Identifying fish scales: The influence of allometry on scale shape and classification

The shape of fish scales is to a considerable extent species-specific and is also useful in the determination of stock membership. Size is also important in distinguishing fish scales, both in absolute terms and potentially through allometric effects on scale shape. The present study uses landmarks and geometric morphometric statistical approaches to address two specific questions: (1) if and how fish scale shape varies with a general measure of size; fish fork length. This is assessed by regressing shape on fork length and (2) To what extent size, expressed as fish fork length (FL), is an important factor that has to be taken into account in the identification of species. The analysis is conducted with five fish scales taken from particular body regions from the longitudinal and transverse axis. Cross-validated discriminant analyses are carried out on shape, form (shape plus size) and allometrically standardised data (i.e. allometrically adjusting scale shape according to FL). Using only shape (without size), identification rates are much better than chance, and by taking size into account, but not allometrically adjusting shape, classification is improved somewhat. Allometric adjustment fails to improve classification markedly beyond what is achieved using shape alone. Scales from the central region perform best for identification purposes. Subsequent analyses ignore anatomical sampling region and still achieve good classification by species, especially when size and shape are combined (form). The findings indicate that in large scale survey studies of fisheries or when only scales are readily available, good identification of species is possible irrespective of allometry and anatomical sampling region. As such, scales offer a cheap, rapid and reliable on-the-spot means of identification of fish that have been caught, even when no other material is available.

Research highlights
► This is an important study; a simple quick and cheap method gives a substantially better identification rate than is expected by chance. This is entirely novel information that may well be of interest to other scholars.
► It is new, until our knowledge there are not available references on this issue. Since normally body size varies considerably within and between species (and local populations) allometric trends in fish scale shape need to be elucidated.
► Knowledge of scale shape-body size scaling relationships within and among fish populations to which scales need to be identified could improve identification rates.
► We expose the percentage of size participation on discrimination rate for different fish scales along the fish body using four species.