bias due to births occurring over too long a period was the prevalent cause for confounding MIR values among I. oxyrhynchus and C. signatus species, whose birth period lasts several months and leads to different ages within the same cohort. Other approaches used for MIA in C. signatus and P. glauca led to distinct times of band formation by age-groups when compared to MIR applied on the entire sample. For the daggernose shark, delays in events related to the reproductive cycle from one year to the next were also found to confound MIR. Requirements for the use of MIR implying a band width that displays a sinusoidal cycle when temporally plotted (month or season) were not fulfilled for any of these species. The method has been of little utility for detecting the periodicity of band deposition among sharks from the tropics. This emphasizes the need for supplying information on the temporal periodicity of pair deposition based on other methods. [ABSTRACT FROM AUTHOR]