Morphological changes in neuromuscular junctions during exercise

Long lasting exercise produces several morphological changes in teleostean neuromuscular junctions (NJs), consisting of progressive synaptic vesicles (SVs) depletion and lamellar branching of the nerve endings. Exercised fishes kept swimming during 1 hr against a 3.5 1/min flow of oxygenated water in spite of the fact that the number of SVs was reduced in about 70% after 10 min of exercise. This observation indicates that the SVs formation fails to restore their original number and consequently, under such circumstances, the transmitter release may occur by a different mechanism.     

Serotoninergic innervation of the cat cerebral cortex

Serotoninergic axons in the cat cerebral cortex were demonstrated immunohistochemically with a monoclonal antibody to serotonin (5-HT). Three types of 5-HT axons are distinguished at the light microscopic level by differences in their morphology. Small varicose axons are fine (less than 0.5 micron) and bear fusiform varicosities that are generally less than 1 micron in diameter. These axons extend throughout the width of the cortex and branch frequently, giving rise to widely spreading collaterals. Nonvaricose axons are smooth, show a relatively large and constant caliber (about 1 micron), travel in straight, horizontal trajectories, and branch infrequently. Large varicose axons are distinguished by large round or oval varicosities (1 micron or more in diameter) borne on fine-caliber fibers. These axons often form basket-like arbors around the somata of single neurons. In the simplest basket-like arbors, several large, round varicosities from a small number of axons contact the soma. In complex baskets intertwining collaterals contact the soma and apparently climb along and outline the cell's major dendrites. The patterns revealed by the climbing axons suggest that a variety of nonpyramidal cell types selectively receive dense 5-HT innervation. Serial reconstructions of the 5-HT axons within the cortex show that the large varicose axons arise as infrequent collaterals from the nonvaricose axons. A single nonvaricose parent axon gives rise to several large varicose axon collaterals that may contribute to different basket-like arbors. Conversely, a single basket-like arbor may be formed by large varicose axon collaterals from more than one nonvaricose parent axon. The small varicose axons do not appear to be related within the cortex to either the nonvaricose or large varicose axon types. The results support the hypothesis that the 5-HT projection to the cortex is organized into two subsystems, one of which may exert widespread influence in the cortex via highly divergent branches, while the other, with a more restricted distribution, acts on specific classes of cortical neurons.     

Microtubules and calibers in normal and regenerating axons of the sural nerve of the rat

The calibers and microtubular content of axons were studied in normal and regenerating fibers of the sural nerve from 17 to 122 days after a lesion of the sciatic nerve of young adult rats. During this period (70-175 days of age), the cross-sectional area of control myelinated axons almost doubled but that of nonmedullated axons did not change. In regenerating nerves, after 122 days of recovery, the cross-sectional area of myelinated fibers was still 38% below that of the normal side. In contrast, the regenerating nonmedullated population was richer in fine (less than 0.2 micron2) and in coarse (greater than 0.9 micron2) fibers than on the control side; the cross-sectional area averages were 0.50 and 0.54-0.70 micron 2 for the normal and regenerating populations, respectively. The microtubular density of normal 3-micron myelinated fibers averaged 24.0 microtubules/micron2. In regenerating fibers of the same size the density varied between 19.2 and 23.2 microtubules/micron2. Microtubular density values of normal and regenerating fibers were not statistically different. In nonmedullated fibers, the microtubular content (expressed as microtubular density or number of microtubules per axon) correlated with the caliber of the fiber. In these correlations, only minor differences were observed between regenerating and uninjured fibers. Our results indicate that nonmedullated fibers terminate their radial growth well before myelinated fibers do, and that axonal microtubular content correlates with the local size of the fiber and is largely insensitive to regeneration.     

Serotonin release variations during recovery of motor function after a spinal cord injury in rats

Current literature suggests that serotonin (5‐HT) release within the ventral horn of the spinal cord plays a role in motor function. We hypothesized that endogenous 5‐HT release is involved in the recovery of motor function after spinal cord injury. To appreciate the functional parameters of regenerating serotonergic fibers, a microdialysis probe was stereotactically implanted in the ventral horn of subhemi‐lesioned rats. Microdialysis in combination with HPLC was used to measure concentrations of 5‐HT in the lumbar ventral horn during periods of rest (90 min), treadmill run (60 min) and postexercise rest (90 min) for a 1‐month time period of recovery following the surgical lesion. Within the same period of time, 5‐HT levels varied significantly. A significant (202%) increase was observed at day 18 postlesion relative to day 8, and a 16.4% decrease was observed at day 34 relative to day 18. Treadmill exercise challenge induced a 10% decrease of 5‐HT release relative to rest at days 18 and 34. In conclusion, overtime treadmill locomotor recovery is parallel to amounts (rest basal levels) and patterns (exercise and postexercise levels) of 5‐HT release suggesting that changes in serotonergic system occurred within the same time frame than locomotor recovery using treadmill challenge. Synapse 64:855–861, 2010. © 2010 Wiley‐Liss, Inc.     

Errors in lamina growth of primary olfactory axons in the rat and mouse olfactory bulb.

In the adult olfactory nerve pathway of rodents, each primary olfactory axon forms a terminal arbor in a single glomerulus in the olfactory bulb. During development, axons are believed to project directly to and terminate precisely within a glomerulus without any exuberant growth or mistargeting. To gain insight into mechanisms underlying this process, the trajectories of primary olfactory axons during glomerular formation were studied in the neonatal period. Histochemical staining of mouse olfactory bulb sections with the lectin Dolichos biflorus-agglutinin revealed that many olfactory axons overshoot the glomerular layer and course into the deeper laminae of the bulb in the early postnatal period. Single primary olfactory axons were anterogradely labelled either with the lipophilic carbocyanine dye, 1,1'-dioctodecyl-3,3,3',3'-tetramethylindocarbocyanine perchlorate (DiI), or with horse-radish peroxidase (HRP) by localized microinjections into the nerve fiber layer of the rat olfactory bulb. Five distinct trajectories of primary olfactory axons were observed in DiI-labelled preparations at postnatal day 1.5 (P1.5). Axons either coursed directly to and terminated specifically within a glomerulus, branched before terminating in a glomerulus, bypassed glomeruli and entered the underlying external plexiform layer, passed through the glomerular layer with side branches into glomeruli, or branched into more than one glomerulus. HRP-labelled axon arbors from eight postnatal ages were reconstructed by camera lucida and were used to determine arbor length, arbor area, and arbor branch number. Whereas primary olfactory axons display errors in laminar targeting in the mammalian olfactory bulb, axon arbors typically achieve their adult morphology without exuberant growth. Many olfactory axons appear not to recognize appropriate cues to terminate within the glomerular layer during the early postnatal period. However, primary olfactory axons exhibit precise targeting in the glomerular layer after P5.5, indicating temporal differences in either the presence of guidance cues or the ability of axons to respond to these cues.     

Treadmill training promotes axon regeneration in injured peripheral nerves

Physical activity after spinal cord injury promotes improvements in motor function, but its effects following peripheral nerve injury are less clear. Although axons in peripheral nerves are known to regenerate better than those in the CNS, methods of accelerating regeneration are needed due to the slow overall rate of growth. Therefore we studied the effect of two weeks of treadmill locomotion on the growth of regenerating axons in peripheral nerves following injury. The common fibular nerves of thy-1-YFP-H mice, in which a subset of axons in peripheral nerves express yellow fluorescent protein (YFP), were cut and repaired with allografts from non-fluorescent littermates, and then harvested two weeks later. Mice were divided into groups of low-intensity continuous training (CT, 60 min), low-intensity interval training (IT; one group, 10 reps, 20 min total), and high-intensity IT (three groups, 2, 4, and 10 reps). One repetition consisted of 2 min of running and 5 min of rest. Sixty minutes of CT resulted in the highest exercise volume, whereas 2 reps of IT resulted in the lowest volume of exercise. The lengths of regenerating YFP⁺ axons were measured in images of longitudinal optical sections of nerves. Axon profiles were significantly longer than control in all exercise groups except the low-intensity IT group. In the CT group and the high-intensity IT groups that trained with 4 or 10 repetitions axons were more than twice as long as unexercised controls. The number of intervals did not impact axon elongation. Axon sprouting was enhanced in IT groups but not the CT group. Thus exercise, even in very small quantities, increases axon elongation in injured peripheral nerves whereas continuous exercise resulting in higher volume (total steps) may have no net impact on axon sprouting.     

Synapsins contribute to the dynamic spatial organization of synaptic vesicles in an activity-dependent manner

The precise subcellular organization of synaptic vesicles (SVs) at presynaptic sites allows for rapid and spatially restricted exocytotic release of neurotransmitter. The synapsins (Syns) are a family of presynaptic proteins that control the availability of SVs for exocytosis by reversibly tethering them to each other and to the actin cytoskeleton in a phosphorylation-dependent manner. Syn ablation leads to reduction in the density of SV proteins in nerve terminals and increased synaptic fatigue under high-frequency stimulation, accompanied by the development of an epileptic phenotype. We analyzed cultured neurons from wild-type and Syn I,II,III(-/-) triple knock-out (TKO) mice and found that SVs were severely dispersed in the absence of Syns. Vesicle dispersion did not affect the readily releasable pool of SVs, whereas the total number of SVs was considerably reduced at synapses of TKO mice. Interestingly, dispersion apparently involved exocytosis-competent SVs as well; it was not affected by stimulation but was reversed by chronic neuronal activity blockade. Altogether, these findings indicate that Syns are essential to maintain the dynamic structural organization of synapses and the size of the reserve pool of SVs during intense SV recycling, whereas an additional Syn-independent mechanism, whose molecular substrate remains to be clarified, targets SVs to synaptic boutons at rest and might be outpaced by activity.     

Ultrastructural immunocytochemical localization of tyrosine hydroxylase in the neostriatum

The morphology and synaptic associations of dopaminergic axons in the n. caudate-putamen (neostriatum) of the adult rat brain are examined. Identification of dopaminergic axons is based upon the electron microscopic immunocytochemical localization of the catecholamine synthesizing enzyme, tyrosine hydroxylase. Immunoreactivity for the enzyme is detected in unmyelinated axons and axon terminals in serial sections collected throughout the neostriatum. The labeled terminals range from 0.1 to 1.5 micron in diameter and have peroxidase reaction product located around closely packed, round vesicles with a diameter of 40-60 nm. The tyrosine hydroxylase containing axon terminals constitute approximately 21% of the total number of terminals in the n. caudatus-putamen and include 3 types which differ in size and synaptic specializations. The most prevalent (82% of total), type I, is small (0.15-0.39 micron in diameter) and forms symmetric junctions with dendrites and dendritic spines. The other two terminal types (II and III) have a medium to large diameter (0.4-1.5 micron) and show either no membrane specializations or asymmetric junctions with dendrites. The axon terminals without observable membrane densities are occasionally oriented so as to suggest an association with dopaminergic and non-dopaminergic axon terminals. These findings indicate that while the dopaminergic terminals may form axoaxonic connections, the primary synaptic contacts are with dendrites of intrinsic neurons in all regions of n. caudatus-putamen.     

Axons regenerated through silicone tube splices. II. Functional morphology

The recovery of axons regenerated through silicone tube splices was studied with electron microscopic and morphometric methods. Regenerated nerves contained both myelinated and unmyelinated axons of near normal morphology. The number and diameter of axons increased with postoperative time, and size-frequency histograms demonstrated that regeneration occurred in all major fiber groups. Remyelination occurred between about 4 and 6 weeks. Some of the smallest regenerated axons had unusually thick myelin sheaths, but overall regenerated axons had a slightly thinner sheath compared with similar-size normal fibers, although the ratio of sheath thickness to axon size was within the normal limits of g = 0.65 to 0.8 by 6 weeks. Axons did not, however, regain their normal size within 10 months of surgery. This deficit was apparently the primary factor limiting conduction velocity in these regenerated axons.     

Synapse‐to‐synapse variation in mean synaptic vesicle size and its relationship with synaptic morphology and function

Synaptic vesicle (SV) size is one parameter that controls the amount of neurotransmitter released from individual SVs and, therefore, is fundamental to our understanding of synaptic function. The recently discovered variability of mean SV size among excitatory hippocampal synapses—if actively regulated—is a potential mechanism for the regulation of transmitter release. Here, we investigated which parameters influence mean SV size. First, we revealed that synapse‐to‐synapse variability of SV size is a general phenomenon in several species and brain regions. In addition, we determined the relationship between mean SV size and synaptic morphology. In three‐dimensional reconstructions from serial ultrathin sections, we found that SV size did not correlate with the area of the postsynaptic density (a measure for synaptic size and synaptic cleft volume) nor with the total number of SVs within a bouton or bouton volume. We tested the long‐held hypothesis that a change in osmotic pressure (potentially caused by a change in neurotransmitter concentration) affects SV size. When we reduced the osmotic pressure, SVs became significantly smaller; however, an increase in osmotic pressure had no effect on SV size. Furthermore, we found that SV size does not adapt to chronic changes in activity and that the SV cycle is capable of providing constant SV size during long‐lasting, high‐frequency stimulation. J. Comp. Neurol. 514:343–352, 2009. © 2009 Wiley‐Liss, Inc.     

Reduced axonopathy and enhanced remyelination after chronic demyelination in fibroblast growth factor 2 (Fgf2)-null mice: differential detection with diffusion tensor imaging

Chronic central nervous system demyelinating diseases result in long-term disability because of limited remyelination capacity and cumulative damage to axons. Corpus callosum demyelination in mice fed cuprizone provides a reproducible model of chronic demyelination in which the demyelinating agent can be removed to test modifications that promote recovery and to develop noninvasive neuroimaging techniques for monitoring changes in myelin and axons. We used the cuprizone model in mice with genetic deletion of fibroblast growth factor 2 (Fgf2) to determine the impact of FGF2 on axon pathology and remyelination after chronic demyelination. We also evaluated the ability of quantitative magnetic resonance diffusion tensor imaging (DTI) to distinguish the corresponding pathological changes in axons and myelin during the progression of demyelination and remyelination. During the recovery period after chronic demyelination, Fgf2-null mice exhibited enhanced remyelination that was detected using DTI measures of radial diffusivity and confirmed by electron microscopic analysis of the proportion of remyelinated axons. Ultrastructural analysis also demonstrated reduced axonal atrophy in chronically demyelinated Fgf2-null versus wild-type mice. This difference in axon atrophy was further demonstrated as reduced immunohistochemical detection of neurofilament dephosphorylation in Fgf2-null mice. Diffusion tensor imaging axial and radial diffusivity measures did not differentiate Fgf2-null mice from wild-type mice to correlate with changes in axonal atrophy during chronic demyelination. Overall, these findings demonstrate that attenuation of FGF2 signaling promotes neuroprotection of axons and remyelination, suggesting that FGF2 is an important negative regulator of recovery after chronic demyelination.     

Microtubules and caliber of central and peripheral processes of sensory axons

The microtubular content and caliber of sensory axons were studied in the L7 dorsal root, at the distal pole of the L7 spinal ganglion, and in the sural nerve of cats. Calibers of myelinated axons were symmetrical about the ganglion. In contrast, nonmedullated axons were strikingly different; 80% of the population at the root were smaller than 0.4 micron2, whilst just across the ganglion the same group was less than 20%. The microtubule densities of myelinated axons of the root were 11.8 and 6.1 microtubules/micron2 for 3- and 10 microns diameter axons, respectively. Across the ganglion the densities of myelinated axons of equal sizes were 24.2 and 14.4 microtubules/micron2, respectively. These values represent an approximate ratio of 1:2 between central and peripheral microtubule densities. Microtubule densities for nonmedullated axons also decreased with the increase in the cross-sectional area. The densities of root nonmedullated axons ranged between 90 and 10 microtubules/micron2; these were smaller, usually by a factor of three, than the densities of peripheral axons of a similar size (range: 367-44). Contrasting with the differences observed across the ganglion, the microtubular content and caliber of sensory axons seems to be quite uniform along their peripheral course. This is supported by the similar values found in the juxtaganglionic and sural nerves. It is estimated that an axon that contains 90 microtubules/micron2 has 26.7 mg of tubulin per ml of axoplasm in its assembled form, and 3.0 mg/ml if it contains 10 microtubules/micron2; these values are the practical limits of assembled tubulin in axoplasms.(ABSTRACT TRUNCATED AT 250 WORDS)     

Optic tectum of the eastern garter snake, Thamnophis sirtalis. IV. Morphology of afferents from the retina

The morphology of single retinal terminals in the optic tectum of the eastern garter snake was demonstrated by orthograde filling from extracellular injections of horseradish peroxidase (HRP) into the optic tract. HRP-filled terminals share a characteristic shape and structure. Their parent axons course caudally in the stratum opticum within fascicles of 200-300 fibers of varying diameters. Single axons exit a fascicle and course into either the stratum fibrosum et griseum superficiale, ventrally, or the stratum zonale, dorsally, where they bifurcate successively two or three times into preterminal branches. Each preterminal branch gives rise to many thin, terminal branchlets laden with boutons. The arbors are ellipsoidal with their long axes oriented mediolaterally and their short axes oriented rostrocaudally. Arbors vary in their overall size (from 45 to 150 micron), in the diameters of their parent axons (from less than 0.5 to 3.0 micron), and in the size of their terminal boutons (from 0.5 to 3.5 micron). Bouton size increased with increasing diameter of the parent axon. The great majority of arbors are confined to one of three retinorecipient sublayers in the superficial tectum. However, the full range of arbor sizes and axon diameters is present in each sublayer.     

Morphology of axons in the human lateral geniculate nucleus: A golgi study in prenatal and postnatal material

A study was made of rapid Golgi preparations from the lateral geniculate nucleus in humans aged from 28 weeks gestation to 70 years in order to identify axon terminals of afferent fibre systems. We describe three main axonal types using, as far as possible, nomenclature already adopted for other species. Type I axons were found only rarely. They are relatively straight with short, stalked side-branches and may represent cortico-geniculate fibres. Type II axons have complex, ball-like arborizations with large, irregular varicosities. They are common at all ages from gestation to maturity and are probably retinal in origin. Type IV axons (Type III was not used as no unequivocally intrinsic axons, for which the term has been used in the past, were identified) are branched, meandering and characterized by many, regular varicosities. Their origin is unclear, but may be related to non-specific brainstem sources. The basic morphology of Type II axons varies little between late gestation and adulthood, but Types I and IV seem to evolve during the perinatal period, perhaps from primitive forms that have similar morphological features. We conclude that the morphology of afferent axons to the human lateral geniculate nucleus is basically similar to that of lower mammalian species.     

Regeneration of frog sympathetic neurons is accompanied by sprouting and retraction of intraganglionic neurites

Regeneration of frog sympathetic neurons (B-cells) was found to be accompanied by sprouting of neurites within the ganglion. Neurons whose axons had been crushed and allowed to regenerate exhibited sprouts that arose mainly from the axon hillock and initial segment of the axon. Sprouting was apparent by 5-7 days and reached maximal values by 14-21 days, but had decreased to control levels by 42-49 days after injury. In contrast, neurons whose axons were prevented from regenerating (by cut and proximal ligation of nerves) exhibited sprouts which did not retract by 42-49 days. These results suggest that successful regeneration to targets may dictate the recovery of normal B-cell morphology in bullfrog sympathetic ganglia.     

Macromolecular structure of axonal membrane in the optic nerve of the jimpy mouse

The macromolecular structure of the axon membrane in 26-28-day-old Jimpy mice and control optic nerve were examined with quantitative freeze-fracture electron microscopy. Premyelinated and myelinated axons were observed in control optic nerves, with axonal diameters of premyelinated axons being generally smaller than that of myelinated axons (approximately 0.2-0.4 micron vs approximately 0.5-1.5 micron, respectively). Axon membrane from control optic nerves exhibited an asymmetrical partitioning of intramembranous particles (IMP). P-faces of internodal membrane displayed nearly twice as many IMP as the premyelinated axolemma (1,731 vs 893 micron-2, respectively). E-faces of internodal and premyelinated axolemma exhibited IMP densities of 124 and 157 micron-2, respectively. Few myelinated axons were apparent in optic nerves from Jimpy mice. The amyelinated axons of Jimpy mice displayed a spectrum of axonal diameters, ranging from approximately 0.2 to 1.5 micron. P-face densities of amyelinated axons, considered as a group, exhibited a wide range (600-2,100 micron-2). However, large diameter (greater than or equal to 0.5 micron) axons exhibited a significantly greater P-face IMP density than that of small caliber (greater than 0.5 micron) axons (1,525 vs 1,032 micron-2, respectively). Aggregations of E-face IMP were not observed along amyelinated axons of Jimpy optic nerves. The results demonstrate that the changes in P-face IMP density that occur during development of normal myelinated axons also occur in developing axons of Jimpy optic nerve, irrespective of a lack of normal glial cell association, and provide further evidence that the primary defect of hypomyelination within Jimpy mice is not attributed to the neuron.     

Ventral root nonmedullated fibers: proportion, calibers, and microtubular content

Proportion, caliber, and microtubular content of the L7 ventral root nonmedullated fibers were studied in the cat. Nonmedullated fibers constituted 28% of the axonal population at the far end of the root. The number of myelinated profiles at the far end of the root and in the vicinity of the ventral surface of the cord (1-2 mm distance) was the same whereas the number of nonmedullated fibers decreased toward the proximal site of the root by 42%. Caliber and microtubular content of nonmedullated fibers were assessed only at the far end of the ventral root. Nearly 90% of the axons were smaller than 0.3 micron2. The average cross-sectional area was 0.15 micron2, a value 35% below that of the L7 dorsal root fibers. The microtubular density was highest in the finest fibers (116 microtubules/micron2 in fibers smaller than 0.1 micron2) and decreased with the increase in cross-sectional area (25 microtubules/micron2 for 0.7-0.8 micron2 axons). The number of microtubules per axon in axons of both roots was similar in fibers smaller than 0.3 micron2; in larger axons, composing about 10% of the population, ventral root fibers had more microtubules than dorsal root fibers. The ventral and dorsal nonmedullated fibers differ slightly but significantly in caliber and microtubule content. However, they are similar in contrast to peripheral nonmedullated fibers, which are three to four times as big and contain two to three times as many microtubules as radicular fibers. Our results confirm the presence of a large admixture of nonmedullated profiles in the L7 ventral root of the cat, support the notion that a number of these fibers make a U turn in the ventral root, and suggest that these arise from the central process of the primary sensory axon.     

Astrocytic glycogen influences axon function and survival during glucose deprivation in central white matter.

We tested the hypothesis that astrocytic glycogen sustains axon function during and enhances axon survival after 60 min of glucose deprivation. Axon function in the rat optic nerve (RON), a CNS white matter tract, was monitored by measuring the area of the stimulus-evoked compound action potential (CAP). Switching to glucose-free artificial CSF (aCSF) had no effect on the CAP area for approximately 30 min, after which the CAP rapidly failed. Exposure to glucose-free aCSF for 60 min caused irreversible injury, which was measured as incomplete recovery of the CAP. Glycogen content of the RON fell to a low stable level 30 min after glucose withdrawal, compatible with rapid use in the absence of glucose. An increase of glycogen content induced by high-glucose pretreatment increased the latency to CAP failure and improved CAP recovery. Conversely, a decrease of glycogen content induced by norepinephrine pretreatment decreased the latency to CAP failure and reduced CAP recovery. To determine whether lactate represented the fuel derived from glycogen and shuttled to axons, we used the lactate transport blockers quercetin, alpha-cyano-4-hydroxycinnamic acid (4-CIN), and p-chloromercuribenzene sulfonic acid (pCMBS). All transport blockers, when applied during glucose withdrawal, decreased latency to CAP failure and decreased CAP recovery. The inhibitors 4-CIN and pCMBS, but not quercetin, blocked lactate uptake by axons. These results indicated that, in the absence of glucose, astrocytic glycogen was broken down to lactate, which was transferred to axons for fuel.     

Reconstructions of corticogeniculate axons in the cat

The terminal arbors of corticofugal axons to the dorsal lateral geniculate nucleus in the cat were filled with horseradish peroxidase and then partially reconstructed through serial sections. The results demonstrate that these arbors are far more complex than was suspected from previous studies of axon segments in individual sections. These axons branch profusely and spread widely within the nucleus. Within laminae A and A1 the terminal arbor of a single axon can be more than 800 micron wide compared with retinogeniculate axons whose terminal arbors range in width from 100 to 410 micron (Sur and Sherman, '82).     

Axon projections of reciprocal inhibitory interneurons in the spinal cord of young Xenopus tadpoles and implications for the pattern of inhibition during swimming and struggling

We have examined the morphology and longitudinal axon projections of a population of spinal commissural interneurons in young Xenopus tadpoles. We aimed to define how the distribution of axons of the whole population constrains the longitudinal distribution of the inhibition they mediate. Forty-three neurons at different positions were filled intracellularly with biocytin and processed with avidin-conjugated horseradish peroxidase. Soma size did not vary longitudinally and only one ipsilateral axon was found. Contralateral axons ascended, descended, or usually branched to do both. Total axon length and the extent of dendritic arborisation decreased caudally. The distributions of ascending and descending axon lengths were different; there were more long ascending (mean 737 ± standard deviation 365 μm) than long descending (447 ± 431 μm) axons. We used the axon length distribution data with existing data on the distribution of commissural interneuron somata to calculate the overall longitudinal density of these inhibitory axons. Axon numbers showed a clear rostrocaudal gradient. Axon length distributions were then incorporated into a simple spatiotemporal model of the forms of inhibition during swimming and struggling motor patterns. The model predicts that the peak of inhibition on each cycle will decrease from head to tail in both motor patterns, a feature already confirmed physiologically for swimming. It also supports a previous proposal that ascending inhibition during struggling shortens cycle period by shortening rostral motor bursts, whereas descending inhibition could delay subsequent burst onset. J. Comp. Neurol. 400:504–518, 1998. © 1998 Wiley-Liss, Inc.