1. The musculature which operates the coxopodite and basi-ischiopodite of the fifth limbs of the common shore crab has been described.

  2. A comparison between the basi-ischiopodite depressor and anterior levator has shown that the levator muscle has a large reserve with which it effects limb severance at autotomy.

  3. Recordings from the nerve to the anterior levator muscle show that it is innervated by a tonic and phasic unit together with another, large spiking unit, which is activated at autotomy.

  4. Transverse sections of the nerve to the anterior levator show that it contains giant axons.

  5. Intracellular recordings from the anterior levator together with whole muscle tension confirm that there are three motor neurones, which produce a slow, fast and ‘giant twitch ‘response in the muscle.

Autotomy is the process by which a part of the body of an animal may be discarded at a point where there are structural adaptations which facilitate severance and reduce subsequent bleeding. Usually it is a damaged limb that is lost, as in the case of spiders (Parry, 1957), mantids (Bordage, 1905), starfish (Abeloos, 1932) and the gastropods Helixarian and Harpa (Abeloos, 1932) while lizards are well known for their ability to drop their tails (Morgan, 1932).

Carcinus can autotomize any limb subjected to damage. The mechanism by which this happens has been documented several times with the most convincing accounts given by Frédéricq (1892) and Paul (1915). Both basi-ischiopodite levators are involved, operating together to break the limb at a preformed plane situated in the fused basi-ischiopodite of each limb. The posterior levator is much the smaller of the two muscles. This muscle contracts when the limb is grossly stimulated, breaking a preformed plane in the tendon of the anterior levator (McVean, 1973). This action switches the application of the large levator so that its tension is concentrated onto a small skeletal plug crossing the limb breakage plane. A powerful contraction of the anterior levator withdraws the plug from its socket distal to the breakage plane so that a relatively small force applied externally to the limb is enough to sever it.

It follows from the above account that autotomy in Carcinus is a precise act involving co-ordination of at least two muscles. While the nature of the stimulus can vary, the neural output to these two muscles is standardized. Thus the central nervous system must be programmed to produce the co-ordinated contraction of these two muscles whenever a particular sensory threshold is achieved. It seems probable that this threshold is centrally determined and can be varied. An animal that has already lost several limbs shows an increased resistance to the loss of yet another (Hoadley, 1934; Gomez, 1964), while Carlisle (1957) found that Maia will not autotomize after terminal anecdysis has been reached.

What is not clear is how the motor commands to the two levators differ from normal walking and elevator commands. There is the choice, as far as the animal is concerned, between employing normal motor neurones at higher firing rates or invoking special motor neurones reserved for this purpose only.

Carcinus maenas (L.) were obtained locally. For straightforward light microscopy, material was fixed in Bouin’s solution and treated with Palmgren’s (1948) silver staining technique. The nerve supplying the anterior levator was fixed in 2·5 % gluter-aldehyde at o °C, buffered with 0·05 M cacodylate to pH 7·4, rinsed in buffer and postfixed in 1% buffered osmium tetroxide. After alcoholic dehydration it was infiltrated with epoxypropane and embedded in Araldite. From this were cut sections for light microscopy which were stained in a solution of 1 % borax and toluidene blue. Areas requiring greater resolution were cut from the same block and examined in the transmission electron microscope.

Electrophysiological investigations were performed with the crab held firmly upside down in a tailor-made clamp. The fifth limbs provide the greatest access with least dissection damage by virtue of their expanded sternal plates. To expose the anterior levator muscle and its nerve the sternal plate of the left side was removed and the ventral surface of the coxopodite was cut away. The basi-ischiopodite depressor and ventral remotor were cleanly removed, exposing the ventral face of the large anterior basi-ischiopodite levator (Text-fig. 1). The nerve supplying this muscle runs ventrally from the limb nerve as soon as it enters the box formed by the endophragmal skeleton around the muscles, but is accessible for only a short distance before it enters the body of the muscle. Activity in this nerve was recorded with glass suction electrodes while the immediate area was bathed in saline. Good preparations lasted several hours.

Text-fig. 1.

Diagram illustrating the form, situation and innervation of the basi-ischiopodite levators of the fifth left limb from a ventral viewpoint. The basi-ischiopodite depressor has been removed. all,al2, anterior and posterior branches of the anterior levator; bp, limb breakage plane ; dr, dorsal coxopodite remotor ; vr, ventral coxopodite remotor ; n, main nerve to limb ; n ali, nerve to posterior branch of the anterior levator; pl, posterior levator.

Text-fig. 1.

Diagram illustrating the form, situation and innervation of the basi-ischiopodite levators of the fifth left limb from a ventral viewpoint. The basi-ischiopodite depressor has been removed. all,al2, anterior and posterior branches of the anterior levator; bp, limb breakage plane ; dr, dorsal coxopodite remotor ; vr, ventral coxopodite remotor ; n, main nerve to limb ; n ali, nerve to posterior branch of the anterior levator; pl, posterior levator.

Intracellular muscle activity was recorded with glass micropipettes filled with 3 M-KCI, suspended by heat-shrinkable plastic tubing so that the tip of the electrode could move with the muscle. To record the tension produced by the anterior levator muscle when the nerve to it was stimulated, the tendon was cut at its insertion onto the basi-ischiopodite and held at rest length with forceps attached to the anode of an RCA 5734 tube. Electromyograms were obtained from both levators with silver wires of 50 μm diameter, insulated except for the tip, inserted through small holes bored through the skeleton in the region of the muscle. The position of the tip was confirmed by dissection after recording. All signals were displayed on either a Tektronix 565 or 502A oscilloscope.

General anatomy

Within the endophragmal chamber associated with each of the fifth limbs are five separate muscles. Three of these operate the coxopodite, two as remotors while the third promotes the limb. These muscles form discrete bundles and tend to have their origins restricted to within a small area. The large basi-ischiopodite levator and depressor muscles also have their origins within the thorax. Distally they penetrate the coxopodite to attach to the dorsal and ventral rim of the basi-ischiopodite respectively. These are powerful muscles containing a large number of fibres all of which attach to a central tendon, thus giving a bifurcated form to the muscle. Such muscles, as in the claw closer, are extremely powerful although their optimum tension range may be restricted because of the shorter length of the individual fibres. Sectioning these muscles gave an estimate of the number of fibres each contained. The results are expressed in Text-fig. 2. The number of fibres in the small posterior levator was estimated by dissection. This is the only muscle in this complex that has its origin outside the thorax, spanning from the inner dorsal face of the coxopodite to the dorsal rim of the basi-ischiopodite.

Text-fig. 2.

Diagram showing the application of each thoracic muscle on the coxopodite and basi-ischiopodite of the fifth right limb viewed from the right-hand side. The arrows indicate the direction in which each muscle moves the segment to which it is attached while the length of the arrow is proportional to the number of muscle fibres in each muscle. The shaded portion represents the basi-ischiopodite.

Text-fig. 2.

Diagram showing the application of each thoracic muscle on the coxopodite and basi-ischiopodite of the fifth right limb viewed from the right-hand side. The arrows indicate the direction in which each muscle moves the segment to which it is attached while the length of the arrow is proportional to the number of muscle fibres in each muscle. The shaded portion represents the basi-ischiopodite.

It is instructive to compare the capabilities of the basi-ischiopodite depressor and its antagonist, the anterior levator. This can be done by calculating the force required in each muscle to fulfil its particular function. Thus the depressor, when contracted, depresses the limb until the tip of the dactylopodite touches the ground. Further contraction will now raise the body of the crab, the dactylopodite acting as the fulcrum. By considering the weight of the crab and by taking moments about the coxo-basal joint, the force exerted by the depressor muscle to hold the animal off the ground can be calculated. For an animal weighing 60 gms this works out at about 130 g wt if all eight walking limbs share an equal load. This figure is calculated for the first two segments being in line. For an emarginated articulation as this is, the force required to hold the animal off the ground varies with the angle subtended between the coxopodite and basi-ischiopodite. If the limb should be imparting an acceleration to the body, then the force achieved by this muscle will be greater than in the static situation. In any case 150 g wt must be a conservative estimate. Under water, the fraction of this force devoted to combating gravity is considerably less.

Assuming that the contribution of each muscle fibre is more or less the same, the force exerted by each fibre of the depressor is in the region of 1 g wt. By a similar argument it can be shown that the force exerted by each fibre in the anterior levator muscle when the limb is held in the same position but with the dactylopodite off the ground is only o-i g wt. Forces of these magnitudes are within the range possible for single muscle fibres of crabs (Atwood, 1967). When the crab is submerged, the depressor muscle fibres have to mount a tension of about 0·2 g wt when all eight walking limbs contribute evenly in supporting the weight. The transition for a crab of this size from walking unsupported in air to walking under water must entail a remarkable re-setting of tension and motor output to the basi-ischiopodite depressor muscles.

It seems likely that the disproportionate number of muscle fibres of the anterior levator, in comparison with the depressor, are used to effect autotomy. If the muscle fibres of the anterior levator are capable of attaining the same order of tension as do the fibres of the depressor when they support the crab out of water, then nine-tenths of the potential power of the levator is reserved for securing a sufficient force to break the plug across the preformed breakage plane when the limb is autotomised. The larger the levator muscle the greater can be the strength of the plug across the breakage plane and the safer the limb from accidental autotomy. There must be a nice balance of strengths between the plug and the levator muscle allowing a reasonable safety factor to prevent the limb being broken off accidentally, yet the muscle must be strong enough to rupture the plug in autotomy. Two kinds of forces produce limb severance; a tensile force along the axis of the plug withdraws it from its socket while a shearing force performs the final separation along the preformed part of the breakage plane. Because the force exerted by the anterior levator during autotomy is applied at the point of insertion, there is a one-to-one mechanical advantage. Similarly, the breakage plane is sited immediately beyond the insertion so that once again the force elicited in the shearing plane is almost equal to that exerted by the muscle.

Structure of the muscles

All five muscles contained fibres whose structure ranged from the two classic extremes of ‘felderstruktur’ and ‘fibrillenstruktur’. The anterior levator contained a majority of ‘fibrillenstruktur’ fibres, with diameters ranging from 580 to 775 μm. A few fibres showing ‘felderstruktur’ and having diameters ranging from 290 to 370 μm are found towards the centre of the muscle. The posterior branch also contained a few fibres with an apparently exaggerated ‘felderstruktur’. Their diameters of about 680 μm are twice as large as any of the other fibres.

In Text-fig. 3 the cross-sectional area of each muscle is shown to increase with the number of fibres contained in it. If the cross-sectional area and the diameter of all the muscle fibres were uniform, then there should be a one-to-one correspondence between the cross-sectional area of the whole muscle and the number of muscle fibres contained within it. Thus if it was assumed that fibre diameter is independent of the crosssectional area of the whole muscle, then a muscle containing half the number of muscle fibres should have half the cross-sectional area of another muscle with twice the number of fibres. Text-fig. 3 shows that this relationship does not hold. Instead, the average cross-sectional area of each muscle fibre diminishes as the area in cross-section of the whole muscle increases. Inspection of individual muscles reveals that there is no uniform diameter for each muscle fibre, so that to be consistent with the described trend, the larger muscles must contain a greater proportion of small-diameter fibres than do the smaller muscles. Bittner (1968), in a table of distinguishing characteristics often attributed to crustacean phasic and tonic fibres, lists the former as having diameters of 300-800 μm and tonic fibres as having diameters of 50-800 μm. It is possible that the larger muscles contain a greater proportion of tonic fibres than do the small muscles. Both the basi-ischiopodite depressor and anterior levator have to be able to exert considerable tonic forces, so this interpretation is consistent with their function. Conversely, during fast walking on land these muscles have to work at rates of up to 1 Hz. Both muscles therefore need to display both phasic and tonic properties.

Text-fig. 3.

Graph showing the relationship between the cross-sectional area of each muscle and the average cross-sectional area and diameter of each muscle fibre contained within any one muscle. 1, basi-ischiopodite depressor; 2, ventral coxopodite remotor; 3′, anterior branch of the anterior basi-ischiopodite levator ; 3″, posterior branch of the anterior levator ; 4, coxopodite promotor; 5, dorsal coxopodite remotor.

Text-fig. 3.

Graph showing the relationship between the cross-sectional area of each muscle and the average cross-sectional area and diameter of each muscle fibre contained within any one muscle. 1, basi-ischiopodite depressor; 2, ventral coxopodite remotor; 3′, anterior branch of the anterior basi-ischiopodite levator ; 3″, posterior branch of the anterior levator ; 4, coxopodite promotor; 5, dorsal coxopodite remotor.

Innervation of the levator muscles

The gross anatomy of the nerves supplying both levators is shown in Text-fig. 1. In cross-section the nerve supplying the posterior branch of the anterior levator, at the point where it leaves the main nerve to the limb, contains twenty-one axons (Pl. 1, fig. 1). The smaller-diameter axons seen in the light microscope were confirmed by electron microscopy (Pl. 1, fig. 2). The five largest axons conform to the concept of a giant axon (Bullock & Horridge, 1965). Their diameters lie between 33 and 50 μm. The diameter of the remaining axons ranges from 24 to 3 μm. On the basis of diameter these axons fall into three groups, a conclusion which is supported by the recordings from the nerve.

Tonic and phasic activity in the nerve to the anterior levator

Two units could regularly be seen to be active. Each had a distinctive spike height. The unit with the smaller spike was tonic in character, with a strong tendency to continue firing whatever the position of the CB joint. Both units displayed the resistance reflex demonstrated by Bush (1965). Activity in the larger unit could always be arrested instantaneously by the smallest elevatory movement, but the tonic unit would continue firing though at a lower frequency. If the limb was rotated about the CB articulation and held at successively more depressed positions, followed by a similar traverse back towards an elevated position, the discharge in the tonic unit was not identical for a repeated position, being consistently less pronounced when the limb was being elevated step by step. Thus the discharge followed a hysteresis curve. The reflex discharge frequency was therefore not absolute for any given limb position, but was influenced by the immediately preceding movement or position. Bush showed that the resistance reflex in this case was generated by the CB chordotonal organ. In addition, passive movement of both ipsilateral and contralateral limbs evoked a similar but reduced discharge in both units.

Tactile stimulation to either the experimental limb, the carapace or the abdomen produced particularly vigorous activity in both units, although the tonic unit had the lower threshold and continued to be active after the stimulus was withdrawn, whereas the phasic unit stopped immediately.

Gentle to moderate stimulation thus produces the same kind of motor output to the anterior levator as is produced by stretching the CB organ. Both share the same motor neurones to the muscle.

Autotomy of a limb in an intact crab can be produced by a variety of stimuli. The most effective of these and the one that probably occurs in natural conditions is crushing of any limb segment except the dactylopodite. Autotomy can be produced in a more controlled manner by bringing a hot iron near to the limb. This has the advantage of not disturbing any electrodes monitoring activity in the nerves.

If either such stimuli is applied to the limb, both tonic and phasic units fire at high frequencies. If the stimulus is prolonged, a third, large spiking unit, becomes active (Text-fig. 4). Initially this third unit fires at a relatively slow rate, but stops if stimulation to the limb is removed. If the stimulation is maintained the frequency of the third unit increases to a critical level beyond which it continues irrespective of the presence or absence of any stimulus. In all cases this unit stopped firing after a few seconds.

Text-fig. 4.

Action potentials in three motor axons in n al2 in response to damaging stimuli applied distally to the limb. The tonic (a) and phasic (b) units fire almost continuously from the moment the stimulus is applied, while the giant axon (c) produces a coherent burst some seconds after the application of the stimulus. During autotomy this burst lasts for up to 20 sec.

Text-fig. 4.

Action potentials in three motor axons in n al2 in response to damaging stimuli applied distally to the limb. The tonic (a) and phasic (b) units fire almost continuously from the moment the stimulus is applied, while the giant axon (c) produces a coherent burst some seconds after the application of the stimulus. During autotomy this burst lasts for up to 20 sec.

Neuromuscular transmission and muscle tension

Preparations mounted with minimum dissection damage showed a variety of junction potentials in the anterior levator muscle. Text-fig. 5,a shows continuous tonic activity of small junction potentials. The resting potential is seen to fluctuate but not obviously in response to the frequency of the junction potentials which individually averaged about 0·5 mV. depolarization. Other fibres sampled showed junction potentials of about this size but with a much lower background frequency. Tactile stimulation of the carapace served to increase the frequency, and these junction potentials then showed some facilitation with marked summation (Text-fig. 5 b). These potentials were also produced by gentle stretching of the CB organ. Activity in the nerve to the muscle simultaneously with the junction potentials suggests that the tonic and phasic motor neurones respectively are responsible for the two kinds of muscle response.

Text-fig. 5.

Intracellularly recorded junction potentials from the anterior levator muscle take three forms. Tonic junction potentials (a) at a steady frequency barely disturb the resting potential whereas phasic potentials (b) produced in response to scratching the carapace, show some facilitation and marked summation. Crushing or heating the limb elicits bursts of over shooting potentials of some 60−75 mV, (c).

Text-fig. 5.

Intracellularly recorded junction potentials from the anterior levator muscle take three forms. Tonic junction potentials (a) at a steady frequency barely disturb the resting potential whereas phasic potentials (b) produced in response to scratching the carapace, show some facilitation and marked summation. Crushing or heating the limb elicits bursts of over shooting potentials of some 60−75 mV, (c).

The intracellular response of the muscle fibres to stimuli which in an intact crab produce autotomy are shown in Text-fig. 5 c. The junction potential gave rise to over shooting spikes of some 60−70 mV.

By monitoring the tension of the muscle and the post-synaptic response to stimulation of the nerve, three systems could be separately distinguished. The tonic system was elicited by reducing the intensity of stimulation until there was no longer any mechanical response to individual stimulus pulses. Initially no mechanical response could be seen, but by increasing the duration of stimulation a slow contraction became apparent. Tension varied with stimulation frequency, the lowest effective rate (20 Hz) requiring half a second to produce any measurable tension, while a stimulation frequency of 100 Hz produced the maximum rate of increase of tension as well as the maximum final tension (Text-fig. 6). This tonic system operates at high frequency, but within a comparatively narrow range of frequency, with a fivefold increase of tension from the minimum to the maximum effective rates. At all effective frequencies tension was maintained for a further 250 ms after stimulation had stopped, then declined slowly.

Text-fig. 6.

By reducing the intensity of stimulation to n a!2 until individual twitches of the anterior levator are abolished, the slow tension response becomes evident for high and prolonged stimulation frequencies. Stimulation frequency: 20, 50, too, 200 Hz.

Text-fig. 6.

By reducing the intensity of stimulation to n a!2 until individual twitches of the anterior levator are abolished, the slow tension response becomes evident for high and prolonged stimulation frequencies. Stimulation frequency: 20, 50, too, 200 Hz.

Within the same neuromuscular unit there are present two twitch systems (Textfig. 7). Stimulation at gradually increasing intensities elicits first the smaller of the twitch systems which shows summation of both junction potentials and tension. This unit differs from the tonic system in that it is considerably activated at frequencies that are only just beginning to produce tension in the tonic system. After stimulation has stopped, tension declines rapidly. At stimulation intensities above those that elicit this first twitch response, a further twitch system is seen which involves an active response of the muscle membrane. The active response follows the junction potential and is associated with a ‘giant ‘twitch of the muscle. As expected, the tension of the muscle when this system is activated is frequency-related, so that between frequencies of 20 and 50 Hz an extremely powerful tetanus is produced.

Text-fig. 7.

Intracellularly recorded junction potentials and tension produced by the whole of the anterior levator muscle in response to stimuli of the same frequency and duration applied at two intensities to n al2. In (a) the small junction potentials are associated with small but distinct twitches in the whole muscle. In (b) the intensity of stimulation was increased to the threshold at which the giant axons were excited, causing overshooting potentials in the muscle associated with much larger contractions.

Text-fig. 7.

Intracellularly recorded junction potentials and tension produced by the whole of the anterior levator muscle in response to stimuli of the same frequency and duration applied at two intensities to n al2. In (a) the small junction potentials are associated with small but distinct twitches in the whole muscle. In (b) the intensity of stimulation was increased to the threshold at which the giant axons were excited, causing overshooting potentials in the muscle associated with much larger contractions.

It is suggested that non-autotomizing activity of the anterior levator muscle is mediated through the tonic unit and smaller of the two twitch units. The tonic neurone maintains a background discharge and thus a continual, gently fluctuating tension unless centrally or reflexly inhibited. This tonic activity may be reflexly excited by external stimuli such as touching the carapace or internally by stretching the CB organ. The small twitch system always operates in conjunction with the tonic unit but requires a higher threshold of excitation to activate it. The response of this unit declines rapidly with withdrawal of the stimulus and also exhibits habituation. Centrally originating activity is also mediated through both these units.

The neurone producing the ‘giant’ twitch in the anterior levator is activated only by gross stimulation to the limb, such as crushing or heating it. This neurone is appar-antly directly influenced by peripheral stimulation but has a much higher threshold than either the tonic or small phasic neurones. Once this neurone is excited and achieves a certain firing frequency, usually about 1·5 Hz, it becomes self generating and runs for a pre-determined period.

It has been shown (McVean, 1973) that autotomy is brought about by the coordinated contraction of both basi-ischiopodite levators. The levators contract together, the posterior levator moving its tendon in such a way that it breaks the tendon of the anterior levator at a preformed breakage plane.

Paul (1915) has already shown that both levators contract simultaneously when the limb is crushed distally. Electromyograms show that the posterior levator is innervated by two units (Text-fig. 8 a), the smaller of which is tonic in nature and responds to non-autotomizing stimuli. The larger, phasic unit, like the autotomizer neurone innervating the anterior levator, responds only to crushing or heating of the same limb.

Text-fig. 8.

Electromyogram from the posterior levator (a) showing the two units active when the limb was crushed distally. The small tonic unit was activated by non-autotomizing stimuli whereas the phasic unit responded to gross stimulation of the limb. In (b) the activity of the phasic unit to the posterior levator is shown to precede the low-frequency potentials of the anterior levator which produce the powerful tetanus employed to sever the limb.

Text-fig. 8.

Electromyogram from the posterior levator (a) showing the two units active when the limb was crushed distally. The small tonic unit was activated by non-autotomizing stimuli whereas the phasic unit responded to gross stimulation of the limb. In (b) the activity of the phasic unit to the posterior levator is shown to precede the low-frequency potentials of the anterior levator which produce the powerful tetanus employed to sever the limb.

Clarac, Wales & Laverack (1971) reported that mechanical deformation of a cuticlestress detector, CSD 1 (Wales, Clarac and Laverack, 1971) whether by external means or internally by artificially pulling on the tendon of the anterior levator, produced a reflex discharge in the phasic unit to the posterior levator.

Paul measured the tension in both levators by cutting their tendons at the point of insertion onto the basi-ischiopodite and attached them to pivoted levers. In his experiment, therefore, CSD 1 could not have initiated the contraction of either levator nor participated in the reflex. This experimental result was confirmed by cutting the tendon of the anterior levator and attaching it to a force transducer. The limb was stimulated with a hot iron held near the meropodite thus ruling out the possibility of unintentionally stimulating CSD 1.

Text-fig. 8,(b) and Text-fig. 9 show that when the limb of an intact animal is crushed the phasic unit to the posterior levator becomes active before the autotomy neurone to the anterior levator and also finishes after it does. It seems to be an inescapable conclusion from these and from Paul’s results that general but gross stimulation anywhere on the limb except the dactylopodite produces a barrage of sensory input which impinges either directly or through interneurones onto the two units in question.

Text-fig. 9.

Activity of the phasic unit to the posterior levator (open circles) brackets the burst bin the autotomy neurone to the anterior levator (closed circles).

Text-fig. 9.

Activity of the phasic unit to the posterior levator (open circles) brackets the burst bin the autotomy neurone to the anterior levator (closed circles).

If strain on CSD 1 were responsible for initiating the autotomy reflex it is difficult to understand how bracing the limb upwards in non-autotomizing situations is not followed by limb severance. Such is not the case. The infliction of gross stimuli to the limb is a sine qua non in the production of autotomy in a normal crab. CSD 1 is not essential to the autotomy reflex although it may play a contributory role. Possibly sensory axons from CSD 1 synapse, together with other sensory axons from the limb, onto an interneurone, which in turn is coupled to both the phasic neurone to the posterior levator and the autotomizer neurone to the anterior levator. If such an interneurone is present it becomes possible to see how the permanent loss of other limbs could exert an inhibitory influence, and in doing so raise the threshold of autotomy for the remaining limbs. These possibilities are summarized in Text-fig. 10. The finer the details are being investigated in terms of the central connexions which influence the autotomizer neurone.

Text-fig. 10.

Two motor neurones to the anterior levator and one to the posterior levator respond at different thresholds to tactile stimuli from all over the body as well as more specific sources such as the CB chordotonal organs. The autotomizer neurone to the anterior levator is excited only by adequate stimulation to the limb it supplies, as is the phasic neurone to the posterior levator. It is postulated that an interneurone forms a gate whereby this sensory information may be modified depending upon the number of legs already lost.

Text-fig. 10.

Two motor neurones to the anterior levator and one to the posterior levator respond at different thresholds to tactile stimuli from all over the body as well as more specific sources such as the CB chordotonal organs. The autotomizer neurone to the anterior levator is excited only by adequate stimulation to the limb it supplies, as is the phasic neurone to the posterior levator. It is postulated that an interneurone forms a gate whereby this sensory information may be modified depending upon the number of legs already lost.

The two levator muscles concerned in autotomy have probably achieved this enhanced function secondarily to being straightforward limb levators. Although the anterior levator sustains a dual role it can only do so by the relegation of the posterior levator to a role solely concerned with autotomy. Originally this muscle was probably involved in simple limb elevation. Thus we might expect to find within the Brachyura or for that matter within other groups of animals that exhibit autotomy, that there are degrees of autotomy, from Carcinus where autotomy appears to be a ‘physiological ‘process to autotomy in the land crab Gecarcinus where the same phenomenon has distinct behavioural overtones (Robinson, Abele & Robinson, 1970). Carcinus can free itself from capture only if the aggressor, usually another crab, squeezes the limb hard enough to activate the autotomy reflex. Escape follows as a byproduct of cuticle deformation. Gecarcinus has incorporated the reflex into a distinct escape-behaviour pattern where the physical stimulus is suppressed and replaced by more complicated situational ones. Escape has become an end in itself.

A comparison between such instances should be rewarding in that whereas the basic system is expected to be similar, it should be possible to trace the development, in terms of central events, both anatomical and physiological, of a behavioural act from a physiological baseline.

It is a pleasure to thank Dr D. A. Dorsett for his help and encouragement and Professor D. J. Crisp, F.R.S., for the facilities of the Marine Station. Mrs Régine Hyde gave expert assistance with the electron microscopy.

Abeloos
,
M.
(
1932
).
La regénération et les problèmes de la morphogénése
,
253
pp.
Collection des Actualities Biologiques
.
Paris
:
Gauthier-Villars
.
Atwood
,
H. L.
(
1967
).
Crustacean neuromuscular mechanisms
.
Am. Zool
.
7
,
527
51
.
Bittner
,
G. D.
(
1968
).
Differentiation of nerve terminals in the crayfish opener muscle and its functional significance
.
J. gen. Physiol
.
51
,
731
58
.
Bordage
,
E.
(
1905
).
Recherches anatomique et biologiques sur l’autotomie et la regénération chez divers arthropodes
.
Bull, scient. Fr. Belg. (Sci)
39
,
307
454
.
Bullock
,
T. H.
&
Horridge
,
G. A.
(
1965
).
Structure and function in the nervous systems of invertebrates
.
San Francisco and London
:
W. H. Freeman and Co
.
Bush
,
B. M. H.
(
1965
).
Leg reflexes from chordotonal organs in the crab, Carcinus maenas
.
Comp. Biochem. Physiol
.
15
,
567
87
.
Carlisle
,
D. B.
(
1957
).
On the hormonal inhibition of moulting in decapod Crustacea 11. The terminal anecdysis in crabs
.
J. mar. biol. Ass. U.K
.
36
,
291
307
.
Clarac
,
F.
,
Wales
,
W.
&
Laverack
,
M. S.
(
1971
).
Stress detection at the autotomy plane in the decapod Crustacea II
.
Z. vergl. Physiol
.
73
,
383
407
.
Frédéricq
,
L.
(
1892
).
Nouvelles recherches sur l’autotomie chez le crabe
.
Archs Biol. (Paris)
12
,
169
97
.
Gomez
,
R.
(
1964
).
Autotomy and regeneration in the crab Paratelphusa hydrodons
.
J. Anim. Morph. Physiol
.
11
(
1
),
97
104
.
Hoadley
,
L.
(
1934
).
Autotomy in the anomuran Porcellana platycheles (Pennant)
.
Biol. Bull. mar. biol. Lab., Woods Hole
67
,
494
503
.
McVean
,
A. R.
(
1973
).
Autotomy in Carcinus maenas (Decapoda : Crustacea)
.
J. Zool., Lond
.
169
,
349
64
.
Palmgren
,
A.
(
1948
).
A rapid method for selective silver staining of nerve fibres and nerve endings in mounted paraffin sections
.
Acta zool. Stockh
.
29
,
377
92
.
Parry
,
D. A.
(
1957
).
Spider leg muscles and the autotomy mechanism
.
Q.Jl Microsc. Sci
.
98
(
3
),
331
40
.
Paul
,
J. H.
(
1915
).
Some new points on autotomy among decapod Crustacea
.
Rep. Dove mar. Lab
.
4
,
44
52
.
Robinson
,
M. H.
,
Abele
,
L. G.
&
Robinson
,
B.
(
1970
).
Attack autotomy: a defense against predators
.
Science, N.Y
.
169
,
300
1
.
Wales
,
W.
,
Clarac
,
F.
&
Laverack
,
M. S.
(
1971
).
Stress detection at the autotomy plane in the decapod Crustacea 1
.
Z. vergl. Physiol
.
73
,
357
82
.

Plate 1

Fig. 1. Transverse 10 μm section of the nerve to the posterior branch of the anterior levator. The giant axons are clearly visible with clusters of smaller diameter axons on the periphery of the nerve. Scale : 30 μm.

Fig. 2. An electron micrograph of a transverse section of the smallest-diameter axons visible in the light micrograph. These small-diameter axons are well supported by enveloping membranes while the membranes around the giant axons are more sparse. Scale: 1·8 μm.

Fig. 1. Transverse 10 μm section of the nerve to the posterior branch of the anterior levator. The giant axons are clearly visible with clusters of smaller diameter axons on the periphery of the nerve. Scale : 30 μm.

Fig. 2. An electron micrograph of a transverse section of the smallest-diameter axons visible in the light micrograph. These small-diameter axons are well supported by enveloping membranes while the membranes around the giant axons are more sparse. Scale: 1·8 μm.