Open wide : the law of physics that propels rockets into space enables an Australian turtle to catch a darting fish

Natural History, March, 2003 by Adam Summers

Slow and steady might win races for tortoises, but it's not clear that the same strategy would work for a pond turtle ambushing its prey. Imagine one of these torpid reptiles trying to hide its awkward shell from a school of minnows: The turtle crouches warily behind a tuft of vegetation. Suddenly ... long pause ... the creature lumbers out from its blind, racing along at inches per second in hot (but clumsy) pursuit of its meal. Favorable comparisons between the turtle and, say, a cheetah lying in wait for a Thomson's gazelle do not spring to mind.

Yet--who'd have thunk it?--several turtles make fine ambush predators. The massive alligator snapper, for one, lures fish into its gaping mouth by twitching the tip of its tongue. Another, the Australian snake-necked turtle, grabs its prey with a quick, serpentine strike. The basic mechanics of its strike are both surprising and surprisingly effective.

The Australian snake-necked turtle (Chelodina longicollis) is a member of the suborder Pleurodira, a group of turtles limited to the Southern Hemisphere. Many, the Australian snake-neck included, have far longer necks than their cousins, the Cryptodira. One consequence is that a pleurodire cannot retract its head into its shell by bending its neck up and back; instead, the animal must fold its neck sideways into a deep hollow at the front of the shell. But the long neck also enables the turtle to ambush fishes and tadpoles by shooting its head far forward, almost as far as the entire length of its body.

The turtle's head lies at the end of eight neck vertebrae, which are connected to the body by more than fifty muscles. Given such a complex anatomy, one might think that making a high-speed stab at a fish would call for neuromuscular coordination worthy of Barry Bonds hitting a slider. Not so. In fact, as Peter Aerts, a biologist at the University of Antwerp in Belgium and his colleagues have found, the turtle's rapid capture of prey paradoxically requires far less motor control than does a slow, deliberate bite.

How can the stimulation of dozens of muscles in the complicated multijoint system that constitutes the turtle's neck be coordinated in just the right sequence and with just the right timing for the turtle to get its head to its quarry? Consider, as Aerts did, a folding carpenter's rule with ten segments (representing the head, the eight vertebrae, and the body). Starting from a relaxed S-bend, similar to the usual starting position for the turtle, imagine extending the "head" of the rule in a straight line towards a target by adjusting each of the hinges a bit at a time. Impossible? No, but certainly extremely tedious.

A turtle relying on vertebral muscles to extend its neck confronts the same problem--and being methodical is no way to catch a darting little fish. But, as Aerts points out, the rule can be quickly and accurately extended to the target if the head is grasped and yanked in the desired direction. The joints move where they will; perhaps they each follow different bending patterns with each new extension. But the head gets where it's going without wavering off course.

What a handy solution to the problem of extending the carpenter's rule! Yet, at first blush, it appears irrelevant to the case of the turtle. After all, why would a hunted fish yank a turtle's head anywhere--when it probably wouldn't want to touch that head with a ten-foot pole? But nature has other ways to get the job done, as Aerts and his colleagues Johan van Damme and Anthony Herrel realized. With a little help from Sir Isaac Newton, a turtle can actually pull its own head towards its prey.

If that action seems about as likely as a fish committing suicide, recall Newton's third law: for every action there is an equal and opposite reaction. Here the action is a sudden suction, caused when the turtle floods its mouth and throat with a large volume of water. The linchpin of the system for controlling this action is a bony structure called the hyoid apparatus. In most vertebrates the hyoid supports the tongue, as it does in the snake-necked turtle. But in the snake-neck, it also pushes down a bone called the hypoglossum (which, as its name suggests, is situated beneath the tongue), thereby expanding the turtle's mouth. The hyoid also moves bones known as branchial arches, which expand its throat.

The change in the width of the neck is twofold. Expanding the neck downward and to the sides causes water to rush into the mouth and flow down the throat. The Newtonian reaction to the rearward-rushing water then snaps the head forward almost instantaneously (the acceleration of the head can be more than four times the acceleration of gravity). The effect is almost exactly the opposite of what happens when an unattended garden hose is turned on: the straight hose, reacting like a rocket to the water shooting out its end, writhes into S-curves. In the case of the turtle, an S-curved neck straightens when water is sucked inside.

To bolster their hypothesis, the Belgian biologists developed a mathematical model that derives the rearward-rushing volume of water from the observed expansion of the neck. From the mass and speed of the moving water, they could calculate the resultant forward motion of the head and neck needed to offset the momentum of the water. Although the predictions of the model break down as the distance between the head and prey becomes minuscule, they closely match the movements observed during much of the turtle's strike. (When the head is close to the prey, the model tends to overestimate head movement because it does not compensate for the neck's connective tissue and muscles, which absorb some of the kinetic energy as they stretch.)