Variable Tensility of the Ligaments in the Stalk of Sea-Lily

FAU Institutional Repository

http://purl.fcla.edu/fau/fauir

This paper was submitted by the faculty of FAU’s Harbor Branch Oceanographic Institute.

Notice: © 1994 Elsevier B.V. This manuscript is an author version with the final publication available by http://www.sciencedirect.com/science/journal/03009629 and may be cited as: Wilkie, I. C., Emson, R. H., & Young, C. M. (1994). Variable tensility of the ligaments in the stalk of sea‐lily. Comparative Biochemistry and Physiology Part A: Physiology, 109(3), 633‐641. doi:10.1016/0300‐9629(94)90203‐8

Variable tensility of the ligaments in the stalk of a sea-lily

I. C. Wilkie,* R. H. Emson”f and C. M. Young1 *Department of Biological Sciences, Gfasgow Caledonian University, 70 Cowcaddens Road, Glasgow G4 OBA, U.K.; tDivision of Life Sciences, King’s College, Campden Hill Road, London W8 7AH, U.K.; and $Harbor Branch Oceanographjc Institution, Box 196, Fort Pierce, FL 33450, U.S.A.

The stalk of isocrinid sea-lilies consists largely of skeletal plates linked by collagenous ligaments. Although lacking contractile tissue, it can bend in response to external stimuli. The stalk of Cctnocrinusaster& was tested mechanically to determine whether the mechanical properties of its ligaments are under physiological control. In bending tests, ligaments at the mobile symplexal junctions showed a limited “slackening” response to high K+ concentrations which was blocked reversibly by the anaesthetic propylene phenoxetol. In bending tests and uniaxial loading tests, ligaments at the normally rigid synostosal junctions ruptured in response to high K+, con~~ing that these junctions are specialized for autotomy. It is concluded that the ligaments are mutable collagenous structures whose presence explains the mechanical versatility of the isocrinid stalk.

Key words: Autotomy; Collagen; Crinoidea; Echinodermata; Mutable collagenous tissue; Sea-lily; Stalk ligaments; Variable tensility. Comp. Biochem. Physiol. lOgA, 633-641, 1994.

In~od~ction The sea-lily stalk consists of a series of the virtually stalkless Cyrtocrinida; see disc-shaped ossicles connected by collage- Grimmer and Holland, 1990) has estab- nous ligaments. It functions primarily as lished that the stalk lacks any form of a strut which holds the main body of the musculature (Grimmer et al., 1984, 1985; animal with its crown of suspension-feeding Holland et al., 1991), it has been conjec- arms above the substratum. However, ob- tured that its flexibility can be altered servation of animals in situ has shown that through changes in the mechanical proper- the stalk can also bend to permit changes ties of the stalk ligaments (Donovan, 1989; in the orientation of the crown (Messing, Baumiller et al., 1991). 1985; Fujita et al., 1987; Baumiller et al., Collagenous tissues showing such vari- 1991). Since an ultrastructural examination able tensility, i.e. mutable collagenous of representatives of the three extant orders tissues (MCTs), are ubiquitous in the of truly stalked sea-lilies (i.e. excluding echinoderms and serve two main functions. Firstly, through reversible changes in tensile stiffness, they provide an energy-sparing Correspondence to: 1. C. Wilkie, Department of Bio- logical Sciences, Glasgow Caledonian University, means of fixing posture. Secondly, through 70 Cowcaddens Road, Glasgow G4 OBA, U.K. irreversible disintegration, they permit Received 3 March 1994; accepted 10 June 1994. autotomy of appendages or parts of the

633 634 I. C. Wilkie et al.

body (Wilkie and Emson, 1988; Candia Materials and Methods Carnevali and Wilkie, 1992). With regard to the crinoids, it has been We collected specimens of the isocrinid demonstrated experimentally that the vari- sea-lily Cenocrinus asterius (L.) by suction able stiffness of the finger-like cirri in un- hose at a depth of around 600 m during two stalked comatulids depends on the presence dives in the Johnson-Sea-Link I submers- of MCTs (Wilkie, 1983). However, despite ible at Egg Island and Chub Cay, Common- work on the mechanical properties of the wealth of the Bahamas in May 1993. The sea-lily stalk (Baumiller and LaBarbera, animals were kept in a darkened, aerated 1993), there has been no direct confirmation aquarium at 10°C on board RV Seward that the tensility of its ligaments is under Johnson and used immediately or up to 4 physiological control. This is an important days after capture. omission, since the sea-lilies are the only The stalk ossicles of isocrinid sea-lilies survivors of a wide diversity of stalked are differentiated into nodals that carry the echinoderms which inhabited the seas of the jointed cirri and internodals that lack cirri Palaeozoic era (Paul and Smith, 1984). (Fig. 1). The flexibility of the stalk depends Although these animals were heavily cal- on the presence of mobile joints, or sym- cified, it has been inferred from their skel- plexies, between most of the ossicles, the etal morphology and microstructure that only exception being the junction between their skeletal elements were connected by each nodal and the internodal distal to it muscles only rarely, even in the jointed which is a rigid joint, or synostosis. feeding appendages (equivalent to the arms During the course of mechanical tests, of modern crinoids) (reviewed by Wilkie short lengths of stalk were subjected to and Emson, 1988; see Ausich and Bau- loads of up to 130 g applied through the miller, 1993). In view of their success during lever of an isotonic displacement transducer the Palaeozoic, it is highly unlikely that these animals were entirely unresponsive to external events such as predator attack or changes in bottom current speed and direction. We have therefore speculated that the collagenous ligaments binding their skeletal elements consisted of MCT, and that the reversible “softening” of these ligaments permitted the posture and orientation of both the stalk and feeding appendages to be adjusted passively by gravity and/or water movements (Wilkie and Emson, 1988). The stalk of modern sea-lilies thus serves as an important model for the non- muscularized appendages of Palaeozoic echinoderms. In this paper, we provide data from experiments on the stalk of a sea-lily which indicate that the mechanical properties of its ligaments are variable Fig. 1. Diagrammatic representation of a short length and under physiological control. We have of the stalk of an isocrinid sea-lily. The stippled also demonstrated functional differen- ossicles are nodals (no), from each of which emerge tiation between different stalk ligaments: the cirri (ci). Nodals are separated by a series of internodal ossicles (in). The articulations between those at movable articulations show only most of the ossicles are mobile symplexies (symp). reversible changes in stiffness; those at im- However, that between each nodal and the internodal movable articulations show only irrevers- distal to it is a rigid synostosis (syno). The stalk ible disintegration and are involved in stalk ligaments are shown as vertical lines and comprise the autotomy. symplexal ligaments (sml), synostosal ligaments (snl) and peripheral through-going ligaments (ptl); these A preliminary report of some of these last are continuous from the internodal on the distal findings has already been published (Wilkie side of a nodal to the next nodal and therefore do not et al., 1993). cross a synostosis. Sea-lily stalk ligaments 635

lever ternodals on either side were clamped vertically and subjected to simple uniaxial tension in the form of a constant load ______.___P_3------(Fig. 2~). All experiments were conducted at room 0 temperature (20°C) with the preparations held in air but kept moist by being flooded frequently with seawater or other media. t Results Internodal preparations When a constant load was first applied, (b) cirrus (c) internodal preparations showed an immedi- Fig. 2. Mechanical testing method. Lengths of stalk ate deflection which stabilized within a few were held rigidly at one end in a clamp and the other seconds. Without subsequent treatment, end was attached via a heart clip and silver chain to the preparations did not bend any further. the lever of an isotonic displacement transducer. However, if they were flooded with a Internodal preparations (a) consisted of a series of internodal ossicles and included only symplexal artic- medium containing an elevated potassium ulations (symp). They were clamped horizontally and ion concentration (either seawater mixed showed an upwards deflection (d) when initially with 0.56M KC1 to give a K+ concen- loaded. Nodal preparations included a synostosal tration of 100 mM, or 0.56 M KC1 alone), articulation (syno) and were clamped either horizon- which would be expected to depolarize tally (b) or vertically (c). neural elements, most of them (73%; II = 15) exhibited a further deflection and (Fig. 2a). Two different preparations were then stabilized again, even under high loads used. (Fig. 3). The deflection began 25-106 set (1) Internodal preparations consisted of after the start of treatment (mean + SD: a series of seven to nine internodals and 49.7 +_31.5; II = 11). No internodal prep- included only symplexal articulations. They aration ruptured in response to K+, even were orientated horizontally and clamped at after prolonged exposure (for 16 min) to one end. The other end was attached via a 0.56 M KCl. Elevated K+ concentrations heart-clip and silver chain to the transducer had no effect on preparations which had lever. Application of a load to the other side been immersed for 30 min or more in a of the lever tended to deflect the preparation 0.1% seawater solution of the anaesthetic upwards (equivalent to the lateral flexion of propylene phenoxetol, although responsive- a stalk in its natural vertical attitude) ness was restored after washing in normal (Fig. 2a). These preparations were subjected seawater. to either a constant load or a load increasing Internodal preparations were also sub- by 1, 5, or 10 g increments every 30 sec. In jected to a stepwise increase in load before the latter case, at the end of the loading and after exposure to elevated K+ con- sequence, the load was removed (except for centrations. Excess K+ increased the flexi- a residue of 1 g) in order to record the elastic bility of the preparations, i.e. the amount of response of the stalk. The residual load deflection under a given load. This effect ensured that gravity could be discounted as could be reversed by returning them to a factor in stalk re-straightening. normal seawater (Fig. 4a). When the maxi- (2) Nodal preparations consisted of a mum load was removed at the end of these nodal ossicle with a number of internodals experiments, untreated preparations tended on either side of it, and therefore included to re-straighten immediately by a variable a synostosal articulation. Such preparations amount, i.e. they showed elastic recoil. with five to eight internodals on either side The magnitude of the recoil was directly of the nodal were clamped horizontally as proportional to that of the deflection above and subjected to either a constant recorded immediately before unloading and load or an increasing load (Fig. 2b). Nodal was 48.4 + 22.4% of the latter (mean k SD; preparations with only two to four in- n = 10). Elastic recovery was not affected 636 I. C. Wilkie er al.

of this response was due to flexion at the symplexal articulations alone and did not involve the synostosis. All responses to high K4 were blocked by pretreatment with 0.1% propylene phenoxetol. In tests in which they were clamped vertically (see Fig. 2c), untreated nodal preparations showed no detectable yield under a constant load of up to 40 g (which subjected the synostosal ligaments to a tensile

W Unseats -t- K-treated

* WashEd 0 1 2 3 4 5 6 7 8 9 10111213 Time (min) Fig. 3. Mechanical behaviour of stalk preparations under constant load. Tracings of curvilinear oscillo- graph recordings of deflection against time for typical internodal and nodal preparations. The internodal preparation (IP) was subjected to a load of 5 g, as in I-\ 0 10 20 30 40 50 Fig. 2a, which caused an immediate deflection, then flooded with 0.56M potassium chloride (K) which caused a further limited deflection. The nodal preparation (NPl) was damped horizontally as in Fig. 2b and loaded with 40 g, then treated with 0.56 M KC1 resulting in rupture at the synostosis. The nodal preparation (NPZ) was clamped vertically as in f Fig. 2c and subjected to a load of 40 g then flooded E* with high potassium (100 mM) sea water which after a delay resulted in its abrupt rupture at the synosto- Z 8 sis. Although recordings NPl and NP2 started at zero 01 deflection, they have been displaced upwards in the a l Untreated figure. (The total number of tests conducted in each 13 K-treated case was IP = 15, NPl = 5, NP2 = 17.) significantly by excess K+ ions (Fig. 4b): K+-treated preparations showed a mean Fig. 4. Mechanical behaviour of stalk preparations recoil of 41.0 f 13.3% (n = 5; using t-test, under increasing load. {a) Excess K + caused a revers- P > 0.1 vs. untreated preparations). ible increase in the ~exibility of internodal preparations. Typical results from a single preparation (a Nodal preparations total of 10 preparations was treated similarly). Before treatment, this was subjected to a preload of 1 g and In bending tests, the initial responses then the load was increased by 10 g increments every of nodal preparations to a constant load 30 sec. This procedure was repeated after the prep- were similar to those of the internodal aration had been immersed in 0.56 M KC1 and then again after a prolonged seawater wash. (b) When preparations: they showed an immediate unloaded, untreated internodal preparations abruptly deflection which stabilized quickIy. When re-straightened by a varying amount. The magnitude they were then treated with a high K+ of this reverse deflection, or elastic recoil, was directly medium, they started to bend further after proportional to the magnitude of the deflection im- 20-70 see (mean _+SD: 47.0 + 22.8; n = 5), mediately before removal of the load. The regression line for this relationship in untreated preparations is but stabiiized again as long as they were given. This relationship was not affected signifi~ntly washed quickly with normal seawater. A by high K+ concentrations. These tests were per- second exposure to high K+ resulted in a formed at the end of loaddeflection sequences like biphasic response: a slow deflection which that described in Fig. 4a. The maximum load was began to stabilize but was interrupted by removed except for the 1 g preload whose presence ensured that gravity could be discounted as a factor abrupt fracture at the synostosal junction in the straightening of the unloaded stalk. Each point (Fig. 3). Visual inspection of the prep- on the graph shows the response of one preparation arations indicated that the slow phase (Fig. 4a from Wilkie et al., 1993). Sea-lily stalk ligaments 631 stress of c. 100 KPa). When treated with sets of ligaments: symplexal ligaments at excess K+, most of them (65%; n = 17) the mobile junctions; synostosal liga- ruptured abruptly and cleanly at the synos- ments at the rigid junctions; and peripheral tosal junction (Fig. 3). Rupture occurred through-going ligaments which are continu- 2.2-16.0 min after the start of treatment ous from the internodal ossicle on the (mean + SD: 5.8 + 4.1; II = 11). In all cases distal side of a nodal to the next nodal and of rupture which were observed, breakage which therefore do not cross a synostosis occurred at the synostosis and never at (for further information on the disposition any of the symplexal junctions in these and ultrastructure of these ligaments, see preparations. Grimmer et al., 1985). Some nodal preparations were not The results reported in this paper demon- attached to the mechanical testing appar- strate that the mechanical behaviour of the atus but were immersed in a high K+ isocrinid stalk can be altered by exposure to medium. When separation at the weakened media containing an elevated K+ concen- synostosal junction was then brought about tration. Such an alteration can be due only by gentle manual traction, the ligaments to changes in the stiffness of the ligaments became visible as five silky strands with that interconnect the stalk ossicles: in view very low tensile strength and enormous of the considerable loads imposed on extensibility (Fig. 5). the stalk in the above experiments, it is Potassium-induced rupture of the nodal highly unlikely that other soft tissues such preparations was blocked reversibly by as epidermis, internal epithelia or neural 0.1% propylene phenoxetol. components could contribute significantly to the recorded mechanical responses. The Discussion stalk ligaments thus appear to include mutable collagenous tissue (MCT). Eflect of K+ on the mechanical properties of Our experiments provided some qualitat- the stalk ive information on the mechanical proper- As shown in Fig. 1, the ossicles of the ties of the stalk ligaments. On imposition of isocrinid stalk are held together by three a constant load in bending tests, the stalk

Fig. 5. Autotomizing nodal preparation. This had been immersed in 0.56 M KCI. The distal end was then held down by the first author’s left thumb while the proximal end was gripped by forceps and pulled gently. The stalk ruptured at the synostosis and the synostosal ligaments became visible as fine silky strands (arrow) with enormous extensibility. 638 1. C. Wilkie et al. pieces showed an immediate deflection sufficiently to increase the flexibility of which stabilized quickly so that the load the stalk while still maintaining its overall was sustained without further displace- integrity. Such a differential capacity for ment. In a bending isocrinid stalk, the variable tensility has been demonstrated fulcrum is located at the edge of the stalk in adjacent MCTs in other echinoderms. ossicles on the flexion side (Donovan, 1989, For example, in the arm of the brittlestar 1990; Baumiller et al., 1991). The initial Ophiura ophiura, the ligament at the proxi- bending of the stalk of C. asterius must mal edge of each oral arm plate shows only therefore have involved mainly stretching reversible changes in stiffness, whereas the of the ligament fibres on the extension side ligament at the distal edge can, in addition, of the stalk. The fact that the stalk did show irreversible disintegration during arm not continue to bend indicates that after autotomy (Wilkie, 1992). undergoing some initial strain, the ligaments can sustain a load without internal Mode of action of K+ shear and therefore that their constituent The action of elevated K+ concentrations collagen fibrils are connected by stable was blocked by pre-treatment with propy- linkages (probably mediated by proteo- lene phenoxetol. Propylene phenoxetol (l- glycan molecules: Trotter and Koob, 1989; phenoxypropan-2-01) as a 0.1% solution in Erlinger et al., 1993). seawater has a reversible anaesthetic action In internodal preparations, elevation of on whole echinoderms and isolated echino- the K+ concentration caused a further lim- derm tissue preparations. It blocks impulse ited flexion which again stabilized. This propagation in squid giant axons (Haydon, could have been achieved only through unpubl.) and ophiuroid giant axons further stretching of the ligaments involving (Cobb, unpubl.). The reversible abolition of abolition or weakening of some of the responsiveness to K + by propylene phen- “stable” linkages that had been maintain- oxetol therefore suggests that these ions ing the original equilibrium condition. depolarize neural elements that influence In effect, high K + “slackened” the liga- the stiffness of the stalk ligaments. Such ments so that they allowed a wider range of nervously mediated changes in mechanical rotation about the stalk joints but still set a properties have been demonstrated in maximum limit to deflection and retained many echinoderm mutable collagenous their elastic properties. These preparations tissues (Wilkie and Emson, 1988) and are contained both symplexal ligaments span- believed to be effected directly by neuro- ning single stalk joints, and peripheral secretory-like juxtaligamental cells which through-going ligaments spanning the are in synaptic contact with motor neurons whole preparation. In the present state of (Wilkie, 1979; Cobb, 1985). Juxtaliga- knowledge, it is not possible to evaluate the mental cells are present in all the liga- relative contribution of these two types of ments of the isocrinid stalk, although their ligament to the overall mechanics of the neural connections have not been identified stalk or to its variable tensility (Baumiller (Grimmer et al., 1985). and LaBarbera, 1993). Whilst the effects of high K+ concen- In nodal preparations subjected to tration may be nervously mediated, the bending tests, limited treatment with elev- long response times to this treatment are ated K+ had similar effects as on inter- disturbing; in bending tests on nodal and nodal preparations. However, prolonged internodal preparations there was a delay of or repeated exposure to high K+ resulted around 1 min between the start of treatment in the fracture of the preparations which and the start of ligament relaxation, and occurred always at the synostosal articula- in uniaxial tests on nodal preparations tion. This was also demonstrated in nodal there was a delay of about 6 min before preparations subjected to uniaxial loading. rupture occurred. These delays may rep- The synostosal ligaments can thus undergo resent the time required for K+ to diffuse an extreme loss of tensile strength and so through stalk tissues and reach a critical bring about stalk fracture, whereas the liga- concentration in the vicinity of undamaged ments that cross the symplexal junctions neural elements that control ligament lack this capacity and can only “slacken” tensility. Furthermore, it has been our Sea-lily stalk ligaments 639 experience with other echinoderm prep- of Chladocrinus decorus and confirmed by arations that their response times to ionic ourselves on anaesthetized stalk pieces from manipulation tend to be shorter when they C. asterius that it is impossible to break the are fully immersed in test solutions than stalk mechanically through a synostosis: the when they are held in air and test solu- stalk always fractures through a symplexy tions are pipetted over them, as in these or partly through a symplexy and a stalk experiments. ossicle. This emphasises the point that, even though the synsostoses may be special- Stalk uutotom.~ ized for autotomy, they are not planes of The finding that the capacity for com- predetermined weakness. At autotomy, a plete rupture is restricted to the ligaments of synostosal junction becomes weakened by the synostosal junctions substantiates the a physiological mechanism which brings long-held view that these articulations are about a drastic drop in the tensile strength specialized for autotomy. This had been and increase in extensibility (Fig. 5) of the inferred previously from various pieces of synostosal ligaments. evidence, e.g. (1) The stalk of captured isocrinids often Role of the ligaments in stalk mobility fragments at synostosal junctions but not Isocrinids rest on the sea-floor with a at symplexal junctions (Baumiller and significant proportion of the distal stalk LaBarbera, 1993; pers. ohs.). lying on the substratum attached by its cirri (2) The stalk ossicles are penetrated by and most of the remainder of the stalk held an axial canal which houses longitudinal vertically (Messing, 1985; Fujita et al., nerves and coelomic and haemal channels. 1987). Since the animals are neutrally buoy- At each synostosis, the axial canal is ant and since it has been calculated that partly infilled by calcite spicules which are hydrodynamic lift could be a factor con- believed to expedite the clean rupture of tributing to their upright posture only in the soft tissues during autotomy and the unusually strong currents, the normal rapid occlusion of the canal by skeletal erect attitude of attached animals must growth after autotomy (Grimmer et al., depend largely on the rigidity of the stalk 1985; Donovan, 1984). (Baumiller, 1992). However, direct obser- (3) It has been found that the distal vation of C. asterius and other isocrinids in extremity of the isocrinid stalk never con- situ has established that the stalk posture sists of any specialized anchoring structure, of individual animals can change. In as is the case in bourgueticrinids and mil- C. aster&s, the distal 20-50% of the staIk lericrinids (Grimmer et al., 1984; Holland lies on the substratum (Messing et al., et al., 1991), but always ends abruptly with 1988). Where there is no bottom current or a normal nodal ossicle (Macurda and only a weak current, the rest of the stalk is Meyer, 1974; Donovan, 1984; Fujita et al., vertical with the oral surface facing up- 1987). This was certainly the case with the wards and the arms drooping downwards captured specimens of C. asterius used in what has been described as the “wilted in this investigation, in which it was obvi- flower posture” (Macurda and Meyer, ous from the sealed axial canal and the 1974). If the current strengthens, localized discoloration of the distal facet of the termi- bending of the proximal stalk allows the nal nodal that fracture had not occurred crown to tilt so that the concave undersur- recently. Autotomy of the distal stalk face of the arms faces the oncoming cur- would appear to be an invariable event in rent, this presumably being a strategy to the isocrinid life-cycle, although whether maximize the particle collecting potential this might represent an energetic strategy of the arms (Macurda and Meyer, 1974; (to reduce overall metabolic demands) or Messing, 1985; Baumiller et al., 1991). It a means of achieving rapid detachment has been deduced from a consideration of from the substratum (e.g. when the cur- hydrodynamics and mechanical properties rent becomes too strong: Roux, 1987) is that flexion of the proximal stalk in such unknown. circumstances can occur only if there is Paradoxically, it was noted by Donovan a reduction in the stiffness of the stalk (1984) with regard to preserved specimens ligaments (Baumiller and LaBarbera, 1993). 640 I. C. Wilkie er al.

Our investigation has provided the first sumably less expensive in energetic terms direct evidence that the stalk ligaments have for such mechanical adaptability to depend this capacity. on external forces (due to gravity and water It can be envisaged that in the upright movements) and the variable tensility of “wilted flower posture”, the ligaments of the MCTs than for it to depend on the active symplexal junctions are in a “stiffened” con- contractions of muscle tissue. The presence dition throughout the length of the stalk. In of MCTs in the stalk and perhaps elsewhere response to a strengthening bottom current, in isocrinids (Birenheide et al., 1993) may the ligaments of the proximal stalk then thus be an energy-saving adaptation which slacken, allowing the crown to tilt and even explains in part the unusually low meta- to rotate around the vertical axis if there is a bolic rates of these animals (Baumiller and change in current direction (Baumiller et al., LaBarbera, 1989) and which may have con- 1991). Our results also suggest that sub- tributed to the early evolutionary success sequent restraightening of the proximal of echinoderms in general (Wilkie and stalk could be assisted by the elasticity of the Emson, 1988). stalk ligaments, although without measure- ments of the forces involved, it is difficult to Acknowledgements-We are grateful for the skilled assess the significance of the elastic be- assistance of the crews of the RV Seward Johnson haviour recorded in these experiments. Ces- and Johnson-Sea-Link I. This work was supported sation or weakening of the bottom current by grants from Glasgow Caledonian University, the may enable elastic recoil of the ligaments to National Science Foundation and the North Atlantic Treaty Organization. partly re-erect the proximal stalk and rotate the crown to a position at which appropriate arm movements could shift the centre of References gravity and return the animal to the full “wilted flower posture”. Ausich W. I. and Baumiller T. K. (1993) Taphonomic method for determining muscular articulations in fossil crinoids. Palaios 8, 477484. Conclusions Baumiller T. K. (1992) Importance of hydrodynamic lift to crinoid autecology, or, could crinoids func- This investigation has provided direct tion as kites? J. Paleont. 66, 658-665. experimental evidence that the ligaments of Baumiller T. K. and LaBarbera M. (1989) Metabolic the isocrinid stalk are mutable collagenous rates of Caribbean crinoids (Echinodermata), with special reference to deep-water stalked and structures. On the one hand, they fulfil stalkless taxa. Comp. Biochem. Physiol. 93A, the mechanical functions of normal liga- 391-394. ments. Those at the symplexal articulations Baumiller T. K. and LaBarbera M. (1993) Mechan- permit limited movement and can sustain ical properties of the stalk and cirri of the sea lily prolonged loading without viscous creep, Cenocrinus asterius. Comp. Biochem. Physiol. 106A, 91-95. thus preventing disarticulation of the stalk; Baumiller T. K., LaBarbera M. and Woodley J. D. in addition, they have a capacity for strain (1991) Ecology and functional morphology of energy storage that may assist passive re- the isocrinid Cenocrinus asterius (Linnaeus) straightening of the stalk. The ligaments (Echinodermata: Crinoidea): in situ and laboratory at the synostosal articulations normally experiments and observations. Bull. Mar. Sci. 48, 73 l-748. act like sutural connective tissue, tightly Birenheide R., Amemiya S. and Motokawa T. (1993) binding adjacent ossicles and preventing Morphology and mechanics of the arm ligament in any movement. On the other hand, as well a stalked crinoid Metacrinus rotundus. Am. Zool. as these conventional properties, the stalk 33, 1ISA. ligaments show variable tensility: whilst still Candia Carnevali M. D. and Wilkie I. C. (1992) Gli straordinari tessuti connettivi degli echinodermi. preventing joint dislocation and retaining Le Scienze 286, 58-70. their elasticity, those at the symplexies can Cobb J. L. S. (1985) Motor innervation of the oral slacken reversibly to permit flexion, and the plate ligament in the brittlestar Ophiura aphiura synostosal ligaments can undergo such an (L.). Cell Tissue Res. 242, 686688. extreme loss of tensile strength that the Donovan S. K. (1984) Stem morphology of the recent crinoid Chladocrinus (Neocrinus) decorus. Palaeon- stalk can be fractured easily. tology 27, 825-841. The isocrinid stalk shows a remarkable Donovan S. K. (1989) The improbability of a muscu- diversity of mechanical responses. It is pre- lar crinoid column. Lethaia 22, 3077315. Sea-lily stalk ligaments 641

Donovan S. K. (1990) Functional morphology of by Keegan B. F. and O’Connor, B. D. S.), synostosal articulations in the crinoid column. pp. 185-193. Balkema, Rotterdam. Lethaia 23, 291-296. Messing C. G., RoseSmyth M. C., Mailer S. R. and Erlinger R., Welsch U. and Scott J. E. (1993) Miller J. E. (1988) Relocation movement in a Ultrastructural and biochemical observations on stalked crinoid (Echinodermata). Bull. Mar. Sci. proteoglycans and collagen in the mutable 42, 480-487. connective tissue of the feather star Antedon Paul C. R. C. and Smith A. B. (1984) The early b$da (Echinodermata, Crinoidea). J. Anat. 183, radiation and phylogeny of the echinoderms. Biol. l-11. Rev. 59, 443-48 1. Fujita T., Suguru 0. and Oji T. (1987) Photographic Roux M. (1987) Evolutionary ecology and bio- observations of the stalked crinoid Metacrinus geography of recent stalked crinoids as a model rotundus Carpenter in Suruga Bay, Central Japan. for the fossil record. Echinoderm Studies 2, l-53. J. oceanogr. Sot. Jap. 43, 333-343. Trotter J. A. and Koob T. J. (1989) Collagen and Grimmer J. C. and Holland N. D. (1990) The struc- proteoglycan in a sea-urchin ligament with mut- ture of a sessile, stalkless crinoid (Holopus rangii). able mechanical properties. Cell Tissue Res. 258, Acta zool., Stockh. 71, 61-67. 527-539. Grimmer J. C., Holland N. D. and Hayami I. (1985) Wilkie I. C. (1979) The juxtahgamental cells of Fine structure of the stalk of an isocrinid sea lily Ophiocomina nigra (Abildgaard) (Echinodermata: (Metacrinus rotundus) (Echinodermata, Crinoida). Ophiuroidea) and their possible role in mechano- Zoomorphology 105, 39-50. effector function of collagenous tissue. Cell Tissue Grimmer J. C., Holland N. D. and Messing C. G. Res. 197, 515-530. (1984) Fine structure of the stalk of the bourgueti- Wilkie I. C. (1983) Nervously mediated change in the crinid sea lily Democrinus conifer (Echinodermata: mechanical properties of the cirral ligaments of a Crinoidea). Mar. Biol. 81, 163-176. crinoid. Mar. Behav. Physiof. 11, l-34. Holland N. D., Grimmer J. C. and Wiegmann K. Wilkie I. C. (1992) Variable tensility of the oral arm (1991) The structure of the sea lily Calamocrinus plate ligaments of the brittlestar Ophiura ophiura diomedae, with special reference to the articula- (Echinodermata: Ophiuroidea). J. Zool., Lond. tions, skeletal microstructure, symbiotic bacteria, 228, 5-26. axial organs, and stalk tissues (Crinoida, Milleri- Wilkie I. C. and Emson R. H. (1988) Mutable crinida). Zoomorphology 110, 115-l 32. collagenous tissues and their significance for Macurda D. B. and Meyer D. L. (1974) Feeding echinoderm palaeontology and phylogeny. In posture of modern stalked crinoids. Nature 247, Echinoderm Phylogeny and Evolutionary Biology 394-396. (Edited by Paul C. R. C. and Smith A. B.), Messing C. G. (1985) Submersible observations of pp. 31 I-330. Clarendon Press, Oxford. deep-water crinoid assemblages in the tropical Wilkie I. C., Emson R. H. and Young C. M. (1993) western Atlantic Ocean. In Echinodermata (Edited Smart collagen in sea lilies. Nature 366, 5 19-520.