Intersegmental Reflex Actions from A Joint Sensory Organ (Cb) to A Muscle Receptor (Mco) In Decapod Crustacean Limbs

In the walking legs of decapod crustaceans, the ‘myochordotonal organ’ (MCO) (Barth, 1934) constitutes with the ‘accessory flexor’ (AF) muscle a neuromuscular-sensory system whose functions can be compared to those of the thoracico-coxal organ (Alexandrowicz & Whitear, 1957; Bush, 1976), the muscle receptor organ of the crustacean abdomen (Alexandrowicz, 1951, 1967; Fields, 1976), and also the mammalian muscle spindle (Matthews, 1972). In all these systems, specialized muscle fibres transmit to the associated sensory cells the mechanical strains consequent upon joint movements, but can also act directly on the sensory discharge by their own contraction. The crustacean myochordotonal organ has been extensively studied, both anatomically (Clarac & Masson, 1969; Alexandrowicz, 1972) and physiologically (Cohen, 1963; Evoy & Cohen, 1971; Clarac & Vedel, 1971, 1975). Its role in mediating ‘resistance reflexes’, analogous functionally to vertebrate stretch reflexes, has also been well established (Bush, 1965b; Cohen, 1965; Evoy & Cohen, 1969; Vedel, Angaut-Petit & Clarac, 1975 a; see also Evoy, 1976; Clarac, 1977).

The existence of a single, discrete muscular structure, exclusively involved in the control of the activity of a sensory organ (MCO), allows a study of the reflex organization and physiological significance of its motor command. The accessory flexor muscle is supplied by only one tonic excitatory motoneurone, whose discharge is modulated by spontaneous or induced central influences (Vedel et al. 1975 a). It is also subject to resistance-reflex-like control, by MCO itself and by the chordotonal organs (MC 1 and MC 2) of its own mero-carpopodite (M-C) joint (Bush, 1965b; Evoy & Cohen, 1969, 1971; Vedel et al. 1975 a).

This paper considers the ‘intersegmental reflex’ actions exerted on the AF muscle by a joint proprioceptor organ located in a more proximal segment of the same leg, namely the coxo-basal chordotonal organ (CB), which monitors movement of the coxo-basipodite (C-B) joint (Bush, 1965 a). These reflex effects are compared with those exerted by the same CB organ on the extensor and flexor muscles producing M-C joint movements.

Thus the present study characterizes the reflex modulation, by one particular joint receptor, of the activity of the single excitatory motoneurone innervating a specific neuromuscular sensory organ. It also aims to extend our present understanding of the overall role of one proprioceptor in intersegmental motor co-ordination within a walking leg. This work represents part of a general study of the CB organ influences on all the joint musculature in the same limb (see Clarac, Vedel & Bush, 1978). Preliminary accounts of some of these results have appeared previously (Bush & Clarac, 1975; Vedel, Clarac & Bush, 1975b).

Four decapod crustacean species were used in this study: the crayfish, Astacus leptodactylus (Astacura); the rock lobsters, Palinurus vulgaris and Jasus lalandii (Palinura); and the shore crab, Carcinus maenas (Brachyura). Work on the three macrurans was carried out in Marseilles, that on the brachyuran in Bristol. The intact animal was strapped dorsal side up in a Perspex dish filled with cooled sea water maintained at about 10°C throughout the experiment.

Experiments were performed on the posterior (fifth), fourth or occasionally third pereiopod (i.e. walking legs), or on the large cheliped (first pereiopod) in the crayfish. The methods of exposing, mechanically stimulating and monitoring the imposed length changes of the coxo-basal chordotonal organ (CB), and of recording the extracellular EMG activity of the limb muscles in the meropodite, were as described in the preceding paper (Clarac et al. 1978).

Intracellular recordings from single muscle fibres of the distal head of the accessory flexor muscle, exposed in situ in the posterior leg of Carcinus, were made using 3 M-KCI filled glass micropipettes. In these experiments, ramp-function CB length changes of 1−1·2 mm (12−16% of its mid-resting length) were applied by means of a servo-controlled puller unit (Bush, Godden & Macdonald, 1975).

Four muscles produce the movements of the C-B joint which levate and depress the distal part of the leg, namely the anterior and posterior levator and depressor muscles (Fig. 1). Between the two levator muscles lies the coxo-basal chordotonal organ (CB), whose sensory cells respond to levation and depression movements of the C-B joint (Bush, 1965 a).

Two muscles are involved in producing the M-C joint movements, the extensor muscle inserting dorsally and the main flexor muscle lying ventrally (Fig. 1). In addition, the accessory flexor (AF) muscle, whose tendon is attached in parallel to that of the main flexor, is characterized by its small size and the fact that it is composed of two separate parts linked by a very thin rod-like tendon. The proximal part is spindleshaped and runs longitudinally within the meropodite, while the distal part is flattened and obliquely orientated. This is not a true ‘effector’ muscle, since it does not contribute any significant flexor power in the limb. Instead it is essentially a receptor muscle, being associated with a two or three component chordotonal organ at its proximal end. The whole complex is therefore aptly termed a ‘myochordotonal’ organ, the only one of its kind in the leg (Cohen, 1963; Clarac, 1968). The myochordotonal receptor organ (MCO) is associated functionally in series with the proximal part of the AF muscle. Distally in the meropodite near the distal head of the AF muscle are two ordinary ‘chordotonal organs’ (joint receptors), MC 1 and MC 2. MC 1, MC 2 and MCO all respond to passive flexion and extension of the M-C joint (Wiersma, 1959; Clarac & Vedel, 1971, 1975).

The M-C extensor and flexor muscles are respectively innervated by two and four excitatory motoneurones, the AF muscle by only one (Wiersma & Ripley, 1952). Each of these muscles, including the accessory flexor, also receives a peripheral inhibitory innervation. The extracellular EMG electrodes used in this study provided a reliable monitor of any activity occurring in the excitatory motoneurones supplying the muscles under observation. For technical reasons considered in the previous paper, however, discharges of the inhibitory motoneurones could not be monitored in this way, so that no conclusions can be drawn about their activity in the reflexes studied here (see Clarac et al. 1978).

In general, the results obtained in the four species studied were similar. Owing to their greater availability and suitability for these experiments, Palinurus vulgaris was used in most cases, and the details described in this paper and all figures apply to this species, except where otherwise indicated. Likewise, the results were identical for the different thoracic limbs, but all results illustrated here were obtained from the posterior two walking legs or, in Astacus leptodactylus, from the cheliped.

In all reptantian Decapoda this muscle is innervated by two excitatory motoneurones (Wiersma & Ripley, 1952). Stretching CB in the stationary limb excites a single extensor motor unit, and releasing CB inhibits it (Fig. 2). This unit often discharges tonically, at a frequency which depends partly upon the degree of flexion of the M-C joint. For any given M-C position, this tonic discharge is generally somewhat greater when CB is held stretched than when it is at a shorter length.

This muscle is exceptional among the leg muscles in receiving four excitatory motoneurones. Two tonic flexor units were sometimes encountered in the present recordings (Fig. 3), their resting discharge frequency increasing with degree of maintained M-C extension. The smaller, more tonic unit discharges at a higher frequency with CB in a released position (Figs. 3−5), but the larger unit usually showed a higher background frequency while CB was held stretched (Fig. 3). Both flexor units, however, were excited by dynamic shortening of CB, and inhibited during CB stretching.

This receptor muscle is innervated by a single excitatory motoneurone, together with a peripheral inhibitory motoneurone (Wiersma & Ripley, 1952; Angaut-Petit, Clarac & Vedel, 1974). Like the two tonic units innervating the main flexor muscle, the single excitatory accessory flexor unit (AF) is reflexly excited by CB shortening (Figs. 2−5; see also Bush & Clarac, 1975). Its generally pronounced tonic discharge, however, commonly is little if at all influenced by the resting length of CB. Nor does this tonic discharge vary consistently with the resting position of the M-C joint. Rather it appears to depend upon unknown factors which affect the degree of central excitability. Stretching CB sometimes inhibits this tonic discharge, though often only slightly and sometimes not at all. Thus the most consistent influence on the accessory flexor muscle seen in the present experiments is that of CB shortening in exciting it.

Since the accessory flexor is a very small muscle within the large volume of the meropodite, the possibility existed that the EMG (or motoneurone) recordings obtained with the relatively gross extracellular needle electrodes used here were in fact from either the main flexor or the extensor muscle (or even from a motor axon supplying a more distal muscle). To counter this potential criticism, we exposed the distal head of the accessory flexor in the posterior leg of several Carcinus preparations, and recorded intracellularly from a number of its muscle fibres, using 3 M-KCI filled glass micropipettes. Some of these preparations showed only weak or inconsistent activity, probably due to damage during the dissection, but in the more healthy, reflexly active ones the response to CB length changes was clearcut.

Stretching CB generally inhibited any ongoing spontaneous excitatory junctional potential (e.j.p.) activity, while releasing CB caused a somewhat variable but definite increase in e.j.p. frequency (Fig. 4). The resulting facilitation and summation of e.j.p.s. led to a small plateau of depolarization, with peak e.j.p. amplitudes of 10 mV or more at the relatively low frequencies obtained. The fact that they were always depolarizing e.j.p.s, with no indication of a reversal potential near resting potential (which was usually around −50 to −60 mV in the more responsive muscle fibres), indicates that they were indeed the intracellular, postsynaptic responses to impulses in an excitatory motoneurone. No evidence was in fact obtained in these experiments of any activation of the peripheral inhibitory motoneurone. Thus it can safely be concluded, on the basis both of these intracellular recordings and of the foregoing extracellular EMGs, that CB shortening reflexly activates the single excitatory motoneurone of the accessory flexor muscle.

Simultaneous recordings from these two muscles frequently exhibit some temporal correspondence between the discharges of the one (or two) ionically active flexor unit(s) and the single accessory flexor unit (Figs. 5, 3). The correspondence is variable and not precise, and in many records appears to be largely absent. Nevertheless almost synchronous firing of these units occurs sufficiently often to suggest a definite functional, central coupling between these two (or three) motoneurones, albeit a rather loose one. Such coupling has been noted previously between the accessory flexor unit and a tonic flexor unit in Astacus leptodactylus (Angaut-Petit et al. 1974; Angaut-Petit & Clarac, 1976). In the present study it has been seen in Palinurus as well as Astacus. No comparable degree of coupling is ever seen between the accessory flexor and extensor excitor units (see Fig. 2).

In preparations showing clear coupling between these two tonic units, the degree of coupling does not appear to be significantly modified by the resting length, or change of length, of CB. During a 23 s period with CB held stretched in the same preparation as that represented in Fig. 5, for example, 90 flexor and 128 accessory flexor impulses occurred, and 73 (i.e. 81 % of the former) were closely coupled-i.e. to within ca. 10% of the flexor inter-spike interval. The corresponding activity for a similar 24 s period with CB relaxed, to a length 3 mm shorter than before, were 115, no, 102 and 89%, respectively. Nor do the ‘spontaneous’ bursts of activity occasionally encountered in these preparations (see below) disrupt this coupling, when present. In some such cases, in fact, the amount of coupling may even increase, due in part to a rise in the discharge frequency of the tonic flexor unit to near that of the accessory flexor unit (Fig. 9D). Even close pairs of impulses in this condition may be fairly well coupled between the two units, though they are rarely absolutely synchronous.

When a second flexor unit was tonically active, this too showed some tendency to fire roughly simultaneously with the accessory flexor unit (Fig. 3). In this case the coupling was more in evidence when CB was held stretched than when it was at a short resting length (Fig. 3D). Here too, however, this difference may have been more apparent than real, owing to the very low discharge frequency when CB was relaxed; there were insufficient occurrences of these flexor motor impulses overall to allow quantitative comparison, as was possible for the smaller flexor unit.

Occasionally a further, much looser form of ‘coupling’ was observed in some of these experiments. A tendency towards gentle rhythmic discharge can sometimes be seen in the accessory flexor motor unit (Fig. 3C, D; see also Fig. 6C, D in the previous paper, Clarac et al. 1978). In such cases the tonic flexor unit may fire single (or groups of) impulses roughly in phase with the AF ‘bursts’. When active, the same can be said of the second, larger, tonically discharging flexor unit (Fig. 3 D). Again, however, too few recordings are available to permit any quantitative analysis of this phenomenon.

When a series of sinusoidal CB length changes begins with a release from a more stretched length, the reflex response of the activated flexor and accessory flexor units to the first release is much stronger than those to subsequent releases (Figs. 5 A, 6). If a stretch precedes the first CB release of a series, however, the response to this first release is commonly not significantly greater than subsequent ones (Figs. 5 B, 6), and sometimes it is smaller. A similar effect, in the opposite direction, is sometimes evident in the extensor response to CB stretch (see Fig. 7). The number of extensor impulses on the first cycle in this preparation is in fact spuriously low, owing to the first stretch in each case having started in the mid-position rather than in the fully released position; cf. Fig. 2 A.

With a long series of sinusoidal CB length changes, the reflex discharges of all responding units, but particularly the accessory flexor unit, commonly continue to adapt over several cycles (Fig. 8 A). The curves in Fig. 8 A represent the mean number of impulses per excitatory half-cycle, in each of ten successive cycles, averaged over five (or nine) velocities of CB movement (cf. Fig. 8B). A small ‘reverse adaptation’, i.e. a progressive increase in successive responses, is sometimes seen in the units responding to the second movement direction of a series, particularly the extensor unit. However, this can probably be accounted for partly by the different starting position of CB, relative to its normal length range in situ.

Most of the sinusoidal changes in CB length illustrated so far have been at approximately the same rate, namely 0·25 cycles/s (4 s/cycle). With a CB excursion of 2 mm, this represents a rate per half-cycle-i.e. a mean velocity of CB stretch or release, averaged over the complete half-cycle –of 1 mm/s. Varying the rate of sinusoidal (or constant-velocity) length changes shows that the resulting reflex discharge frequencies vary directly with stretch or release velocity. This has been described previously for the classical intra-segmental resistance reflexes (see Bush, 1962, 19656), but it also clearly applies to the intersegmental reflexes studied here, particularly those of the three M-C joint muscles.

If the sinusoidal stretch is changed smoothly during one cycle from a cycle period of ca. 12 s to one of ca. 3·5 s, the amplitude of the instantaneous frequency (i.e. reciprocal of inter-impulse interval) envelope, and the peak values of the AF motor frequency, increase markedly at the higher CB velocity. Furthermore, they adapt with cycle repetition at the higher velocity to a mean peak level still well above the adapted level at the lower velocity. Comparison of the responses of all three muscles in the meropodite, at four different velocities of sinusoidal CB length change (Fig. 7), reveals a clearcut positive relationship between their reflex discharge frequencies and CB velocity.

This relationship is illustrated graphically in Fig. 8B, which plots mean discharge frequencies of the active motoneurones of all three muscles in the meropodite, for each of five different velocities (AF 2, E, F), and also those of the accessory flexor unit alone for nine different velocities, recorded 3 h earlier in the same preparation (AF 1). The mean frequency values were obtained by dividing the total number of impulses (i.e. EMG ‘spikes’) occurring during the appropriate (excitatory) half-cycle by the half-cycle duration, and averaging this value over ten successive cycles. Each series started with a CB stretch from approximately mid-way between the minimum and maximum CB lengths attained. Similar but less regular curves are obtained if the frequency occurring in any one cycle alone (e.g. the first, or fourth, cycle) at each velocity is selected for plotting.

In any event, it may be concluded that in the passive, quiescent animal, the reflex response frequencies of these three tonic motor units vary directly and fairly systematically with velocity of CB length change, in addition to showing a rather less readily defined or consistent ‘adaptation’ to successive sinusoidal length changes. It should be noted that although increased velocity of CB movement does result in an increase in the mean frequency of the AF (and F and E) motor impulses, the total number of impulses per half-cycle remains roughly constant, or may even decrease slightly with the decrease in cycle duration. However, it seems highly probable that the observed differences in AF as well as F and E mean frequencies will, as in other crustacean neuro-muscular systems, result in corresponding differences in AF (and F and E) tension development, and hence also in the afferent feedback from MCO (see Discussion).

If the animal’s eyes are suddenly illuminated, particularly after being in darkness for a while, the active flexor unit(s), and especially the accessory flexor motoneurone, respond with a greatly increased discharge. The latency of this ‘light reflex’ is long, commonly 1·2−1·8 s, indicating a polysynaptic pathway. The intensity of the effect varies directly with the change in intensity of the light falling on the eyes. Responses to sinusoidal changes in CB length summate with the light response so as to modulate the high frequency discharge (Fig. 9 A–C). This enhanced discharge of the accessory and main flexor units following illumination declines progressively over several CB cycles, often lasting for 10 s or more before the discharge frequencies return more or less to their ‘normal’ pre-illumination levels.

That the initial light response is not an indiscriminate increase in general central excitatory state, however, is shown by the fact that the response of the single active extensor unit to CB stretch decreases, in parallel with the activation of the flexor units. This observation is consistent with a general ‘alerting’ effect of the light stimulus, since this might be expected to involve activation of the accessory flexor muscle (cf. Evoy & Cohen, 1971), in association with some flexion of the M-C joint, as the restrained animal attempts to stand more upright. Other modalities of stimulation which appear to have a general arousal effect behaviourally, e.g. sharply tapping or stroking the animal on its dorsal surface, also influence the flexor and accessory flexor units in a somewhat similar way.

As in most experiments on ‘lively’, viable crustacean preparations with intact CNS, sporadic, apparently spontaneous, bursts of motoneurone activity occasionally occurred in the present study (Fig. 9D). Such centrally originating activity often tends to override, and sometimes completely block, the normal reflex responses to proprioceptive stimuli. An example of this is illustrated in Fig. 10B, where both the flexor and the accessory flexor units were evidently centrally activated while the extensor unit was inhibited (cf. Fig. 10A). This might have represented part of a general ‘arousal’ reaction, similar to that referred to above in response to exterocep-tive stimuli.

In contrast to the other two muscles, the accessory flexor unit still appeared to be modulated by CB length changes in a more or less normal manner. This again is reminiscent of the situation during a strong light reaction (e.g. Fig. 9C), when the accessory flexor unit was more clearly modulated than the main flexor units or the totally inhibited extensor unit. Slight modulation of the main flexor unit by CB was nevertheless still discernible, although owing to its relatively low discharge frequency, it was less obvious here than in the accessory flexor. It would seem, then, that the reflex action of CB upon the accessory flexor motoneurone in particular, does constitute a fairly persistent influence upon this receptor muscle, an observation which must have important functional significance.

The major new finding of this investigation is the clearcut reflex modulation by one leg joint proprioceptor of the motor output to a muscle receptor in another segment of the leg. Imposed changes in length of the coxo-basal chordotonal organ (CB) alone evoke, in the resting limb, pronounced dynamic alterations in the tonic discharge of the single excitatory motoneurone (AF) innervating the accessory flexor muscle of the myochordotonal organ (see also Bush & Clarac, 1975). At the same time, CB movement also reflexly influences the more tonically active excitatory motoneurones of the main flexor and extensor muscles of the same, M-C joint (see Clarac et al. 1978). Evidence is also presented that, as would be expected, all these reflexes - including those of the AF motoneurone - can be overridden or substantially modified by ‘central drive’ and other, ‘extrinsic’ factors, including illumination of the eyes and general tactile stimulation.

Of particular interest in the present results is the broadly parallel nature of the dynamic reflex actions of CB length changes upon the AF and tonic main flexor (F) motoneurones. Thus both AF and one or two of the four excitatory motoneurones of the flexor muscle are reflexly excited by CB shortening and inhibited by stretching CB, while the tonic extensor unit (E) is excited during stretching. Moreover, both AF and F units respond at frequencies which increase together with velocity of CB shortening; they both also adapt to repetitive sinusoidal movements with a similar ‘time-course’ (Fig. 8, cf. AF 2 and F). Further, central drive and other extrinsic factors also tend to affect the AF and F neurones similarly, and the extensor unit in opposite sense.

The ‘static’ reflex actions on these motor units of different resting CB lengths, in contrast to the dynamic reflexes referred to above, are less predictable. Firstly, the AF motoneurone’s tonic background frequency is not consistently related to CB length, being greater at times with CB held stretched and at others with CB relaxed, even over quite short recording periods. Possibly this reflects the variety of influences evidently impinging upon the AF motoneurone (cf. Evoy & Cohen, 1971 ; Vedel et al. 1975 a). More surprising is the observation that in most cases where two main flexor (F) units are tonically active, the resting length of CB evidently affects them in opposite ways (Fig. 3C, D). The smaller, most tonic flexor unit is more active when CB is relatively short than when it is held stretched, as would be expected by analogy with the intrasegmental reflexes, whereas the reverse applies to the larger of these two units. The functional significance of this reciprocal influence of CB length upon two units of the same muscle is as yet obscure.

Apart from the general parallel activation of the AF and F motoneurones discussed above, a more specific form of ‘coupling’ between individual impulses in these two motoneurones was sometimes observed. The degree of this coupling does not seem to be modified by CB position or movement, nor by spontaneous central activation of these units. This observation is consistent with the neuronal model previously advanced by Angaut-Petit et al. (1974), with a single excitatory interneurone driving both units. The indication of some degree of synchrony between occasional rhythmic discharges of the AF motor unit and impulses in the small tonic flexor unit is also compatible with a common excitatory interneuronal input to these two motoneurones. These observations suggest that both the CB reflex and the central activation of these two tonic motor units may normally be mediated by a common driver interneurone. In some of the present experiments a second, larger but at times tonically active flexor unit sometimes also tended to fire roughly simultaneously with the single accessory flexor unit (Fig. 3C, D), though owing to its lower discharge frequency this coupling was less obvious than with the smaller unit. Since CB length appeared to have a partially antagonistic effect on this unit (see above), additional interneurones may be involved here.

Since the accessory flexor is demonstrably a ‘receptor muscle’ rather than a power muscle, any variation in AF motor discharge frequency, whether central or peripheral in origin, is likely to influence the sensory signal emanating from the associated myochordotonal receptors (Clarac & Vedel, 1971 ; Cohen, 1965 ; Evoy & Cohen, 1971). The strength of any such efferent influence on the afferent signal will presumably depend upon the magnitude of the frequency changes. Thus the increase in AF frequency with velocity of CB shortening, or on eye illumination, for example, or its progressive adaptation with repetitive CB movements, will probably result in corresponding changes in feedback from the MCO.

Like the two M-C joint chordotonal organs, MC 1 and MC 2, the myochordotonal (MCO) receptors contribute to the intrasegmental resistance (or ‘myotatic’) reflexes of the main flexor and extensor muscles resulting from passive M-C joint movement, albeit in a more labile manner (Bush, 1965b; Evoy & Cohen, 1969; Vedel et al. 1975a). Any variation in AF motor activity might therefore be expected to affect the ‘gain’ of the intrasegmental MCO reflex loop. Under appropriate experimental conditions, isometric contraction of the accessory flexor muscle tends to cause reflex contraction in the main flexor muscle (Clarac & Vedel, 1975), and accordingly to enhance the reflex resistance to any imposed M-C extension - i.e. to increase the gain of this myotatic reflex. Thus the intersegmental reflex excitation of AF by CB shortening will tend to reinforce any concurrent intrasegmental flexor resistance reflex, in addition to the direct, intersegmental reflex excitation of the main flexor muscle caused by CB shortening.

In vivo, when variations in discharge rate of the flexor and extensor motoneurones result in overt movements at the M-C joint, co-activation of the AF and F unit would tend to adjust the AF length to that of the flexor muscle. That is, such AF activity could serve essentially to maintain the gain of the MCO receptors, and thereby of the whole MCO →M-C servo loop, similar to the role postulated for the motor control of the thoracico-coxal muscle receptor (Bush, 1977). Whether this servo loop gain was actually increased or decreased, rather than simply maintained at a more or less constant level, would depend quantitatively upon the relative degrees of excitation (or inhibition) of the AF and F motoneurones. Certainly where the variation in flexor and extensor motor activity results, not in overt M-C movement but simply in tonus changes, any variation in AF motor discharge might indeed be expected to modulate the loop gain accordingly. From the present experiments it is clear that, although the CB reflex drive to AF often does appear to exceed the flexor drive, particularly in Astacus (Figs. 7, 8), this can vary, even in the same preparation (Fig. 5, cf. A, B, C). Different stimulating factors, moreover, can affect AF and F to different extents (Fig. 9, cf. A, D).

The results presented in this and the preceding paper strongly suggest that the CB organ, and very probably each chordotonal organ in the leg, contributes to the regulation and co-ordination of posture and movement, not only of its own joint but of the entire leg. An important element in this regulatory control is the specific reflex modulation by CB of the motor output to, and hence probably of the proprioceptive re-afference from, the myochordotonal muscle receptor organ in the meropodite. The apparently greater ‘flexibility’ of this reflex, compared to the accompanying reflexes from CB to the main M-C power muscles, probably reflects its susceptibility to other interacting or potentially overriding influences, both peripheral and central.

In these terms, a plausible general hypothesis is that the MCO might be critically involved in determining the optimal overall load distribution, and hence posture, of the leg for a particular stance (cf. Evoy, 1976). The ‘pure’ chordotonal organs like CB (i.e. those without efferent control) might then provide, through the various intra- and intersegmental reflexes described, a ‘fine adjustment’ of the position and movement of each joint in the leg. Such an hypothesis would be consistent both with previous results from behavioural studies, e.g. those involving receptor ablation or limb immobilization (Evoy & Cohen, 1971; Clarac, 1977), and with the present evidence of highly specific (yet possibly widespread) proprioceptive regulation of AF motor activity. Further experiments are planned to test this hypothesis, and to seek answers to the other questions raised by this investigation.