On the error involved in high-speed film when used to evaluate maximum accelerations of fish

Abstract

An analysis of hypothetical, film-derived, and accelerometer-measured acceleration–time data of fish fast starts indicates that the total error in film studies is the sum of sampling frequency error (i.e., the error due to over-smoothing at low film speeds) and measurement error. Measurement error depends on the digitizing apparatus, the image magnification, and the magnitude of the acceleration in question. Total error approaches a minimum at an optimal film speed determined by the magnification of the image. This error is reduced by increasing image magnification and by selecting a new (higher) film speed. Theoretical expectations of total error were tested experimentally by comparing statistically differentiated film records with accelerometer records of fast starts in fish. Average experimentally determined error, for a range of film speeds, was found to be within 13% of the theoretical average. Accelerometer data recorded during fast starts of pike indicate that previous assessments have underestimated performance. The maximum acceleration rate here, for a C-type escape, was 241.2 m∙s−2, more than twice that previously reported for any fish.