ABSTRACT

Back-calculation is commonly used in age and growth studies to increase the sample size when few fish are captured from a population; however, which structure to use for back-calculation has been often disputed. We compared back-calculated lengths of crappie (Pomoxis spp.) scales and sagittal otoliths from seven Midwestern lakes using linear regression. To examine this relationship on a finer scale, a residual sum of squares (RSS) analysis was also run to determine whether the two structures suggested different mean lengths at age for one population (Cedar Lake, Illinois). A total of 150 fish, age-1 to age-10 (mean 2.8 years), was collected from all lakes; a total of 296 back-calculated lengths was obtained for each structure. Thirteen fish were discarded because of lack of age agreement between structures. Overall for all lakes, the otolith versus scale back-calculated lengths regression yielded an intercept of-22.19 mm and a slope of 1.00 (R²=0.77, df=1, 296, P<0.0001). The slope was not different from 1.00 (95% CI of 0.94–1.07), but the intercept was slightly different from zero (95% CI of-34.66 to −9.72). Furthermore, the RSS analysis suggested that the growth curves for Cedar Lake did not differ between structures (F_3,2=0.29, p>0.75), indicating that the pattern seen across all lakes held true for one population. Either structure could be used for back-calculation studies. However, perfect agreement of ages between structures (i.e., scales and otoliths) did not occur. Crappie length at age should be back-calculated using otoliths when more precise age estimates are needed and with scales when fish are too few to be killed.