After injecting the posterior pituitery of the albino rats with cholera-toxinconjugated HRP,retrogradely labeled neurons were found in the area of the bednuclei of the stria terminalis (BST),which were then analyzed,based on therecent cytoarchitectonical study of the BST.In the anterior subdivision of the BSTdendrites could be seen in the parastrial nucleus with a few labeled cells surround-ing it;cells of the magnocellular nucleus were found scattered in the subdivision;some labeled cells were also identified in the ventral nucleus.Surrounding the dor-sal half of the posterior division of the BST there were several fairly distinct labe-led cell groups,viz.,the anterior fornical nucleus,medial and lateral dorsal accessorygroups,and the subinterventricular group.Though these cell groups are contiguouswith the BST,they do not seem to belong to the BST itself,but their dendritesoften extend into the posterior division of the BST,thus sharing some commonafferents with the latter.A few less constant cells were sometimes seen scattered invarious parts of the BST.

Horseradish peroxidase was injected into C_(6,7) or L_(5,6) spinal cord in 21 adult cats and the morphology and distribution of the labeled cells in the dorsal columu nuclei (DCN) were studied. Cells projecting to the cervical cord were found to be distributed mainly in a region from 2.5mm below to 0.5mm above the obex, while cells projecting to the lumbar cord were found within 2.5mm below the obex. Medio-Iaterally, the cervical projecting cells were located chiefly at the junctional area between the gracile and medial cuneate nuclei and at the ventromedial part of the latter. The lumbar projection cells were located more medially, concentrating at the junctional area and the ventrolateral part of the gracile nucleus. In both cases scattered cells were found in the two dorsal column nuclei. The difference in cellular distribution between cervical and lumbar injection cases is consistent with a somatotopical organization.The labeled cells in the dorsal column nuclei varied in shape and size. Small cells, mostly fusiform, were concentrated at the junctional area between the gracile and medial cuneate nuclei. Large cells were found scattered in the two nuclei. Many of them were triangular or multipolar with long and straight dendrites. A few were round and their dendrites bushy.

HRP was injected into the cervical or lumbar spinal cord in 8 adult cats. The anterogradely labeled terminal arborization and its relation with the retrogradely labeled cells in the DCN were studied.The non-primary afferent fibers were distributed diffusely in the DCN. Injection of HRP in one spinal segment, either cervical or lumbar, resulted in labeling of terminal branches throughout the rostrocaudal extent of the DCN. In C6 or C7 injection cases large amount of labeled terminal branches were found in the medial cuneate nucleus, chiefly in its extraclusteral regions, viz. in the rostral part and in the ventral area of the caudal two thirds of the nucleus. A small celled area at the dorsolateral brim of the middle part of the medial cuneate nucleus was found to approach and to be contiguous with the external cuneate nucleus at higher level. This area was likewise densely labeled. Labeled terminal branches were also found in the gracile nucleus in lesser amount.After L5 or L6 injections the labeled terminals were much less in amount than in cervical injection cases. They were distributed mainly in the gracile nucleus. In the lower part of the nucleus the labeled terminals were found in its dorsal, medial, and ventrar areas. In the upper part of the nucleus they projected diffusely. A slight degree of labeling was also found in the ventromedial angle of the medial cuneate nucleus.The areas of distribution of the labeled terminals and labeled cells overlapped but did not coincide with eachother. No or merely a few labeled terminals were found around some of the labeled cells, while other cells were located right in the center of heaviest terminal labeling with abundant terminal branches surrounding their somata and dendrites. This proved that the non-primary afferent fibers of the DCN might form monosynaptic feed-back loop with the DGN-spinal cells.

HRP was injected into the cervical or lumbar spinal cord in 8 cats and the spinobrainstem projections were traced with the anterograde HRP technique. The labeled terminal branches were found most densely concentrated in the ipsilateral dorsal column nuclei and lateral reticular nucleus, the contralateral medial and dorsal accessory olivary nuclei and the bilateral pontobulbar bodies and dorsolateral part of pontine nuclei. In the reticular formation and a number of brainstem nuclei labeled terminals were also found to varying degrees. The most remote site found labeled was the hypothalamus.It was also found that the dorsal accessory olivary nucleus could be further subdivided into a caudal and an oral part, cells in the caudal part being smaller than that in the oral one.In the accessory olivary nuclei labeled terminal branches were more numerous in lumbar injection cases. Clearcut somatotopical localization was demonstrated in the oral part of the dorsal accessory olivary nucleus.In the lateral reticular nucleus the subnucleus magonocellularis and the lateral wing of the subnucleus parvocellularis were the major sites of labeling and the labeling in cervical injection cases greatly out numbered that in lumbar injections.The cervical and lumbar injections were equal in their amount of projection to the pontobulbar bodies and the dorso-lateral part of the pontine nuclei.The distribution of spinal afferents in the inferior olivary nucleus, the lateral reticular nucleus, the pontobulbar body and the pontine nucleus was discussed.

Different parts of the cochlea were destroyed in 6 cats (11 sides). The degenerating fibers and preterminal degenerations were studied by the Nauta technique with the following conclusions: 1. The overwhelming majority of the fibers from the spiral ganglion project to within the limit of the cochlear nucleus complex. No fiber degeneration could be found in the trapezoid body. A few degenerating fibers from the basal turn of the spiral ganglion, however, were found in the dorsal acoustic stria. 2. No preteriminal degeneration was found in the “granular cell masses” and “dorsal cell complex of the acoustic tubercle”of G. Fuse. 3. In the dorsal cochlear nucleus, degenerating fibers were found to terminate in the II-V layers. 4. Contrary to Lorente de No’s finding, the present study showed that the fibers from the basal portion of the spiral ganglion bifurcate at the dorsal part of the ventral cochlear nucleus, those from the apical portion at the ventral part. 5. The spiral ganglion projects in a clear-cut topographical way to the 3 parts of the cochlear nucleus complex, viz, the basal portion and the successively more apical portions of the spiral ganglion project (a) to the dorso-medial and successively more ventro-lateral parts of the dorsal cochlear nucleus; (b) to the anterior part of the ventral cochlear nucleus along the dorso-caudo-medial to ventro-rostro-lateral axis; and (c) to the posterior part of the ventral cochlear nucleus along the dorso-rostro-medial to ventro-caudo-lateral axis.

It is well known anatomically and physiologically that the raphe nuclei project to the spinal cord. No definite data on somatotopic raphe-spinal relationship, however, are available. The raphe nuclei, especially the nucleus raphe magnus, were found to inhibit nociceptive transmission and pain reflexes in the spinal cord, and they presumably play an important role in acupuncture analgesia. Thus, the question on whether there is a somatotopic raphe-spinal projection has much to do with the analysis of the function of the raphe nuclei and the evaluation of their possible role in acupuncture analgesia. 11 cats were used in the present study. Horseradish peroxidase (HRP) was injected into the dorsal horn of the 7th cervical or 4,5th lumbar spinal cord and labelled cells were traced in the raphe nuclei. In the cases of cervical cord injection, labelled cells were found in the nucleus raphe magnus, nucleus raphe pallidus and nucleus raphe obscurus. In the lumbar injections only the nucleus raphe. magnus and nucleus raphe pallidus were labelled. The labellcd cells were distributed over relatively wide areas in the raphe nuclei after HRP injection into single spinal segments. Nevertheless, a certain degree of somatotopic relationship existed. In the nucleus raphe magnus cells projecting to G7 were distributed more rostrally and those projecting to L4,5, more caudally. In the nucleus raphe pallidus, on the contrary, cells to C7 were located more caudally while those to L4,5 more rostrally. A definite, though diffuse to a certain degree, somatotopie raphe-spinal projection is consistent with the extent of analgesia during acupuncture or electrical stimulation of the nucleus raphe magnus, thus favours the hypothsis of the role of descending inhibition from the raphe nuclei in acupuncture analgesia, and provides a possible explanation for the relative specificity of acupuncture points in their fields of analgesia.

A case of HRP injection into laminae Ⅰ-Ⅲ of the cervical spinal cord of the cat was reported. The labeled cells were found mainly in the dorsal column nuclei, nucleus raphe magnus, reticular nuclei, vestibular nuclei and nucleus of tractus solitarius.

Usually Rexed's lamina Ⅱ of the spinal cord is taken as the substantia gelatinosa Rolandi. This, however, is not based on unequivocal evidence. We have proved in goat ('81) that the substantia gelatinosa Rolandi actually corresponds to laminae Ⅰ, Ⅱ and Ⅲ. The investigation was extended to the human being in the present study.Lumbar_(3,4) spinal segments were cut parasagittally in 3 cases (an adult, a child and a new-born baby). Sections were first examined under a stereomicroscope. The substantia gelatinosa Rolandi could be clearly identified as a semitransparent gelatinous layer. Dorsal to it was a spongy layer with islands of gelatinous material scattered in white matter. The dorsal border of the spongy layer, the boundary between it and the ventral lying gelatinous layer and the ventral border of the latter were marked by razor cuts.The sections were then stained with Nissl method and the cytoarchitecture of the tissue between the cuts was studied. It has been found that the substantia gelatinosa Rolandi of the human being corresponds principally to Rexed's laminae Ⅱand Ⅲ. The overlying spongy layer (lamina Ⅰ) however, composing of scattered, broken gelatinous substance, should also be considered a part of the substanstia gelatinosa.

Retrograde HRP technique was used to study the somatotopical organization of the raphe—spinal projection of the rat. HRP or wheat germ agglutinin conjugated HRP was injected into unilateral gray matter or dorsal horn of the cervical cord or lumbar cord. The following conclusions have been reached.The caudal part of NRM projects to areas ventral to the dorsal horn. Projection to the dorsal horn was found to originate from the rostral part of NRM, which is somatotopically localized. Cells projecting to the cervical dorsal horn are distributed more rostrally than those to the lumbar dorsal horn, although an extensive overlapping of these two parts is evident. Somatotopical localization in the form of "gap" as suggested by Watkins et al. could not been verified in our experiment.The functional significance of the sornatotopical organization of NRM is discussed.

The localization of the neurons which control the inferior oblique muscle in the oculomotor nucleus and their dendritic architecture were studied by injecting the conjugated cholera toxin-horseradish peroxidase (CT-HRP) into the inferior oblique muscle of 7 rabbits.The oculomotor nucleus could be divided into oral, middle and caudal parts. The middle part was further divided into dorsomedial and ventrolateral parts, and the caudal part divided into dorsal and ventral parts. The labeled neurons innervating the inferior oblique muscle were mainly distributed ipsilaterally and occupying two thirds. of the rostrocaudal extent of the oculomotor nucleus, a few were scattered contralaterally.The labeled cells were found in the dorsomedial part of the nucleus orally, and shifted in successive caudal sections to the medial and then to the ventral part. No labeled cells in the oral and caudal ends of the nucleus could be identified.The dendritic branches of the labeled neurons covered the whole nucleus, but densest in its dorsomedial part. Many of them extended beyond the boundary of the nucleus into the central gray matter dorsally, some even approacheding the aqueduct, or through the medial longitudinal fasciculus into the reticular formation laterally and ventrally. A few dendrites crossed the midline into the contralateral nucleus. Therefore the receptive field of the oculomotor nucleus is presumably much larger than the area of the nucleus itself.