Mechanical Diversity of Connective Tissue of the Body Wall of Sea Anemones

The passive mechanical response of an organism to an imposed load depends upon both the shape of the organism and the mechanical properties of the materials composing its body (cf. Koehl, 1977 b). The mechanical behaviour of a material in turn depends upon the molecular and microscopic structure of that material. I would therefore expect to find structural features on the molecular and microscopic levels of organization of an organism which equip that organism for the particular mechanical activities it performs and the environmental forces it encounters.

The connective tissues of the body walls of hydrostatically supported organisms are an obvious place to look for examples of molecular and microscopic features which suit tissues to the particular mechanical life-styles of organisms. There are species of hydrostatically supported sea anemones representing different extremes in mechanical behaviour. Metridium senile fimbriatum (Verrill) are calm-water anemones (Koehl 1976,1977a) that are capable of changing their body dimensions considerably (Batham & Pantin, 1950), and Anthopleura xanthogrammica (Brandt) are rough-water anemones that do not undergo extreme shape and size changes (Koehl, 1976,1977 a). I chose to investigate two questions concerning the body-wall material of these species of anemones: (1) in what ways do the mechanical properties of body-wall materials relate to the mechanical behaviour of these anemones? and (2) what macromolecular and microscopic features of the body-wall materials of these anemones are responsible for differences in their mechanical behaviour?

The body wall of sea anemones is a thick layer of connective tissue, mesogloea, sandwiched between single layers of ectodermal cells and endodermal epithelio-muscle cells (Stephenson, 1935). Chapman (1953 b) observed that the passive mechanical properties of the body wall of M. senile and Calliactis parasitica were similar to that of mesogloea from which the muscle fibres had been removed.

Mesogloea is a composite material consisting of fibres contained in an amorphous matrix. Mesogloeal fibres have been judged to be collagen on the basis of ultrastructure as well as various chemical and physical properties for several species of sea anemones: M. senile (Chapman, 1953a; Grimstone et al. 1958; Gross, Dunsha & Glazer, 1958; Piez & Gross, 1959; Pikkarainen et al. 1968; Gosline, 1971a), M. dianthus (Katzman & Kang, 1972), M. canum (Batham, 1960), Metridium sp. (Kulonen & Pikkarainen, 1970), C. parasitica (Chapman, 1953a), Aiptasia sp. (Gosline & Lenhoff, 1968), and Actinia equina (Nordwig & Hayduk, 1969; Nordwig, Hieber-Rogall & Hayduk, 1970; Nordwig, Nowack & Hieber-Rogall, 1973). Elastin-like fibres have not been found in the mesogloeas of Anthozoans, including the anemone A. equina (Elder, 1973). The matrix surrounding the collagen fibres has been found to contain highly hydrated protein and polysaccharide polymers in two species of anemones: M. dianthus (Katzman & Jeanloz, 1970 a, b; Katzman & Oronsky, 1971; Katzman, 1972; Katzman et al. 1972; Katzman & Kang, 1972), and M. senile (Gross et al. 1958; Gosline, 1971a, b). Some of the polysaccharides appear to be covalently bound to the collagen in M. dianthus (Katzman & Jeanloz, 1970a; Katzman & Oronsky, 1971). The polysaccharide chains in most collagenous connective tissues are polyelectrolytes having a high density of negative charges; because charged segments of such polysaccharide chains repel each other, the chains assume extended conformations (Schubert & Hammerman, 1968). Sea anemone mesogloea is unique in that it contains neutral rather than charged polysaccharides (Gross et al. 1958; Katzman & Jeanloz, 1970b; Gosline, 1971a). Gosline (1971b) suggests that the neutral polysaccharide chains in mesogloea take on a more nearly random-coil conformation than can charged chains, and that electrostatic binding between matrix molecules and between matrix and collagen molecules is limited by the dearth of charged polysaccharides in mesogloea.

Various measurements of the mechanical properties of sea anemone mesogloea indicate that they are time-dependent (Chapman, 1953b; Alexander, 1962; Gosline, 1971b; Koehl, 1977b) (i.e. mesogloea’s response to a load depends not only on the magnitude of the force, but also on the rate, duration, and past history of force application to the mesogloea). Alexander (1962) found that M. senile mesogloea could be stretched uniaxially to three times its original length by a small stress (force per cross-sectional area) applied for periods of 10-20 h, and that mesogloea would nearly return to its original unstretched length over the same time period after the stress was removed. Such time-dependent large-deformation elastic behaviour is typical of networks of random-coil polymer molecules (Treolar, 1970 ; Ferry, 1970 ; Aklonis, MacKnight & Shen, 1972; Wainwright et al. 1976). Collagen fibres are crystalline in structure and are relatively inextensible (they can only be stretched elastically by 4% of their resting length and break at extensions of 8%) (Wainwright et al. 1976). Gosline (1971b) therefore suggested that the extensibility and elasticity of M. senile mesogloea is due to the random-coil polymers in the matrix rather than to the collagen fibres.

On the basis of mechanical, chemical, and structural evidence, Gosline (1971b) proposed that the mesogloea of M. senile is a pliant composite material with a discontinuous fibre system of collagen in a highly hydrated protein-polysaccharide matrix. The matrix of mesogloea is a dilute network of random-coil polymers lightly cross-linked to each other and to the collagen fibres. This viscoelastic matrix is responsible for the ability of M. senile mesogloea to be extended slowly by as much as 200% and to recover slowly by configurational entropy elasticity (explained in Treolar, 1970). The collagen fibres, which are not cross-linked to each other, act to reinforce the matrix. As the mesogloea is stretched, the collagen fibres tend to align with the stress axis and slide past each other and in so doing, they deform the matrix molecules between them. The closer the collagen fibres are to each other and the closer they are to being parallel with the stress axis, the greater is the rate of deformation of matrix molecules between them for a given extension rate of the mesogloea. The greater the rate of deformation of the viscoelastic matrix, the more rigid it will be.

Collagenous connective tissues vary greatly in their mechanical properties (Viidik, 1972) in ways which can often be correlated with their mode of functioning in an animal in its environment. I applied theories and techniques of polymer mechanics (discussed in Ferry (1970), Aklonis et al. (1972) and Wainwright et al. (1976) (1) to describe quantitatively the mechanical properties of the mesogloea of two species of sea anemones in relation to their diverse activities and habitats, and (2) to indicate ways in which microscopic and molecular structures of these collagenous connective tissues might vary that would result in different mechanical properties.

Strips of body wall for mechanical testing were cut from M. senile and A. xanthogrammica which had been collected and maintained in aquaria as described by Koehl (1077b). To measure the passive mechanical properties of anemone body wall without the complication of contraction by the circular and longitudinal muscles of the anemone column, I trimmed away the mesenteries from a strip and then narcotized the remaining muscles by soaking the strip in a solution of 20% MgSO_4.7H₂O one to one with Instant Ocean (33 ‰) (Pantin, 1964) or in a 0·3% solution of propylene phenoxitol in Instant Ocean. A strip was judged to be anaesthetized if it did not visibly contract when pinched with forceps and trimmed with a razor blade. Gentamicin (50 µg.ml^-1) was added to the MgSO₄ anaesthetic and to the Instant Ocean for the propylene phenoxitol anaesthetic and these solutions were then autoclaved in order to minimize bacterial decay of the body-wall strips. The strips were gripped in the various testing devices as described by Koehl (1977b). The length between grips, the width, and the thickness of the body wall strips were measured to the nearest o-1 mm before each test using vernier callipers.

Dynamic stress-strain tests were made using a non-resonance, forced-vibration instrument (Fig. 2) constructed by J. Gosline. A specimen was sinusoidally pulled by a vibrator (Ling Dynamic Systems, Type 203) at frequencies from 01 to 30 Hz. Extension of the specimen (± 0·05 µm) was measured by a LVDT (Model 100 MHR, Schaevitz Engineering), and force (S.E. = ±2%) was measured by a BLH silicon strain gauge mounted on beam A. Signals from the LVDT and the strain gauge were amplified by carrier amplifiers (SE Laboratories, Type 4300). The amplitude and phase of the force and extension signals were then monitored by a transfer function analyser (SE Laboratories, Model SM 272DP). Specimens were immersed in a controlled-temperature bath (LAUDA K-2/RD, Brinkman Instruments) of MgSO₄ anaesthetic. The frequency-dependent storage modulus (E′) is the elastic modulus (an index of rigidity) of a material in phase with the stress,

Cryostat sections (Cryo-Cut Cryostat Microtome, Model 830–C) were made of fresh anaesthetized pieces of body wall from A. xanthogrammica and M. senile. These sections were immediately mounted on microscope slides in sea water and observed using a compound polarizing microscope (Wild M-21) with a first-order red compensator to indicate positive and negative birefringence.

Other anaesthetized pieces of body wall were fixed in Baker’s formaldehyde-calcium fixative for marine invertebrates (Pantin, 1964), embedded in paraffin according to the procedure outlined by Grimstone & Skaer (1972), sectioned using a 820 Spencer microtome, and stained using the various techniques listed in Table 1.

Pieces of anaesthetized anemone body wall were fixed in 2% glutaraldehyde in Instant Ocean and prepared for SEM according to the procedure outlined by Mariscal (1974) for sea anemones. Some specimens were dried using a critical point drying system (Sorvall 49300) and others using a manifold freeze-dryer (Virtis Model 10-117-A). The lengths, widths, and thicknesses of the specimens were measured with vernier callipers to the nearest 0·1 mm after each step in preparation to determine that tissue shrinkage was not significantly anisotropic. Dried specimens were cut along various axes using disposable histo-knives, were shadowed with gold-palladium, and observed using a JEOL JSM-SM Scanning Electron Microsope. The orientations of fibres (selected using a random number table) were measured to the nearest1° using a protractor on scanning electron micrographs of tangential sections of body wall from A. xanthogrammica and M. senile.

Wet and salt-free dry weights, and dry weights of autoclave-soluble material were made according to the procedures outlined by Gosline (1971a) for pieces of mesogloea from which the epithelium and muscle had been removed.

The hydrostatic skeleton of a sea anemone is basically a tension-resisting container (the body wall) filled with compression-resisting fluid under pressure. Sea anemone mesogloea is subjected mainly to tensile stresses of three origins: (1) muscular contractions, which increase the pressure of the water in the coelenteron and, hence, subject the mesogloea to tensile stresses, (2) flow forces, which tend to bend the anemone and produce tensile stresses in the mesogloea (Koehl, 1977 b) and (3) ciliary pumping by the siphonoglyphs, which inflates the animal with water and thus subjects the-mesogloea to tensile stresses. Tensile creep tests can be used to simulate these stresses in mesogloea so that the response of the mesogloea can be measured.

A typical creep curve for M. senile mesogloea is shown in Fig. 3. Note that M. senile mesogloea extends by a small amount (about 2–5 % of resting length) when stressed for short times (less than a min). If a stress is applied to M. senile mesogloea for an hour, it extends by about 50%. When M. senile mesogloea has extended by 150% or more after 12-24 h of being stressed, it reaches an equilibrium length beyond which it does not extend. (Three of the eight specimens of M. senile mesogloea did not reach an equlibrium length after 24 h (like the ‘Type II’ results of Alexander (1962)), but had local areas where the tissue had yielded and undergone plastic flow.

M. senile mesogloea is in essence two different materials which depend on the time scale of stress application as follows. (1) When stressed for short times, as it would be during muscular contractions involved with rapid shape changes, M. senile mesogloea is fairly rigid. This rigidity allows the animal to bend, straighten, sway, etc., without just ballooning out its body wall in areas where muscles are not contracting. Alexander (1962) and Gosline (1971 b) have suggested that the rigidity of M. senile mesogloea during short periods of stress application allows the animal to resist waves. I tend to disagree with this interpretation because M. senile do not usually encounter waves (Koehl, 1976, 1977a), and because they would undergo local buckling if they were subjected to a wave (Koehl, 1977 b). (2) When stressed for periods of several hours to a day, as by the tidal currents which cause bending (Koehl, 1976,1977 a) and by muscular contractions and ciliary pumping which produce the extreme changes in a M. senile’s shape and size, the mesogloea is very extensible and thus requires small forces. to accomplish large movements.

A typical creep curve for 17 specimens of A. xanthogrammica mesogloea is shown in Fig. 3. The mesogloea of this species is fairly rigid when stressed for short times, but (unlike M. senile mesogloea) remains fairly rigid even after being stressed for several hours. The mean compliance (D(t)) of pieces of A. xanthogrammica mesogloea is not significantly different from that of pieces of M. senile mesogloea subjected to the same stress for 0·5 _s (F_{[1 4]} = 0·26, 0·50 < P < 0·75), but is significantly lower than that of pieces of M. senile mesogloea subjected to the same stress for 1 min (F_[1,8] = 7·99, 0.01 < P < 0·025) or longer, for example, 9 h (F_[1,8] = 1·42, 0.01 < P < 0·025) (see Figs. 9-11). This low compliance of A. xanthogrammica mesogloea for long times is consistent with the observation that an A. xanthogrammica does not undergo extreme shape and size changes and is not noticeably deformed after hours of exposure to wave action.

Short-term creep and recovery tests on A. xanthogrammica mesogloea using stresses and periods to simulate wave action (Koehl, 1977 a, b) showed that the material stretches by about 3 % of its resting length and returns to its original length before the ‘next wave hits’ (Fig. 4A). This complete recovery occurs even at stresses of as much as an order of magnitude greater (Fig. 4B, C) than those which are produced by non-storm waves. In contrast, M. senile mesogloea subjected to similar short-term creep and recovery tests does not recover completely to resting shape before the ‘next wave hits’ (Fig. 4D).

The time-dependent response to stress of each species’ mesogloea can thus be related to the behaviour of whole anemones when subjected to environmental forces and when performing mechanical activities.

The body wall of A. xanthogrammica reacts in the same way as that of M. senile to a variety of histological stains for collagenous connective tissues (Table 1) which suggests that its basic composition is similar to that of anemones (including M. senile) for which detailed chemical analyses have been performed (i.e. collagen fibres in a highly hydrated protein-polysaccharide matrix).

Both M. senile and A. xanthogrammica have definite patterns in the arrangement of fibres in their body walls. As noted by Gosline (1971 a), polarized light microscopy reveals two distinct layers in M. senile body wall mesogloea: (1) an inner layer of densely packed parallel fibres oriented circumferentially, and (2) an outer layer of fibres arranged in a three-dimensional feltwork. A. xanthogrammica mesogloea does not appear, in polarized light, to have these two distinct layers, but rather to be composed of a feltwork of swatches of fibres running parallel to each other (Fig. 6B). It appears that more of these tracts of fibres towards the inner surface of the body wall are oriented circumferentially than are those towards the outer surface.

The observation that the fibres in the inner layer of mesogloea are more circumferentially oriented than those in the outer layer are corroborated by SEMs of M. senile (Fig. 5 A, B) and A. xanthogrammica (Fig. 5C, D) body walls. The mean fibre angl (that between a fibre and the long axis of the body) of the outer layer of M. senile mesogloea (mean angle,⁠) is significantly lower (F_[1,70]= 259·62, P < 0·001) than that of the inner layer ⁠. Similarly, the mean fibre angle of A. xanthogrammica outer mesogloea is significantly lower (F_[170] = 141·72, P< 0.001) than that of inner mesogloea ⁠. The fibres of the inner layer of M. senile mesoglea have a significantly higher (F_[1,71] = 10·16, 0.001 < P < 0·05) mean fibre angle than do those of the inner layer of A. xanthogrammica mesogloea (i.e. more inner fibres are closer to being circumferentially oriented in M. senile than are those in A. xanthogrammica).

This structural anisotropy observed in mesogloea of M. senile and A. xanthogrammica can be related to the mechanical anistropy of these tissues. Only collagen fibres aligned parallel to stress axes will slide past each other and shear the matrix when the tissue is stretched; if collagen fibres have preferred orientations in the tissue, the tissue will be less extensible in directions parallel to them (Gosline, 1971 b). For short periods of stress application, both A. xanthogrammica and M. senile mesogloeas appear to be equally rigid in the longitudinal direction and the circumferential direction (Koehl, 1977b). However, when stressed for long times (as they would be with constant internal pressure) they are more extensible in the longitudinal than the circumferential direction (Fig. 7). This mechanical anisotropy is more pronounced in the mesogloea of M. senile than in that of A. xanthogrammica; this is not surprising in light of the observation that the anisotropy of collagen fibre arrangement is more pronounced in M. senile. Such circumferential reinforcement is no doubt advantageous for M. senile which often inflate themselves to the tall, slim shape illustrated in Koehl (1977 a).

The circumferential stresses due to internal pressure in hydrostatically supported cylinders are twice the longitudinal stresses (Wainwright et al. 1976). In thick-walled hydrostatic cylinders these circumferential stresses are greatest at the inner surface (Timoshenko, 1956), which is where the circumferential fibres are located in these anemones. The longitudinal stresses in a bending anemone are greatest at the outer surface (Koehl, 1977 b), which is where the fibres at the smallest angles to the body long axis are found in these animals. Thus, the reinforcing collagen fibres are arranged parallel to the major stress axes in different regions of the body walls of both M. senile and A. xanthogrammica.

Creep tests. A. xanthogrammica mesogloea is not as extensible as that of M. senile when subjected to tensile stresses for periods of a minute to a day. Based on the polymer mechanics illustrated in Fig. 8, I predicted that A. xanthogrammica was less extensible than M. senile mesogloea either because (1) A. xanthogrammica mesogloea has a longer entanglement plateau than that of M. senile, or because (2) A. xanthogrammica mesogloea is more highly cross-linked and hence reaches equilibrium length without undergoing much extension.

A typical log D(t) versus log t curve for A. xanthogrammica mesogloea at the biologically relevant temperature of 6 °C is presented in Fig. 9. A. xanthogrammica mesogloea behaves like a polymer solution in its entanglement platea (Fig. 8). Other creep tests at higher temperatures (analogous to running the tests for longer times) showed a rise in compliance after 12-24h (Fig. 9). This rise probably represents the transition where entanglements in the matrix break down (perhaps by slippage of tangled molecules or by disruption of hydrogen bonds). The possibility that mesogloea of A. xanthogrammica (unlike that of M. senile (Gosline, 1971 b)) contains mechanically continuous collagen fibres which undergo creep after long periods of stress has not been ruled out. However, if A. xanthogrammica mesogloea is assumed to be a composite material containing continuous fibres with elastic moduli of 10⁹ N.m⁻² (that of collagen (Wainwright et al. 1976)), the calculated modulus of the composite (Wainwright et al. 1976) is three orders of magnitude greater than its measured modulus (Koehl, 1977b). Thus, rearrangement of matrix molecules is a more likely explanation for the creep behaviour of A. xanthogrammica mesogloea than is stretching of continuous collagen fibres. Therefore, it appears that the mesogloea of A. xanthogrammica is less extensible than that of M. senile because the former remains in its entanglement plateau longer than the latter.

M. senile mesogloea reaches an equilibrium compliance after 12-24 h (Fig. 10), indicating the polymers in this mesogloea are cross-linked to each other to form a network (Gosline, 1971b). M. senile stressed at 21 °C for 24 h maintains an equilibrium compliance, indicating that the continuity of the network has not broken down and suggesting that the rise in compliance of A. xanthogrammica mesogloea under similar conditions is not due to tissue decomposition. Strips of mesogloea did begin to decay after 48 h of testing, however, thus A. xanthogrammica creep tests could not be runppong enough to determine whether or not this mesogloea is also a cross-linked network. A. xanthogrammica mesogloea displays the same sort of slow elastic recovery towards resting shape as does that of M. senile on cessation of prolonged stress. Because A. xanthogrammica mesogloea is in its entanglement plateau over the time scales involved, entanglements must be acting as the cross-links necessary for configurational entropy elasticity.

The creep of mesogloea anaesthetized in propylene phenoxitol was the same as that anaesthetized in MgSO₄ (Fig. 11). The similarity of these results indicates that the anaesthetics probably do not grossly alter the passive behaviour of mesogloea. The creep behaviour of unanaesthetized body wall, however, is quite different from anaesthetized because the muscles sporadically contract and relax throughout the experimental period (Fig. 11). This illustrates that an anemone can prevent or alter the extension of its body wall merely by contracting its muscles -a trick that manmade polymeric composite materials are incapable of doing.

The mean logs of the frequency-dependent storage moduli (E′) for both A. xanthogrammica and M. senile are plotted against the log of the inverse of frequency (roughly analogous to log time in a creep test) in Fig. 12. Over the frequency range used, the mesogloea of both species behaves like a polymer solution its entanglement plateau with storage moduli of the order of10⁵ N.m⁻². E’ is analogous to the inverse of D(t) measured for short times in creep tests; note that D(t) for A. xanthogrammica and M. senile mesogloeas at short times is 10⁻⁵ m².N⁻²¹ (Figs. 9–11). These storage moduli are consistent with those for M. senile mesogloea measured by Gosline (1971 i). With these short periods of stress application, the rigidity (E′) of A. xanthogrammica and M. senile mesogloeas are not significantly different.

The E’ of mesogloea pulled circumferentially in dynamic tests was not significantly different from the E’ of mesogloea pulled longitudinally for both species of anemones. The range of extension rates used in the dynamic tests (0·15–75·2% extension, s⁻¹) is roughly the same as the range of extension rates used by Koehl (1977 ft) in short-term force-extension tests of anemone mesogloea. The mean of the elastic moduli (E′) measured in those force-extension tests for mesogloea pulled circumferentially is also not significantly different from the mean of the E’s for mesogloea pulled longitudinally. These observations show that for short periods of stress application mesogloea from M. senile and from A. xanthogrammica..is mechanically isotropic. This is consistent with the observation that circumferential and longitudinal D(t)’s are not different from each other at short times during creep tests for both species.

The measurements of viscous energy loss (tan δ) for M. senile mesogloea are consistent with those of Gosline (1971b) (Fig. 12). The tan δ’s for A. xanthogrammica and M. senile mesogloea are not significantly different at high frequencies. However, at frequencies of 0-3 Hz and lower, M. senile mesogloea has a significantly higher tan δ than does A. xanthogrammica mesogloea (for example, F_{(1 10]} = 10·73, 0·005 < P < 0.001, when Hz = 0–3). What does such a difference between the tan δs of the mesogloeas of M. senile and A. xanthogrammica mean for these anemones in nature? At a frequency of 0·2 Hz, which roughly approximates the stress cycle due to a wave, the tan S of A. xanthogrammica mesogloea is significantly lower (F_[1,12]= 11·84, P⁼ 0·005) than that of M. senile mesogloea (i.e. A. xanthogrammica mesogloea stores as potential energy a greater proportion of the energy used to stretch it than does that of M. senile). Since such stored energy is used in the elastic return to resting shape of the mesogloea, it is not surprising that A. xanthogrammica mesogloea in short-term creep and recovery tests could return to resting length before the ‘next wave hit’ whereas M. senile mesogloea could not.

The mechanical behaviour of mesogloea from A. xanthogrammica and M. senile, when subjected to force-extension tests (Koehl, 1977 b) and to creep and dynamic tests, can be summarized as follows. (1) When stressed for short times, mesogloea is mechanically isotropic and in A. xanthogrammica and M. senile does not differ in rigidity (E, E’, or I/D(t)) or viscous energy loss (tan δ). (2) When stressed for longer times, the mesogloea of either species is more rigid when pulled circumferentially than longitudinally. (3) The rigidity of M. senile mesogloea is more time-dependent than that of A. xanthogrammica; when stressed for longer than 1 min M. senile mesoglea becomes less rigid than that of A. xanthogrammica. (4) When stressed at frequencies lower than 0·3 Hz, A. xanthogrammica mesogloea stores more energy for elastic recoil (i.e. has a lower tan δ) than does that of M. senile.

Thus, A. xanthogrammica and M. senile mesogloea, stressed circumferentially and longitudinally, behave in the same manner at short times when extension in response to stress is due to local straightening of polymer chains. The above differences arise because A. xanthogrammica mesogloea remains in the entanglement plateau longer than does that of M. senile, and because circumferentially stressed mesogloea remains in the entanglement plateau longer than longitudinally stressed mesogloea from the same species. This indicates that macromolecular differences that affect the length of the entanglement plateau are responsible for the observed differences in mechanical behaviour of mesogloeas.

There are several possible macromolecular differences which might increase the lengths of the entanglement plateaux of mesogloea. For example, they would be longer the higher the concentration and/or the molecular weight of protein-polysaccharide polymers in the matrix (Ferry, 1970). Entanglement plateaux would also be longer the more closely packed in parallel arrays along the stress axis are the collagen fibres in mesogloea. Such close parallel packing increases the deformation of the matrix between fibres, hence more extensive molecular rearrangements are required in the matrix for a given extension of the tissue (Gosline, 1971b; Wainwright, et al. 1976). Furthermore, the more closely packed the fibres, the greater the probability of entanglements between them and the matrix polymers. As described above, such a difference in collagen packing appears to be responsible for the differences observed in the rigidity of circumferentially versus longitudinally stressed mesogloea.

Several observations indicate that perhaps both of the macromolecular differences listed above are responsible for the observed differences in mechanical behaviour of mesogloea of M. senile and A. xanthogrammica. Estimates of the composition, by weight, indicate that there is nearly twice as much material other than water (i.e. collagen and matrix polymers) in A. xanthogrammica mesogloea as in that of M. senile (Table 2). The concentration by weight of the polymer solution in A. xanthogrammica mesogloea matrix is about twice that of M. senile. This indicates that there are probably more and/or larger molecules in A. xanthogrammica matrix than in that of M. senile. There is also almost twice as much collagen by weight in A. xanthogrammica mesogloea as in that of M. senile. If collagen fibres account for a greater proportion of the mesogloea in A. xanthogrammica than they do in M. senile, then the fibres would be more densely packed in the former than in the latter.

The suggestion that the mesogloea of A. xanthogrammica has more parallel collagen fibres than that of M. senile is corroborated by the observation that longitudinal, transverse, and tangential (Fig. 6 A, B) sections of the outer mesogloea of A. xanthogrammica are more highly birefringent than equivalent sections of M. senile. The higher birefringence in the mesogloea of the former species probably indicates the presence of more collagen fibres parallel to each other in the planes of the sections. It also appears in polarized light micrographs that wider swatches of parallel fibres occur in outer mesogloea of A. xanthogrammica than in that of M. senile. The fibres visible in the light microscope and in low-magnification SEMs of A. xanthogrammica and M. senile mesogloea are actually bundles of collagen fibres. These individual fibres can be resolved in TEM (Grimstone et al. 1958; Gosline, 1971a). The fibre bundles in SEMs of A. xanthogrammica mesogloea appear to be larger than those in SEMs of M. senile mesogloea (Fig. 6C, D).

My work suggests that A. xanthogrammica mesogloea is more rigid than that of M. senile, when stressed over the time scale of a minute to a day perhaps because it has a higher concentration of and/or larger polymers in its matrix, and perhaps, because it has more densely packed parallel arrays of reinforcing collagen fibres. These suggestions should be tested by TEM and by more detailed chemical analyses of the mesogloeas of both species of anemones.

The time-dependent response to stress of the body-wall mesogloea of two species of anemones can be related to the behaviour of the anemones when subjected to environmental forces and when performing mechanical activities. M. senile mesogloea is relatively rigid when stressed for short periods (as it would be during rapid muscular movements of the animals) but is very extensible when stressed for long times (as it would be by the slow tidal currents that bend these anemones over and as it would be during slow extensive postural changes).

In contrast, A. xanthogrammica mesogloea remains relatively rigid even after being stressed for several hours; these anemones do not undergo extreme postural changes nor do they deform noticeably when exposed to wave action for hours. A. xanthogrammica mesogloea stores for elastic recoil a greater percentage of the energy used to stretch it than does that of M. senile. The mesogloea of an A. xanthogrammica, when stressed by a wave, can thus elastically return to resting length before the impact of the next one.

Polarized light microscopy and SEM reveal that reinforcing collagen fibres in the anriv walls of the hydrostatically supported sea anemones, M. senile and A. xantho-grammica, are aligned parallel with the major stress axes in different regions of the body wall.

Mechanical tests of the mesogloeas of A. xanthogrammica and M. senile indicate that a basic way in which connective tissues composed of rigid fibres in pliant matrixes may vary in their mechanical behaviour is in the length of time that molecular entanglements retard extension of the tissues. Observations of composition and microscopic structure of the mesogloeas of these two species of anemones indicate that the duration of this ‘entanglement plateau’ is great if (I) the fibres are closely packed and oriented parallel to the stress axis, and (2) the concentration and/or molecular weight of the polymers in the matrix is high.

This study illustrates how structural features on the macromolecular and microscopic levels of organization of an organism can equip that organism for the particular mechanical activities it performs and the environmental forces it encounters.

This work was supported by a Cocos Foundation Training Grant in Morphology, a Graduate Women in Science Grant, and a Duke University Graduate School Travel Award. I gratefully acknowledge the extensive advice, enthusiasm, and support of S. A. Wainwright who introduced me to biomechanics. I also thank R. Fernald for the use of facilities at Friday Harbor Laboratories, University of Washington, and J. Gosline for his advice and for the use of facilities in his laboratory in the Department of Zoology, University of British Columbia. I appreciate the discussions with and help with instrumentation from M. Denny, M. LaBarbera, and S. Vogel. I am grateful to T. Suchanek for supplying me at Duke with freshly collected anemones from the Pacific, and to C. Levesque, L. Loveless, A. Painter, and M. Tomlinson for their technical assistance.