Navigating nature's terrain: Jumping performance robust to substrate moisture and roughness by blackspotted rockskippers (Entomacrodus striatus)

Escape responses are vital for the survival of prey. The high speeds and accelerations needed to evade predators successfully require exerting forces on the environment. Unlike water, terrestrial habitats can vary in ways that constrain the forces applied, requiring animals to adjust their behavior in response to variable conditions. We evaluated the terrestrial jumping of an amphibious fish, the blackspotted rockskipper (Entomacrodus striatus), to determine if substrate roughness and wetness influence jumping performance. We predicted that rockskippers produce a greater force output as substrate roughness increased and wetness decreased. Using a novel waterproof force-plate capable of detecting milliNewton loads, we collected ground reaction forces from rockskippers jumping on wet and dry sandpapers of varying grits. We also used micro-CT scans to quantify muscle mass as a relative fraction of body mass to determine if these jumps could be performed without power amplification. Mixed-model analysis of jumps revealed significantly higher maximum horizontal forces, jump duration, and maximum power on dry vs. wet substrates, but no effect of substrate roughness. However, final jump outcomes (takeoff speed and angle) were unaffected. Peak jump power was within the range of typical fish muscle. Thus, these fish display a jumping behavior which is robust to substrate property variation.

Methods
Data Collection: Force data were recorded from blackspotted rockskippers (Entomacrodus striatus) using a custom 3D-printed force plate with strain gauges. The plate was designed to detect forces as low as 2 mN along three axes (vertical, fore-aft, and medio-lateral) and was sampled at 1000 Hz using a NIDAQ USB-6002. As rockskippers jumped from the platform, strain gauges recorded deformations as voltage changes, which were then amplified and stored.

Calibration and Filtering: The recorded voltage data were converted to force values through a least-squares linear regression calibration, correcting for cross-talk between channels and calibrated with weights applied across the force range (2, 10, and 20 grams). A Savitzky-Golay filter (window length = 6, polynomial order = 3) was applied to the data to minimize noise and smooth the force traces.

Data Analysis: Post-processing involved calculating takeoff angle, acceleration, velocity, and power for each jump using the calibrated 3-axis force data, time, and the mass of each rockskipper. Acceleration was derived by dividing force by mass. Velocity was calculated by integrating acceleration over time. Takeoff angle was determined as the arctangent of the vertical and horizontal velocities at takeoff.Power was computed by taking the rate of change of kinetic and potential energy over time.

Statistical Analysis: The processed data were analyzed using mixed-model ANCOVA, with individual rockskipper identity as a random factor to account for repeated measures, and substrate wetness and roughness as fixed factors. Rockskipper mass was included as a covariate to control for size differences across individuals. Dependent variables included takeoff angle, maximum power, maximum work, maximum horizontal force, maximum vertical force, and jump duration. Additional tests, such as Levene’s test and Ryan-Joiner tests, confirmed the homogeneity of variances and normality of residuals, with transformations applied to non-normal distributions as needed.