Diab ABSTRACT Spines of the sea urchin Strongylocentrotus fransiscanus were deflected through varying angles and the resistive forces offered by their ligaments were recorded. Data obtained for spines out of catch and spines in catch show basic differences in shape as well as amplitude. This difference in shape was also observed in the data obtained for dif¬ ferent spines in catch. The contractile force of the muscle layer at the base of the sea urchin spines was determined and found to be sig¬ nificant in comparison to the resistive forces upon 10° deflection offered by the spine ligaments when in a state of catch. A survey of a population of spines, both agitated and non-agitated, was conducted to obtain an objective definition of the state of catch versus non-catch. An angular limit to deflection was found to exist for spines in catch, beyond which their ligaments would become irreversibly damaged. Ex¬ perimental and mathematical calculation of Young's Modulus was done for a spine ligament in catch, and this was found to be similar to that given for vertebrate tendon at the same strain. The relationship be¬ tween resistive force offered and angular deflection was examined, and it was found to be non-linear. The possibility of nervous control of catch was considered, although results obtained make this seem unlikely. That the setting of catch is Ca"' dependent was here confirmed. Further¬ more, the reversibility of the effects of Ca depletion were confirmed, although the process of reversal remains unclear, for recovery time was found to be very long. The effects of protein inhibitors on catch were also studied, and the preliminary results provide strong evidence for an enzymatic mechanism in the control of catch. INTRODUCTION The spines of the sea urchin are connected to the animal's test by means of a ball and socket joint. Three concentric cylinders constitute this joint: an outer epithelial layer, a middle muscular layer, and an inner ligamental layer. The structure of this joint enables the sea urchin to move its spines freely in all directions by contraction of the muscular layers in the direction of deflection. This is apparently ac¬ companied by active shortening of the inner ligamental layer (Smith, Wainwright, Baker and Cayer, 1981). Agitation of an urchin, however, causes its spines to "freeze" and their movement to be inhibited. This is known as "catch." Although the mechanisms for setting and maintaining catch are not well understood, cross-linking of the collagen fibrils which make up the inner ligamental layer appears to be involved (Smith et al., 1981). Also involved seems to be the ionic composition of the inter-fibrillar matrix of the col¬ lagenous tissue (Wilkie, 1978). Unfortunately, good data on the rates at which catch can be set and released in vivo are entirely lacking. This catch should not be confused with the superficially similar phenomenon which occurs in the non-striated muscle of molluscs, as first described in Pecten maximus by von Uexkull (1912) and extensively studied in Mytilus edulis by Twarog (1954). This catch is due to neurally con¬ trolled muscle activity, whereas that of the sea urchin spine ligaments does not involve muscle at all (Kawaguti and Kamishima, 1965), although it would appear to be under neural control in the living urchin. In this paper, an attempt has been made to shed light on this non-muscular type of catch. In particular, an examination of the biomechanical properties of the spine ligaments has been conducted, and an exploration of possible control mechanisms, especially those involving nervous and enzymatic processes, has been undertaken. METHODS Strongylocentrotus fransiscanus and S. purpuratus were obtained from the subtidal and intertidal regions, respectively, of Monterey Bay. Specimens were kept in running sea water (15°C) and well oxygenated tanks. The experimental preparation consisted of an urchin test with the Aristotle's Lantern removed. The inside of the test was scraped out to eliminate any participation of the internal nervous system (e.g. radial nerves). The ligaments to be studied were stripped of their surrounding muscle and epithelial layers. The entire preparation was held steady by pushing a rubber stopper through that portion of the test previously occupied by the Aristotle's Lantern and firmly mounting this on a plexi¬ glass base placed in a glass dish filled with sea water. Test spines were connected to a Pixie element force transducer (Endevco Corp.) by means of cylindrical caps made of stainless steel hypodermic tubing fitted over their tips (Fig la). The connexion be¬ tween spine cap and force transducer was rigid stainless steel wire which could swivel at the spine cap to maintain horizontal transducer pull. The force transducer was attached to a micromanipulator, and lateral movement of the transducer produced spine deflections through desired angles. Angle was measured by a protractor attached to a second micromanipulator and brought up immediately behind the spine. Any resistance to deflection by a spine would bend the legs of the trans¬ ducer and thereby cause an increase in resistance in one of its semi¬ conductor elements and a decrease in resistance in the other (Fig. 1b). The ratio of the resistances was linearly transformed into a voltage by an operational amplifier circuit. The measurement of force was essential¬ ly isometric; that is, the transducer bending was a very small fraction of spine deflection (O.lu/g compared to approximately 5 m.m./10°). Data was recorded on a Brush 220 pen oscillograph. Stress versus strain was measured by a "tensometer" (Fig. 2). The preparation here used consisted of a ligament in catch with spine and single ossicle attached. The preparation was held between one clamp mounted on an anchored double cantilever beam and another clamp mounted on a plate moved by a lead screw. Force was sensed by deflection of the double beam and was monitored by a linearly variable differentiated transducer (LVDT) on the anchored plate. Deformation of the sample was monitored by an LVDT mounted on a moving plate. The voltage output of the LVDT was registered on a chart recorder. The maximum stress measur¬ able was 0.85 N and the maximum strain 408 (cf. Tansey, 1983). The most effective way to induce catch in the spine ligament was found to be physical agitation. The preparation to be studied would be placed in a glass dish filled with sea water. Through this sea water would be bubbled a constant stream of air for ten minutues, following which time experimentation would begin. All chemicals used were applied in drops by means of a Pasteur pipette directly on the stripped ligaments in 2 ml volumes per trial and given three minutes to take effect (except DTT, NEM, dansylcadaverine and putrescine, all of which were given ten minutes). The tests examin¬ ing the effects of acetylcholine and adrenaline were done four times, and all other tests twice. The chemicals used were stored in opague vessels to avoid any light-catalyzed oxidations, and were kept at a temperature of 4°C. Finall, Ca" -free sea water was made up in the fol¬ lowing proportions (M): 0.445 Nacl, 0.060 Mgcl., 0.010 KCl, 0.015 EGTA, and 0.010 Tris buffer (pH 7.8). RESULTS Typical data for spines out of catch and in catch Figure 3 (a) shows a record of resistive force offered for a 10° deflection (F (10) by a spine out of catch. Figure 3(b) shows the E (10) for the same spine in catch. Figure 3(c) shows a record of E (10) of a different spine in catch obtained from the same test. The difference in record amplitude between the spine out of catch and the two spines in catch is clear. Also significant is the shape difference of the two types of record: the spine out of catch offers an F (10) which takes the form of a step function, whereas the spines in catch offer F (10)'s which show an initial, sharp peak followed by a "sag" that eventually levels off to a steady state value. Finally worthy of note is the shape difference within the catch records them¬ selves, namely, the extent of sag, in this case being more pronounced in (c) than in (b). The basis of this sag is not completely understood, but it seems to be characteristic of the catch state, and may well be an indication of the quality of catch, more pronounced sag implying a weaker state. For all experiments resistive force was defined as the highest point of the curve recorded. Thus, for spines in catch, resisitive force was taken to be the amplitude of the initial, sharp peak. Determination of the contractile force of the muscle layer surrounding ligament. Prior to detailed examination of the physical properties of the spine ligament, it was necessary to obtain a measure of the contractile force (F.) of which its surrounding muscle was capable. A comparison of this force with that offered upon 10° deflection of a spine in catch would determine whether the preparation was to consist of a stripped ligament or whether the muscular layer around the ligament could be left intact. A test spine with muscle intact was connected to the force trans¬ ducer as shown in Fig. la. F was measured by using a probe to mechan¬ ically stimulate the portion of the test immediately before that region of the muscle layer facing away from the transducer. This caused the muscle to contract and attempt deflection of the spine, thereby pulling against the transducer and yielding discreet peaks on the chart re¬ corder (Fig. 4, top arrows). The test spine was subsequently stripped of its muscle and set into catch. Fy(10) was recorded (Fig. 4, bottom arrow) and this was compared with the F The results showed that 2/3 F, a force too great to be ignored. All experiments were therefore run on ligaments stripped of their surrounding muscle. Definition of catch and non-catch in a population of spines. F (10)'s were measured for 88 spines, some agitated and some not, to obtain an objective definition of catch versus non-catch. Spine length for all trials was relatively constant (approximately 2.5 cm) and so no normalization of results by lever arm was considered necessary. Nor was ligament cross-sectional area taken into account for the same reasons. The forces recorded and the number of spines offering each force were plotted in the form of a histogram (Fig. 5). It was observed that most spines fell into two categories, one having a range for F (10) of 0.00—0.10g, the other of 0.34—-064g. The former category was deemed characteristic of spines out of catch, the latter of spines in catch. Any F (10)'s recorded which were greater than 0.64g were also considered characteristic of spines in catch. Whether these larger values might have resulted from slight errors in angular deflection measurements (which would have resulted in large differences in resistive forces recorded-see next two sections), differences in spine length, or some other variable is unclear. Critical deflection angle for irreversible damage. Having defined the state of catch, it was now possible to look more closely at its mechanical properties. First to come to light was that there seemed to be a limit to the extent a ligament in catch could be stretched before its constitution was unalterably changed. It was observed that repeated spine deflections of 10° or less yielded re¬ sistive forces of similar magnitude (Fig. 6a), whereas repeated deflections beyond this (in this case 20° and 30°--Fig. 6b) yielded resistive forces of decreasing magnitude. These results suggest that a ligament in catch begins to be damaged at spine deflections of between 10° and 20°. Computation of Young's Modulus (E) for a ligament in catch. Young's Modulus is a measure of the tensile stiffness of a material. Measurement of stress (force/cross-sectional area) at varying strain (change in ligament length/original ligament length) was done by using the "tensometer" of Figure 2 (see Methods), and the data obtained were plotted in the form of a stree-strain curve (Fig. 7). From the slope of this curve at 0.58 strain, Young's Modulus (stress/strain) was calcu¬ lated to be 2.8 x 10° N/m", a value similar to that given for vertebrate tendon at the same strain, namely, 3.5 x 10° N/m (Wainwright, Biggs, Currey and Gosline, 1976). Young's Modulus was also calculated theoretically by a mathematical derivation of E based on a schematic diagram of the spine-ligament system (Fig. 8). The value obrained was 1.1 x 10° N/m", which, consider¬ ing the potential sources of error in ligament dimension measurements, is certainly consistent with that obtained experimentally (see Appendix A). Relationship between resistive force and angular deflection for a spine in catch. A spine in catch was deflected incrementally to different angles and the resistive force offered at each angle was recorded (Fig. 9). It was found that the relationship between these two variable was not linear. Below deflections of approximately 8°, the curve resembled that obtained for stress versus strain (Fig. 7). Above deflections of 8°, however, the curve began to rise sharply, an increase in resistive force coincid¬ ing with the region of angular deflection in which the critical angle for damage has been found to lie. The increase in resistive force is probably due to the application of deflective force directly on to the inter-fibrillar cross-links which cause catch in the collagenous ligament. Consequently, excessive force applied here would tend to break these cross-links and thereby abolish catch. Hence the occurence of the critical angle for damage in this region. Consideration of possible nervous control of catch. Takahashi (1967) reported that the tonus of an isolated strip of ligament connected to spine and single ossicle was increased by the addition of 10 Macetylcholine (ACh) and decreased by the addition of 10 "M adrenalin (Adr). Tonus was assessed by the rate of isotonic lengthening of ligament preparations under a constant load. The rate decreased upon the addition of ACh and increased upon the addition of Adr, although the extent of length change is unclear, for his figure has no vertical scale bar (cf. Boolootian, 1966). An attempt to verify these results was made here by measuring the F (10)'s offered by ligaments before and after chemical treatment: sim¬ ilar pre- and post-treatment Fy(10)'s constituted negative results and indicated that the chemicals had been ineffective, whereas dissimilar Fy(10)'s (a post-treatment increase showing the setting of catch, a post-treatment decrease showing the releasing of catch) constituted positive results and indicated that the chemicals had been effective. For all chemical analysis, the following control was used. Methylene blue was applied by Pasteur pipette to a stripped ligament. The ques¬ tion to be addressed was whether chemical solutions applied to a lig¬ ament could penetrate it and thus exhibit their maximum effect, or whether the density of the ligament prevented their diffusion and restricted them to the ligament's outer surface, thereby dampening their potential effects. After one minute, the ligament was cut open to re¬ veal its inner surface. This was found to be blue in colour, showing that the ligament was indeed permeable to liguids. ACh was used in 10 'M concentrations and Adr in 10 M concentra¬ tions. ACh was applied to both ligaments out of catch, to see if catch could be induced, and ligaments in catch, to see if some extreme state of catch, otherwise unattainable by mere physical agitation, could be induced. Adr was applied to ligaments in catch to see if it could abolish this state. The results obtained in all cases were negative. As another test of cholinergic control of catch, curare was applied in concentrations of 10 "M to a ligament in catch to see if this catch could be abolished. If catch were ACh dependent, then curare, by blocking any ACh recepors present within the ligament, might abolish catch. The results obtained were again negative, lending further testimony to the catch state's independence of ACh. Other catecholamines were also tested, namely, serotonin and dopamine. These are known to release catch in the non-straiated muscle of molluscs (Florey, 1966). Concentrations used were 10 M and 10 application was to ligaments in catch to see if effects similar to those in molluscs could occur in the sea urchin, that is, the release of catch. Once more, the results obtained were negative (all chemical analysis results are summarized in Table I). The effects of Ca on catch. Wilkie (1978) proposed that the ionic composition of the matrix sur¬ rounding the collagen fibrils in echinoderm connective tissue is re¬ sponsible for the maintenance of catch. A variation in inter-fibrillar ionic composition might, therefore, influence catch in the spine liga¬ ments of sea urchins. Smith et al. (1981) have demonstrated that the absence of divalent cation prevents the setting of catch in these liga¬ ments. This has been confirmed in this study for the divalent cation Ca". A ligament in a state of non-catch was bathed in Ca -free sea water (see Methods) for five minutes and then bubbled in the same medium for a further ten minutes to see if the physical agitation which normal¬ ly sets catch in regular sea water could do so here, in the absence of Ca". Positive results were obtained, indicating that the setting of catch is Ca dependent. Smith et al. also state that the prevention of catch in their ex¬ -free and Ca" -Mg -free sea waters was reversible, but periments by Ca they give no indication of the time course for these reversals. Follow¬ ing the above described experiment, the preparation was bathed in regular sea water to determine the extent of time required for the recovery of catch. It was found that bubbling the preparation in this new medium after 30 minutes could not induce catch-indeed, the F (10) recorded was equal to that recorded when the preparation was bubbled in Ca"'-free sea water (0.05g). A similar result was obtained upon bubbling after one hour in the regular sea water. Bubbling after three hours yielded an F.(10) of twice the amplitude of that yielded after one hour, yet it was still in the force range previously deemed to be char¬ acteristic of spines out of catch (0.lg). Finally, the preparation was left in the regular sea water overnight (nine hours) after which time an F.(10) characteristic of spines in catch (cf. Fig. 5) was yielded: O.38g. Thus, the spine had recovered. Ca'-free sea water was then applied to a ligament in catch to see if it could abolish catch. The results obtained were negative, showing that the maintenance of the catch state was not Ca dependent. Are disulphide bonds involved in catch? Ascaris collagen contains 27/1000 residues of 4 cystine and is known to form cross-links by means of disulphone bonds between these residues (Wainwright et al., 1976). The possibility of the occurence of such disulphide cross-linking between the collagen fibrils in the spine ligaments of sea urchins was here considered. The chemicals used to test for this were ditiothreitol (DTT) and N—-ethyl maleimide (NEM). These were used in 5 x 10 'M concentrations. DTT breaks disulphide linkages, and was therefore applied to ligaments in catch to see if it could abolish this catch. NEM prevents the formation of disulphide linkages, and was therefore applied to ligaments out of catch; sub¬ sequent bubbling and recording of F (10) would determine whether it had prevented the setting of catch. F (10)'s offered before and after 12 treatment with DTT were similar, showing that this chemical had had no effect on the maintenance of catch. F (10)'s offered upon bubbling after NEM had been added were considerably greater than those offered before bubbling (having an amplitude, 0.43g, characteristic of spines in catch), showing that NEM had not prevented the setting of catch. Both these sets of results suggest that disulphide cross-linking is not involved in the setting or maintenance of catch in the spine ligaments of sea urchins. The effects of protein inhibitors on catch. Dansylcadaverine and putrescine act as substrates for the enzyme transglutaminase, and thereby inhibit its biological functions when added to a system in excess. Transglutaminase will catalyze, for ex¬ ample, the incorporation of dansylcadaverine into a convenient protein receptor such as casein, and measurement of the protein bound fluores¬ cence of this reaction has been used to ascertain the potency of this enzyme (Lorand, 1976). In vertebrates, transglutaminase is known to catalyze the cross-linking of the fibrin clot during haemostasis, and the production of the vaginal plug by postejaculatory clotting of rodent seminal plasma (Folk, 1980). It is also known to catalyze the in vitro formation of cross-links between fibronectin and collagen in cultured fibroblasts of rat kidney (Folk, 1980). In light of these observations, it seemed possible that trans¬ glutaminase or a transglutaminase-like enzyme (TGL) might be involved in the cross-linking of collagen fibrils which constitutes the state of catch. Dansylcadaverine and putrescine were applied to both lig¬ aments out of catch. The application to ligaments already in catch was a control for any reactions other than the suspected enzymatic one, The control application gave negative results, showing that the in¬ hibitors did not abolish catch. On the other hand, the application to ligaments out of catch produced positive results, for post-treatment bubbling could not induce catch in these ligaments. This showed that the inhibitors could prevent the setting of catch, and in so doing presented evidence for a potential enzymatic mechanism for the control of the catch state, one which would involve possibly transglutaminase or an enzyme similarly sensitive to dansylcadaverine and putrescine inhibition. DISCUSSION This paper presents measurements of the mechanical properties of the spine ligaments in sea urchins and considers some possible physiolog¬ ical control mechanisms. It confirms and extends, on a different urchin genus, the basic findings of Smith et al. It also offers suggestive preliminary evidence for a Ca dependent enzymatic mechanism of collagen cross-linking as the basis for setting catch. One new and important finding reported here is the occurence of some sort of structural damaging or weakening when a spine in catch is deflected through an angle of between 10° and 20°. A similar alteration in ligament constitution at approximately the same amount of stretch appears to have occured in the experiments of Smith et al. Although they do not consider this idea per se, they do present a plot of ex¬ tension versus time for ligaments stretched vertically along the long axes of their spines (see their Fig. 3). Extensibility at a stretch 14 (measured under constant load) of 0.3--0.4 mm in the spines they term "movable" decreased dramatically relative to extensions on either side of this range. This means that in this region, the ligaments were suddenly offering an increased resistance to stretch. When extension was carried out beyond 0.4 mm, an increased rate of extension per time was observed, as though the ligaments had become damaged or structurally weakened. Smith et al. do not discuss this result, but it may reflect functional passage through the 10°--20° critical angle range described above. By using the schematic diagram of Fig 8, the range of angular deflection which corresponds to their "critical" 0.3--0.4 mm extension range can be calculated. If a ligament radius (x) of 1.0 mm is taken for the ligaments they were using, as estimated from their micrographs (see their Fig. 4), 0.3--0.4 mm ligament extensions (1) are seen to be equivalent to angular deflections of 16.9°--21.8°. This was precisely the range in which the critical angle for damage was found to lie in this study. Exactly what happens at this angle requires a model for the liga¬ ment system and the mechanism of catch maintenance by the collagen fibrils. According to the ideas presented by Smith et al., this critical angle may represent the limiting amount of stretch which the cross-links between the collagen fibrils making up the ligament can withstand before suffering breakage. Since the breakage appears to be largely irreversible (cf. Fig. 6), it can be concluded that the cross¬ links are not capable of reformation once broken by excessive stretch. Another index of the stress-strain relationship of the spine liga¬ ment is given by the relationship between resistive force offered (F and angular deflection (ad) presented in Fig. 9. At first sight, these graphs appear not to be similar, although their variables would suggest that they ought to be: F is directly proportional to stress (through cross-sectional area) and ad is an indirect, but linear (for small angles), way of measuring strain. From Fig. 8, conversion of strain to equivalent ad is possible. This shows that the graphs are indeed simi¬ lar, and that they differ only in extent. The maximum strain to which the spine ligament was subjected was 2.58 (Fig. 7), and this corresponds to an ad of 5.4° (using a ligament length (L) of 4 mm and a ligament radius (x) of 1 mm). Thus, the range of the stress-strain curve is equivalent to that part of the F-ad curve over angles of 0°--5.4°. Moreover, the slope of the stress-strain curve would have probably risen steeply had more strain (that is, greater than 2.58) been applied to the caught ligament, in harmony with the F-ad curve. Experiments were also conducted to test possible nervous control of the catch state. The findings of these experiments would suggest that nervous control is either not involved, or is only a part of a greater mechanism. On the other hand, experiments conducted to test possible enzymatic control offer strong (although it must be emphasized still preliminary) evidence for the role of transglutaminase or TGL in the mechanism of catch. Transglutaminase has been observed to play a part in several clotting-type processes in a wide variety of organisms, and in particu¬ lar in the cross-linking of collagen fibres in both vertebrates and in¬ vertebrates (Folk, 1980). Therefore, it is reasonable to suspect that it might also play a part in the cross-linking of collagen fibrils in the spine ligaments in sea urchins. Now it has been demonstrated that the setting of catch is Ca dependent. Ca could participate in at least two important control processes: the conduction of nervous activity, especially neurosecretion, and the activation of an enzymatic reaction chain. Concerning the latter, it is known that transglutaminase requires Ca' to convert it from its zymogen form to its active state (Lorand, 1976). Thus, if catch were controlled by transglutaminase or TGL, then the absence of Ca"' should block its setting by preventing the activation of this enzyme. An important question which must now be addressed concerns the whereabouts of the transglutaminase or TGL store. Smith et al. state that in the ligament, "axons contain large numbers of opaque secretory droplets." Could the contents of these vesicles be the enzyme, and therefore these vesicles the enzyme store? If so, then the following mechanism could be conceived: Ca" ions trigger the neurosecretory release of transglutaminase or TGL and their subsequent activation, enabling them to catalyze the cross-linking of the collagen fibrils and thereby set catch. This would explain the requirement of Ca" for the setting of catch. However, this model would not be able to explain the observed long recovery time (many hours) for a ligament when its Ca supply is depleted and then replenished. It would seem that once the Ca" supply had been replenished, the ligament should immediately regain its ability to catch. Bearing this problem in mind, a more satisfactory mechanism is the following. Ca" again permits neurosecretory activity of the nerve fibres which run through the ligament, but the released substance is not enzyme, rather, it is an essential cofactor. This cofactor activates the enzyme which is already present in the medium surrounding the collagenfibrils. Ca must play one further role, however: it must somehow prevent the loss of the enzyme by diffusion in a stripped liga¬ ment surrounded by regular sea water. One possibility is that Ca mediates or stabilizes the binding of the enzyme to some extracellular structure, such as the collagen fibrils themselves or the external sur¬ face of the nerve endings. The latter of these two seems the more likely, for it would give the ligament system more control of the catch mechanism by bringing the point of release and the point of action of the cofactor closer. It would also for the same reason tend to conserve cofactor and avoid any spontaneous cross-linking of the collagen fibrils resulting in spontaneous catch. Taking the surface of the nerve endings then as a likely place for the location of the transglutaminase of TGL, released factor might thus "activate" the Ca -enzyme complex by causing its release from its anchoring structures, allowing it then to cross¬ link the collagen fibrils and set catch. This mechanism clearly incorporates the observation that the setting of catch id Ca dependent. In addition, it also explains the long recovery time after treatment with Ca" -free sea water. Removal of Ca" from the medium surrounding a ligament would not only prevent the neurosecretion of cofactor, it would also result in the release of inactive enzyme into the inter-fibrillar matrix. Within a short time, this inactive enzyme would diffuse out of the ligament and be lost to the surrounding medium. Resupply of Ca would immediately reestablish the ability of the neurosecretory vesicles to release cofactor, but would not immediately reestablish the ability of the ligament to catch; rather, a considerably longer time would be required to regenerate the lost enzyme. ACKNOWLEDGMENTS First and foremost, endless gratitude is due Professor William Frank Gilly. His seemingly indefatigable drive and attention, which would pick up my spirits when all else was crumbling, his sharp-edged wit, which at the best of times was only just bearable, and his acute but truly constructive criticism, which would restrain the energy of an oftentimes overzealous undergraduate and firmly warn him against im¬ pulsiveness and the dangers of getting "too big for one's boots," all contributed tremendously to my development as a thoughtful, cautious and honest scientist. Thanks must also go to Mark Denny, without whose insight and warm encouragement I may never have ventured into the wild world of biomechanics, and to David Epel, whose marksmanship may have hit the bull even "in the dark." Let me also express appreciation to Stuart Thompson for his valuable feedback on directions for further research and to all those who had the courage to inquire, the patience to listen, and the interest to criticize. Finally, let me thank the sea urchin itself for providing me with such an exciting but at the same time accessible system to study, a system which I find more baffling now than when I started and which I hope to come back to soon. APPENDIX A From a schematic diagram of the spine-ligament system (Fig. 8), a mathematical derivation of Young's Hodulus (E) is possible: 1/y - X cos 1 2 y x cos V Strain = 1/L cos where V - 15 yost/os where stress - F/A E a stress/strain where A =x dx d 2 F L cos/1 A F = E (x dx df) y x cosd/L cos y Moment (M)= Fx cos -E (x dx do) yx cos 9/Y L cos Hotal - 4 e t eosh 2 00 40 E/V L cost) 26/4 (1/2 + 1/4 sin 209 Motal (1/2 + 1/4 sin 29) F a (y E/V2 L CoS 4 C A simple rearrangement then gives the final equation for f. E - F 21 cos 4/ 2/ (1/2 t 1/4 sin 20 When one considers that working with physically measured ligamental radius means working with numbers taken to the fourth power, one can appreciate that the mathematically calculated Value for Young's Modulus is quite consistent with that obtained experimentally. REFERENCE BOOLOOTIAN, R.A., editor, 1966. Physiology of echinodermata, Inter¬ science Publishers, pp. 521-524. FLOREY, E., 1966. An introduction to general and comparative animal physiology, W.B. Saunders Company, pp. 590-594. FOLK, J.E., 1980. Transglutaminases. Ann. Rev. Biochem. 49: 517-531. KAWAGUTI, S. and KAMISHIMA, Y., 1965. Electron microscopy on the spine muscle of the echinoid. Biol. J. Okayama Univ. 11: 31-40. LORAND, L., editor, 1976. Fibrin stabilizing factor (factor XIII). Meth. Enz. 45: 177-191. SMITH, D.S., WAINWRIGHT, S.A., BAKER, J., and CAYER, M.L., 1981. Structural features associated with movement and 'catch' of sea urchin spines. Tissue & Cell 13: 299-320. TAKAHASHI, K., 1967. The catch apparatus of the sea urchin spine. J. Frac. Sci. Univ. Tokyo 11: 109-130. TANSEY, K., 1983. Comparative analysis of starfish skeletal properties in Pycnopodia helianthoides, Pisaster ochraceus, and Pateria miniata. Unpublished. TWAROG, B.M., 1954. Responses of a molluscan smooth muscle to acetylcholine and 5-hydroxytryptamine. J. Cell Comp. Physiol. 44: 141-163. von UEXKULL, J., 1912. Studien über den tonus. 6: Die Pilgermuschel. Z. Biol. Munchen 58: 305-302. WAINWRIGHT, S.A., BIGGS, W.D., CURREY, J.D., and GOSLINE, J.M., 1976. Mechanical design in organisms, Edward Arnold, London. WILKIE, I.C., 1978. Nervously medicated change in the mechanical properties of a brittlestar ligament. Mar. Behav. Physiol. 5: 289-306. FIGURE LEGENDS Fig, la: Diagram of the apparatus used to measure resistive forces offered by spines in catch and spines out of catch when deflected to desired angles. The functional core of the force transducer shown above. Fig. 1b: The "tensometer" used to measure stress at varying strains Fig. 2: exerted longitudinally on a ligament in catch. Typical chart recordings of F (10) offered by a spine out of Fig. 3: catch (a), the same spine in catch (b), and a different spine obtained from the same test in catch (c). F of the muscle layer surrounding a spine ligament measured Fig. 4: m by mechanical stimulation (top arrows) compared to F (10) offered by the same ligament after it was stripped of its muscle and set into catch (bottom arrow). Histogram of the F (10)'s offered by a population of spines, Fig. 5: some agitated, some not, giving an objective definition of catch versus non-catch. F (10) recorded repeatedly from a spine in catch showing a Fig. 6a: relatively constant amplitude per deflection. F (20) (bottom arrows) and F (30) (top arrows) recorded Fig. 65: répeatedly from the same spine in catch showing a gradual decrease in amplitude per deflection. Stress-strain curve for a spine ligament in catch, as measured Fig. 7: by the tensometer in Fig. 2. Schematic diagram of spine-ligament system (a) and cross¬ Fig. 8: sectional area of a spine ligament (b) used in the calculation of strain equivalence of angular deflection and mathematical computation of Young's Modulus. The relationship between resistive force offered and angular Fig. 9: deflections over a range of 0°--11.5°. The results of chemical analysis carried out on spine ligaments Table I in catch and out of catch. 24 SPINE - LIGAMENT TEST TRANSDUCER (a) SILICON V BRASS (b) AMPLIFIER CHART RECORDER. EPOXY — STARFISH SKELETON STRETCHING MACHINE LVDT ANCHORED LEVER PLATE MOVING PLATE LVDT Fig. 2 DOUBLE CANTILEVER O CLAMP ......... - . . . . J. § O CLAMPS 0.50 0.2. 0.50 0.25 (b) 8 0.50 025 — 100 ata — 50 100 150 TME (sec) 300 200 TIE (sec) —— 200 100 TIME (sec) — 400 Fig. 3 — — S — 10 (5) aouo — — — — SANIdS 40 AZGWON — (6) oo 9 — — — (5) oo 8 0 300 50 Fig. 7 0.5 1.5 1.0 % STRAIN 2.0 2.5 0 + O.3 S O. L O. 8 ANGULAR DEFLECTION (degrees) — 10 SPINE — LIGAMENT (a) Fig. 8 0 (b)