Consequences of differences among readers in age estimations of Baltic cod larvae and early juveniles for growth rate and hatch date analysis. Is more experience always better?

Abstract

We evaluated the precision (i.e., reproducibility of estimations) and accuracy (i.e., reflection of true age) of ageing larval and early juvenile Baltic cod, Gadus morhua, (4.1–33 mm SL) from sagittae and lapilli by inexperienced and experienced readers and an expert. For all three readers, lapilli provided higher precision and accuracy of age estimates than did sagittae. Moreover, with the used method of unpolished otoliths analysis, lapilli allowed ageing juveniles larger (up to 33 mm SL) than did sagittae (up to approximately 18 mm SL). With regard to the level of experience of the reader ageing larval fish, the highest precision in ageing was achieved by the expert and experienced reader, and the lowest by the inexperienced reader. Interestingly, reader experience was not reflected in the accuracy of estimates as it was higher with the inexperienced reader than with the experienced one. Differences in ageing among readers had significant consequences for growth rate analysis (slope of regression line fitted to SL-at-age data). The results obtained covered a wide range of values, from 0.11 mm day−1 to 0.45 mm day−1 for sagittae and from 0.21 mm day−1 to 0.68 mm day−1 for lapilli. The consequences of ageing error for back-calculated hatching dates were statistically significant as well. It may be therefore recommended to use lapilli for cod larvae and early juveniles ageing. Moreover, it is recommended not to make a quick assumption that ageing done by more experienced reader must provide more accurate results that better reflect true age, even if precision of the increment counts is for the more experienced reader high.