Morphometric minefields—towards a measurement standard for chondrichthyan fishes

Size measurements are crucial for studies on the growth, maturation, maximum size, and population structure of cartilaginous fishes. However, researchers use a variety of measurement techniques even when working on the same species. Accurate comparison of results among studies is only possible if the measurement technique used is adequately defined and, if different techniques are used, a conversion equation can be derived. These conditions have not always been met, leading to invalid comparisons and incorrect conclusions. This paper reviews methods used for measuring chondrichthyans, and summarises the variety of constraints that influence the choice of a measurement technique. Estimates of the variability present in some measurement techniques are derived for shortfin mako shark, Isurus oxyrinchus, porbeagle shark, Lamna nasus, blue shark, Prionace glauca, Antarctic thorny skate, Amblyraja georgiana, and Pacific electric ray, Torpedo californica. Total length measured with the tail in the natural position (sharks) and disc widths (batoids) have higher variability than other methods, and are not recommended. Instead, the longest longitudinal axis should be measured where possible and practical; i.e., flexed total length for sharks, total length for batoids (excluding suborder Myliobatoidei), pelvic length for batoids of the suborder Myliobatoidei, and chimaera length (snout to posterior end of supracaudal fin) for chimaeroids (except for Callorhinchus, for which fork length should be measured from the anterior edge of the snout protuberance). Straight-line measurements are preferred to measurements over the curve of the body. Importantly, measurement methods must be clearly defined, giving information on the anterior reference point, the posterior reference point, and how the measurement was made between these two. Measurements using at least two different methods are recommended on at least a subsample of the fish in order to develop conversion regression relationships.