SUBMERGED WALKING IN THE EPAULETTE SHARK HEMISCYLLIUM-OCELLATUM (HEMISCYLLIDAE) AND ITS IMPLICATIONS FOR LOCOMOTION IN RHIPIDISTIAN FISHES AND EARLY TETRAPODS

The small reef-dwelling orectolobiform shark Hemiscyllium ocellatum us es both sets of paired fins to walk over submerged substrates. On flat , horizontal surfaces, the diagonal fin-pairs function almost in synch rony, with each pelvic fin being activated slightly ahead of the contr alateral pectoral fin. This yields a gait close to a walking trot. As with extant tetrapods, the durations of both the locomotor cycle and i ts contact phase decrease with increasing locomotor speed. Unlike exta nt tetrapods, the phase relations of ipsilateral paired appendages do not vary with speed, but the pectoral fins remain around 0.6 of a cycl e out of phase with the pelvics. Bottom-walking in Hemiscyllium is ass ociated with waves of lateral bending of the body axis. In slow walkin g, lateral bending is more pronounced at the head than the tail. With increasing walking speed, there is a marked increase in bending in the terminal portion of the tail. Fast walking gives way to a mode of pro gression that is transitional to swimming at higher speeds. During sub merged walking, both girdle relation and rotation of the paired fins a bout their girdles contribute to stride. Preliminary measurements indi cate that rotation of the pectoral girdle contributes between 65% and 95% of each stride, whereas pelvic-girdle rotation contributes only 15 -30%. The remainder of each stride comes from rotations of the paired fins on their girdles. Modeling the vertical forces operating on a ''s tandard'' quadruped that progresses using a slow-walking-trot indicate s that this gait is unstable in a terrestrial environment. However, if such an animal is submerged, its effective weight is greatly reduced and the drag forces resisting toppling are much increased; thus, stati c instability does not present a serious problem. Moreover, because He miscyllium ocellatum has an elongate tail, the animal's centre of mass lies posterior to its pelvis. Because the tail serves as a fifth supp ort, H. ocellatum is statically, as well as dynamically, stable throug hout the locomotor cycle when it uses a gait close to a slow-walking-t rot in water. When removed from water and placed on a sandpaper-covere d horizontal substrate, individuals of H. ocellatum progressed effecti vely using a similar mode of walking. Data from non-teleost fishes on the relationship between the length of propulsive waves used during sw imming and the distance separating the pectoral and pelvic girdles, su ggest that if rhipidistian fishes undertook submerged walking and if t heir progression was generated substantially through lateral undulatio ns of the body axis, they likely would have used a walking trot. Such data also support the suggestion that rhipidistians might have activat ed their paired fins in a trotlike pattern if they undertook overland journeys, even though this may have been statically unstable. Several lines of evidence suggest that submerged walking in early amphibians p robably was achieved using a walking-trot. For terrestrial locomotion, the situation is more complicated. If most early amphibians were prop ortioned like Ichthyostega, overland progression with the body lifted above the ground would have been ineffective, unless the ability to ac tivate the paired appendages in a lateral sequence had already been ac quired. For longer-tailed forms like Acanthostega, stable overland pro gression using a walking-trot may have been possible. It is difficult to determine when tetrapods (or their immediate ancestors) first becam e capable of using both the lateral sequence walk and trot. However, i f all Late Devonian amphibians prove to be very similar in their trunk proportions, Australian trackways of this age would provide evidence for such capability. Less equivocal evidence for the two gaits is prov ided by Late Carboniferous trackways from North America.