Consistent size-independent harvest selection on fish body shape in two recreationally exploited marine species

Abstract

Harvesting wild animals may exert size-independent selection pressures on a range of morphological, life history, and behavioral traits. Most work so far has focused on selection pressures on life history traits and body size as morphological trait. We studied here how recreational fishing selects for morphological traits related to body shape, which may correlate with underlying swimming behavior. Using landmark-based geometric morphometrics, we found consistent recreational fishing-induced selection pressures on body shape in two recreationally exploited marine fish species. We show that individuals with larger-sized mouths and more streamlined and elongated bodies were more vulnerable to passively operated hook-and-line fishing independent of the individual's body size or condition. While the greater vulnerability of individuals with larger mouth gapes can be explained by the direct physical interaction with hooks, selection against streamlined and elongated individuals could either involve a specific foraging mode or relate to underlying elevated swimming behavior. Harvesting using passive gear is common around the globe, and thus, size-independent selection on body shape is expected to be widespread potentially leaving behind individuals with smaller oral gapes and more compact bodies. This might have repercussions for food webs by altering foraging and predation.

Keywords: Behavior; ecomorphology; fisheries-induced selection; geometric morphometrics; predator–prey interactions; recreational fishing.

Figure 1

Body shape landmarks (n = 13) acquired in the two study species. The upper panel shows an individual of Serranus scriba and the down panel an individual of Diplodus annularis. In both cases, the coordinates (landmarks) acquired for this study are shown as blue points (labels from i to xiii).

Figure 2

Approach to removing the arching effect following Valentin et al. (2008). Each panel represents the two extreme shapes of the main gradient of shape variation (first principal component axis (PC 1) of the shape descriptors; the amount of shape variability explained is indicated by a percentage). Each landmark of one of the two extreme shapes is indicated by points, and the other is represented by lines connecting the two shapes (note that is arbitrary which of the two shapes is represented by points). Before Burnaby's projection, the landmarks (i) and (vi) point toward one direction and the four central landmarks toward the opposite direction, thus suggesting that the fish is not correctly aligned but bent. After projection, (i) and (vi) point at opposite directions and, in the case of S. scriba (the two panels below), the four central landmarks suggest a deeper/compressed pattern.

Figure 3

Box plots derived from the linear discriminant analysis (LDA) and the mean geometric body shape predicted for an average individual sampled either of the two methods: fished and population sample (note that these shape changes correspond to the sampling method only; the effects of “fish size,” “fish condition” and “study site” have been statistically removed). In both species, the main shape differences were localized at the mouth (landmarks i and x), the insertion of the dorsal and the pelvic fins (landmarks iv and ix) and the posterior extreme of the lateral line (landmark vi).

Figure 4

Patterns of covariation between the head and the trunk (light and solid black dots, respectively) in D. annularis (the two panels above) and S. scriba (the two panels below). The coordinates of the landmarks of the first and the second axis of the partial least squares (PLS 1 and 2) carried out for each species are represented. The maximum (black line) and the minimum (gray line) values observed for each axis of the PLS have been superimposed to improve visualization.