Joint Receptors in the Antennule of Panulirus Argus Latreille

The major proprioceptors at the distal joints of the walking legs of Crustacea are chordotonal organs (Whitear, 1962; Pringle, 1961; Cohen & Dijkgraaf, 1961). These take the form of elastic strands or elastic sheets which are distorted by joint movement, and in which dendritic processes of associated sensory neurons terminate in scolopales. Electrophysiological observations (Burke, 1954; Wiersma & Boettiger, 1959; Wiersma, 1959; Mendelson, 1963) have demonstrated both tonic position receptors and phasic movement receptors in the organs. These receptors are normally unidirectional, responding to either flexion or extension of the joint but not both. The organs, therefore, are capable of conveying information about direction and rate of movement, and about joint position.

The present work shows that similar receptor organs occur at the joints of the antennule in the spiny lobster. One joint is unusual in that it has two axes of movement. It is of particular interest that a single elastic strand organ is used to detect both direction and axis of movement around this joint.

Adult Panulirus argus Latreille were obtained from commercial sources on Bermuda and maintained in running sea water in large tanks at the Bermuda Biological Station. Antennules were removed by cutting across the first segment just distal to the statocyst. For morphological work antennular segments were cut open and stained with methylene blue (see Whitear, 1962). Fresh preparations were drawn or photographed, and/or the proprioceptive organ was dissected out, fixed in ammonium molybdate and sucrose, and mounted in Permount. Receptor systems at each of the three distal joints were prepared. The basal joint was not examined and will not be considered in this paper.

Measurements of cell locations at varying joint positions were made on fresh, stained preparations.

For electrical recording antennules were removed and the segment distal to the joint under study was cut across. Tubing (Intramedic, polyethylene, Clay-Adams, Inc., New York, or rubber tubing) was slipped over the cut end of the distal segment, and the antennule preparation was perfused with a modification of Homarus fluid (Cole, 1941). Such perfusion extended the life of the preparation. The antennular nerve was exposed at the proximal end of the preparation, often in the segment just proximal to the joint studied. Fine bundles were split off from the gross nerve, or from natural bundles adjacent to the organ (see Wiersma, 1959). Nerve impulses were led from these with conventional silver or platinum wire electrodes and after amplification (Tektronix 122 pre-amplifier) were displayed on an oscilloscope (Tektronix 502) and photographed (Grass Oscilloscope Camera C4). Preparations of single units or of a few units were usually obtained.

Movement about an antennular joint was monitored on the lower beam of the oscilloscope by means of a photocell system (see Cohen, 1963) in which a shutter attached to the moving antennular segment was interposed or removed from between a photocell and an illumination source. The area of the photocell shadowed, and consequently the potential displayed, varied monotonically but non-linearly with angular joint movement. The monitor trace in the following records therefore gives the onset and duration of the stimulus, but does not provide reliable quantitative information in terms of angular velocity or position. In several cases position was estimated by sketching the preparation and subsequently measuring angular positions on the sketch. This gave approximate angular values to the nearest 5°. Although crude, such monitoring methods were adequate for the purposes of the present paper.

A Prior manipulator was normally used to produce smooth, relatively slow, joint movements. It was attached to the antennule by thread or wooden dowelling. For more rapid movements the thread was pulled by hand.

As in other decapods the antennule consists of three basal segments and two long distal flagella (Text-fig. 1). Active movement occurs at the joint between the third or distal segment and the outer flagellum, third joint. The inner flagellum is immobile. Movement also occurs between the second segment and third segment, second joint’, between the first segment and the second segment, first joint ; and between the carapace and first segment, basal joint. Only the three peripheral joints are considered in this paper. Sandeman (1963) recently described structures at the first and second joints in Squilla mantis.

A single proprioceptor organ occurs at each of the three peripheral joints (Text-fig. 1). In general form these resemble the crab leg-joint receptors described by Burke (1954) and Whitear (1962). A compact elastic strand runs across the joint—or between joint and apodeme—in such a way that it is stretched or otherwise distorted by joint movement. The cell bodies of bipolar sensory neurons lie along the strand in looser connective tissue. The distal, dendritic processes of the sensory neurons appear to enter the strand proper where they presumably terminate in scolopales (Whitear, 1962). The proximal axons enter the antennular nerve in bundles and run toward the brain.

Although there are intermediate sizes, the cell bodies often appear to fall into two classes—one of 50−70 μ diameter and one of 10−30 μ diameter. The detailed structure and distribution of the cell bodies in the organs will be discussed for each joint in turn.

The outer flagellum moves about the third joint through an angle of 30−35° (Text-fig. 1). It is depressed by a single muscle, M. reductor 4, which inserts on a single apodeme arising from the nearer ventro-lateral edge of the proximal flagellar annulus. The flagellum pivots about the dorsal, hinge-like diagonal crease of flexible connective tissue joining it with the distal segment. Elevation of the outer flagellum is entirely passive, brought about by a ventral elastic cushion at the joint. The elevated position is therefore the ‘rest’ position.

The elastic strand of the receptor organ at the third joint originates on the flagellar apodeme. It inserts ventrally on the distal segment just proximal to the joint, and between the bases of the two flagella. In addition several weaker elastic filaments— accessory strands—split off near the distal end of the organ and apparently attach to muscle or connective tissue near the origin of the apodeme (Text-fig. 2). Tensions exerted on the strand by joint movement are therefore complex (see Text-fig. 3). Although most of the main strand is generally stretched by flagellar depression or muscular contraction and shortened by elevation or muscular relaxation, the effects on the secondary filaments are reversed, and these in turn react on neighbouring regions of the main strand.

The position of cell bodies and sensory terminals is somewhat variable. Generally three groups can be identified: (1) a proximal group of large and small cells located about half-way along the strand, (2) another mixed group located near the origin of the weaker accessory filaments, and (3) a final group, usually of small cells spread linearly along the main strand between the accessory strands and the terminal insertion. In some preparations, as shown in Text-fig. 2, these cell groups may be linearly arranged, while in others, even in the opposite antennule of the same animal, the groups overlap more extensively. In all, however, depression of the flagellum appears to cause stretch of the dendrites of the first and third groups, but relaxation of those of the second group, while flagellar elevation causes relaxation of the first group, greater relaxation of the third group (see Text-fig. 3), and stretch of the second group. The exact details of these changes in the normal animal are difficult to obtain because even the limited dissection used to expose the elastic strand in Text-fig. 2 undoubtedly removes some accessory connexions or constraining tissue and thus alters the distribution of tension. Moreover, the exact effects of tension changes on the dendritic receptor structures were not observed. Nevertheless, as a qualitative description, the changes noted in Text-fig. 3 are presumably valid.

At least sixty cell bodies have been counted in methylene blue preparations of the receptor organ in the third joint.

The second joint is more conventional. The distal segment moves dorso-ventrally about two lateral pivots (30° up and about 65•70° down). There are two apodemes and two muscles, an elevator and a depressor, so movement in both directions is ‘active’. ‘Rest* position is considered to be ‘straight ahead’.

The main elastic strand originates on the ventral apodeme and neighbouring muscle and inserts dorso-medially just distal to the joint (Text-fig. 4). An accessory branch containing the distalmost group of cell bodies splits off the main strand and attaches separately just ventro-medial to the main strand insertion. Depression of the third segment stretches the main strand ; elevation shortens the strand. Undoubtedly more complex effects occur in the accessory strand, but these were not examined in detail.

Cell bodies are usually arranged linearly along the main and accessory strands. The more distal elements are generally small cells, the more proximal are mixed large and small neurons.

Seventy to eighty cell bodies have been counted in methylene blue preparations of the second joint organ.

The first joint has a single dorso-medial pivot but two apodemes, ventro-medial and ventro-lateral (Text-fig. 1). There are folds of elastic connective tissue at the joint which partially restrict movement so that contraction of the muscle attached to the ventro-medial apodeme, M. reductor 2, causes ventral movement of the second segment (up to about 40°) while contraction of the muscle inserting on the ventro-lateral apodeme, M. productor 2, causes primarily lateral movement (up to 65−70°)—see Textfigs. 1 and 6. In both cases extension is apparently passive and is probably mediated largely by the elastic tissue of the joint. At rest the second segment is in the extended position. It is worth emphasizing that the first joint differs from the second and third in that movement occurs in two primary axes rather than in one.

The single primary elastic strand of the first joint originates on the dorso-medial wall of the first segment among the fibres of the depressor muscle. It inserts on a boss on the ventro-lateral margin of the second segment about two-thirds of the way between the ventro-medial and ventro-lateral apodeme origins (Text-figs. 5 and 6). An accessory strand with numerous branching elastic elements extends medially from the main strand and terminates ventrally in diffuse attachments at a locus between the primary strand and the ventro-medial apodeme. Both depression and lateral flexion of the second segment cause shortening of the major strand. In this respect it differs somewhat from the second and third joint organs where active muscular flexion causes extension of the major elastic strand. A more detailed discussion of movements of the strand at this joint will be deferred until its physiology has been described.

The cell bodies of the first joint organ occur in roughly three groups: (1) there is a group of large and small cells at the point the more proximal of the two proprioceptor nerve bundles leaves the strand (see Text-fig. 5), (2) there are large and small cells arranged linearly along the strand up to the point of branching of the accessory strand, and (3) there are small cell bodies along the accessory strand. The axons of the more distal sensory neurons leave the elastic strand in a second nerve bundle slightly anterior to the one indicated above.

Between sixty-five and seventy sensory neurons have been counted in a single methylene blue preparation of the first joint organ. Of these, eighteen were large cells in the first or second group, and twenty-two were small cells in the accessory strand.

Many of the large and small cell bodies of the first joint organ appeared to be paired (Plate 1). This was also occasionally suggested in the other organs, but was most obvious in the second cell group of the first joint organ. The large cell body is normally the more proximal of the two, its distal process passing close to the small cell body and continuing on into the elastic strand in close proximity to the distal process of the small cell. Presumably the two processes terminate together in a single scolopale (see Whitear, 1962). Not all small cells, however, were associated with the large cells, and in those cases, as described by Whitear (1962) in the crab leg, cells of similar size often seem to be paired together.

Proprioceptive responses were readily obtained from all three joint organs. Both movement receptors and position receptors were found at each joint, with movement receptors predominating. As commonly found elsewhere (Burke, 1954; Cohen, 1963 ; Mendelson, 1963; Wiersma, 1959; Wiersma & Boettiger, 1959) each movement receptor responded to movement in one direction only, e.g. some fibres at the second joint responded to ventral but not dorsal movement, while others responded only to dorsal movement. Following the practice of Wiersma & Boettiger (1959), such receptors are termed unidirectional. Response frequency of the movement receptors generally increased with the rate, but not with the extent of movement. The response frequency also seemed relatively independent of initial or terminal angular position (Text-fig. 7). In some multi-unit preparations activity in several identifiable fibres appeared in repeatable successions as the joint moved over its entire range, suggesting some ‘range fractionation’ among the movement fibres (see Cohen, 1963). These, however, were not common, and the range within which movement was an effective stimulus was often large relative to the total range of joint movement. If the absence of such position specificity occurs in the normal animal, and is not the result of changes in sensitivity during preparation for recording, then these movement receptors must convey relatively imprecise information about joint position.

Tonic position receptors were found less frequently and tended to have smaller action potentials, as was also found in the crab leg (Wiersma & Boettiger, 1959). Position receptors were most active at one or the other extreme position of the joint. Typically they did not respond over the entire range of joint movement but, beginning with maximal activity at one extreme, progressively declined in frequency as the joint angle decreased. At the mid-position most were silent or only slightly active, and over the remaining half of the range were completely inactive (Text-fig. 8).

Some fibres were intermediate between movement and position receptors (Textfig. 9). A few movement fibres had relatively long time courses for adaptation, and many position fibres were not completely insensitive to movement per se. Firing frequencies in the latter class temporarily increased with movement toward the most active position and decreased with movement away from it.

The properties outlined are very like those described for joint-receptor organs in the legs of crabs (Burke, 1954; Cohen, 1963; Mendelson, 1963; Wiersma, 1959; Wiersma & Boettiger, 1959). Taken together, the combined properties of individual sensory neurons can provide unambiguous information about position, rate, and direction of movement at either of the two distal antennular joints in which there is one axis of movement. At the first joint, however, there are two axes of movement, and the mechanisms of discrimination are less clear. This is emphasized by the observation that the gross movement of the single major elastic strand—shortening—is similar for both ventral and lateral flexion. Particular attention was therefore given to the problem of axial discrimination at the first joint.

Most of the receptor elements at the first joint gave similar responses to both axes of movement. Movement receptors usually responded either to both ventral and lateral flexion, or to both dorsal and medial extension. Many position receptors responded to both lateral and ventral flexed positions, at least one responded to the extended ‘rest’ position. Examples of such responses are shown in Text-fig. 10. It is estimated that 80−90% of the fibres recorded from the first joint organ showed this lack of axis specificity. No fibres were found to respond to both flexion and extension in only one axis. The apparent ambiguity of responses from the organ is therefore not solved by specific ‘axis receptors’.

Some movement receptors, however, were completely specific, responding to only one of the four possible directions of movement. Some of the position and intermediate receptors also responded to flexion in one axis but not the other. Examples of such completely specific fibres are also shown in Text-fig. 10.

Table 1 gives a summary of the fibres found at the first joint. Most of the movement receptors were non-specific and responded to both axes of flexion, or to both axes of extension. Some movement receptors responded specifically to ventral flexion, while others responded specifically to medial extension. Movement receptors specific to lateral flexion and to dorsal extension have not been found, but in view of the limited sample of fibres examined this observation is of dubious significance. Non-axis-specific position receptors were found that responded to both flexed positions; one responded at the extended position. Axis-specific position receptors responded to ventral flexed position only, while others responded to lateral flexed position only. Although only a few of the movement fibres were completely specific, nearly half of the position fibres differentiated between the two axes.

There is no doubt that these differentially sensitive fibres were associated with the organ described. Most of the specific responses were recorded from nerve bundles immediately leaving the organ, and in other cases the bundles of the antennular nerve were subsequently traced back to the organ.

The mechanisms of axis-specific responses are not known with certainty. Detailed observations of relative movements of the organ of the first joint suggest, however, that differential gross mechanical distortion of the receptor elements may be involved.

The major elastic strand of the first joint shortens by about 10% of its total rest length for a 40−45° flexion in either ventral or lateral directions. This alone clearly provides no mechanical basis for axis discrimination. At the distal end of the first joint organ, however, the accessory strands contain a group of small cell bodies and fan out to an attachment near the ventro-lateral apodeme (Text-figs. 5 and 11). During ventral movement, the origin of the ventro-lateral apodeme moves only very slightly. Conversely, during forced lateral movement, the origin of the ventro-medial apodeme is nearly stationary. Thus during lateral flexion the accessory branches terminating near the ventro-lateral apodeme stretch as the major elastic strand shortens (Text-fig. 11). During ventral flexion, the accessory branches shorten relative to the major strand.

Text-fig. 12 a shows the configuration and linear distance between sensory cell bodies located in the accessory branch at rest, at ventral flexion, and at lateral flexion. There is some stretching at the extreme terminus of the accessory branches, and there is an unquestionable differential distortion of the accessory branches with the two axes of flexion. Positions of identifiable sensory neuron cell bodies within the accessory strand in relation to points on the major strand are also plotted in Text-fig. 12. There is differential movement of the accessory and major strands at and near their point of junction. Either differential stretch of elements in the accessory strands or bidirectional shearing stresses of processes extending between the accessory and major strands would presumably be sufficient to produce the axis-specific responses observed.

In major characteristics the proprioceptors at each of the three joints of the antennule appear essentially identical. They are also like the elastic strand organs of decapod pareiopods as described by Burke (1954), Wiersma & Boettiger (1959), and Whitear (1962). The differences in detail only serve to emphasize the basic structural similarities. For example, both insertion and origin of the major elastic strand may vary—in joints two and three it originates on an apodeme, in joint one, on the wall of the first segment ; in joints one and two it inserts on the distal segment of the joint, in joint three, on the ventral wall of the proximal (third) segment—and the strand may (joints one and two) or may not (joint three) cross the joint. In all three cases, however, the major strand stretches on movement of the joint in one direction—normally flexion—and shortens on movement in the other. Another invariant property, very apparent in the antennular organs but less emphasized in descriptions of pareiopod organs, is the distal accessory strand or strands associated with a group of cell bodies. These accessory strands are apparently responsible for differential stresses during joint movement, and their invariance lends support to the argument that such stresses play a significant role in normal proprioceptor function. A third invariant, already recognized in the parieopod organs (Wiersma & Boettiger, 1959; Whitear, 1962) is the proximal location of large and the distal location of small neurons in any one organ. The tendency of large and small neurons to pair where possible, however, does not seem to have been recognized, and may well be a specific property of antennular organs. It is of some interest that only one chordotonal or proprioceptor organ occurred at each antennular joint; in some pareiopod joints there were at least two organs (Whitear, 1962).

One of the unexpected findings of this study was that a single elastic strand organ mediated proprioception at the biaxial first joint. It apparently does this not by adding a class of specific axis-sensitive receptors, but by utilizing combined axis-position or axis-movement receptors. There are, therefore, at least twelve potential classes of receptors at the first joint. In both general classes of position and movement receptors there may be elements sensitive to: (1) lateral and ventral flexion, (2) medial and dorsal extension, (3) lateral flexion only, (4) ventral flexion only, (5) medial extension only, and (6) dorsal extension only. Members of only eight of the twelve possible classes have been identified (see Table 1), but they are so distributed that both movement and position receptor classes contain elements specific for each of the two axes. There is little doubt, therefore, that the total input from the first joint organ is capable of giving unambiguous information about the position of the second segment. Although there are undoubtedly differences in the relative number of each receptor class in the organ, our exploration of receptor distribution was too cursory to permit speculation about the significance of distributions given in Table 1. It is worth noting, however, that if the estimate of seventy cells in the first joint organ is reasonably accurate, each receptor class must contain relatively few elements, and the range of parallel or redundant input channels is limited.

The mechanisms of biaxial discrimination seem to be closely associated with the mechanisms of the unidirectional response. Unfortunately, there is no clear agreement about the nature of sensory transduction in the elastic strand organs. The early observations of Wiersma & Boettiger (1959) suggested that stretch of the sensory neuron dendrite might be the effective stimulus, and this has been persuasively argued by Cohen (1963) in an analysis of movements in a myochordotonal organ. If this view is accepted, then the presence of unidirectional flexion and extension receptors in one organ must imply that the gross anatomical arrangement of the receptor processes and elastic ligaments is such that some cells are stretched during flexion while others are stretched during extension. This indeed was shown by Wiersma & Boettiger (1959) and by our own observations described above.

Whitear (1962) and Mendelson (1963), however, have suggested alternative bases for unidirectional sensitivity which require that the nature of the terminal sensory process or its position within the scolopale determine whether the receptor responds to stretch or relaxation of the elastic strand. Whitear’s argument is largely anatomical and is derived from the observation of heterodynal scolopales within the elastic organs. As she indicates in a footnote, however, physiological observations by B. Bush demonstrate that organs containing only isodynal scolopales also have undirectional receptors for both directions, so the anatomical argument is currently unconvincing. Mendelson’s argument is largely physiological, and relies on the claim that bi-directional discrimination remains after cutting accessory strands in the PD organ of the crab leg, and that completely isolated organs continue to give unidirectional responses on both stretch and relaxation. At first glance this would suggest that the grosser movements observed by Wiersma & Boettiger (1959), Cohen (1963), and ourselves have little direct relation to unidirectional discrimination. As Cohen observed, however, removal of accessory attachments may grossly alter the response characteristics of a given cell, even shifting a relaxation-sensitive element to a stretch-sensitive element. Mendelson does not give data for response properties of single elements before and after isolation, nor does he describe the effects of stretch on cell and dendrite position in the isolated strand. In the absence of such controls, one cannot exclude the possibility of local stretch during gross shortening of the isolated organ, and consequently cannot conclude that accessory strands and differential stretch are unimportant for normal function of the elastic strand organs. The argument for intra-scolopale direction discrimination consequently loses much of its force. We return to the hypothesis that the effective stimulus of the receptor neurons is stretch of some portion of the dendrites (see Cohen, 1963) and that differential responses to direction of joint movement and to axis of joint movement result from specific arrangements of elastic and accessory strands that produce differential stretch of the dendrites. This not only is consistent with our observations of changes in cellular position at the first and third joint, but providess a rational explanation for the invariant occurrence of accessory strands and associated cell groups.

Contribution number 362 from the Bermuda Biological Station. Supported in by U.S.P.H.S. grant NB-03271.

Paired cell bodies in the elastic strand organ of the first joint. Each of the two large cell bodies is paired with a more distal small cell. The distal process of the large cell runs past the small cell body and continues close beside the distal process of the small cell. Methylene blue stain, eS, elastic strand ; I ax, axon of large cell; LC, large cell body; Id, distal process of large cell; t ax, axon of small cell; SC, small cell body; td, distal process of small cell.