Proprioception by the Coxo-Basal Chordotonal Organ, Cb, in Legs of the Crab, Carcinus Maenas

In each joint of decapod crustacean legs there are one or two chordotonal proprioceptor organs (Alexandrowicz & Whitear, 1957; Wiersma, 1959), whose scolopidia each receive the distal processes of two bipolar sensory cells (Whitear, 1960, 1962). These processes are of two structural types : one has a ciliary segment, with the fine structure of a sensory cilium, interposed between the axial filament proximally and the paraciliary and terminal segments distally; the other does not. The two sensory neurone types are accordingly termed ‘ciliary cell ‘and ‘paraciliary cell ‘respectively. The scolopidia of the propo-dactylopodite joint organ, PD, and of the two carpo-propodite receptors, CP ₁ and CP ₂, and the two mero-carpopodite receptors, MC₁ and MC ₂, are all ‘heterodynal’, each scolopidium having a ciliary and a paraciliary ending. The coxo-basal organ, CB, on the other hand, has ‘isodynal’ scolopidia, both their distal processes being of the ciliary type. Each receptor’s scolopidia lie in an elastic strand or sheet of connective tissue, attached proximally to a muscle tendon—except in CB— and distally to the integument of, usually, the next segment of the leg.

Burke (1954), recording from the sensory nerve of PD in Carcinus maenas, observed small tonic spikes and large phasic spikes in response to passive and active movements of the joint and to stretching and relaxing of the isolated organ. Wiersma & Boettiger (1959) found four main types of fibre in the PD nerve, position-sensitive afferents and movement-sensitive afferents for each direction—adduction (‘closing’) and abduction (‘opening’). Similar afferent fibre-types occur in CP₁ and CP₂, and in MC₁ and MC₂ (Wiersma, 1959). Hitherto no analysis of the responses of the coxo-basal receptor has been reported.

Since the sensory endings of CB are all of the ciliary type, it was thought that an electrophysiological study of the receptor might indicate the function of the ciliary cell in these chordotonal organs, and hence also, indirectly, of the paraciliary cell. An analysis has therefore been made of its afferent responses to imposed (passive) movements of the coxo-basal joint, with the receptor in situ, and to stretching and releasing of the excised receptor strand. With the latter procedure any possible extrinsic influences on the fundamental sensory response should be eliminated. A subsequent paper presents a further comparative study on the more distal receptors (Bush, 1964).

The shore crab, Carcinus maenas (L.), was used for all experiments, this being the species employed by Whitear (1962) for electron-microscope studies, and by Wiersma & Boettiger (1959) for PD afferent recordings. Pantin’s (1946) physiological saline solution for Carcinus, brought to pH 7·5 by adding 1·5 mM./l Na₂HPO₄, was used both for the experimental preparation and for keeping animals alive after removal of one or two legs.

A walking leg, generally the first or second on one side or the other, was excised, taking with it some of its intra-thoracic basal musculature to avoid damage to the receptor and to ensure a sufficient length of nerve for recording purposes (4-5 mm.). The leg was severed 2-3 mm. distal to the basi-ischiopodite articulation. The isolated portion of the leg was then fixed in a small clamp in an upright position, with the exposed central end of the coxa uppermost and the basipodite downwards. The CB nerve was cut about 1 cm. from the thoracic ganglion, and traced distally to the receptor strand. This could then be more or less completely exposed by dissecting away the surrounding muscles. In some later experiments the coxal musculature was left as intact as possible, so as to avoid any possibility of modifying the receptor response by removal of muscle and connective tissue attachments ; however, no obvious differences in response were evident. The preparation was usually almost completely immersed in physiological saline.

By means of a thin pin inserted through the anterior and posterior faces at the relatively stiff basi-ischiopodite joint the preparation was coupled to a manually operated Palmer rack-and-pinion. With this device the basipodite could be moved passively through any angle of its normal arc of movement, up to the maximum physiological angle of about no°. Usually the joint was not moved through more than 90° of this arc, so as to avoid excessive stimulation in the extreme positions; no essential part of the total response was thereby omitted. This was achieved by means of appropriately placed stops on the rack-and-pinion, whose total excursion was then 8—9 mm. The displacement of the rack-and-pinion—and hence of the moving joint—was monitored by a simple lever carrying a vane which interrupted a light beam focused on a phototransistor, whose amplified output was fed into one beam of the oscilloscope. The sensory response of the receptor was recorded on the other beam via an a.c. coupled preamplifier, using a micromanipulated monopolar platinum wire electrode.

In each experiment the response of the whole CB nerve was recorded first, the cut end of this nerve being lifted above the surface of the saline bath for recording. Subsequently the nerve was split into small bundles with a fine stainless steel needle. To facilitate this procedure the cut end of the whole nerve was held at the saline surface under slight tension by means of a small micromanipulated forcep-clamp. Each isolated bundle was lifted on the electrode above the fluid surface for recording, the nerve clamp then forming the indifferent electrode.

The receptor strand was exposed in the manner indicated above. Then it was clamped at both ends in micromanipulated forceps, proximally by the endoskeletal rod on which the strand originates and distally near the basipodite insertion of the strand. The receptor rod and distal insertion were severed, together with the few fine auxiliary connective tissue attachments and the hypodermal branches of the CB nerve. The whole organ was lifted from its position in the coxopodite and re-immersed in saline. The elastic strand could now be stretched and released by means of the micromanipulators. Usually the distal clamp was moved, since, being further from the origin of the sensory nerve, its movement caused less disturbance of the fluid around the recording lead and hence less fluctuation of the oscilloscope base-line. The excursion of the manipulator (2-5 mm. maximum) was monitored on the oscilloscope with the photo-transistor transducer already described. As with the rack-and-pinion device, fairly smooth movements could be achieved and the transducer gave a reasonably linear record of the movements.

As Whitear (1962) has described, the CB receptor organ lies dorsally in the coxopodite, its elastic strand running parallel to the leg axis from a small endoskeletal peg at the proximal end of the coxopodite, between the anterior and posterior levator tendons, to the proximal integumental rim of the basipodite. Unlike the more distal chordotonal organs, it is not attached to a muscle tendon. CB is thus stretched on extension of the coxobasal joint (i.e. on depression of the leg) and relaxed on flexion of this joint (limb elevation).

Recordings from the whole sensory nerve of CB in situ showed that this receptor responds to both directions of movement and also to position of the basipodite on the coxopodite, and this result was confirmed by isolated fibre responses. These are summarized in the left-hand sections of Tables 1 and 2 (the right-hand sections give the results of the stretch/release experiments considered later). Passive movement of the coxo-basal joint was effected by hand-probe in preparation I, and by the rack-andpinion in preparations II-VII.

Table 1 shows the similarity in numbers of extension and flexion fibres found in each preparation (movement and position fibres for the same direction being lumped together). The failure to find any flexion fibres in preparation III was no doubt due to injury to the majority of sensory neurons in the course of the experiment rather than to a real absence of such fibres. A similar explanation underlies the small number of responsive fibres isolated in each experiment, and the large difference in numbers of flexion and extension fibres found in preparation VI.

In Table 2 fibres responding primarily to joint movement are separated from those predominantly signalling joint position. A third category of ‘intermediate fibres’ is included here, to take account of a few fibres which could not readily be classified as primarily movement-sensitive or position-sensitive. (The numbers in brackets indicate units which for one reason or another were somewhat doubtful.) The greater number of movement fibres than position fibres for each direction does not necessarily reflect a real difference in numbers, but is probably largely due to the movement fibres being thicker and therefore more easily isolated for recording.

Thus all of the four main functional types of sensory fibre previously recognized in the PD organ (Wiersma & Boettiger, 1959), and in other similar decapod chordotonal receptors (Wiersma, 1959), are present in the CB organ. These are: (a) position fibres for (i) extended position, and (ii) flexed position; and (b) movement fibres for (iii) extension and (iv) flexion. Furthermore, from Tables 1 and 2 it appears that extension and flexion fibres are about equally common in CB, and this applies separately to the movement fibres and probably also to the position fibres. The operational characteristics of these sensory types in CB will now be considered.

These are characterized by their tonic responses to a fixed position of the joint, generally at or near one or other extreme position of its total arc. They are commonly fibres of small diameter, and are therefore more difficult to isolate than the larger movement fibres. Pl. 1A shows the response of a small bundle containing two small elevated (flexed) position fibres and one somewhat larger depressed (extended) position fibre. The two elevated position fibres are more or less ‘pure’ position fibres, being almost completely independent of movement and having a tonic discharge of almost constant frequency with little or no apparent adaptation. The larger fibre firing in the depressed position, on the other hand, is somewhat influenced by movement, at least at higher velocities, and shows distinct adaptation in the fixed position. In these respects, as well as in spike size (and hence probably axon diameter), this fibre is somewhat intermediate in its response, although still primarily a position fibre. A similar, not-quite-pure, position fibre is illustrated in Pl. 1B, which also shows the lower response frequency at positions slightly removed from the extreme position, a typical feature of position fibres. This property is represented in Text-fig. 1, showing the decreasing frequency of response towards the mid-position of two position fibres for opposite ends of the joint’s arc. The left-hand curve is for a ‘pure’ elevated position fibre, similar to, and from the same preparation as, the two shown in Pl. 1A ; the right-hand curve is for the depressed position fibre in Pl. 1B. Note also the difference in response frequencies between the two fibres for comparable positions. This is also evident in Pl. 1A and is a common difference between pure and less pure position fibres.

The ideal decapod crustacean movement fibre responds only during movement of the joint, in one direction, with a constant frequency of discharge for movement at a steady rate (Wiersma & Boettiger, 1959). In addition, however, many movement fibres have a low-frequency discharge in or near the extreme position at the end of the movement in the effective direction. Pl. 1C, D shows the responses of two CB movement fibres for opposite directions of movement of the coxo-basal joint. The smaller spike fibre fires on depression of the leg at this joint, the larger on elevation. The latter fibre also shows a maintained discharge at or near the fully elevated position. However, this tonic ‘position-sensitive’ discharge, and also that in the fully depressed position in Pl. 1E, are at a frequency of about 50 cyc./sec., which was the frequency of the a.c. mains supply to the recording apparatus. Possibly, therefore, these apparent position responses were in fact responses to slight ‘mains-feedback’ vibrations in the mechanical (or electrical) recording system. This would indicate very high movement-sensitivity. Indeed, when quiescent, fibres having such responses in or near an extreme position could often be induced to discharge at these or other positions at 50/sec. by a tap on the bench. Appropriate manipulation of the apparatus sometimes succeeded in eliminating or at least greatly reducing responses of this kind. The occasional impulses at other positions in Pl. 1D may also have been due to imperceptible vibrations of one kind or another.

A good movement fibre tends to fire at a preferred ‘saturation’ frequency at different supra-threshold rates of movement (Wiersma & Boettiger, 1959). Unfortunately none of the imposed movements in the present study was perfectly steady, so this property could not be properly tested. Nevertheless, each of the two fibres in Pl. 1 C, for example, shows some tendency to fire at a certain characteristic frequency during movement at different speeds. This is also evident in the depression fibre represented in Pl. 1 E, where even at the lower velocity groups of impulses approach the saturation frequency of this fibre. By contrast, there is no indication of a saturation frequency at comparable velocity in either of the two elevation fibres in Pl. 1 F, though the small spike fibre does show saturation at higher velocities.

The large spike elevation fibre in F represents a less common type of movement fibre. Its response, though also more or less independent of joint position, is strongly dependent upon the speed of movement, having a moderate threshold and a high saturation velocity. Except with fast movements such fibres commonly cease firing well before the end of full-arc movements, particularly if the movement slows down a little, and resume firing only if the movement speeds up again—e.g. the last burst of (six) large spikes during the slow second elevation movement illustrated. (The other intermittent spikes at this slow speed probably also resulted from slight irregularities in the imposed movement.) When recording whole-nerve activity the large spikes of fibres of this kind may often be seen contributing towards a relatively dense (i.e. high-frequency) discharge at the beginning of a movement.

When the average response frequency for each movement is plotted against the average rate of each movement curves of the form seen in Text-fig. 2 are obtained. The ideal pure movement fibre of this type of receptor would have a curve similar to a rectangular hyperbola, with a steep initial slope flattening off rapidly. Of the three fibres represented in Text-fig. 2, namely the depression fibre of Pl. 1E, and the two elevation fibres of F, the small elevation fibre evidently approaches closest to this ideal form, while the depression fibre is not far removed. The large elevation fibre, however, is well removed from this ideal. The curves for such fibres approach straight lines of steep slope, with higher threshold (and saturation?) velocities than normal movement fibres, and usually no spontaneous resting or position discharge. These fibres should probably be regarded as phasic movement fibres, or acceleration fibres.

A very different and uncommon fibre is illustrated in Pl. 1 G, H. This appears to be intermediate between a slow movement fibre and a position fibre, but with the unusual feature of being most sensitive not in either extreme position, but in and around a position of about 75 % flexed. As this and one or two other somewhat similar fibres were found at a fairly late stage of the preparation, these particular responses might have been aberrant, a result perhaps of ‘fatigue’ of the sensory endings. However, the occurrence still later in this experiment of at least one typical positionfibre response (Pl. 1B), indicates that the preparation as a whole had not begun to deteriorate. In any event the present response recalls the phenomenon of ‘range fractionation’ amongst movement and position fibres of the myochordotonal organ in the meropodite of Cancer (Cohen, 1963), and in limb proprioceptors of Limulus (Barber, 1960); the response of the mid-sector position fibres of the latter receptors would appear to be quite similar to the present response from CB.

Since extension of the coxo-basal joint (i.e. limb depression) stretches the CB receptor strand and flexion relaxes it, the same afferent fibres could be expected to respond to extension of the joint and stretch of the strand, and others to flexion and relaxation. This was established in two early experiments, in which the response of a fibre bundle containing one or two movement fibres for each direction was recorded, first during hand-produced movements of the joint, then during stretch and release of the intact receptor with hand-held forceps. Oscillograms from three further experiments, in which the transducer system was used to monitor the micromanipulator-produced displacements of the distal end of the excised CB receptor, confirmed the presence of individual unidirectional sensory fibres, some responding to the dynamic or to the static phase of applied stretch, and others to release. The numbers of fibres isolated in these three preparations are summarized in the right-hand sections of Tables 1 and 2.

As before, position fibres were more difficult to obtain in isolation than movement fibres. A bundle comprising about 20 % of the thickness of one CB nerve contained many position fibres but no movement fibres. In this case the marked difference in spike heights of the stretched and the relaxed position fibres indicated the presence of distinct fibres for the two extreme positions. A similar overall difference is to some extent evident in Pl. 2E, showing the response of a complete CB nerve.

Another portion of the former CB nerve contained several primarily movement-sensitive fibres (Pl. 2 A). The very small spikes in these records represent two distinct pure movement fibres, one for each direction (this is more readily ascertainable in the original oscillograms). They show no tonic position discharge and no effect of position on their movement responses. In contrast, the two large-spike movement fibres responding to stretch both show a pronounced influence of ‘position’, or length of the receptor strand, in their responses. This is evident both in their tendency, during movement, to fire at increasing frequency with increasing length of the strand, and in their sustained discharge in the fully stretched positions. However, the last of the three such positions illustrated represents about 140 % of the physiological maximum length, while in the other two records this position discharge-frequency was quite low in both fibres. At the receptor length corresponding to that in the fully extended joint (i.e. 100% length), this tonic discharge approached zero, particularly in the larger fibre. Furthermore, faster stretching than here shown elicited fairly strong responses in both these fibres from the beginning of the movement.

The two influences of receptor strand length on the movement responses of each of the two preceding large-spike stretch fibres, were exhibited separately by two movement fibres responding to release from another preparation (Pl. 2B). The larger of these responded to movement increasingly strongly towards the extreme relaxed position, but displayed no sustained discharge in this position. The smaller fibre, on the other hand, while responding at constant frequency throughout the complete release movement, even at low velocity, showed a strong sustained discharge in the fully relaxed position. Again, however, the high frequency of the latter activity, in this case about equal to the frequency during movement (and possibly almost saturation frequency for this fibre), resulted from the receptor being relaxed well beyond its normal limit, at which limit the frequency was distinctly lower.

An adjoining bundle in this preparation yielded a high-threshold movement fibre responding to stretch, which was unusual in showing a very high-frequency, and often rhythmic, discharge of 150-200 impulses/sec. when the CB strand was stretched well beyond its normal limit (Pl. 2C). Since this was the only fibre found with such a response this should probably be regarded as an aberration due to overstretch rather than as a true position response. In any event, the high frequency seen here may well represent this fibre’s saturation frequency, of a high value appropriate to an acceleration fibre rather than an ordinary movement fibre.

The overall response of the isolated CB organ to stretch-release, recorded from its whole sensory nerve, is shown in Pl. 2D, E. (Similar oscillograph records were obtained in the previous experiments with CB in situ, in response to joint movement.) Movement fibres responding both to stretch and to relaxation, and also stretched and relaxed position fibres, are evident in good numbers (E). The tonic, static positionfibre discharges are seen to be considerably weaker in the intermediate positions than at either end. This applies to each of the individual fibres active on the intermediate steps (D), as well as to the over-all response. However, the apparently greater density of impulses in the stretched rather than in the relaxed positions does not necessarily reflect a corresponding difference in the numbers of active fibres, but may be due largely to fortuitously differing average spike heights—particularly since the degree and direction of this recorded effect varied from one preparation to another. By contrast with the position responses, the stepwise movement responses of whole CB nerves showed relatively little variation with receptor length. Nevertheless, some increase in density of step-movement responses towards one or both ends of the total excursion is not uncommon in these records, as in the series of stretch steps in Pl. 2D. This is probably a combined effect of an influence of position on the activity of movement fibres, and of recruitment of relatively large-spike position fibres.

Records such as Pl. 2D, E of the response of the whole receptor nerve, made early in a preparation before any deterioration could set in and before splitting the nerve into small fibre bundles thereby inevitably damaging many fibres, give a rough indication of the number of responding fibres in the sensory nerve. In some of these records it is possible to make out at least six or seven different spike heights for each of the four main types of fibre, while in a few cases up to ten to twelve different repetitive heights are discernible for movement fibres in one or each direction. These estimates would give a total of, say, thirty to fifty different responding fibres. Whitear (1962) figures seventy-six cells in a drawing of a methylene-blue preparation of CB, which together with others not stained and/or more distally located, might give a total of, say, 90 (80-100)—that is, two or three times as many cells as fibres on the above estimates. Possible implications of this are considered below.

Whitear (1962) suggested that the structural differentiation between ciliary and paraciliary cells which she has described in the chordotonal organs of Carcinus legs, might directly underlie the differential unidirectional sensitivity of both movement and position fibres, the ciliary cell being responsive to relaxation (i.e. shortening) of the receptor strand, and the paraciliary cell to stretch (lengthening). As noted in a subsequent paper (Bush, 1964), this hypothesis would seem to fit PD, CP₂ and MC₂, and possibly also, with some modifications, CP₁ and MC₁. However CB, despite its isodynal scolopidia, has now been shown to possess fibres responding to lengthening as well as others responding to shortening. Further-more, the invariable pairing of the sensory cells in these chordotonal organs, and the presence of the ephapse between the two distal processes of a pair, suggests that some form of functional interaction between the two cells may occur at this area of contact (Whitear, 1962).

In order, then, to fit Whitear’s hypothesis to the experimental evidence now available, three postulates might be made, (i) The distal process of the paraciliary cell is activated by a high tube-scolopale tension or by an increase in this tension, such as presumably occurs with the receptor strand in a stretched position or during stretching movement, (ii) The ciliary cell, at least at low or decreasing tension, has an inherent spontaneity of discharge, which is suppressed by high or increasing tension ; in isodynal scolopidia one of the two ciliary cells is more readily suppressed than the other. (Conceivably the ciliary segment might actually beat, rather like a motile cilium, so generating its own discharge I) (iii) Interaction occurs at the ephapse in all scolopidia, such that an active cell tends to inhibit its partner; in isodynal scolopidia the ephapse is polarized in such a way that only the cell which is more readily suppressed during stretch can inhibit its partner. These hypothetical relationships might be summarized thus:

As before, PD, CP2 and MC2, and probably also CPi and MCi, fit into this scheme with relative ease. CB remains the major difficulty; here, its directional sensitivity would depend upon the polarized ephapse, and the purely quantitative difference in sensitivity of the two distal processes of a pair. A further possibility is that while some scolopidia, firmly embedded in the surrounding connective tissue matrix, might respond only to relaxation, other, more loosely embedded scolopidia, owing to repeated slipping between tube and matrix during stretching of the strand, might thereby be sensitive to overall stretching, being inactivated during relaxation by ephaptic inhibition. However, this leads to additional complications in the theory which are probably unwarranted at this stage.

In the preceding hypothesis the adequate stimulus for these receptors was assumed to be an increase or decrease in the longitudinal tension on the scolopidia. Instead, bending of the scolopidium might be the adequate stimulus. This is the basis of Mendelson’s hypothesis (1963). He postulates that, as a result of differential displacement of longitudinal elastic elements during stretching or relaxation of the receptor strand, the distal tip of each scolopidium is displaced at right angles to its long axis, so bending the scolopidium; then, if the two distal processes lie in the plane of curvature, one will be stretched during shortening of the receptor strand and the other during elongation. However, this hypothesis depends upon the assumption that the scolopidia lie roughly at right angles to the long axis of the receptor strand,whereas in fact they appear to lie parallel to it (Whitear, 1962).

Given the latter observation the hypothesis might be modified along the lines indicated in Text-fig. 3. During stretch of the receptor strand the attachment point A_l would move distally, bending the tube to the left. Conversely, during relaxation the tube would move to the right—that is, assuming the elastic fibre, E, still to be under tension. This hypothesis would, of course, depend upon the presence of elastic fibres with appropriate attachments to the scolopidia. Elastic fibres have not yet been demonstrated histologically, though they are presumably present in some form to permit the elasticity of the receptor strand as a whole. Attachment plaques, desmosomes, and ‘dark bodies’ are present in various regions of the scolopidia, including the outer membranes of the scolopale cell and its distal prolongations around the tube (Whitear, 1962). The degree of bending would depend upon the precise morphology and localization of the elastic fibre attachments and of the whole receptor strand or sheet, as well as on the dynamic properties of the system. If the structure were such as to permit roughly equal bending in either direction, then isodynal scolopidia as in CB would suffice. If, however, greater bending occurred in one direction, then a paraciliary type of distal process, as in PD, might be necessary to provide adequate sensitivity to the opposite direction. An enlarged paraciliary cell might be necessary in receptors having a sheet-like form like CP₁ and MC₁, while the much reduced paraciliary cell of CP ₂ and MC ₂ would correspond with their apparent lack of response to one direction (Bush, 1964).

Of the two modified hypotheses presented here the latter is perhaps the simpler and accordingly more attractive, relying as it does on fewer basic assumptions. Which, if either, proves the more tenable must await further experimental evidence, in particular from intracellular electrodes.

In addition to the problem of unidirectional movement sensitivity, discussed above, there is the question of the differentiation between movement and position fibres. This evidently does not depend upon any obvious finestructural differentiation, since Whitear (1962) finds no constant differences of form between the scolopidia in different regions of a given organ. Wiersma & Boettiger (1959) traced their movement fibres to the larger, proximal cell bodies of PD, and the position fibres to smaller, more distal cells ; and Whitear (1960, 1962) showed the distal scolopidia to be somewhat smaller than the proximal ones. Assuming a perpendicular orientation of the scolopidia Mendelson (1963) suggested that the tubes of movement-sensitive scolopidia may be more loosely coupled to the surrounding matrix than those of the position-sensitive ones, so that they slip, ratchetwise, during change of length of the receptor strand; a continuing length change would then be necessary to produce a continuous sensory discharge. Even assuming a correct, parallel scolopidial orientation, however, a similar mechanism would still seem feasible. Moreover, the same effect could presumably occur in reverse, thus permitting differentiation of relaxation endings into position-sensitive and movement-sensitive ones too. The CB flexion fibre illustrated in Pl. 1 G, H is suggestive of a hypothetical transitional form between position and movement fibres, which would fit in with the foregoing suggestions on the differentiation of these two primary types.

Two experimental observations suggest that different scolopidia may be involved in the responses to movement in opposite directions. First, Wiersma & Boettiger (1959) noted that flexion movement fibres from PD appeared to come from one side of the organ, while extension fibres apparently came from the opposite side. Secondly, the rough estimate from the oscillograph records of one CB preparation in the present study suggested that there may be only half the number of responding fibres as cells (from Whitear’s estimate of the latter from a methylene-blue preparation). These observations, if substantiated, would suggest that only one of the two cells per scolopidium has an actively conducting axon. In heterodynal scolopidia, either the ciliary or the paraciliary cell, depending upon whether it was a relaxation-sensitive or a stretch-sensitive scolopidium, might conduct; or the ciliary cell might always be the active one. In either case the second cell could presumably still have an essential, moderating influence, possibly in a purely inhibitory capacity.

Should this suggestion, that only one sensory cell of a scolopidial pair transmits information to the C.N.S., prove correct, both preceding hypotheses to account for the differential direction sensitivity among movement fibres would require modification, but not necessarily abandonment. The hypothesis suggested above to account for differentiation into movement and position fibres, since it depends upon different scolopidia anyway, would not necessarily need alteration.

Whitear (1962) suggests that the function of the CB organ may be to provide information about the angle of the coxo-basal joint so that the levator muscles do not contract beyond a safe point, save when the autotomy reflex overrides the information from CB. However, the presence in CB of movement and position fibres for both elevation and depression of the leg indicates that CB is not concerned solely with elevation of the limb, still less with elevated position only, which is probably all that would be required for the function suggested. The similarity of the afferent responses of CB with those of PD and the other chordotonal organs of the crab’s leg, together with the fact that CB is the only receptor monitoring activity of this joint alone, suggests that it is probably involved in reflex regulation of the movements of this joint, in a manner similar to PD for the propo-dactylopidite joint (Bush, 1962, 1963). Experimental evidence for this view is presented in a subsequent paper (Bush, 1965). This is not to say that CB may not also subserve the function suggested by Whitear.

I would like to thank Dr Mary Whitear for persuading me to undertake this study of CB. The work was carried out during tenure of a Research Fellowship at Jesus College, and was aided by U.S. PHS Research Grant B-3994 for equipment, both of which I gratefully acknowledge.

PLATE I

Responses of afferent fibres isolated from the CB receptor nerve to stationary position and passive movement of the coxo-basal joint. Lower beams represent the imposed positions and movements : joint extension (i.e. limb depression) upwards; flexion (= elevation) downwards. Large spikes in A and F are re-pointed ; small spikes in F are indicated by bars and dots above record.

PLATE 2

CB afferent responses to stretch and release of the receptor strand : (A-C) isolated fibre bundle responses ; (D, E) response of a whole CB nerve. Lower beams : stretch upwards, release downwards (vice versa in A). Large spikes in D and E re-pointed ; small spikes in A indicated by bars.