The validity of different methods in the backcalculation of the lengths ofroach - a comparison between scales and cleithra

The backcalculated lengths of roach Rutilus rutilus obtained with three dif ferent proportional methods [body proportional (BPH) and scale proportional hypotheses (SPH), Fraser-Lee] were compared. In the BPH and SPH methods bo th scales and cleithra were used, while the Fraser-Lee method was used with scales only. The validity of the results was evaluated by comparing the re sults with observed lengths within one cohort. Both body-scale and body-cle ithrum relationships were best described by a power function (r(2)=0.83). T he backcalculated lengths of roach at age 1 varied from 51 mm (SPH, scales and cleithra) to 72 mm (Fraser-Lee), a 30% difference. In within-cohort com parisons, the difference compared with observed lengths was smallest in BPH (<4% with few exceptions). The linear Fraser-Lee method overestimated (up to 5%) the lengths at young ages and underestimated the lengths at older ag es (3-5%). In SPH the difference with observed length was 12% at highest. T he length estimates by SPH were more affected by the age-structure of the s ample than the estimates by BPH, which suggested that Lee's phenomenon was present. This was tested by backcalculating the lengths at the age of 2 yea rs for roach, separately from different age-groups. The backcalculated leng ths declined with the increasing age of fish used in the calculations. The phenomenon was strongest in SPH and weakest in Fraser-Lee. The results sugg ested that in the backcalculation of the lengths of roach, BPH method gives the most reliable results. Considering the other methods, Fraser-Lee shoul d be used rather than SPH. (C) 1999 The Fisheries Society oi. the British I sles.