Homologies Between Neuromuscular Systems Serving Different Functions in Two Decapods of Different Families

Although swimming is performed in different ways by the sand crabs Emerita (family Hippidae) and Blepharipoda (family Albuneidae), the component movements have been shown to have similarities of form and phase that suggest an evolution from tailflipping in a common macruran ancestor (Paul, 1981). If the two types of swimming are homologous, then there should be homologies between the muscles and motoneurones of the tailfans in the two species, and this is examined in the current paper.

Comparison of the tailfan neuromusculature in these sand crabs is complicated because the form of the uropod’s movements as well as the gross anatomies of the tailfans are quite different; and, furthermore, Emerita has apparently evolved two movements that are not present in Blepharipoda’s, repertoire: the return stroke and power stroke of the uropod (Paul, 1981). However, comparison is also made easier by the relatively small number of neurones (five) that innervate the principal muscles responsible for these movements, the return-stroke and power-stroke muscles (Paul, 1971a). The structural similarity found in this study between homologous motoneurones performing different functions in the two species suggests that morphological changes in nervous systems may be quite conservative even during diversification of neuromusculature to serve different behaviours.

Sand crabs were collected and maintained as described elsewhere (Paul, 1981). The articulations of the uropods were examined after the following procedure, recommended by Robert Hessler. Whole abdomens were immersed in 5% KOH and left at 55° C for 5 days, by which time all tissue was dissolved except the exoskeleton, arthrodial membranes, and partially calcified tendons. The abdomens were then rinsed in acid-alcohol, stained in Solution 2 of Mallory’s triple stain (3– 4 min), transferred to 80% ethanol until appropriately destained, dehydrated in 100% ethanol, and cleared and stored in terpineol. Arthrodial membranes remain flexible and are completely transparent after this treatment, so that one can view joints and articulations ‘from the inside’ without the structural disturbances inherent in dissected material while manipulating appendages and body segments through their normal range of excursion.

The neuromuscular anatomy of the sixth abdominal segment and telson of Blepharipoda was examined mostly in ventral dissections. The innervation of particular muscles was observed repeatedly with methylene-blue staining and in some cases by recording synaptic potentials in muscle fibres with 2·7 M-KCl-filled microelectrodes following delivery of single or brief trains of electrical pulses to the nerve through a suction electrode. The neuromuscular anatomy of Emerita has been described in detail (Paul, 1971a, b). All of the features important in the comparison with Blepharipoda have been reconfirmed in this study.

The wet weights of the muscles in the sixth abdominal segment and the telson were obtained after carefully dissecting out each muscle and blotting it to remove excess sea water. Three sets of muscles were weighed for each species. The percentage of the total weight contributed by each bilateral pair of muscles was calculated and the means and standard deviations of the three sets of percentages for each species determined.

Motoneurones were backfilled with CoCl₂ (Pitman, Tweedle & Cohen, 1972; Wine, Mittenthal & Kennedy, 1974). Many of these preparations were subsequently intensified with silver (Bacon & Altman, 1977). The positions and central morphologies of the filled motoneurones were examined in wholemounts, drawn with the aid of a camera lucida, and photographed. All of the motor roots discussed in this study have been backfilled in 10-30 animals with the exception of the lateral rotator nerve, which has been backfilled four times in Blepharipoda and three times in Emerita.

Examination of the articulation of the uropods in Blepharipoda and Emerita in undissected tailfans following the KOH treatment revealed considerable differences in skeletal structures and their arrangements in the two sand crabs. From this comparison alone one could predict the different vectors and amplitudes of uropod movement observed during swimming (Paul, 1981). Fig. 1 shows the dorsal aspect of the tailfans of Blepharipoda and Emerita (A, D) and the articulation of the right uropod viewed both from the side (C, F) and from a position within the telson from which the base of the uropod is seen end-on (B, E). A right-angle bend in the propodite of the uropod brings this terminal appendage forward in the sand crabs, along the side of the abdomen, rather than flanking the telson as in the conventional crustacean tailfan. Articulation of the uropod with the sixth segment is at a single point; a shallow socket in the anterior side of the uropod rides over an articular process which projects laterally and posteroventrally from the tergum. This articular process is slender in Emerita, and is further forward and projects further ventrally than its stubby counterpart in Blepharipoda; the latter is close to the inward-projecting flanges of the articulation between the sixth segment and telson (Fig. 1 B, C, E, F). The proximal portion of the propodite is also longer in Emerita than in Blepharipoda (Fig. 1 A, D). The effect of these skeletal differences is that the uropod in Emerita is displaced away from the body, to the side of segment 6, and is free to move in all three planes, most particularly in a sizeable arc along the antero-posterior axis. In contrast, the uropod in Blepharipoda, as in macrurans, is wedged in close to the body, beside the junction of segment 6 and the telson, where its movement is restricted, particularly along the long axis of the body; in these animals abduction and depression of the uropod are the principal movements that are made.

An estimate of the relative importance of different muscles in the execution of particular movements may be obtained from comparisons of their relative masses. The relative wet weights of muscles in the tailfan of Blepharipoda and Emerita (excluding those muscles that are contained entirely within the uropod) are compared in Fig. 2. Blepharipoda are larger animals than Emerita: the mean carapace length of the Blepharipoda was 4· 6 ± 0· 5 cm (n = 3) and of the Emerita 2· 1 ± 0· 3 cm (n = 3). However, only 40% of the tailfan muscles is contained in the telson of Blepharipoda compared to 80% in Emerita (Fig. 2 A).The histograms in Fig. 2B and C show the distribution of muscle mass among the muscles contained in segment 6 and the telson respectively; the terminology used is that suggested for Emerita (Paul, 1971a), and is, at least provisionally, appropriate for the muscles in Blepharipoda. Homologous muscles in segment 6 are relatively the same size in the two species (Fig. 2B), while homologous muscles in the telson differ greatly in size (Figs. 2C, 4, 7, 10). The largest muscle in Blepharipoda is one of the smallest in Emerita (the dorsomedial muscle, DM), whereas the second largest in Emerita is very small in Blepharipoda (the return-stroke muscle, RS) There appears to be no homologue in Blepharipoda for Emerita’s massive powerstroke muscle (PS); the functional analogue is DM (see below and Paul, 1981). Each muscle, PS in Emerita and DM in Blepharipoda, makes up about 50% of the telson muscle mass. The lateral muscles, LA, are power-stroke synergists and of relatively equal size in both animals. Thus, the muscles primarily responsible for the power phase of the uropod movement have similar weights relative to tailfan size in both species.

The external appearance of the terminal ganglion is very different in the two sand crabs (Figs. 3, 5). It is broad and flat in Emerita and almost cylindrical in Blepharipoda. In both crabs, three roots exit from the posterior face of the ganglion to innervate uropods, telson and, along with the second and third roots of ganglion 5, the sixth abdominal segment. (Intestinal nerves exit dorsally from the posterior part of the ganglion.) The branching and destination of these roots are shown in Fig. 3. Apart from the different dimensions, the major difference between the two crabs is the course of the axons innervating RS: in Blepharipoda they leave the large lateral root along with the LR motoneurones in the sixth segment; in Emerita they continue into the telson in close association with the PS and LA motoneurones, until near their destination.

The return-stroke (RS) muscle in the telson and the lateral muscle (LR) in the sixth segment are synergists in the return-stroke phase of the uropod movement. Their combined action is to abduct and perhaps slightly elevate the uropod in Blepharipoda and to retract it in Emerita. This difference is due to the different axes of RS contraction and different positions of their common insertion on the dorsal rim of the uropod, medial to the articular process in Blepharipoda and posterior to it in Emerita (Fig. 4).

Three neurones innervate RS in Emerita (Paul, 1971a). Their large somata (diameter ⩾40 μ m) are ventral and ipsilateral to the axons, two being in the posterior quadrant and the third near the middle of the ganglion (Fig. 5 A, E). The neurite of the latter (Fig. 5A,₂) forms an arch closer to the dorsal surface of the ganglion than do those of the pair of cells, so that its integrating segment and first-order branches extend through a more dorsal plane than the corresponding parts of the other two motoneurones. The finer dendritic branches of all three cells, however, appear to intermingle and terminate in the same regions, principally along the ipsilateral margin of the ganglion, but each also has a few processes that cross the midline in the dorsal neuropil (Fig. 5 A). The posterior two RS motoneurones have nearly identical shapes and positions, including even the finer branches of their dendrites.

Three cells stain in cobalt-backfills of the RS nerve in Blepharipoda (Fig. 5). One of these has not been filled completely; its axon in the root and soma are shown in Tig. 5 C, D. The other two are closely associated, synchronously branching neurones that are much smaller (soma diameters = 10– 20μm), but have a form and occupy a position in the ganglion extremely similar to the posterior pair of RS motoneurones in Emerita (Fig. 5 B-D, F).

In summary, the same number of RS motoneurones occurs in each crab; they occupy equivalent positions in the sixth abdominal ganglion despite the latter’s very different proportions; and there is strong resemblance in shape between at least two of the three RS motoneurones in the two sand crabs. The much larger size of these neurones in Emerita than in Blepharipoda correlates well with the difference in size of the muscles they innervate (Figs. 2, 4).

The RS muscle is peculiar to sand crabs, being absent from macrurans (Daniel, 1931; Schmidt, 1915; unpublished observation) as well as the more distantly related penaeids (Young, 1959). How does a new muscle and its innervation evolve? The common tendon of RS and the lateral muscle, LR, in the sixth segment in both sand crabs (Fig. 4) and the close association of the innervation of these two muscles in the less specialized Blepharipoda (Fig. 3) suggest that RS might have been derived from sixth segment musculature, that is from part of LR in an ancestral macruran (probably equivalent to uropod remotors in modern macrura). Comparison of the central morphology of motoneurones innervating these two muscles in each crab reveals resemblances that may be indicative of ancestral affinity (Figs. 5, 6). The somata positions, neurite orientations and dendritic fields of RS and LR motoneurones in Emerita are nearly identical (compare Fig. 5 A and 6 A) as are those of the RS motoneurones and the ipsilateral, posterior cluster of LR neurones in Blepharipoda (compare Figs. 5 B-D and 6B, C). Note also the similarity in position and morphology of LR motoneurones in the two sand crabs. These morphological observations are consistent with the possible derivation of both sets of motoneurones from a single pool of motoneurones serving a single progenitor muscle in a common ancestor, although other explanations of these similarities are possible.

These muscles are much more substantial in Blepharipoda than in Emerita (see Fig. 2C). The dorsomedial muscle (DM) has two massive heads in Blepharipoda with broad insertions adjacent to each other on the ventral surface of the propodite (Fig. 7). This muscle depresses the uropod and is, therefore, the major uropod ‘power-stroke’ muscle in the telson of Blepharipoda. DM in Emerita is rather small Imd inserts via a discrete tendon on the medial ventral edge of the uropod, close to the insertion of two sensory strands (Paul, 1971b, 1972, 1976). (The medial muscle in the sixth segment, MR, also inserts here (Fig. 7) as does VM (see below).) The insertion of the lateral muscle, LA, is medial to the articular process in Blepharipoda and lateral to it in Emerita, so that LA would be expected to adduct as well as depress the uropod in Blepharipoda, but abduct and protract the uropod in Emerita (Fig. 7). From an evolutionary perspective it may be more accurate to include LA with those muscles that have ventral (or posterior) origins (PS, VM, ME, see next sections) because the entire group may have evolved from telson musculature innervated by the sixth root of ganglion 6 in a macruran ancestor, an hypothesis suggested by the arrangement of muscles in the crayfish (unpublished observations).

In Emerita LA and DM are innervated by widely separated branches of the large lateral root of the terminal ganglion, whereas in Blepharipoda the motoneurones of these muscles separate from the same branch only in close proximity to their destinations (Fig. 3). Preliminary data on positions and morphologies of these groups of motoneurones are in partial agreement with the suggestion that motoneurones innervating homologous muscles in different species resemble each other morphologically and in their position within the central nervous system. That is, in both sand crabs the somata and integrating segments of LA motoneurones lie anterior in the ganglion; in both, long neurites extend across the midline, proliferating and terminating in anterior contralateral neuropil (Fig. 8 A, B). In contrast to the similar location and branching pattern of the neurites, the positions of the somata of LA motoneurones are rather different, being anterior to the integrating Segment in Emerita and lateral to it in Blepharipoda. Does this difference reflect adoption of different ways of packing homologous cells into differently proportioned ganglia during the divergent evolution of the two sand crabs or are these motoneurones not homologous to each other? Might the similarity of their dendritic morphologies be the result of convergent evolution, stemming perhaps from similarly organized functional connexions? Analysis of the properties and roles of these motoneurones may provide answers to these questions.

The axons of DM motoneurones in Emerita are contained in a large ventral branch of the lateral root that includes the dendrites of four non-spiking stretch receptive cells and also a number of small axons associated with the muscular strand (Paul, 1971a, b; 1972). Backfills of this nerve that exclude the nonspiking sensory cells reveal two clusters of neurones, one posterior and one central, in the ipsilateral posterior quadrant of the ganglion. The dispositions of these cells in Emerita is similar to those of the additional cells revealed in Blepharipoda when the nerve branch containing both LA and DM motoneurones is backfilled (compare C and D, Fig. 8).

The power-stroke muscle (PS) arises far posterior in the elongate telson of Emerita (Fig. 7; see also Fig. 2C) and is innervated by one excitatory and one inhibitory motoneurone (Paul, 1971a). Both of these are large cells with somata (diameter ⩾ 40 μ m) near the midline; the proximal portion of their integrating segments are transversely orientated and the branching of their neurites rather sparse (Fig. 9). Their axons exit through the large, lateral root and are adjacent to axons innervating LA until the nerve bifurcates near these muscles in the telson (Fig. 3).

Blepharipoda does not have a PS muscle and the homologue of this muscle does not appear to be the posterior head of Blepharipoda’s DM (Paul, 1979 b). (1) The origins do not match. DM arises dorsally from the anterior-medial surface of the telson (in both crabs), whereas the origin of PS is far posterior in the telson of Emerita. (2) The two heads of Blepharipoda’?, DM share their innervation in part (unpublished observation), whereas the innervations of Emerita’s DM and PS are completely separate (Paul, 1971a; unpublished observation). (3) Cobalt backfilling of the roots innervating telson muscles in Blepharipoda has revealed no motoneurones that resemble the two PS motoneurones in Emerita. (4) One of the many heads of flexor muscles in the telson of crayfish and spiny lobster is located in a position analogous to that of PS (unpublished observation), although the insertions are different. The innervation of the musculature in the crayfish telson has not been fully investigated (see Larimer & Kennedy, 1969), but cobalt backfills of root 6, which innervates nearly all the musculature in the telson, reveal that some of the motoneurones have large somata near the midline (M. Bastiani, personal communication; J. J. Wine, personal communication). No counterparts for all of these motoneurones send axons out of the corresponding (VM) nerve in either sand crab. If Emerita’s power-stroke muscle has evolved from part of the telson musculature of ancestral macrurans that was innervated by root 6, then the two PS motoneurones, with their large somata near the midline, could be homologues of crayfish root 6 motoneurones that have redirected their axons out of a different root (see discussion).

In Blepharipoda the two muscles with ventral origins and insertions are the terminal components of the flexor musculature that spans the length of the abdomen (Fig. 10). The ventromedial muscle, VM, inserts on the posterior side of a partial skeletal rib at the juncture of segment 6 and telson; the flexors in the sixth segment insert on the anterior side of this rib. The smaller medial muscles, ME, insert on the midline of the anterior telson on a small calcified area in the ventral arthrodial membrane. Both muscles arise in the posterior telson, primarily from the ventral arthrodial membrane; deeper fibres of VM arise from the posterior dorsal surface of the telson. The VM flex the telson-in conjunction with the flexors of the 6th segment; the ME cup the telson by deflecting ventrally its flexible margins.

In Emerita, both VM and ME arise entirely from the middle region of the ventral arthrodial membrane (Fig. 10). The ME insert on the midline in a manner similar to their homologues in Blepharipoda, but their function is obscure since the telson is rigid. Each VM inserts on a uropod, its tendon attaching to the medial extension of the ventral side of the propodite, directly ventral to the DM tendon (see above). The long axis of this muscle is parallel to that of PS (Figs. 7, 10), but since its insertion on the uropod is medial to the point of articulation it tends to adduct and pronate as well as protract the appendage.

Although the VM muscles perform different functions in the two sand crabs, their homology to each other and to part of the telson flexion musculature in crayfish and spiny lobster (unpublished observation) is clear from their positions relative to other anatomical features, their morphology, and their innervation. Each is innervated by the most posterior motor nerve (excluding intestinal nerves) of the terminal abdominal ganglion; and the positions and morphologies of the motoneurones are similar in the two sand crabs (below) and resemble the sixth root motoneurones in crayfish (M. Bastiani, personal communication; J. J. Wine, personal communication).

The positions and morphology of VM motoneurones of Emerita and Blepharipoda are compared in Fig. 11. The six motoneurones in Emerita (, 1971 a, 1979 a) can be divided into four groups based on position of somata and orientation of principal neurites (A-D, 11A-C). The somata of the pair of ‘a’ motoneurones are generally ar posterior in the ganglion, close to the root through which the axons exit, and their neurites tend to be closely intertwined (A_lt B₁, E, Fig. 11); although the soma of The single ‘c’ neurone is often in the same region, its neurite is separate from those of the ‘a’ motoneurones and is distinctively shaped (Fig. 11A2, C). The somata of the two motoneurones in group ‘b’, often slightly larger than the others, lie on or just across the midline from their axons (Fig. 11, A1, B2). The dendritic fields of these five neurones, groups a-c, overlap extensively over most of the ipsilateral neuropil and extend to or across the midline in the anterior part of the ganglion (Fig. 11E). The soma and proximal part of the integrating segment of the remaining cell, ‘d’, is contralateral to its axon (Fig. 11A2, C, E).

VM in Blepharipoda is innervated by at least eight cells that are grouped similarly to Emerita’s VM motoneurones (Fig. 11D, F, G). The camera lucida drawings include only those processes that could be traced unambiguously to the somata shown, each of these motoneurones clearly having a much larger dendritic tree (Fig. 11E-G). Although these morphological data do not allow suggestions for specific homologies between individual cells, the VM motoneurones of the two crabs are very similar to each other in position and morphology (Fig. 11E, F). In fact, homologues of each of the six VM motoneurones in Emerita may not occur in Blepharipoda, since the muscles may be derivatives of different heads of flexion musculature in the telson, which is complex in crayfish and spiny lobster and probably was also in ancestral macrurans.

The uropods of sand crabs are unlike those of other Crustacea. The propodite makes a right-angle bend so that the appendage extends forward rather than rearward from its articulation with the sixth segment as it does in other crustaceans with tailfans. Blepharipoda, however, uses the tailfan as an extension of the abdomen, in much the same way as do macrurans, and is able only to extend (abduct) the uropod in the horizontal plane and depress it in the vertical plane (Paul, 1981). The inventions of Emerita that made possible its unique swimming behaviour were the return-stroke and power-stroke movements of the uropod. In order to have evolved uropod beating, at least two additional changes had to occur in the macruran-type tailfan besides the altered shape of the uropod itself: skeletal changes to permit the large excursions of the uropod in the longitudinal axis, and the subversion, or invention, of neuromuscular systems to perform these new movements.

Mechanical constraints inherent in the uropod articulation and spatial constraints inherent in the flattened, folded abdomen were probably the major determinants leading to expansion of the telson and its musculature in Emerita. Movement of the uropod through a large arc necessitated elongation of the retraction and protraction muscles since their insertions had to remain close to the pivotal point of the appendage where leverage is poor. The dorso-ventral flattening of the folded abdomen in sand crabs precluded vertical orientations of the muscles such as is typical for most arthropod appendages that perform large movements, e.g. the leg muscles in the thorax or the protractor and retractor muscles of macruran swimmerets. In addition, the force of the power muscles needs to be applied predominantly in the vertical one of the articulation, not medial to it. This excludes muscles of the sixth segment because the articular process projects laterally (Fig. 1). The telson, on the oth hand, could not only elongate but broaden so that muscles could insert on botff sides of the pivotal point, resulting in the unusual mobility of the uropod.

Is it possible to identify in less-specialized tailfans muscles that might be homologous to Emerita’ massive RS and PS? RS is apparently a sand crab invention; it is absent from macrurans (Daniel, 1931; Schmidt, 1915; unpublished observation) as well as from penaeids (Young, 1959). PS is peculiar to Emerita (Hippidae). Derivation of RS from LR musculature (probably corresponding to crayfish’s remotor musculature) in the sixth segment of an ancestral macruran appears to be the most likely origin of this sand crab muscle. In both sand crabs RS and LR share one insertion on the uropod. Moreover, the close affinity of these two muscles is suggested by their innervation: in Blepharipoda, a less-specialized animal than Emerita, the return-stroke muscle is innervated by a branch of the LR nerve (Fig. 3); and furthermore, in both sand crabs, not only are the somata of the RS motoneurones located among those of the LR neurones but the dendritic morphologies of all these motoneurones are very similar (Fig. 5, 6). These features are most simply explained as reflecting derivation of both neuromuscular systems from a single muscle and pool of motoneurones in an ancestral macruran (see Neuronal Homology, below). Greater conservatism in motoneurones than in either peripheral structures or central connexions has been described for several other neuromuscular systems (Bent & Chapple, 1977; Chapple, 1977b; Kaars, 1979; Kahan, 1971; Mittenthal & Wine, 1978; Wiersma & Ripley, 1952). Whatever the derivation of RS, its appearance in sand crabs was probably concomitant with the altered morphology of the uropod that presumably evolved in association with the habit of burrowing backward into sand (Paul, 1981). Passive depression of the uropods must occur during this behaviour and should be opposed in part by contraction of RS.

In the crayfish telson there are two candidates for possible homology with Emerita’s massive PS, for which Blepharipoda has no homologue. Both of the crayfish muscles appear to be innervated by root-6 motoneurones (Larimer & Kennedy, 1969; unpublished observation). However, the PS axons do not exit through the homologue of root 6, the VM nerve. In arthropods even more distantly related than sand crabs and macrurans the root through which an axon exits often appears to be the same, although most of the neurones studied have been in the middle ganglia and innervate muscles that have roughly equivalent functions in the different species (Table 5, Mittenthal & Wine, 1978). Serially homologous leg motoneurones in locusts also exit by the same root even when their physiological properties are different (Wilson, 1979a, b). But developmental studies have demonstrated considerable variability in the course axons take to make more or less invariant connexions, which suggests that the route taken by nerve fibres is less critical than the destination (Cowan, 1979). And since Emerita has apparently evolved a ‘new’ muscle and considerably rearranged the telson anatomy, it seems plausible that motor axons could have become rerouted in order to take the most economical path to their targets during development.

Muscles in the crayfish telson include a number of more or less distinct heads that insert on the stout, transverse, membrane-like tendon between segment 6 and telson (unpublished observation). The telson flexor, VM, in Blepharipoda probably corresponds to one of the more medial of these; the homology of Emerita’s VM with part of the macruran telson flexor Systems was not obvious without the intermediate comparative step provided by Blepharipoda, since functionally it belongs to the uropod and not to the axial musculature. Whether the ventromedial muscles of the two sand crabs represent derivatives of the same or of different muscles innervated by root 6 in a macruran-like ancestor can not be distinguished without a better understanding than we have at present of the functional relationships and innervation of the various telson muscles in macrurans.

The considerable difference in ventral musculature in the telsons of the two sand crabs suggest that their families separated from a macruran-like, sand crab ancestor, already possessing RS, before the musculature in the telson had become very modified. Emerita rerouted parts of the ventral muscular system to serve the appendage whereas Blepharipoda reduced the mass but retained the flexor function of these muscles.

Homologies have been described between neurones with qualitatively similar functions in different species (e.g. Bent & Chapple, 1977; Chapple, 1977 a, b, Dickinson, 1979; Kaars, 1979; Kahan, 1971; Keyser & Lent, 1977; Mittenthal & Wine, 1978; Silvey & Wilson, 1979). In these cases morphological similarities by themselves often appear predictive of neuronal homology, although evidence for homology is the more compelling the more information is available concerning a whole constellation of additional features, such as synaptic properties, membrane characteristics, input resistance, ontogeny (see Rowe & Stone, 1977; Tyner, 1975). One might anticipate greater difficulty in distinguishing on the basis of morphology alone homologous neurones in related animals who differ in musculature or behaviour because experimental as well as descriptive data suggest that function and many aspects of neurone morphology are interdependent variables (Hoy & Moiseff, 1979; Kennedy & Davis, 1977; Kimmel, Schabtach & Kimmel, 1977; Murphy et al. 1975). However, in these cases, also, similarity in cell shape, in location within the central nervous system, and particularly in location of soma appear to be strong criteria for homology. Wilson & Hoyle (1979; Wilson, 1979 a, b) have reported retention of ‘primeval’ morphologies, i.e. same position and shape of major neurites (compared with their serial homologues) of locust metathoracic leg motoneurones which have altered their functional properties. Another example is the homologous Tritocerebral Commissure Giant interneurone in a locust, a cricket, and a mantid that bear strong anatomical resemblances to each other, particularly in location of cell body, of dendritic arborizations, and of output branches (Bacon, 1980).

In the sand crabs, in addition to functional changes of some of the motoneurones themselves, the very different shapes of the terminal ganglia, into which presumably approximately the same number of cells are packed, might impose shape ‘distortions’ that are not directly related to either ancestry or function, thereby making more difficult the application of morphological criteria of homology. Although this mag occasionally occur, it appears not to be a general problem. Rather, the results OF this comparison of motoneurones in Emerita and Blepharipoda encourage attempts to identify homologies between single neurones in members of related taxa with divergent behaviour. Whether or not the swimming patterns of the two sand crabs are homologous to each other and to macruran tailflipping, many of the individual motoneurones certainly are. Future analyses of the physiological properties of these cells in sand crabs and their putative homologues in a macruran, such as crayfish, should reveal how tailfan motoneurones have changed during the evolution of Emerita’s new mode of swimming. Comparison of these homologous motoneurones may suggest which features of neuronal structure and function are plastic and which relatively immutable through evolutionary diversification of behaviour.

I thank C. Pittendrigh, Director, and S. Thompson and D. Epel of Hopkins Marine Station, J. J. Wine of Stanford University and B. Mulloney of the University of California, Davis, for providing facilities for this work. I also thank the following people for their helpful discussions of various aspects of this study: M. Bastiani, G. Hagiawara, L. A. Miller, B. Mulloney, M. Paul and J. J. Wine. Supported in part by N.S.F. grant BNS 78-10516 (to B. Mulloney).