A locomotor innovation enables water-land transition in a marine fish

Abstract

Background: Morphological innovations that significantly enhance performance capacity may enable exploitation of new resources and invasion of new ecological niches. The invasion of land from the aquatic realm requires dramatic structural and physiological modifications to permit survival in a gravity-dominated, aerial environment. Most fishes are obligatorily aquatic, with amphibious fishes typically making slow-moving and short forays on to land.

Methodology/principal findings: Here I describe the behaviors and movements of a little known marine fish that moves extraordinarily rapidly on land. I found that the Pacific leaping blenny, Alticus arnoldorum, employs a tail-twisting movement on land, previously unreported in fishes. Focal point behavioral observations of Alticus show that they have largely abandoned the marine realm, feed and reproduce on land, and even defend terrestrial territories. Comparisons of these blennies' terrestrial kinematic and kinetic (i.e., force) measurements with those of less terrestrial sister genera show A. arnoldorum move with greater stability and locomotor control, and can move away more rapidly from impending threats.

Conclusions/significance: My results demonstrate that axial tail twisting serves as a key innovation enabling invasion of a novel marine niche. This paper highlights the potential of using this system to address general evolutionary questions about water-land transitions and niche invasions.

Figure 1. Tail twisting capability and ecological groupings of blenny genera examined in this study.

Field observations showed that Alticus and Andamia are on land during both low and high tide. In contrast, Paralticus and Praealticus exit the water infrequently during low tide (pers. observ.). Istiblennius, Blenniella, and Entomacrodus all are fully aquatic in their habits and only periodically emerge from water, despite being capable of breathing air. Tail twisting behavior in Paralticus is unknown because individuals were not available for examination. The phylogeny used here was obtained from Spring and Williams .

Figure 2. Dorsal midline splines taken from high-speed video of representative, stereotypical terrestrial jumps.

Presented midlines include a a, terrestrial (Andamia tetradactyla), b, amphibious (Praealticus labrovittatus), and c, aquatic (Blenniella gibbifrons) blenny. The head is indicated by the filled circle, and the darkest midline corresponds to initial body position before movement. Lighter solid midlines indicate the body is contact with the ground, whereas the lightest dashed midlines indicate that the body is raised off the ground. Note the large yaw of the aquatic blenny, c, as its body leaves the surface. Dotted arrows indicate tail movement or overall body movement.

Figure 3. Comparison of aquatic escape and terrestrial jump maneuvers in aquatic, amphibious, and terrestrial fishes.

a, An aquatic C-start escape response by a fully-aquatic fish, Polypterus senegalensis, modified from Tytell and Lauder . b, An aquatic blenny (Blenniella gibbifrons) jumping on land. Note the lack of tail twisting and similar body position to the fish in panel a. c, An amphibious blenny (Praealticus labrovittatus) showing the stereotyped tail to head movement used for terrestrial locomotion. d, A terrestrial blenny, Alticus arnoldorum, demonstrating axial tail twisting. The numbered vertical bars to the right of panel d correspond to jump phases 1, 2, and 3, as shown in panels c and d. See text for descriptions of the phase kinematics. Blennies shown in panels b–d jumped off the same balsa wood surface. The terrestrial blenny never slipped, whereas all others did. Sketches were produced for greater visible clarity, and were traced from high-speed video frames. The gray circle serves as a fixed point. Five mm scale bars are provided at the bottom of panels b–d; panel a serves as a generic kinematic reference for a C-start escape response.

Figure 4. Three terrestrial locomotor modes performed by Alticus arnoldorum.

a, Hopping, b, jumping, and c, climbing. d, Results of a discriminatory analysis on four principal components derived from two principal components analysis (PCA) models of preparatory and propulsive kinematics. All three locomotor modes are distinct (Wilks' Lambda: P<0.0001; ANOVA: F = 159.37, P<0.0001) and categorized with 93.7% accuracy (14 of 15 correct). See Tables 1 and S1 for a list of variables included in the PCA models.