Combination of cholinesterase staining of nerves and stereoscopic viewing for three-dimensional study of skin innervation on whole mounts.

A histochemical staining technique for cholinesterases is reported for the visualization of skin innervation on whole mounts in the chick. A method for realization of stereoscopic pictures showing the nerve fiber pattern in the full thickness of preparations is described. The combination of these 2 techniques allows an excellent three-dimensional demonstration of skin innervation.

Immunohistochemical demonstration of vasopressin nerve fibers in the cerebral artery.

Vasopressin-immunoreactive nerve fibers were demonstrated in the cerebral pial arteries by peroxidase immunohistochemistry. In the large pial artery (proximal part of the middle cerebral artery), they ran longitudinally to the long axis of the vessel. They ran in a spiral pattern in the distal part of the middle cerebral artery. Even in small arteries, vasopressin nerve fibers were found arranged in a longitudinal fashion. The present morphological data suggest that vasopressin nerve fibers in the cerebral artery may play a role in cerebral circulation.

Retrograde tracing of nerve fibers to the rat middle cerebral artery with true blue: colocalization with different peptides.

The origin of nerve fibers to the rat middle cerebral artery was studied by retrograde tracing with the fluorescent tracer True Blue (TB) in combination with immunocytochemistry to known perivascular peptides. Application of TB to the middle cerebral artery labeled nerve cell bodies in the ipsilateral superior cervical ganglion, the otic ganglion, the sphenopalatine ganglion, the trigeminal ganglion, and the cervical dorsal root ganglion at level C2. A few labeled nerve cell bodies were seen in contralateral ganglia. Judging from the number and intensity of the labeling, the superior cervical ganglion and the trigeminal ganglion and dorsal root ganglion at level C2 contributed most to the innervation. A moderate number of nerve cell bodies were labeled in the sphenopalatine and otic ganglia. The TB-labeled nerve cell bodies were further examined for the presence of neuropeptides. For that purpose antibodies raised against neuropeptide Y (NPY), vasoactive intestinal polypeptide (VIP), substance P (SP) and calcitonin gene-related peptide (CGRP) were used. A considerable portion of the TB-labeled nerve cell bodies in the superior cervical ganglion contained NPY. About half of the labeled nerve cell bodies in the sphenopalatine and otic ganglia contained VIP. In the trigeminal ganglion and in the dorsal root ganglion at level C2, one-third of the TB-labeled nerve cell bodies were CGRP-immunoreactive, while only few nerve cell bodies contained SP. The study provides direct evidence for the origin of cerebrovascular peptidergic nerve fibers and demonstrates that not only ipsilateral but also contralateral ganglia contribute to the innervation of the cerebral circulation.

Origins and pathways of cerebrovascular vasoactive intestinal polypeptide-positive nerves in rat.

In order to clarify the origins and pathways of vasoactive intestinal polypeptide (VIP)-containing nerve fibers in cerebral blood vessels of rat, denervation experiments and retrograde axonal tracing methods (true blue) were used. Numerous VIP-positive nerve cells were recognized in the sphenopalatine ganglion and in a mini-ganglion (internal carotid mini-ganglion) located on the internal carotid artery in the carotid canal, where the parasympathetic greater superficial petrosal nerve is joined by the sympathetic fibers from the internal carotid nerve, to form the Vidian nerve. VIP fiber bridges in the greater deep petrosal nerve and the internal carotid nerve reached the wall of the internal carotid artery. Two weeks after bilateral removal of the sphenopalatine ganglion or sectioning of the structures in the ethmoidal foramen, VIP fibers in the anterior part of the circle of Willis completely disappeared. Very few remained in the middle cerebral artery, the posterior cerebral artery, and rostral two-thirds of the basilar artery, whereas they remained in the caudal one-third of the basilar artery, the vertebral artery, and intracranial and carotid canal segments of the internal carotid artery. One week after application of true blue to the middle cerebral artery, dye accumulated in the ganglion cells in the sphenopalatine, otic and internal carotid mini-ganglion; some of the cells were positive for VIP. The results show that the VIP nerves in rat cerebral blood vessels originate: (a) in the sphenopalatine, and otic ganglion to innervate the circle of Willis and its branches from anterior and caudally and (b) from the internal carotid mini-ganglion to innervate the internal carotid artery at the level of the carotid canal and to some extent its intracranial extensions.

Reconstruction of trajectory of primary afferent collaterals in the dorsal horn of the cat spinal cord, using Golgi-stained serial sections.

Using Golgi-stained serial sections obtained at the sacro-caudal levels of the cat spinal cord, it was possible to reconstruct the trajectory of primary afferents. They were classified into two groups: reliable primary afferents directly traced from the dorsal root and probable primary afferents traced from the dorsal funiculus or Lissauer's tract. The diameters of the reliable primary afferents vary from 0.88-1.88 mum. According to their courses, reliable primary afferents as well as probable primary afferents were classified into three groups: the first is distributed to both medial and lateral halves of the dorsal horn, the second to the medial half, and the third to the lateral half. Commissural fibers were also observed among the probable primary afferents. The rostro-caudal and medio-lateral extents of reliable primary afferents are found to be between 250 and 950 mum and 270 and 700 mum respectively, while those of the probable primary afferents were between 125 and 670 mum and 270 and 1,640 mum respectively. These primary afferent fibers are connected with at least two or more laminae of the dorsal horn gray matter.

Separate origins for the dynorphin and enkephalin immunoreactive fibers in the inferior mesenteric ganglion of the guinea pig.

By different denervation procedures the origin of dynorphin-(1-17) and enkephalin immunoreactive fibers in the guinea pig inferior mesenteric ganglion was investigated. It was found that the dynorphin-(1-17)-positive fibers reached the ganglion predominantly via the colonic nerves and to a lesser extent via the hypogastric and intermesenteric nerves whereas the enkephalin-positive fibers reached the ganglion via the lumbar splanchnic nerves. These findings show that the dynorphin-(1-17) and enkephalin systems are separate in this ganglion.

Somatotopic organization of the thoracic spinal nerve in the dorsal horn demonstrated with transganglionic degeneration.

Transganglionic degeneration in primary sensory neurons (TGD) has been studied with the Fink-Heimer and cupric silver methods in the adult rat after transection of thoracic spinal nerve branches. Degeneration was found in the ipsilateral dorsal horn after 13 to 53 days postoperative survival. It was observed in laminae III and IV, inconsistently in lamina I, but not in lamina II (substantia gelatinosa). After transection of the dorsal ramus of the spinal nerve degeneration was seen in the lateral third of the laminae. Transection of the intercostal nerve (ventral ramus) at about the costal angle gave rise to degeneration in the medial two-thirds and transection more distally, at about the mid-clavicular line, in the most medial part of the dorsal horn. The results are in accordance with previous anatomical and physiological studies of the somatotopical organization of the dorsal horn. They clearly show the usefulness of TGD as a tool for analysis of central projections of primary sensory neurons.

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.

Sprouting in the avian brainstem auditory pathway: dependence on dendritic integrity.

The brainstem auditory pathway of the chicken were used to examine the relationship between the maintenance of dendrites following denervation and the successful reinnervation (sprouting) by surviving afferents. In the system the third-order cells in n. laminaris receive spatially segregated binaural innervation from n. magnocellularis. Afferents from the ipsilateral n. magnocellularis innervate the dendrites on the dorsal aspect of n. laminaris cells, while afferents from contralateral magnocellular neurons innervate ventral dendrites via the crossed dorsal cochlear tract. Denervation of the ventral dendrites of n. laminaris cells by transection at the midline results in rapid and severe atrophy of the denervated dendrite. Unilateral cochlea removal induces transneuronal degeneration of 30-45% of the ipsilateral magnocellular cells, thereby partially denervating one dendrite of the n. laminaris cells on each side of the brain. In animals with long-standing transections of the crossed dorsal cochlear tract there is no evidence of sprouting the fibers from the ipsilateral n. magnocellularis when the projections of the surviving magnocellular neurons are traced with degeneration methods after a secondary cochlea removal. However, when dendrites of n. laminaris are partially denervated dendrites do not disappear. Furthermore, secondary lesions of the crossed dorsal cochlear tract or secondary cochlea removal reveal that these denervated dendrites are reinnervated by the afferents from the opposite n. magnocellularis which are normally restricted to the opposite dendrite of the n. laminaris cells.

Fibers in spinal nerves of tree frogs: Eleutherodactylus and Hyla.

This is a study on numbers and sizes of nerve fibers as determined by electron microscope mapping of the spinal nerves of two species (in different families) of free frogs. The study provides data for comparison of fiber populations in the grass frog, Rana pipiens. In both free frogs the greatest number of fibers in the ten pairs of spinal nerves are in the second and ninth rami. The combined counts of myelinated and unmyelinated fibers in the rami on one side of the cord in Eleutherodactylus and Hyla are respectively: to the forelimb, 3,147 and 4,460 fibers; to the trunk, 334 and 1,234 fibers; and to the hindlimb, 4,335 and 6,213 fibers. With a few notable differences this distribution of fibers to regional body parts correlates closely with that found in Rana. The arboreal frogs have a relatively smaller number of medium and especially large size fibers in comparison with the aquatic-terrestrial living grass frog. The importance is stressed of examining the spinal nerves of an entirely aquatic frog with different feeding habits, such as Xenopus, for further comparison.

Somatotopic organization of cutaneous afferent terminals and dorsal horn neuronal receptive fields in the superficial and deep laminae of the rat lumbar spinal cord.

The somatotopic organization of A- and C-afferent fibre terminals in the dorsal horn of the rat lumbar spinal cord was compared with the spatial location of second-order dorsal horn neuronal mechanoreceptive fields. The central terminal fields of the sural, saphenous, and tibial nerve were mapped by labelling the nerves with horseradish peroxidase (HRP). A previous study used the transganglionic transport of wheat germ agglutinin conjugated to horseradish peroxidase (WGA-HRP) to produce a somatotopic map of high-threshold C-fibre terminal fields in lamina II (Swett and Woolf: J. Comp. Neurol. 231:66-77, '85). In the present study the terminal fields of low-threshold A beta afferents that terminate in laminae III and IV were mapped by using unconjugated HRP at prolonged survival times (72 hours). Unfixed tissue was used to increase the sensitivity of the tetramethylbenzidine reaction, thus allowing these afferent terminals to be clearly seen. The general spatial arrangement of the terminal fields in laminae III/IV closely resembled that found in lamina II in the mediolateral and rostrocaudal planes but because of a dorsoventral obliquity of the afferent terminals, the superficial and deeper fields are not in strict vertical register. The input to laminae II-IV of the dorsal horn may therefore be viewed as two horizontally arranged sheets of afferent terminals both accurately representing the skin surface, the more superficial sheet representing the high-threshold C-afferents and the deeper sheet, low-threshold A-beta afferents. The spatial organization of high-threshold A-delta afferents in laminae I and V appears to be quite different, with a transverse rather than a longitudinal orientation. To study dorsal horn cell receptive field organization two single units with mechanoreceptive fields were recorded extracellularly in each of 87 vertical tracks in the lumbar spinal cord, one unit in the superficial dorsal horn and the second in the deep dorsal horn. In general the somatotopic organization of the receptive fields of both sets of units followed that of the afferent terminal fields but there were cells with receptive fields that were anomalous relative to the recording site. No evidence of any vertical relation or columnar arrangement in receptive field size, threshold, or location on the body surface was found when comparing the two units in a pair. Furthermore, no laminar functional specialization was found, the majority of neurones having both low- and high-threshold inputs.(ABSTRACT TRUNCATED AT 400 WORDS)

The length of cerebellar parallel fibers in chicken and rhesus monkey.

The cerebellar parallel fibers, which course through the molecular layer parallel to the long axes of the cortical folds known as folia, originate from ascending granule cell axons and relay the mossy fiber input to dendrites of Purkinje cells. Purkinje cell axons in the cerebellar white matter collect into sheets or zones oriented at right angles to the folia. Each of these zones, which are approximately 0.5-1 mm wide, innervates a different portion of the deep cerebellar and the vestibular nuclei. An experimental light microscopic study was carried out to determine the maximal length of parallel fibers in long folia of avian and primate cerebellar cortex. With a fine surgical knife, vermal folia were cut perpendicular to their long axes in four adult White Leghorn hens and in three adult rhesus monkeys deeply anesthetized with sodium pentobarbital. The animals were Killed 3-5 days after the operation. Sections of the transected folia were stained with the Fink-Heimer or the DeOlmos-Ingram methods, which revealed the anterogradely degenerated parallel fibers as darkly stained dots. In both species, the pattern of parallel fiber degeneration in the molecular layer had a trapezoidal configuration with the shorter base bordering the Purkinje cell layer and the longer base bordering the pia mater. In both species, the length of parallel fibers averaged approximately 6 mm, although the range was 4-8 mm in chickens and 4.8-6.6 mm in monkeys.(ABSTRACT TRUNCATED AT 250 WORDS)

Autoradiographic localization of newly synthesized octopamine to retinal efferents in the Limulus visual system.

The biogenic amine octopamine is synthesized from both tyrosine and tyramine in the lateral, median, and ventral eyes of Limulus. The autoradiographic studies presented here were designed to locate the sites of octopamine synthesis in the ventral and lateral eyes. We found that efferent fibers, which project to ventral and lateral eyes from the central nervous system, became intensely and selectively labeled during in vitro incubations with 3H-tyramine. In the ventral eye, more than 95% of the efferent fibers were labeled. Results of biochemical analyses suggested that most of the radioactive substance within these efferent fibers was newly synthesized octopamine. The selective labeling of efferent fibers during incubation with 3H-tyramine was used as an anatomical tool to study the number and distribution of efferent fibers within the ventral eye. Light microscopic (LM) reconstructions of the distribution of label in serial longitudinal sections through ventral optic nerves together with electron microscopic (EM) autoradiographic analyses revealed between 70 and 200 efferent axons. The results of these studies and of reconstructions of efferent innervation to photoreceptor somata suggest that each ventral photoreceptor cell or each small cluster of cells is innervated by a separate efferent fiber. Both LM reconstructions and EM analyses showed that efferent fibers ramify extensively and specifically in and near the internal rhabdom of ventral photoreceptor cells. In EM autoradiographs of lateral eyes incubated with 3H- tyramine, the silver grains that were located over ommatidia were concentrated exclusively over efferent fibers. All of these efferent fibers, which lay near rhabdoms and in partitions between retinular cells, were labeled. The results of our present studies support our hypothesis that octopamine is a neurotransmitter in Limulus retinal efferent fibers. This amine may modulate the biochemistry and physiology of ventral photoreceptor cells and may mediate many of the known effects of circadian efferent innervation to the lateral eye.

Columnar organization of corticocortical projections in squirrel and rhesus monkeys: similarity of column width in species differing in cortical volume.

To compare the size and pattern of the terminal distribution of corticocortical projections in two primate species with brains of different size, tritiated amino acids were injected into the prefrontal cortex of New World squirrel monkeys (Saimiri sciureus) and Old World rhesus monkeys (Macaca mulatta), and their brains were processed for light microscopic autoradiography. In both species, prefrontal efferents are directed to a number of cortical targets in the same and opposite hemispheres, where in coronal sections, they generally terminate as radially oriented columns. In the rhesus monkey, the median width of the columns in transverse sections is 685 micrometers. In squirrel monkey, corresponding columns have a median with of 555 micrometers. Considering that the volume of the neocortex in rhesus monkey is approximately 4.5 x larger than that of squirrel monkey, the dimensions of cortical columns in the two species are surprisingly similar. This finding suggests that phylogenetic expansion in cortical surface area is accompanied by an increase in the number, rather than the width of afferent fiber columns. The increase in number of modular units may be relevant to the increasing computational and information processing capacity of the cerebral cortex in the course of evolution.

The laminar distribution of intracortical fibers originating in the olfactory cortex of the rat.

In this study, the autoradiographic method for tracing axonal connections was used to identify the laminar distribution of intracortical fibers originating in the olfactory cortical areas of the rat. Most of the projections can be divided into two major fiber systems with different laminar patterns of termination. The first of these, termed the layer Ib fiber system, arises in the anterior olfactory nucleus, the anterior and posterior piriform cortex, and the lateral entorhinal cortex, and terminates predominantly in layer Ib and, in many cases, layer III of the entire olfactory cortex. The second system, termed the layer II-deep Ib fiber system, originates in three relatively small olfactory cortical areas--the dorsal peduncular cortex, the ventral tenia tecta, and the periamygdaloid cortex--and terminates in and around the cells of layer II in most parts of the olfactory cortex. There is significant overlap in the laminar distribution of the two systems, although the distinction between them is readily apparent. Within the layer Ib fiber system there are relatively slight but consistent differences in the lamination of fibers from different areas. The fibers from the anterior olfactory nucleus are concentrated in the deep part of layer Ib while those from the anterior piriform cortex are concentrated in the superficial part of this layer. The fibers from the posterior piriform cortex tend to be densest in the middle of layer Ib. These differences are maintained in all areas of termination of each set of fibers, both ipsilaterally and contralaterally. In addition, intracortical fibers from the anterior cortical nucleus of the amygdala are distributed throughout layer I, including layer Ia and Ib. Fibers from the nucleus of the lateral olfactory tract terminate bilaterally around the cells of the islands of Calleja and the medial edge of the anterior piriform cortex.

The optic nerve of the brush-tailed possum, Trichosurus vulpecula: fibre diameter spectrum and conduction latency groups.

The principal findings of this report on the morphology and electrophysiology of the possum optic nerve are: (i) There are about 230,000 fibres in the optic nerve. This fibre count, based on electron microscopy, is slightly less than a previously reported estimate of the total number of ganglion cells in the possum retina. (ii) The majority (greater than 98%) of the fibres of the optic nerve are myelinated axons of retinal ganglion cells. The diameters of these fibres range from 0.4--4.6 micrometer (axon diameter range: 0.3--3.8 micrometer) and the frequency distribution of the fibre diameters (and axon diameters) is positively skewed and unimodal. (iii) The antidromic compound action potential of the possum optic nerve shows four negative peaks following stimulation of the optic chiasm. These peaks are associated with four conduction latency groups of fibres which have been designated t1, t2, t3 and t4 in order of increasing conduction latency. (iv) The mean peak conduction velocities of the fibres in the conduction latency groups are 13.1 ms-1 (t1), 8.1 ms-1 (t2), 5.7 ms-1 (t3) and 3.1 ms-1 (t4). (v) There is no direct correlation between the frequency distribution of fibre (or axon) diameters as measured by electron microscopy of transverse sections of fixed optic nerve and the conduction latency groups. (vi) The reconstruction of the possum optic nerve compund action potential on the basis of either axon or fibre diameter frequency distribution does not provide an acceptable, indirect correlation between the morphology and the electrophysiology of this optic nerve.

Primary afferent and propriospinal fibers in the rat dorsal and dorsolateral funiculi.

The purpose of this study is to determine the numbers of primary afferent and propriospinal fibers in the dorsal and dorsolateral funiculi of the rat. The reason for concentrating on these areas is that they contain large numbers of unmyelinated axons. Our data are axonal numbers from the S2 segment of spinal cord in animals that had unilateral dorsal rhizotomies or spinal cord isolations. The major conclusions are 1) that 23% of the primary afferent fibers in the dorsal funiculus are unmyelinated; 2) that there are approximately 12,500 unmyelinated primary afferent fibers in the dorsolateral funiculus, which is more than the number of primary afferent fibers in the dorsal funiculus and tract of Lissauer combined, and 3) that approximately 25% of the axons in the dorsal funiculus and 44% of the axons in the dorsolateral funiculus are propriospinal. These data modify and extend previous ideas of the organization of spinal white matter.

The central nucleus of the inferior colliculus in the cat.

The central nucleus of the inferior colliculus in the cat is distinguished by its unique neuropil. In Golgi-impregnated material, it is composed primarily of neurons with disc-shaped dendritic fields arranged into parallel arrays, or laminae, complemented by the laminar afferent axons from the lateral lemniscus. Large, medium-large, medium, and small varieties of disc-shaped cells are distinguished on the basis of the size of the dendritic field and cell body size, dendritic diameter, and dendritic appendages. A second major class of neurons in the central nucleus are the stellate cells with dichotomously branched, spherical-shaped dendritic trees. Simple, complex, and small stellate cells can be distinguished by their size and by the complexity of the dendritic and axonal branching. Laminar afferent axons are recognized by the nests of collateral side branches and the grapelike clusters of terminal boutons--thick, thin, and intermediate-sized varieties are apparent. Other axon types include local collaterals of central nucleus neurons, some of which are distinguished by their frequent and complex collaterals. In the central nucleus, the configuration of the fibrodendritic laminae, the presence of subdivisions, and the banding of afferent axons suggest levels of organization which are superimposed on the synaptic arrangements of the individual cell and axon types. The laminar pattern, as studied in serial Golgi-impregnated sections, differs from previous reports. The central nucleus contains subdivisions which can be distinguished by their laminar pattern, different proportions of cell types, and the packing density of the cell bodies and axonal plexus. The patterns of degeneration observed in Nauta-stained material after lesions of caudal auditory pathways show that thick and fine afferent fibers form dense bands of degeneration separated by sparse, fine-fiber degeneration. The bands are thicker than individual laminae but smaller than the subdivisions. The intrinsic organization of the neurons and axons, combined with the laminar organization, subdivisions, and banding patterns, each may contribute different aspects to the processing of auditory information in the central nucleus.

Sites on termination on the saccular macula of auditory nerve fibers in the goldfish as determined by intracellular injection of procion yellow.

Procion yellow dye was injected intracellularly into large auditory fibers of goldfish to investigate the relationship between the response type of these fibers and their site of termination in the saccular macula. Fibers which responded to the rarefaction phases of sound were found to terminate on the hair cells located in the ventral half of the saccular macula, while those which responded to the compression phases were found to terminate on the dorsally located hair cells. A third group of fibers, which responded to both sound phases, was found to bifurcate outside the saccular macula and terminate on both dorsally and ventrally located hair cells. Some of the fibers of this group were found to trifurcate before entering the saccular macula. Results obtained in the present study, in conjunction with the findings on the orientation of hair cells within the sensory macula, can explain the mechanism underlying the presence of three different response types of fibers in the auditory nerve of the goldfish.

Comparative study of the posterior red nucleus in baboons and gibbons.

The posterior red nucleus (PRN) was studied in two species of primates by the technique of retrograde degeneration of rubrospinal cells following transection of the spinal cord at different levels. The form of the PRN was reconstructed for both a quadruped monkey (baboon) and an anthropoid with erect posture (gibbon). The PRN contains polymorphic cells characterized by their very chromophilic and granular Nissl substance. These neurons vary in diameter from 25 micrometer to 70 micrometer. Some of them give rise to the rubrospinal tract. Baboon: The approximately 1,300 rubrospinal cells in this species are divided into two equal groups, one related to the contralateral forelimb, with axons ending between the second cervical and third thoracic segment, and the other related to the contralateral hindlimb, projecting caudally beyond T3. Following a high cervical lesion, nondegenerated cells of similar description remain throughout the nucleus. A significantly large group of these cells occurs medially and may be the source of fibers ending in the brain stem or cerebellum. Gibbon: In this species, the number of rubrospinal cells controlling the hindlimb is less than half that found in the baboon. This reduction in the gibbon is much greater for medium-sized cells, but is also significant for the giant cells. These results obtained from primates are compared with those reported for the cat. A possible function for the PRN in the control of limb movements is discussed from the viewpoint of phylogeny.