Thread trailing of turtles: methods for evaluating spatial movements and pathway structure

Abstract

Although many authors have used thread trailing to monitor movements of turtles and other vertebrates, most have evaluated only a portion of the information contained in these pathways. We describe ways of extracting information from thread trails by determining length of trail, mean distance moved per unit time, net displacement, area encompassed by the trail, mean turning angle, and mean vector length. We further describe and compare four measures of trail sinuosity, the ratio of greatest distance to length of path (d/L), the square root of the ratio of area to path length [Formula: see text], fractal dimension (D), and sinuosity (S); we then applied these methods to the trails of a population of ornate box turtles (Terrapene ornata) from the Nebraska Sand Hills. Though preliminary, these data suggest that vegetation cover affects the sinuosity of turtle trails (based on D), that temperature affects the mean distance moved per day, that T. ornata ordinarily show little or no directional bias, and that injury can affect distance moved, trail sinuosity, and mean turning angle. Among these measures of trail sinuosity, D and S are highly correlated and provide similar information. Though less closely correlated with D and S, or with each other, [Formula: see text] and d/L also reflect trail sinuosity; of these indices, however, d/L best reflects trail directionality. This analysis suggests that a more comprehensive evaluation of pathways at this fine scale will provide new insights into the routine movements of animals.