Intrinsic mechanics and control of fast cranio-cervical movements in aquatic feeding turtles

Intrinsic Mechanics and Control of Fast Cranio-Cervical Movements in Aquatic Feeding Turtles1

SYNOPSIS. Aquatic feeding strikes on agile prey in snake-necked turtles involve fast neck extension, bucco-pharyngo-oesophageal expansion, and head retraction. The ultimate, rectilinear acceleration of the head towards the prey requires complex vertebral rotations, that vary widely from strike to strike. This poses complex motor control issues for the numerous intrinsic neck-muscles, which are the sole neck extensors. Mathematical modelling reveals that extensor activity might be superfluous for this phase of the strike. The ultimate acceleration of the head at the end of the strike always coincides with forceful oropharyngeal expansion. The momentum of the induced flow of water is sufficient to pull the head (and the neck) straight towards the prey. This buccal expansion proceeds identically to that observed in primary aquatic feeders: a rostro-caudal expansion sequence characterized by an optimal timing of the functional components supporting the expansion wave. Yet distinct structural solutions, both at the skeletal, and muscular level, are involved. This points towards prominent hydrodynamic constraints. Head and neck are retracted by extrinsic neck muscles. Given the high number of degrees of freedom, this musculo-skeletal system is obviously under-determined, which compromises control. We propose that erroneous folding of the neck (Le., diverging from the highly persistent retracted configuration) might be avoided through the presence of a subtle click system at the level of the joint between cervical vertebrae 5 and 6.

INTRODUCTION

Despite the recent debate on the phylogeny of turtles (for a comprehensive synopsis see Pough et al., 2001), there is little doubt that the characteristic 'shell' was present in the earliest representatives of this taxon (Benton, 1996; Pough et al., 2001). According to Lee (1997), the evolutionary development of the shell, and the broad flattened body are linked to the herbivorous diet of the ancestral forms. This led (through a network of self-consolidating selective interactions; see Lee, 1997) to the slow, clumsy, and heavily armoured recent turtles. At first glance, such a design seems to lock turtles into herbivory or a scavenging diet. In the best case only non-elusive prey could be taken.

This holds true for most terrestrial turtles (Pritchard, 1979), but in the aquatic environment many species show adaptations allowing them to exploit another feeding niche: active predation. In most cases (except, for instance luring by the alligator snapper), active predation requires a certain amount of compensatory and/or inertial suction (Van Damme and Aerts, 1997). In this way, even elusive prey like fishes can be captured (e.g., mata mata, snapping turtles, snake necked turtles, etc.; see for instance Pough et al., 2001; Lauder and Prendergast, 1992; Summers et al., 1998; Van Damme and Aerts, 1997). (For an extended bibliography on turtle feeding we refer to Schwenk, 2000).

Long necked turtles further compensate for their reduced mobility (imposed by the rigid, heavy shell) by combining suction feeding with a sudden fast dart of the head towards the prey. In snake necked turtles (e.g., Chelodina longicollis), a feeding bout typically consists of a slow approach of the prey through head/neck as well as body movements; a fast strike of the head/neck during which the prey is taken; and subsequent head retraction (Van Damme and Aerts, 1997; Van Damme et aL, submitted). During the first stage, the prey is targeted and the long axis of the head is aligned with it. The fast strike always shows a quasi linear displacement of the head (irrespective the position of the prey with respect to the body) and goes along with oropharyngeal expansion for suction. Afterwards the head is retracted.

Feeding in this way might be associated with (at least) three control-problems posed to the musculo-skeletal system of head and neck. (1) Many joints must be moved simultaneously, and in a highly coordinated way, by a multitude of intrinsic neck muscles to obtain the fast, precisely directed final dart towards the prey. (2) Oropharyngeal expansion must generate a proper suction flow. (3) After the strike, the flexible neck must be folded in a particular configuration by a limited number of extrinsic neck muscles.

In the next sections we will focus on each of the above mentioned issues. Special attention will be paid to the manner in which the intrinsic properties of the head-- neck system can influence (simplify?) its muscular control. We will argue how the hydrodynamics of suction can play a role in both neck extension and retraction.

FAST STRIKE

The head-neck-body system of Chelodina (turtles in general), is an open kinematic chain of 10 links (body + eight neck vertebra + head), in extension controlled by a multitude of small mono- and multi-articular muscles (more than 50; for a morphological description of the neck muscles of Chelodina see Shah, 1963). This implies a redundancy problem at both the kinematic and kinetic level (cf., Van den Berg, 2000). Theoretically, an indefinite number of combinations of intervertebral rotations, each potentially driven by several alternative muscle activation patterns, is available to strike a target (at any position) by means of a linear head displacement. If it is further taken into account that at any instant the net muscle moments on each vertebral segment (also those showing no rotation) must equal the angular inertia (i.e., moment of inertia times angular acceleration) minus the external moments acting on the segment (see Fig. 1; dynamic equilibrium; see for instance Enoka, 1994; Nigg and Herzog, 1999) it is obvious that the motor programs to apply for fast, accurate strikes are inevitably extremely variable and highly unpredictable. In other words, it is doubtful that these fast ballistic neck movements can be the expression of fixed or general motor patterns, or that these movements could rely on feed-forward information gained during the pre-strike only. If true, this implies that the fast neck extension during the strike should be steered by instantaneous feedback. But, given the complexity of the dynamic interactions, the redundancy of the system and the high speed of the neck motion (accelerations up to 44 rn/sec2 are recorded for Chelodina longicollis; Van Damme and Aerts, 1997), this seems a rather troublesome task.