Central projections of the brachial nerve in bullfrogs: muscle and cutaneous afferents project to different regions of the spinal cord

Adjustments of motor pattern for load compensation via modulated activations of muscle synergies during natural behaviors

It has been suggested that the motor system may circumvent the difficulty of controlling many degrees of freedom in the musculoskeletal apparatus by generating motor outputs through a combination of discrete muscle synergies. How a discretely organized motor system compensates for diverse perturbations has remained elusive. Here, we investigate whether motor responses observed after an inertial-load perturbation can be generated by altering the recruitment of synergies normally used for constructing unperturbed movements. Electromyographic (EMG, 13 muscles) data were collected from the bullfrog hindlimb during natural behaviors before, during, and after the same limb was loaded by a weight attached to the calf. Kinematic analysis reveals the absence of aftereffect on load removal, suggesting that load-related EMG changes were results of immediate motor pattern adjustments. We then extracted synergies from EMGs using the nonnegative matrix factorization algorithm and developed a procedure for assessing the extent of synergy sharing across different loading conditions. Most synergies extracted were found to be activated in all loaded and unloaded conditions. However, for certain synergies, the amplitude, duration, and/or onset time of their activation bursts were up- or down-modulated during loading. Behavioral parameterizations reveal that load-related modulation of synergy activations depended on the behavioral variety (e.g., kick direction and amplitude) and the movement phase performed. Our results suggest that muscle synergies are robust across different dynamic conditions and immediate motor adjustments can be accomplished by modulating synergy activations. An appendix describes the novel procedure we developed, useful for discovering shared and specific features from multiple data sets.

Immunohistochemical localization of calbindin-D28k and calretinin in the spinal cord of Xenopus laevis

Immunohistochemical techniques were used to investigate the distribution and morphology of neurons containing the calcium-binding proteins calbindin-D28k (CB) and calretinin (CR) in the spinal cord of Xenopus laevis and determine the extent to which this organization is comparable to that of mammals. Most CB- and CR-containing neurons were located in the superficial dorsal gray field, but with distinct topography. The lateral, ventrolateral, and ventromedial fields also possessed abundant neurons labeled for either CB or CR. Double immunohistofluorescence demonstrated that a subpopulation of dorsal root ganglion cells and neurons in the dorsal and ventrolateral fields contained CB and CR. By means of a similar technique, a cell population in the dorsal field was doubly labeled only for CB and nitric oxide synthase (NOS), whereas in the ventrolateral field colocalization of NOS with CB and CR was found. Choline acetyltransferase immunohistochemistry revealed that a subpopulation of ventral horn neurons, including motoneurons, colocalized CB and CR. The involvement of CB- and CR-containing neurons in ascending spinal projections was demonstrated combining the retrograde transport of dextran amines and immunohistochemistry. Cells colocalizing the tracer and CB or CR were quite numerous, primarily in the dorsal and ventrolateral fields. Similar experiments demonstrated supraspinal projections from CB- and CR-containing cells in the brainstem and diencephalon. The distribution, projections, and colocalization with neurotransmitters of the neuronal systems containing CB and CR in Xenopus suggest that CB and CR are important neuromodulator substances with functions conserved in the spinal cord from amphibians through mammals.

CB1 cannabinoid receptors in amphibian spinal cord: relationships with some nociception markers

The role of cannabinoids in spinal analgesia has so far been investigated in mammals and the interactions between cannabinoid receptors and markers involved in nociception have been described in the rat spinal cord. An endocannabinoid system is well developed also in the amphibian brain. However, the anatomical substrates of pain modulation have been scarcely investigated in anamniotes, neither is there reference to such a role for cannabinoids in lower vertebrates. In the present paper we employed multiple cytochemical approaches to study the distribution of CB1 cannabinoid receptors and their morphofunctional relationships with some nociception markers (i.e. Substance P, nitric oxide synthase, GABA and mu opioid receptors) in the spinal cord of the anuran amphibian Xenopus laevis. We found a co-distribution of CB1 receptors with the aforementioned signaling molecules, as well as a more limited cellular co-localization, in the dorsal and central fields of the spinal cord. These regions correspond to the mammalian laminae I-IV and X, respectively, areas strongly involved in spinal analgesia. Comparison of these results with those previously obtained in the mammalian spinal cord, reveals a number of similarities between the two systems and suggests that cannabinoids might participate in the control of pain sensitivity also in the amphibian spinal cord.

Sciatic nerve transection decrease substance P immunoreactivity in the lumbosacral spinal cord of the frog (Rana catesbeiana)

Using immunohistochemistry and optical densitometry, substance P (SP) was investigated in the lumbar spinal cord of the frog Rana catesbeiana after sciatic nerve transection. In control animals, there was a high density of SP fibers in the Lissauer's tract and in the mediolateral band of the dorsal gray matter. Other SP immunoreactive fibers were observed in the dorsal part of the lateral funiculus and in the ventral horn. No SP label was found in any cell bodies. After axotomy, SP immunoreactive fibers decreased in the Lissauer's tract on the same side of the lesion. The other regions remained labeled. The changes were observed at 3 days following axonal injury and persisted at 5, 8 and 15 days. At 20 days, there was no significant difference between the axotomized side and the control one, thus indicating a recovery of the SP expression. These results indicate that the frog may be used as a model to study the effects of peripheral axotomy, contributing to elucidate the SP actions in the pain neuropath.

Distribution and ultrastructure of tachykinin-like immunoreactivity in the frog (Rana esculenta) spinal cord, notably, the dorsal horn

Tachykinins are involved in pain transmission at the spinal level. In frog, at least four tachykinins [TK] have been isolated from the brain, but their organization in the dorsal horn of the spinal cord is still poorly known. We have reexamined TK distribution by immunocytochemistry using an antibody recognizing the sequence common to all tachykinins in the spinal cord and dorsal root ganglia of the green frog Rana esculenta. A dense tachykinin-like immunoreactivity (TK-LI) was observed in the dorsolateral fasciculus or Lissauer's tract running ventromedial to the entry of the dorsal root and in numerous small and medium-sized dorsal root ganglion cells showing a primary afferent origin for part of TK-LI of the dorsal horn. The observation of numerous cell bodies in the dorsal horn, in addition, suggested a local or propriospinal origin. One group of cells was localized at the entrance of the Lissauer's tract TK-LI fibers into the dorsal horn, and another group was localized in the upper dorsal horn, a region with a low density of TK-LI fibers. It was suggested that the latter group may correspond to neurokinin B. Electron microscopic examination of the Lissauer's tract showed numerous immunoreactive axons, some located at the center of glomerular-like arrangements, suggesting that the information brought by these fibers may be transmitted and most probably modulated before their entry in the dorsal horn. In conclusion, the functional organization of tachykinins in the frog spinal cord seems to be similar to that of mammals, albeit with a different morphological organization.

Primary and secondary somatosensory projections in direct-developing plethodontid salamanders

In three species of plethodontid salamanders (Plethodon jordani, Hydromantes italicus, and Bolitoglossa subpalmata), primary and secondary somatosensory pathways were investigated by means of tract-tracing in vivo and in vitro using biocytin, horseradish peroxidase, and neurobiotin. Afferent sensory fibers of cranial nerves V, VII, and X and the brachial nerve run in the dorsal funiculus of the medulla oblongata and spinal cord. Fibers ascend to the level of, but do not enter, the cerebellum. In the caudal medulla oblongata, sensory tracts of the cranial nerves descend in a dorsal and a dorsolateral bundle and reach the level of the fourth spinal nerve. Two bundles are likewise formed by spinal afferent fibers, which descend to the level of the seventh spinal nerve. Secondary somatosensory projections ascend in contralateral ventral, contralateral lateral, and ipsilateral lateral tracts, the latter two corresponding to the spinal lemniscal tracts of Herrick. These tracts reach the cerebellum, mesencephalic, and diencephalic targets (tegmentum, torus, tectum, tuberculum posterius, pretectum, and ventral thalamus) ipsi- and contra-laterally. The projection to the tectum is confined to fiber layer 4. Fibers of the ascending tracts cross in the cerebellar and tectal commissure. Our study demonstrates that the ascending secondary somatosensory pathways of plethodontid salamanders differ remarkably from those of other amphibians.

A critical period for the influence of peripheral targets on the central projections of developing sensory neurons

During development, the projections that sensory neurons make within the spinal cord are influenced by the specific targets they contact in the periphery. If sensory ganglia normally supplying principally cutaneous targets are forced to grow into limb muscles, in early stage tadpoles, many sensory neurons within these ganglia innervate limb muscles and subsequently develop spinal projections appropriate for muscle spindle afferents. If the same procedure is performed with adult frogs, however, these novel projections do not form. In this study, we have determined the developmental stages at which this sensitivity to peripheral targets exists. Axons from sensory neurons in thoracic (largely cutaneous) dorsal root ganglia were re-routed into the front leg at various stages through metamorphosis, and the central spinal projections of these re-routed fibers were assessed with HRP labeling. We found that thoracic sensory axons could be made to project to limb muscles throughout development, but that the central projections of these neurons were only appropriate for spindle afferents if the fibers were re-routed before stage XVIII, shortly before metamorphic climax. Because sensory neurons can regenerate specifically into the appropriate spinal laminae even in adult frogs, these results suggest that changes in either the DRG or the arm musculature occur by stage XVII so that DRG neurons cannot respond to novel peripheral targets.

Specificity of identified central synapses in the embryonic cockroach: appropriate connections form before the onset of spontaneous afferent activity

The mechanisms by which neurons recognize the appropriate postsynaptic cells remain largely unknown. A useful approach to this problem is to use a system with a few identifiable neurons that form highly specific synaptic connections. We studied the development of synapses between two identified cercal sensory afferents and two giant interneurons (GIs) in the embryonic cockroach Periplaneta americana. By 46% of embryonic development, the axons of the filiform hair sensory neurons have entered the terminal ganglionic neuropil and grow alongside the GI primary dendrites, although they do not form synapses. From 50% of development, the GI dendrites grow outward from the center of the neuropil to contact the presynaptic axons and their branches. The sensory neurons begin to spike at 52% of development, and, from 55% of development, these action potentials evoked excitatory postsynaptic potentials in the GIs. Synaptic contacts were first seen at this time. The pattern of synaptic connections was highly specific from the outset. G12 had strong input from the medial (M) afferent and had almost negligible input from the lateral (L) afferent, whereas G13 had input from both. This specificity was present before bursts of spontaneous activity began in the sensory neurons at 59% of development. G12 filopodia selectively formed synaptic contacts with the M axon rather than the L axon. The few contacts made by G12 with the L axon had a normal morphology but fewer presynaptic densities. Filopodial insertions were not involved in selective synapse formation. In this system, highly specific synaptic recognition appears to be activity independent.

Anuran dorsal column nucleus: organization, immunohistochemical characterization, and fiber connections in Rana perezi and Xenopus laevis

As part of a research program on the evolution of somatosensory systems in vertebrates, the dorsal column nucleus (DCN) was studied with (immuno)histochemical and tract-tracing techniques in anurans (the large green frog, Rana perezi, and the clawed toad, Xenopus laevis). The anuran DCN contains some nicotinamide adenine dinucleotide phosphate diaphorase-positive neurons, very little calbindin D-28k, and a distinct parvalbumin-positive cell population. The anuran DCN is innervated by primary and non-primary spinal afferents, by primary afferents from cranial nerves V, VII, IX, and X, by serotonin-immunoreactive fibers, and by peptidergic fibers. Non-primary DCN afferents from the spinal cord appear to arise throughout the spinal cord, but particularly from the ipsilateral dorsal gray. The present study focused on the efferent connections of the DCN, in particular the targets of the medial lemniscus. The medial lemniscus could be traced throughout the brainstem and into the diencephalon. Along its course, the medial lemniscus gives off collaterals to various parts of the reticular formation, to the octavolateral area, and to the granular layer of the cerebellum. At mesencephalic levels, the medial lemniscus innervates the lateral part of the torus semicircularis as well as various tegmental nuclei. A striking difference between the two species studied is that while in R. perezi medial lemniscal fibers do not reach the tectum mesencephali, in X. laevis, intermediate and deep tectal layers are innervated. Beyond the midbrain, both dorsal and ventral thalamic areas are innervated by the medial lemniscus. The present study shows that the anuran "lemniscal pathway" is basically similar to that of amniotes.

Distribution and morphology of sacral spinal cord neurons innervating pelvic structures in Xenopus laevis

Relatively little is known about the organization of neural input to pelvic viscera in amphibia. In this study, sacral spinal efferent neurons were labeled in Xenopus laevis frogs by application of horseradish peroxidase (HRP) to the tenth spinal nerve, to pelvic musculature, or to the pelvic nerve. DiI was applied to the pelvic nerve with similar results. Labeled spinal neurons were located in the intermediate gray or in the ventral horn. Neurons in the tenth dorsal root ganglion, but not in the spinal cord, were labeled after application of HRP or DiI to the pudendal nerve. The labeled neurons in the spinal cord intermediate gray were in a position comparable to that of the mammalian sacral parasympathetic nucleus (SPN). Two apparent subdivisions included 1) a medial cluster of cells with mediolaterally oriented dendrites and 2) a lateral group with dorsoventrally oriented dendrites. An intermediate group, not clearly classed with the other two, was also identifiable. In some cases, labeled tenth nerve primary afferents were seen in contact with efferent neurons of the intermediate gray. Labeled neurons in the ventral horn medial to the lateral motor column were small, with dendrites oriented mediolaterally, in a position comparable to that of the mammalian Onuf's nucleus. The peripheral targets of DiI-labeled pelvic nerve axons were the compressor cloaca muscle, cloaca, and bladder. DiI-labeled pudendal nerve axons distributed peripherally to cloacal lip and medial thigh integument. These data suggest that the pudendal nerve in amphibians is purely sensory and that both somatic and autonomic motor axons traverse the pelvic nerve.

Development of specific muscle and cutaneous sensory projections in cultured segments of spinal cord

Development of sensory projections was studied in cultured spinal segments with attached dorsal root ganglia. In spinal segments from stage 30 (E6.5) and older chicken embryos, prelabeled muscle and cutaneous afferents established appropriate projections. Cutaneous afferents terminated solely within the dorsolateral laminae, whereas some muscle afferents (presumably Ia afferents) projected ventrally towards motoneurons. Development of appropriate projections suggests that sufficient cues are preserved in spinal segments to support the formation of modality-specific sensory projections. Further, because these projections developed in the absence of muscle or skin, these results show that the continued presence of peripheral targets is not required for the formation of specific central projections after stage 29 (E6.0). Development of the dorsal horn in cultured spinal segments was assessed using the dorsal midline as a marker. In ovo, this midline structure appears at stage 29. Lack of midline formation in stage 28 and 29 cultured spinal segments suggests that the development of the dorsal horn is arrested in this preparation. This is consistent with earlier reports suggesting that dorsal horn development may be dependent on factors outside the spinal cord. Because dorsal horn development is blocked in cultured spinal segments, this preparation makes it possible to study the consequences of premature ingrowth of sensory axons into the spinal cord. In chicken embryos sensory afferents reach the spinal cord at stage 25 (E4.5) but do not arborize within the gray matter until stage 30. During this period dorsal horn cells are still being generated. In spinal segments, only those segments that have developed a midline at the time of culture support the formation of midline at the time of culture support the formation of specific sensory projections.(ABSTRACT TRUNCATED AT 250 WORDS)

Trigeminal primary afferent projections to the spinal cord of the frog, Rana ridibunda

The distribution in the spinal cord of the trigeminal primary projections in the frog Rana ridibunda was studied by means of the anterograde transport of horseradish peroxidase (HRP). Upon entering the medulla via the single trigeminal root, a conspicuous descending tract that reaches the cervical spinal cord segments is established. This projection arises in the ophthalmic (V1), maxillary (V2), and mandibular (V3) trigeminal nerve subdivisions. In the spinal cord, only a minor somatotopic arrangement of the trigeminal fibers was observed, with the fibers arising in V3 terminating somewhat more medially than those from V1 and V2. A dense projection to the medial aspect of the spinal cord, above the central canal, primarily involves V3. Each trigeminal branch sends projections at cervical levels to the contralateral dorsal field, and those from V2 are most abundant. Bilateral experiments with HRP application show convergence of primary trigeminal and spinal afferents within the dorsal field of the spinal cord. The pattern of arrangement of the trigeminal primary afferent fibers in the spinal cord of this frog largely resembles that of amniotes. However, the organization seems simpler and the slight somatotopic distribution of V1, V2, and V3 fibers is similar to the condition in other anamniotes.

Reticulospinal actions on primary afferent depolarization of cutaneous and muscle afferents in the isolated frog neuraxis

The effects of the brainstem reticular formation on the intraspinal excitability of low threshold cutaneous and muscle afferents were studied in the frog neuraxis isolated together with the right hindlimb nerves. Stimulation of low threshold fibers (less than two times threshold) in cutaneous nerves produced short latency, negative field potentials in the ipsilateral dorsal neuropil (200-400 microns depth) that reversed to positivity at deeper regions (500-700 microns). Stimulation of low threshold fibers (less than two times threshold) in muscle nerves produced, instead, negative response that acquired their maximum amplitude in the ventral neuropil (700-900 microns depth). These electrophysiological findings suggest, in agreement with observations in the cat, that low threshold cutaneous and muscle afferents end at different sites in the spinal cord. Intraspinal microstimulation applied within the dorsal neuropil produced antidromic responses in low threshold cutaneous afferents that were increased in size following stimulation of the dorsal or ventral roots, as well as of the brainstem reticular formation. This increase in excitability is interpreted as being due to primary afferent depolarization (PAD) of the intraspinal terminals of cutaneous fibers. Antidromic responses recorded in muscle nerves following intraspinal stimulation within the ventral neuropil were also increased following conditioning stimulation of adjacent dorsal or ventral roots. However, stimulation of the bulbar reticular formation produced practically no changes in the antidromic responses, but was able to inhibit the PAD of low threshold muscle afferents elicited by stimulation of the dorsal or ventral roots. It is suggested that the PAD of low threshold cutaneous and muscle afferents is mediated by independent sets of interneurons. Reticulospinal fibers would have excitatory connections with the interneurons mediating the PAD of cutaneous fibers and inhibitory connections with the interneurons mediating the PAD of muscle afferents. Although our results provide no direct information on whether the reticulospinal depression of the PAD elicited in low threshold muscle afferents is due to inhibition along the pathways producing PAD of muscle spindle or of tendon organ afferents, it seems likely-by analogy with what has been seen in the cat spinal cord-that these inhibitory actions are mostly restricted to the pathways producing PAD in the terminal arborizations of muscle spindle afferents. These results emphasize the specificity of the descending control of the synaptic efficacy of low threshold cutaneous and muscle afferents which could be of importance for motor performance.

Effects of dorsal root cut on the forces evoked by spinal microstimulation in the spinalized frog

Spinalized frogs were microstimulated in the intermediate grey layers of the lumbar spinal cord; the forces evoked in the hindlimb were measured at several limb positions. The data were expressed as force fields. After the collection of many force fields, the dorsal roots were cut with the stimulating electrode in place, and the position-dependent stimulation-evoked forces were again measured repeatedly. We found that the position-dependent pattern of evoked forces--the force fields--did not change after the dorsal roots were cut. In other words, the postcut evoked forces pointed in the same direction as the precut evoked forces. This result was predicted and confirmed by the muscle activations (EMGs): Before and after the dorsal roots were cut, the same muscles were activated in the same proportions. In all limb positions, the rank ordering of the muscle activations remained fixed. The stimulation needed to evoke forces was increased by deafferentation, and there were subtle changes in the force magnitudes that were consistent with a linearization of the muscle stiffness by the afferents. We conclude that the microstimulation activated specific muscle synergies that resulted in limb forces pointing toward a particular posture. The patterns of evoked forces were predominantly attributable to feedforward activation of these muscle synergies.

Projection patterns of primary sensory neurons studied by transganglionic methods: somatotopy and target-related organization

The anatomical organization of the centrally projecting branches of different peripheral sensory nerves was not possible to investigate efficiently until the development of the axonal tracing methods. Horseradish peroxidase applied peripherally could be visualized in central projection areas provided a sensitive histochemical method was used; this created the basis for transganglionic tracing from the periphery. This has permitted the investigation of large-scale projections from peripheral sensory nerves. The use of conjugates of horseradish peroxidase and lectins with affinities for different populations of primary sensory neurons, as well as the use of different postoperative survival times, has offered the possibility for selective visualization of projections from subsets of primary sensory neurons. For detailed studies of single afferent fiber projections, a combined physiological-anatomical approach using single-unit recording followed by intraaxonal application of horseradish peroxidase, has become the method of choice. This chapter will focus on results which have been achieved by transganglionic tracing methods, in regard to the organization of the central projections of peripheral sensory nerves.

Development of serotoninergic system in the brain and spinal cord of the chick

(1) Development of serotonin positive cells and fibers was immunohistochemically studied by the use of an antibody against serotonin. (2) Serotoninergic neurons were first observed in the immature rohmbencephalon raphe nuclei on embryonic day (E)4, where two clusters of serotonin positive neurons were located: one observed at the rostral part of the rohmbencephalon corresponding to the dorsal raphe nuclei had many serotonin positive cells: the other located at the caudal part of the rohmbencephalon corresponding to the medullary raphe nuclei of the adult animals had only a small number of serotoninergic cells. (3) By E8 the number of serotonin positive cells in the brain stem increased, and virtually all the raphe nuclei found in an adult animal were located. (4) Serotonin positive fibers in the marginal layer reached up to the diencephalon and telencephalon on E6 and E8, respectively. (5) Serotonin positive cells were found beside the midline regions in the ventral part of the spinal cord of the embryonic as well as posthatching chick. (6) Because almost all the serotoninergic fibers in the spinal cord originated from the brain stem raphe nuclei, propriospinal serotonin positive cells were considered as phylogenetic vestiges. (7) Serotoninergic fibers were first found in the marginal layer of the cervical and lumbar spinal cord on E6 and E8, respectively. (8) There was a waiting period of a few days before they penetrated into the mantle layer. (9) Terminal arbolization of the serotoninergic fibers started from late embryonic periods (E16 less than), and was maximized within one week of hatching. (10) Thereafter the density of serotonin positive fibers decreased in all the regions of the spinal cord. (11) Developmental changes of the density of serotonin determined with a high performance liquid chromatography were the same as those determined through immunohistochemistry. Namely the density of serotonin increased linearly from E6 to hatching period, and reached the maximum value one week posthatching. (12( The density of the serotonin in the adult spinal cord was about half of the maximum value. (13) It is to say that the densities of serotonin and serotoninergic fibers transiently increased around one week posthatching. (14) Following the transient increase serotoninergic fibers were eliminated from the neuropil, the fibers were localized in the specific regions of the motor nucleus: motor neuron pools of extensor muscles of the hip joint in the lumbosacral spinal cord.(ABSTRACT TRUNCATED AT 400 WORDS)

Alternatives to the use of mammals for pain research

The study of pain and analgesia is an important area of biomedical research which has led to a number of significant advances in the treatment of acute and chronic pain in the clinic. This area of research examines the physiology of pain transmission and the pharmacology of analgesic drugs by employing a variety of in vitro and in vivo animal models. To date, the vast majority of in vivo models for pain research have used mammalian species, primarily rodents and, to a lesser extent, canines, felines, and primates. The present review summarizes the special considerations of animal use in pain research and the philosophic and scientific basis for developing adjunct models using lower vertebrates. Existent literature on pain research using non-mammalian vertebrates is reviewed, with a special focus on amphibian species. Given the ethical concerns of experimental animal use and the importance of a comparative approach to the basic understanding of pain-processing, the further development of non-mammalian models for pain research should be encouraged.

Synapses on motoneuron dendrites in the brachial section of the frog spinal cord: a computer-aided electron microscopic study of cobalt-filled cells

Cobalt-labelled motoneuron dendrites of the frog spinal cord at the level of the second spinal nerve were photographed in the electron microscope from long series of ultrathin sections. Three-dimensional computer reconstructions of 120 dendrite segments were analysed. The samples were taken from two locations: proximal to cell body and distal, as defined in a transverse plane of the spinal cord. The dendrites showed highly irregular outlines with many 1-2 microns-long 'thorns' (on average 8.5 thorns per 100 microns 2 of dendritic area). Taken together, the reconstructed dendrite segments from the proximal sites had a total length of about 250 microns; those from the distal locations, 180 microns. On all segments together there were 699 synapses. Nine percent of the synapses were on thorns, and many more close to their base on the dendritic shaft. The synapses were classified in four groups. One third of the synapses were asymmetric with spherical vesicles; one half were symmetric with spherical vesicles; and one tenth were symmetric with flattened vesicles. A fourth, small class of asymmetric synapses had dense-core vesicles. The area of the active zones was large for the asymmetric synapses (median value 0.20 microns 2), and small for the symmetric ones (median value 0.10 microns 2), and the difference was significant. On average, the areas of the active zones of the synapses on thin dendrites were larger than those of synapses on large calibre dendrites. About every 4 microns 2 of dendritic area received one contact. There was a significant difference between the areas of the active zones of the synapses at the two locations. Moreover, the number per unit dendritic length was correlated with dendrite calibre. On average, the active zones covered more than 4% of the dendritic area; this value for thin dendrites was about twice as large as that of large calibre dendrites. We suggest that the larger active zones and the larger synaptic coverage of the thin dendrites compensate for the longer electrotonic distance of these synapses from the soma.

Potassium channel blockade differentially affects the relative refractory period of frog afferent terminals and axons

1. The effects of potassium channel blockade on afferent axons and terminal regions in frog dorsal roots and spinal cords, respectively, were investigated in vitro. 2. A condition-test (C-T) protocol was used to assess the population relative refractory period. Characteristics of main axons were evaluated by stimulation at the proximal end of transected dorsal roots (DR). Characteristics of terminal regions were tested by stimulation at the base of the dorsal horn (DH). 3. DH recovery of excitability was delayed by low concentrations of 4-aminopyridine (4-AP) and tetraethylammonium (TEA) alone or combined. The same treatments did not affect recovery to DR stimulation. 4. DH recovery of excitability was not delayed by solutions suppressing terminal calcium influx. 5. We conclude that sensitivity of the relative refractory period to potassium channel blocking agents differs between main axons and axon terminal regions. This may indicate differences between axon terminals and main axons in the mechanism of action potential repolarization. 6. We hypothesize that rapid action potential repolarization by pharmacologically sensitive potassium channels in presynaptic terminal regions keeps terminal action potentials short. Terminal action potential brevity would limit calcium influx, thus preventing terminal calcium overload but contributing to transmission failures at spinal synapses.

Axonal guidance of adenosine deaminase immunoreactive primary afferent fibers in developing mouse spinal cord

This study examined the precision of central fiber growth in a subpopulation of dorsal root ganglion neurons in developing mouse spinal cord. Immunohistochemical techniques using a monospecific, polyclonal antiserum to mouse adenosine deaminase (ADA) were utilized to label a population of primary sensory afferents that have been found to exclusively innervate laminae I and II of the dorsal horn in adult mice. Initial growth of ADA-immunoreactive (ADA-IR) primary afferents occurred very early in development, embryonic day 10 (E10), a time coincident with the earliest settling time of dorsal root ganglion neurons. Adenosine deaminase immunoreactive primary afferents were observed throughout the cross-sectional area of the primordial dorsal funiculus (DF) as early as E10. Immunostained fibers remained quiescent in the DF during its growth and separation into the tract of Lissauer and dorsal column pathway. By E15, the two pathways had formed and ADA-IR fibers were observed exclusively in the tract of Lissauer. This segregation of fibers remained throughout development and reflected the adult pattern. Growth was reinitiated at E16 when the fibers advanced into the dorsal horn and proceeded directly to laminae I and II mimicking their adult distribution. Exuberant fiber growth was not detected throughout their development. These results strongly suggest that ADA-IR fibers exhibit precise fiber guidance to a preferred pathway, the tract of Lissauer, and accurate laminar innervation of the dorsal horn.

Specification of synaptic connections between sensory and motor neurons in the developing spinal cord

Experimental studies of mechanisms underlying the specification of synaptic connections in the monosynaptic stretch reflex of frogs and chicks are described. Sensory neurons innervating the triceps brachii muscles of bullfrogs are born throughout the period of sensory neurogenesis and do not appear to be related clonally. Instead, the peripheral targets of these sensory neurons play a major role in determining their central connections with motoneurons. Developing thoracic sensory neurons made to project to novel targets in the forelimb project into the brachial spinal cord, which they normally never do. Moreover, these foreign sensory neurons make monosynaptic excitatory connections with the now functionally appropriate brachial motoneurons. Normal patterns of neuronal activity are not necessary for the formation of specific central connections. Neuromuscular blockade of developing chick embryos with curare during the period of synaptogenesis still results in the formation of correct sensory-motor connections. Competitive interactions among the afferent fibers also do not seem to be important in this process. When the number of sensory neurons projecting to the forelimb is drastically reduced during development, each afferent still makes central connections of the same strength and specificity as normal. These results are discussed with reference to the development of retinal ganglion cells and their projections to the brain. Although many aspects of the two systems are similar, patterned neural activity appears to play a much more important role in the development of the visual pathway than in the spinal reflex pathway described here.

Spatial distribution of pre- and postsynaptic sites of axon terminals in the dorsal horn of the frog spinal cord

Axon terminals which could be interpreted as dorsal root boutons, were photographed from a series of 98 ultrathin sections with a Jeol 100B electron microscope. A total of 13 boutons were recovered for computer reconstruction. Two of them were terminal boutons, eight en passant boutons and three boutons were only partially recovered. All boutons contained multiple synaptic sites (maximum 33 and minimum seven) at which axodendritic and axoaxonic synapses were established. Axodendritic synapses were of the asymmetric type and they were directed toward adjacent dendrites. In axoaxonic synapses, which were of the symmetric type, the boutons were invariably on the postsynaptic side. Among the presynaptic profiles axons with spherical and pleomorphic vesicles and dendrites with flattened vesicles could be discerned. On average, each 2.67-microns2 bouton surface area contained one presynaptic site at which an axodendritic synapse was established, and each 7-microns2 surface area contained one postsynaptic site for an axoaxonic (or dendroaxonic) contact. A tendency of grouping of synaptic sites was observed. Distance measurements between the closest neighbours of all synaptic sites were made in four combinations in boutons with the original and with a random distribution of synaptic sites. The arithmetic mean of distances measured between the presynaptic and the closest postsynaptic sites was almost twice as big as that measured in the reverse direction. The difference between these values became greatly reduced in the case of random distribution. The arithmetic mean of distances between the closest neighbours of presynaptic sites was about the same as that between the closest neighbours of postsynaptic sites. This latter value was considerably increased with randomly distributed synaptic sites. The results suggest a non-random distribution of synaptic sites on the surface of boutons. The analysis of cluster formation of synaptic sites performed with a numerical taxonomy technique revealed that the majority of the 153 synaptic sites were comprised in 27 clusters containing both pre- and postsynaptic sites within the 1-micron similarity level. All postsynaptic sites were within 1 micron of one or more presynaptic sites. On the basis of the assumption that the postsynaptic sites are occupied by inhibitory axoaxonic synapses, it is suggested that the transmitter release from the presynaptic sites can be individually controlled in this structural arrangement. A probable mechanism of this function may be the passive invasion of the bouton by the impulse propagating actively along the dorsal root fibre.

Anatomical specificity of regenerated muscle sensory afferents in the spinal cord of the bullfrog

The specificity of central projections made by regenerated muscle sensory fibers in the brachial spinal cord was studied with anatomical tracing methods. Sensory fibers were interrupted by freezing dorsal roots in postmetamorphic bullfrogs. After several months, regenerated sensory fibers were labeled with horseradish peroxidase applied to the triceps brachii muscle nerve, and their arborizations within the spinal cord were reconstructed from serial cross sections. Most of the regenerated projections from triceps muscle sensory afferents ended in or near their normal terminal field. A few branched and appeared to terminate more dorsally than normal, however, sometimes within the region where cutaneous afferents normally terminate. In contrast to the normal pathway followed by muscle afferents within the spinal cord, many regenerated afferents grew along the circumference of the spinal cord, just under the pial surface, and then turned abruptly toward the midline and into their appropriate terminal region. This suggests that regenerating afferents may actively seek out their appropriate targets and are not simply passively guided to them.

Development of spinocerebellar afferents in the clawed toad, Xenopus laevis

The development of spinocerebellar projections in the clawed toad, Xenopus laevis, was studied with horseradish peroxidase as an anterograde and retrograde tracer. Early in development cells of origin of spinocerebellar projections were found, contralaterally, in or close to the medial motor column. In older tadpoles ipsilaterally projecting spinal neurons were also labeled from the cerebellum. These are virtually indistinguishable from the large primary motoneurons that occupy a very similar position in the spinal cord. Most of the labeled spinal cells were found in the thoracic spinal cord; they lie halfway between the brachial and lumbar secondary motor columns. Surprisingly, no primary spinocerebellar projection arising from dorsal root spinal ganglion cells could be demonstrated in X. laevis tadpoles and adult toads. Therefore, fibers in the cerebellum that were labeled anterogradely from the spinal cord can be expected to originate exclusively from the secondary spinocerebellar tract cells. These fibers appear to cross the cerebellum in or at the border of the granular layer. The present data suggest that in X. laevis early in the development of the cerebellum a distinct secondary spinocerebellar projection is already present, originating in neurons that can be compared with the "spinal border cells" in mammals. The relative sparseness of this secondary spinocerebellar projection and the apparent absence of primary spinocerebellar afferents probably indicate that spinocerebellar pathways are only of minor importance in X. laevis. The possibility remains, however, that the expansion of the secondary spinocerebellar pathway only starts when metamorphosis has been completed.

Immunohistochemical localization of substance P, somatostatin, enkephalin, and serotonin in the spinal cord of the northern leopard frog, Rana pipiens

Using the indirect antibody peroxidase-antiperoxidase method of Sternberger, we localized substance P (SP), somatostatin (SOM), enkephalin (ENK), and serotonin (5HT, 5-hydroxytryptamine) in the spinal cord of Rana pipiens. This is the first study to demonstrate all four substances in adjacent sections of frog spinal cord. The distribution patterns of ENK, SP, SOM, and 5HT in our study differ from that described for laminae I and II in amniotes. A high density of ENK, SP, and SOM fibers is present in a band ventral to the dorsal terminal field of cutaneous primary afferent fibers and slightly overlapping the ventral terminal field of muscle primary afferent fibers. However, a high density of 5HT fibers is present in the dorsal terminal field.

Specificity of sensory projections to the spinal cord during development in bullfrogs

Sensory neurons in dorsal root ganglia of frogs project to areas of the spinal cord they do not normally innervate following removal of adjacent ganglia at tadpole stages (Frank and Westerfield, J. Physiol. (Lond.) 324:495-505, '82b). A possible explanation of this phenomenon is that sensory neurons project to wider areas of the spinal cord in tadpoles than in adult frogs and that partial deafferentation causes the retention of these widespread projections. Therefore, the specificity of sensory projections to the spinal cord in tadpoles was assessed by staining individual dorsal roots with horseradish peroxidase. Thoracic sensory neurons project to thoracic segments of the spinal cord and to the brainstem in tadpoles, like thoracic sensory neurons in adult frogs. They rarely arborize in the brachial region even at stages when no other sensory fibers arborize at this level. Furthermore, their projections are restricted to the dorsal horn at all stages. Conversely, hypoglossal sensory neurons, which project into the intermediate gray matter in the adult, also project to this area in tadpoles. The finding that sensory neurons in tadpoles only project to areas of the spinal cord that they innervate in the adult suggests that the novel projections observed following partial deafferentation of the spinal cord are actually induced by the operation. An additional finding was that forelimb afferents, which project to an area extending from the obex to midthoracic levels in adult frogs, arborize at rostral spinal levels and at thoracic levels several stages before they form projections to the region around their own dorsal root. These differences in the stages at which projections to different levels of the spinal cord develop suggest that local properties of the spinal cord may control the timing of sensory fiber arborization.

The development of primary afferents to the lumbar spinal cord in Xenopus laevis

The development of dorsal root ganglia (DRG) and the primary afferent system of the hindlimb was studied during metamorphosis in Xenopus laevis larvae. The first DRG cells appeared at stage 40 and at stage 48 the first primary afferent fibers entered the lumbar spinal cord where they bifurcated into ascending and descending branches. Primary afferent fibers and the dendrites of the secondary motoneurons contacted first in a lateral (state 56) and later (stage 58) in a dorsomedial terminal field. Reflexogenic hindlimb behaviour at stage 56 concurred with the presence of the lateral terminal field and many unipolar dorsal root ganglion cells.

Spinal cord projections from hindlimb muscle nerves in the rat studied by transganglionic transport of horseradish peroxidase, wheat germ agglutinin conjugated horseradish peroxidase, or horseradish peroxidase with dimethylsulfoxide

The spinal cord projections of four different groups of hindlimb muscle nerve branches--the medial and lateral gastrocnemius nerves, muscle branches of the deep peroneal nerve, muscle branches of the femoral nerve, and a nerve to the hamstring muscles--were studied with transganglionic transport of horseradish peroxidase (HRP) in the rat. The influence of varying the postoperative survival (3, 6, and 10 days) and of using wheat germ agglutinin-HRP conjugate (WGA-HRP), or HRP with dimethylsulfoxide (DMSO) instead of free HRP was studied for the gastrocnemius nerves. After 3 days' survival following application of HRP to the gastrocnemius nerves, fine granular labeling was found mainly in lamina V in L4-5, and coarse granular labeling was found in Clarke's column as far caudally as L2, and in laminae VI and VII predominantly in Th12-L2. After 6 or 10 days' survival, the fine labeling in lamina V was sparse or absent, whereas the coarse labeling appeared to remain or to be only slightly reduced in Clarke's column and in laminae VI and VII. No labeling suggestive of terminals was observed in laminae I-III from the gastrocnemius nerves. Except for sparse labeling in lamina I in some of the cases and some minor differences rostrocaudally, the spinal distribution of labeling was similar to that from the other nerves investigated. The distribution of labeling obtained after application of WGA-HRP or HRP with DMSO to the gastrocnemius nerves was very similar to that obtained with free HRP after 3 days' survival. The results indicate that the spinal cord projections of hindlimb muscle nerves in the rat distribute mainly in the deep part of the dorsal horn and in the intermediate zone. Furthermore, the lack of labeling suggestive of terminals in laminae I-III from the gastrocnemius nerves suggests, in conflict with earlier findings in the cat, that primary afferent fibers from muscles do not necessarily terminate in these laminae in the rat. The results suggest, furthermore, that fine granular labeling found in lamina V represents fine-calibered afferent fibers. Finally, the similar spinal projection patterns of the different muscle nerves investigated suggest either a less developed or an essentially different somatotopic organization for muscle afferents compared to cutaneous afferents, as revealed in earlier studies.

Development of synaptic connections between muscle sensory and motor neurons: anatomical evidence that postsynaptic dendrites grow into a preformed sensory neuropil

The anatomical development of muscle sensory arbors and dendrites of brachial motoneurons in the spinal cord of the bullfrog was studied by labeling both types of cells with horseradish peroxidase. Sensory and motoneurons were labeled in tadpoles (stages XV-XVIII) by backfilling the triceps nerve in vivo with HRP throughout the stages in development when functional monosynaptic connections between these cells are first being formed. Individual triceps motoneurons were injected with HRP in other tadpoles at the same developmental stages. By stage XV, triceps sensory afferents already projected to and arborized in the ventral sensory neuropil region of the spinal cord where sensory-motor connections are made. In contrast, the dendrites of triceps motoneurons rarely were present in this region until stage XVI. By stage XVII, triceps dendrites in this region were common and they intermingled with the collaterals of muscle sensory axons. Thus, sensory axons supplying limb muscles grow into the future neuropil region well in advance of the arrival of motoneuronal dendrites. Electrophysiological studies have shown that the connections between triceps sensory and motor neurons are already specific at stage XVII, as soon as monosynaptic potentials between these cells can be detected (Frank and Westerfield: J. Physiol. (Lond.) 343:593-610, '83). The present anatomical results demonstrate that the processes of sensory and motor cells are not in close anatomical proximity before this time; thus the selection of appropriate synaptic partners must occur from the outset.

Reformation of specific synaptic connections by regenerating sensory axons in the spinal cord of the bullfrog

The regrowth of sensory axons into the spinal cord of juvenile bullfrogs was studied after disruption of these fibers in the dorsal root. Within 9 d after the root had been frozen, regenerating sensory axons had reached the spinal cord, as revealed by labeling with horseradish peroxidase. Growth into the spinal cord, however, was much slower. Even several months after denervation, very few fibers had reestablished any of their normal longitudinal projections within the dorsal funiculus. Eventually, however, sensory axons grew across the region and into the dorsal horn. Intracellular recordings from motoneurons revealed that these axons made functional reconnections with spinal neurons. Muscle sensory axons established direct, monosynaptic inputs to motoneurons, whereas cutaneous fibers innervated these neurons polysynaptically. Moreover, sensory afferents from a particular muscle distinguished among different classes of motoneurons, just as in normal frogs. Thus, specific synaptic pathways can be reestablished by regenerating sensory axons if they can reach their appropriate target region within the spinal cord.

Dorsal root axonal regeneration in the adult frog spinal cord. A model of vertebrate CNS regeneration

The frog dorsal root provides a useful model for the study of axonal regeneration in an adult vertebrate CNS. We have used the model to compare the regeneration of two very different types of axons within the same CNS environment and have found that regenerating dorsal root, as well as rerouted motoneuron axons, display similar growth patterns in the spinal cord. Both sensory and motor axons grow preferentially in some regions and not in others. They both regenerate effectively longitudinally as well as radially within the dorsolateral fasciculus (DLF). By contrast, fewer sensory and motor axons regenerate longitudinally or radially in the dorsal funiculus (DF). This similar preferential growth of two very different populations of axons suggests that the growth patterns reflect regional differences in the cellular environment of the cord. The DLF has fascicles of unmyelinated axons separated by radial glial processes and, after dorsal root injury, is mildly gliotic. By contrast, DF has very large myelinated axons, which widely separate the radial glial processes that traverse the region. After dorsal root injury, this region is markedly gliotic and contains myelin, debris and oligodendroglia, and microglial macrophages. Our data suggest that unmyelinated axons and radial glial processes are more preferred substrates for axonal growth than myelin debris, oligodendroglia and macrophages. It is not surprising, then, that regions of the adult mammalian CNS that are characterized by large myelinated axons fail to support axonal growth. Moreover, there is some evidence that regions of the adult mammalian CNS that are characterized by unmyelinated axons support axonal growth.

Restoration of neuromuscular specificity following ventral rhizotomy in the bullfrog tadpole, Rana catesbeiana

The specificity of hindlimb reinnervation following transection of lumbar ventral roots was investigated in adult and larval bullfrogs (Rana catesbeiana). Five to 6 weeks following ventral rhizotomy, the retrogradely transported marker horseradish peroxidase (HRP) was applied to circumscribed regions of the hindlimb. The location of labeled motoneuron somata within the lumbar lateral motor column was compared with that obtained in unoperated tadpoles. Reinnervation of the hindlimb was largely specific in tadpoles operated during the first third of larval life. However, localization was largely lost in older tadpoles and adult frogs. Repeated applications of 3H-thymidine combined with retrograde labeling with HRP failed to provide evidence that newly born motoneurons contribute to reinnervation of the hindlimb. Hindlimb reinnervation thus appears to result from regeneration of transected motor axons. In contrast to the lack of neuromuscular specificity seen in older animals after transection of ventral roots, motoneuron axons disconnected from their targets by crush injury regenerate to the appropriate limb regions.

The organization of the motoneurons innervating the axial musculature of vertebrates. II. Florida water snakes (Nerodia fasciata pictiventris)

The motor pools of axial muscles in Florida water snakes (Nerodia fasciata pictiventris) were studied by applying horseradish peroxidase (HRP) to branches of spinal nerves innervating individual muscles or groups of muscles. Motor pools of different muscles or muscle groups were located in characteristic positions in both the transverse and the longitudinal extent of the motor column. Epaxial pools were located ventromedially in the column, segregated from most hypaxial ones, which were dorsolateral. The only exception to this general rule was the motoneurons innervating the levator costae muscle. Some of the motoneurons innervating this hypaxial muscle were located in the ventral part of the motor column, like epaxial motoneurons, but they were segregated longitudinally from epaxial ones. The arrangement of the motor pools was strikingly similar to the motor pools of presumptive homologous muscles in rats (Smith and Hollyday: J. Comp. Neurol. 220:29-43, '83), even though the locomotor mechanics in the two animals are very different. The similarities may reflect a comparable relationship between the location of motoneurons in the motor column and the location, in embryonic life, of the muscles they innervate. They also suggest that differences in the locomotor mechanics in the two species are accomplished without any dramatic reorganization of the medial motor column, in marked contrast to the substantial reorganization necessary to account for differences in the motor columns of amniotes and anamniotes.

Regeneration of motoneuron axons into the adult frog spinal cord after ventral-to-dorsal-root anastomosis

Motoneuron axons routed into the adult frog spinal cord via a ventral-to-dorsal-root anastomosis regenerated into the white and the gray matters. The distribution, growth patterns, and arborizations of regenerated ventral root axons were compared to those of regenerated dorsal root axons within the same environment. Within the spinal white matter, regenerating ventral root axons behaved very similarly to regenerating dorsal root axons. Here, the regenerating ventral root axons grew longitudinally beneath the pia and radially toward the spinal gray matter, particularly within the dorsolateral fasciculus. The location of the regenerating axons and the patterns of their growth within the white matter suggest that glial endfeet and radial glial processes play a major role in the determination of these axonal growth patterns. When motor axons entered the gray matter, their arborizations were very similar to those of regenerated dorsal root axons, suggesting that these two very distinct populations of axons respond similarly to local cues within the spinal gray matter. One difference between the arborizations of these two populations of axons was the relative number of varicosities along axonal branches. Regenerated motoneuronal arborizations within the spinal gray matter had fewer en passant varicosities than regenerated dorsal root axonal arborizations. This difference may reflect the synaptogenetic response of the two types of axons to targets within the gray matter. The low number of en passant varicosities associated with the ventral root axonal aborizations suggests that these axons do not synapse with all available targets and that the rules governing synaptic specificity during development may apply during regeneration in the adult frog spinal cord.

The development of the relationship between dorsal root afferents and motoneurons in the larval bullfrog spinal cord

The relationship of dorsal root afferents to motoneuron somata and dendrites was studied by labelling dorsal and ventral roots of the tadpole lumbar enlargement with HRP at different stages of hindlimb development. Procedures were used which allowed for sequential light and electron microscopic analysis to determine whether close appositions between labelled elements represented synaptic contacts. Lateral motor column (LMC) motoneuron dendrites grow first into the lateral funiculus, and later begin arborizing within the spinal gray, concurrent with the arrival of developing dorsal root afferent fibers. Mature-appearing synaptic contacts between dorsal root afferents and motoneuron dendrites are established first on distal dendrites, and are observed on progressively more proximal dendrites as hindlimb development proceeds. Migrating motoneurons were also labelled in some animals. Distinct dorsal and ventral migratory pathways were noted; cells migrating dorsally were contacted by developing dorsal root afferents. Migrating motoneurons were associated with radially oriented processes, and were often closely apposed to other cells. The coincident development of dorsal root projections and the motoneuron dendrites which these fibers innervate in the adult, as well as the interaction between these two systems during cell migration, suggest that these two systems may be interdependent in establishing their normal relationship during development.

Regeneration of lumbar dorsal root axons into the spinal cord of adult frogs (Rana pipiens), an HRP study

Lumbar dorsal roots of adult frogs were crushed or cut and reanastomosed. Following survival times of up to 75 days, the regenerating dorsal roots were recut and anterogradely injury-filled with horseradish peroxidase. This revealed that in the adult frog, regenerating axons re-enter the spinal cord. Comparison of the distribution of these axons with that of normal dorsal root axons showed that there is a partial restoration of the segmental distribution in the gray matter. However, the long ascending sensory tract of the dorsal funiculus was not restored. The dorsal funiculus was markedly gliotic and had relatively few labelled, regenerated axons. The labelled axons that were seen in the dorsal funiculus either extended longitudinally for a distance just beneath the pia, apparently in association with the glia limitans, or traversed the region to enter the dorsal gray matter. Most of the large and small diameter axons that entered the gray matter did so by passing through the region of the dorsolateral fasciculus. Within the gray matter, small diameter, regenerated axons arborized in the region of the dorsal terminal field, a region that has been shown in the normal frog to receive cutaneous afferents only. Many large diameter axons, presumably muscle afferents, arborized in the ventral terminal field, a region shown in the normal frog to receive muscle afferents exclusively. However, many of these large diameter axons had arborizations that extended to both terminal fields, thus suggesting that some abberant connections are made during dorsal root regeneration in the adult frog.

Morphology of primary afferents to the spinal cord of the turtle Pseudemys scripta elegans

The morphology of primary afferents to the spinal cord of the turtle Pseudemys scripta elegans was studied by means of intra-axonal injections of horseradish peroxidase. A total of 74 collaterals arising from 34 different afferents in 22 animals was investigated. Within this sample, a division into three morphologically distinct collateral types appeared possible. Collaterals of the same parent axon could always be classified to the same type. Type A collateral arborizations could be found within area I-II and III of the spinal grey matter. The number of presynaptic boutons per collateral varied considerably. However, collaterals of the same parent axon usually possessed a similar general appearance. Type B collaterals terminated within area IV and V-VI. The general shape and number of boutons could differ considerably between collaterals of different parent fibers but also between collaterals of the same axon. Type C collaterals formed terminal arborizations in the lateral parts of areas IV, V, VI and VII-VIII and demonstrated a fair constancy in general appearance and number of presynaptic boutons. Type A collaterals are thought to be derived from fibers innervating various cutaneous receptors. Terminal arborizations of type C collaterals are fully overlapping with the dorsal dendritic trees of turtle lumbar motoneurons. It is suggested that type C collaterals form contacts with these motoneurons and arise from muscle spindle innervating afferents. The origin of type B collaterals is less clear, attractive possibilities may be found in joint and/or tendon organs.