TRPV4 channels in the human urogenital tract play a role in cell junction formation and epithelial barrier.

AIM: The molecular interactions between transient receptor potential vanilloid subtype 4 channels (TRPV4) and cell junction formation were investigated in the human and mouse urogenital tract. MATERIALS AND METHODS: A qualitative study was performed to investigate TRPV4 channels, adherence junctions (AJs) and tight junctions (TJs) in kidney, ureter and bladder tissues from humans and wild-type and transgenic TRPV4 knockout (-/-) mice with immunohistochemistry, Western blotting, immunoprecipitation and reverse trasnscription-PCR. Cell junction formation in the wild-type and TRPV4 knockout (-/-) mouse was evaluated with immunohistochemistry and transmission electron microscope (TEM) techniques. RESULTS: TRPV4 channels are predominantly located in membranes of epithelial cells of the bladder, ureter and the collecting ducts of the kidney. There is a molecular interaction between the TRPV4 channel and the AJ. TEM evaluation showed that AJ formation is disrupted in the TRPV4 -/- mouse resulting in deficient intercellular connections and integrity of the epithelium. CONCLUSIONS: TRPV4 is believed to be a mechanoreceptor in the bladder. This study demonstrates that TRPV4 is also involved in intercellular connectivity and structural integrity of the epithelium.

Microscopic localization of PEG-liposomes in a rat model of focal infection.

In the present study the microscopic localization of polyethylene glycol (PEG) liposomes in infected tissues was studied with both light microscopy (LM) and transmission electron microscopy (TEM) in rats with focal intramuscular Staphylococcus aureus infection. PEG-liposomes containing colloidal gold were prepared and injected intravenously in rats with focal S. aureus infection and tissues were dissected at 24 h post injection. Sections were cut and liposomes were visualized for microscopic evaluation using silver enhancement. Uptake of PEG-liposomes was visualized by both scintigraphy and LM in the abscess, liver and spleen. In the infected area, the liposomes were mainly found in the vicinity of blood vessels. TEM showed that the liposomes were localized in the macrophages and to a lesser extent in endothelial cells in the infectious tissue. In the liver, the liposomes appeared mainly localized in Kupffer cells. In the spleen, uptake was only seen in cells of the red pulp and in cells around the central arteries. Our microscopic observations indicate that uptake and retention of PEG-liposomes in the infectious focus is a result of enhanced extravasation due to increased vascular permeability and subsequent phagocytosis of PEG-liposomes by macrophages in the infected tissue.

Myelinated dendrites in the mormyrid electrosensory lobe.

This is the third paper in a series on the morphology, immunohistochemistry, and synaptology of the mormyrid electrosensory lateral line lobe (ELL). The ELL is a highly laminated, cerebellum-like structure in the rhombencephalon that subserves an active electric sense: Objects in the nearby environment are detected on the basis of changes in the reafferent electrosensory signals that are generated by the animal's own electric organ discharge. This paper concentrates on the intermediate (cell and fiber) layer of the medial zone of the ELL and pays particular attention to the large multipolar neurons of this layer (LMI cells). LMI cells are gamma-aminobutyric acid (GABA)ergic and have one axon and three to seven proximal dendrites that all become myelinated after their last proximal branching point. The axon projects to the contralateral homotopic region and has ipsilateral collaterals. Both ipsilaterally and contralaterally, it terminates in the deep and superficial granular layers. The myelinated dendrites end in the deep granular layer, where they most likely do not make postsynaptic specializations, but do make presynaptic specializations, similar to those of the LMI axons. Because it is not possible to distinguish between axonal and dendritic LMI terminals in the granular layer, the authors refer to both as LMI terminals. These are densely filled with small, flattened vesicles and form large appositions with ELL granular cell somata and dendrites with symmetric synaptic membrane specializations. LMI cells do not receive direct electrosensory input on their somata, but electrophysiological recordings suggest that they nevertheless respond strongly to electrosensory signals (Bell [1990] J. Neurophysiol. 63:303-318). Consequently, the authors speculate that the myelinated dendrites of LMI cells are excited ephaptically (i.e., by electric field effects) by granular cells, which, in turn, are excited via mixed synapses by mormyromast primary afferents. The authors suggest that this ephaptic activation of the GABAergic presynaptic terminals of the myelinated dendrites may trigger immediate synaptic release of GABA and, thus, may provide a very fast local feedback inhibition of the excited granular cells in the center of the electrosensory receptive field. Subsequent propagation of the dendritic excitation down the myelinated dendrites to the somata and axon hillocks of LMI cells probably generates somatic action potentials, resulting in the spread of inhibition through axonal terminals to a wide region around the receptive field center and in the contralateral ELL. Similar presynaptic myelinated dendrites that subserve feedback inhibition, until now, have not been described elsewhere in the brain of vertebrates.

Systemic administration of the propargylamine CGP 3466B prevents behavioural and morphological deficits in rats with 6-hydroxydopamine-induced lesions in the substantia nigra.

The ability of CGP 3466B to attenuate the behavioural and morphological consequences of experimentally induced cell death was investigated in a recently updated animal model of Parkinson's disease. 6-Hydroxydopamine was infused bilaterally into the substantia nigra pars compacta of rats that were pretreated with desimipramine. Treatment with CGP 3466B (0.0014-1.4 mg/kg, injected subcutaneously) or its solvent was begun 2 h after the 6-OHDA injection, and maintained twice daily for 14 days. After a washout period of 14 days, changes in motor behaviour were evaluated, using the open field test (analysis of normal and abnormal stepping, e.g.) and the paw test (analysis of retraction time of limbs). Changes in learning and memory were evaluated with the help of the Morris water maze task. Following immunocytochemical staining of tyrosine hydroxylase, the extent of the lesion was quantified using a computerized system. CGP 3466B prevented all deficits produced by 6-hydroxydopamine (6-OHDA), though at different doses. It prevented: abnormal stepping (0.0014-0.014 mg/kg); increased forelimb and hindlimb retraction time (0.014-0.14 mg/kg and 0.0014-0.14 mg/kg, respectively); delayed learning (1.4 mg/kg); and reduced tyrosine hydroxylase immunoreactivity in the substantia nigra (0.0014-0.014 mg/kg). CGP 3466B (0.0014-0.14 mg/kg) induced no deficits in sham-treated rats. CGP 3466B (1.4 mg/kg), however, did not show any benefit on motor deficits in 6-OHDA-lesioned rats, and induced abnormal movements and decreased the tyrosine hydroxylase immunoreactivity in the substantia nigra pars compacta and the ventral tegmental area of sham-lesioned animals. It is concluded that CGP 3466B prevents all 6-OHDA-induced behavioural and immunocytochemical deficits, though at different doses. CGP 3466B is suggested to be a valuable agent for inhibiting the dopaminergic degeneration in patients with Parkinson's disease.

Plasticity in the stress-regulating circuit: decreased input from the bed nucleus of the stria terminalis to the hypothalamic paraventricular nucleus in Wistar rats following adrenalectomy.

The bed nucleus of the stria terminalis is involved in the stress-regulating circuit by funnelling limbic information to the hypothalamic paraventricular nucleus. Since adrenalectomy influences both limbic structures (by inducing cell death in the hippocampus) and the hypothalamic paraventricular nucleus (by increased corticotrophin-releasing hormone synthesis), we investigated whether the bed nucleus of the stria terminalis is also influenced by adrenalectomy. For this purpose, we analysed and compared the projections from the bed nucleus of the stria terminalis to the hypothalamic paraventricular nucleus in normal and adrenalectomized rats by anterograde tracer injections in the bed nucleus of the stria terminalis. Quantitative analysis of the fibre pattern in the hypothalamic paraventricular nucleus of normal rats revealed a homogeneous distribution of fibres of the bed nucleus of the stria terminalis over the different subdivisions of the hypothalamic paraventricular nucleus. In adrenalectomized rats, the absolute fibre density was significantly lower in the whole hypothalamic paraventricular nucleus (1.17 +/- 0.27 10(-3) microm/microm3 in adrenalectomized rats versus 2.59 +/- 0.24 10(-3) microm/microm3 in normal rats; P < 0.01) and all its subdivisions. The largest decrease of fibre density was found in the corticotrophin-releasing hormone-rich part of the hypothalamic paraventricular nucleus (relative fibre density; adrenalectomized rats: 0.602 +/- 0.106, versus 1.095 +/- 0.019 in normal rats, P < 0.01). These results show a loss of input from the bed nucleus of the stria terminalis to the hypothalamic paraventricular nucleus, and particularly to the corticotrophin-releasing hormone neurons, following adrenalectomy. The data suggest that this pathway within the stress-regulating circuit is functionally affected by corticosteroids in adult rats and may imply that human disorders associated with corticosteroid imbalance are allied to a changed circuitry in the brain.

Projection neurons of the mormyrid electrosensory lateral line lobe: morphology, immunohistochemistry, and synaptology.

This paper describes the morphological, immunohistochemical, and synaptic properties of projection neurons in the highly laminated medial and dorsolateral zones of the mormyrid electrosensory lateral line lobe (ELL). These structures are involved in active electrolocation, i.e., the detection and localization of objects in the nearby environment of the fish on the basis of changes in the reafferent electrosensory signal generated by the animal's own electric organ discharge. Electrosensory, corollary electromotor command-associated signals (corollary discharges), and a variety of other inputs are integrated within the ELL microcircuit. The organization of ELL projection neurons is analyzed at the light and electron microscopic levels based on Golgi impregnations, intracellular labeling, neuroanatomical tracer techniques, and gamma-aminobutyric acid (GABA), gamma-aminobutyric acid decarboxylase (GAD), and glutamate immunohistochemistry. Two main types of ELL projection neurons have been distinguished in mormyrids: large ganglionic (LG) and large fusiform (LF) cells. LG cells have a multipolar cell body (average diameter 13 microns) in the ganglionic layer, whereas LF cells have a fusiform cell body (on average, about 10 x 20 microns) in the granular layer. Apart from the location and shape of their soma, the morphological properties of these cell types are largely similar. They are glutamaterigic and project to the midbrain torus semicircularis, where their axon terminals make axodendritic synaptic contacts in the lateral nucleus. They have 6-12 apical dendrites in the molecular layer, with about 10,000 spines contacted by GABA-negative terminals and about 3,000 GABA-positive contacts on the smooth dendritic surface between the spines. Their somata and short, smooth basal dendrites, which arborize in the plexiform layer (LG cells) or in the granular layer (LF cells), are densely covered with GABA-positive, inhibitory terminals. Correlation with physiological data suggests that LG cells are I units, which are inhibited by stimulation of the center of their receptive fields, and LF cells are E units, excited by electric stimulation of the receptive field center. Comparison with the projection neurons of the ELL of gymnotiform fish, which constitute another group of active electrolocating teleosts, shows some striking differences, emphasizing the independent development of the ELL in both groups of teleosts.

Interneurons of the ganglionic layer in the mormyrid electrosensory lateral line lobe: morphology, immunohistochemistry, and synaptology.

This is the second paper in a series that describes the morphology, immunohistochemistry, and synaptology of the mormyrid electrosensory lateral line lobe (ELL). The ELL is a highly laminated cerebellum-like structure in the rhombencephalon that subserves an active electric sense: Objects in the nearby environment of the fish are detected on the basis of changes in the reafferent electrosensory signals that are generated by the animal's own electric organ discharge. The present paper describes interneurons in the superficial (molecular, ganglionic, and plexiform) layers of the ELL cortex that were analyzed in the light and electron microscopes after Golgi impregnation, intracellular labeling, neuroanatomical tracing, and gamma-aminobutyric acid (GABA) immunohistochemistry. The most numerous interneurons in the ganglionic layer are GABAergic medium-sized ganglionic (MG) cells and small ganglionic (SG) cells. MG cells have 10-20 spiny apical dendrites in the molecular layer, a cell body of 10-12 microns diameter in the ganglionic layer, a single basal dendrite that gives rise to fine, beaded, axon-like branches in either the plexiform layer (MG1 subtype) or the deeper granular layer (MG2 subtype), and an axon that terminates in the plexiform layer. Their apical dendritic tree has 12,000-22,000 spines that are contacted by GABA-negative terminals, and it receives, 1,250-2,500 GABA-positive contacts on the smooth dendritic surface between the spines. The average ratio of GABA-negative to GABA-positive contacts on the interneuron apical dendrites (14:1) is significantly higher than that for the efferent projection cells that have been described previously (Grant et al. [1996] J. Comp. Neurol., this issue). The somata and basal dendrites of MG cells receive a low to moderate density of GABAergic synaptic input, and their axons make GABAergic synaptic contacts with the somata and cell bodies of MG as well as with large ganglionic (LG) cells. SG cells probably represent immature, growing MG cells. Other interneurons in the superficial ELL layers include GABAergic stellate cells in the molecular layer, two types of non-GABAergic cells with smooth dendrites in the deep molecular layer that are named thick-smooth dendrite cells and deep molecular layer cells, and horizontal cells that are encountered particularly in the plexiform layer. Comparison with the ELL of waveform gymnotiform fish, which is another group of active electrolocating teleosts that has been investigated thoroughly, shows striking differences. In these fish, no GABAergic interneurons are found in the ganglionic (pyramidal) layer of the ELL, and GABA-negative interneurons with smooth dendrites in the molecular layer also seem to be lacking. At present, the phylogenetic origin of the described superficial interneurons in the mormyrid ELL is uncertain.

The hypothalamic paraventricular nucleus in two types of Wistar rats with different stress responses. I. Morphometric comparison.

The present study evaluates the role of the hypothalamic paraventricular nucleus (PVH) in stress regulation by a morphometric comparison of the vascular, neuronal and synaptic properties of this nucleus in two lines of Wistar rats. It has been previously reported that these two lines of rats, indicated as APO-SUS (apomorphine-susceptible) and APO-UNSUS (apomorphine-unsusceptible) rats on the basis of their reactivity to a subcutaneous injection of apomorphine, display a variety of pharmacological and behavioral differences, including differences in their stress-coping mechanisms (Cools et al., Neuropsychobiology, 28 (1993) 100-105). The results show a similar vascular and neuronal organization of the PVH in both lines, but distinct synaptic differences. The PVH (0.12 mm3 volume with about 15,000 neurons on one side) has an overall vascular density of 5.6%, with significant differences between subdivisions (parvocellular central part: 8.3%, parvocellular dorsal/ventral/posterior part: 4.6-5.3%), which means that vascularity is a useful tool to delineate subdivisions in the parvocellular PVH. The neuronal density of 132 x 10(3)/mm3 as found in the present study is two times higher than reported in a previous study Possible reasons for this discrepancy are extensively discussed. The most significant finding of the present study is the observation that APO-SUS rats have a significantly higher synaptic density (158 x 10(6)/mm3) in the PVH than APO-UNSUS rats (108 x 10(6)/mm3). It is discussed in which way this synaptic difference may be correlated with the different activity of the hypothalamo-pituitary-adrenal axis in both lines of Wistar rats.

The hypothalamic paraventricular nucleus in two types of Wistar rats with different stress responses. II. Differential Fos-expression.

The present study investigates the role of corticotropin-releasing hormone (CRH) neurons in stress regulation by a comparison of stress induced Fos-immunoreactivity and CRH-immunoreactivity in the hypothalamic paraventricular nucleus (PVH) of APO-SUS (apomorphine-susceptible), APO-UNSUS (apomorphine-unsusceptible), normal Wistar and adrenalectomized Wistar (ADX) rats. The first two types represent a good model to study the role of the PVH in stress regulation, since they show different stress responses and a differential synaptic organization of the PVH. After placement on an open field for 15 min all rats showed an increase in the number of Fos-immunoreactive nuclei compared to control handling. Interestingly, open field stress, but not control handling, induces significantly fewer Fos-immunoreactive nuclei in the PVH of APO-SUS rats (1255 +/- 49) compared to APO-UNSUS rats (1832 +/- 201). Experiments with ADX rats revealed that 93% of the CRH-immunoreactive neurons contained a Fos-immunoreactive nucleus, which suggests that the differential Fos-expression in APO-SUS and APO-UNSUS rats represents a differential activation of the CRH neurons. This hypothesis is discussed in relation to reported differences in stress responses, stress-induced ACTH levels and synaptic organization of the PVH.

Distribution of noradrenaline-immunoreactivity in the brain of the mormyrid teleost Gnathonemus petersii.

The distribution of noradrenaline-immunoreactivity in the brain of the mormyrid fish Gnathonemus petersii was studied in order to evaluate the noradrenergic innervation of a number of specialized mormyrid brain regions, including electrosensory centers and a gigantocerebellum. Noradrenaline-immunoreactive (NAi) neurons occur in the hypothalamic paraventricular organ (PVO), the locus coeruleus, and the caudal rhombencephalon. In the PVO, NAi cerebrospinal fluid (CSF)-contacting neurons are located in the same regions where dopamine- and serotonin-containing CSF contacting neurons occur. The locus coeruleus consists, on each side, of at least 30 rather large NAi neurons with ventrolaterally directed dendrites and dorsolaterally coursing axons. In the caudal rhombencephalon, NAi neurons are located in the transition region between the ventromedial motor zone and the dorsolateral sensory zone. The density of NAi fibers is very high in the efferent tract of the locus coeruleus, the medial forebrain bundle, and two telencephalic, one preoptic, and one rhombencephalic subependymal axonal plexus. A marked NAi innervation is present in the dorsomedial and ventral telencephalon, the preoptic region, periventricular hypothalamic and thalamic regions, the midbrain tectum, cerebellar granular layers, the electrosensory lateral line lobe, the rhombencephalic transition region between the sensory and motor zones, and the area postrema. Other regions are more sparsely innervated by NAi fibers, but regions completely devoid of NAi fibers were not observed. Interestingly, NAi fibers form large club endings in some subdivisions of the precerebellar nucleus lateralis valvulae, and parallel fibers in the cerebellar granular layer. Comparison with the distribution of NAi or dopamine-beta-hydroxylase-immunoreactivity in other species shows that all teleosts studied to date have noradrenergic cells in the locus coeruleus and the caudal rhombencephalon. However, NAi CSF-contacting PVO cells have been described only in the teleost Gnathonemus petersii and the lizard Gekko gecko (Smeets and Steinbusch: J. Comp. Neurol. 285:453-466, '89). It is possible that they might pick up catecholamines as well as serotonin from the CSF, into which monoamines might be released by telencephalic and preoptic subependymal axonal plexuses.

An intriguing pitfall in chemical neuroanatomy: specific populations of unspecifically immunoreactive neurons in the brain of the mormyrid fish Gnathonemus petersii.

The present paper describes the location, morphology, ultrastructure and immunocytochemical properties of neurons in the brain of the mormyrid fish Gnathonemus petersii, that appear to be unspecifically immunoreactive to a number of secondary or tertiary antibodies used in immunohistochemical procedures, including rabbit-anti-mouse immunoglobulins (IGGs), rabbit peroxidase-anti-peroxidase IGGs, and rabbit-anti-sheep or sheep peroxidase-anti-peroxidase IGGs. Unspecifically immunoreactive (UI) cells have typically neuronal morphological and ultrastructural characteristics, and occur at four specific locations in the mormyrid brain. A small rhombencephalic group is located rostrolateral to the efferent octaval nucleus, between the fasciculus longitudinalis medialis and the decussation of the lateral lemniscus. A mesencephalic cluster of cells is located in the dorsal midbrain tegmentum against the tractus telencephalo-mesencephalicus. In addition, dispersed UI neurons were observed in the nucleus lateralis of the torus semicircularis and in the preoptic region above the optic chiasm. Remarkably, UI cells are clearly present in a substantial number of brains investigated, but not detectable in others. The present findings point to a curious pitfall in chemical neuroanatomy, the functional significance of which is unknown at present. In several previous studies using the brain of G. petersii, UI cells were abusively included in the description of monoaminergic cell groups. Similar cells have until now not been reported in other vertebrate brains.

Distribution of zebrin II in the gigantocerebellum of the mormyrid fish Gnathonemus petersii compared with other teleosts.

Immunocytochemistry has demonstrated unexpected heterogeneity among cerebellar Purkinje cells. For example, monoclonal antibody Mab anti-zebrin II reveals parasagittal bands of immunoreactive Purkinje cells in the mammalian cerebellum, but reveals a non-sagittal cerebellar compartmentation pattern in goldfish and gymnotiform fish. The present paper investigates the cerebellar compartmentation pattern, as reflected in the zebrin II distribution, in two other teleosts, the electric mormyrid fish Gnathonemus petersii with its large and regularly built gigantocerebellum, and the electrosensory osteoglossomorph teleost Xenomystis nigri, by using light as well as electron microscopic immunohistochemical techniques. Zebrin II is expressed only in Purkinje cells, where it is present in the cytoplasm of all neuronal compartments, including spines, distal and proximal dendrites, the cell body, and the initial part, as well as terminal boutons of the axon. Other types of cerebellar neurons, including the eurydendroid projection neurons, are zebrin II-negative. In Gnathonemus, zebrin II-positive Purkinje cells are present in the large caudolateral part of the valvula, in lobes C2, C3, and C4 of the corpus, and in the anterior as well as the posterior part of the caudal cerebellar lobe. Zebrin II-negative Purkinje cells are present in a continuous region encompassing the rostromedial part of the valvula, the lobus transitorius, lobe C1 and the ventral part of lobe C2, and in a small, lateral zone of the posterior part of the caudal lobe. In Xenomystis, all Purkinje cells, including those in the medial valvula and the posterior part of the caudal lobe, appear to react with mab anti-zebrin II. This more widespread distribution may be due to the presence of a second antigenic polypeptide in this species. On the basis of the present findings, it is concluded that the mormyrid lobus transitorius, lobe C1, and the ventral part of lobe C2 probably belong to the valvula, while the corpus is restricted to the dorsal part of lobe C2, lobe C3, and lobe C4. The functional significance of zebrin II expression for different subsets of teleostean Purkinje cells remains unclear, since comparisons of different teleosts reveal no general correlation with particular afferent or efferent connections, nor with special morphological features such as a dendritic palisade pattern or different arrangements of the Purkinje cell bodies. A comparison between mammals and teleosts suggests that a distinct parasagittal cerebellar zonation in teleosts is absent, and the major part of the teleostean cerebellum may be considered as a single (midsagittal) cerebellar zone, with about the same width as one mammalian parasagittal zone.

Demonstration of dopamine in electron-dense synaptic vesicles in the pars intermedia of Xenopus laevis, by freeze substitution and postembedding immunogold electron microscopy.

The presence of dopamine in the pituitary of the clawed toad Xenopus laevis was studied by light and electron microscope immunocytochemistry, using pre- and postembedding techniques. Light microscopy showed the presence of an intricate, anti-dopamine-positive fibre network throughout the pars intermedia. In preembedded stained material, dopamine appeared to occur in varicosities which make synaptic contacts with both folliculo-stellate cells and melanotrope cells. Post-embedding immunogold staining of freeze-substituted material permitted the localization of anti-dopamine reactivity in electron-dense vesicles in these varicosities. This finding supports the hypothesis that dopamine is involved in the (inhibitory) control of melanotrope cell activity in X. laevis.

Synaptic morphometry and synapse-to-neuron ratios in the superior colliculus of albino rats.

The superior colliculus of mammals is generally divided into seven layers on the basis of the distribution of myelinated fibers, which are densely packed in layers III, V, and VII but sparse in the other layers. The laminar distribution of afferents and efferents allows, in addition, for the distinction of a superficial visual zone (layers I-III) and a deeper multimodal and premotor zone (layers IV-VII). Collicular neurons, however, do not show a lamination pattern, but are rather homogeneously distributed with only gradual transitions (Albers et al.: J. Comp. Neurol. 274:357-370, '88). The present study analyses whether the distribution of collicular synapses is correlated with the laminar organization of collicular axons or rather with the more homogeneous distribution of collicular neurons. For this purpose, the size and density of synaptic terminals and contacts as well as synapse-to-neuron ratios were determined in all collicular layers of albino rats by means of quantitative analysis of electron microscopic pictures. The size of presynaptic terminals and contacts does not differ significantly between individual collicular layers. On average, presynaptic terminal diameter is 1,079 nm, and synaptic contact size 338 nm, while 23% of all contacts are of the symmetrical type with pleiomorphic vesicles. The average numerical synaptic density is 422 million per mm3. This value is significantly higher in layers I and II (on average 670 million per mm3) than in layers III-VII (on average 370 million per mm3). The synapse-to-neuron (S/N) ratios calculated show that collicular neurons have on average 6,120 synaptic contacts on their receptive surface. The S/N ratio is lowest in layer III (4,330), while this ratio is highest in layers I and VII (i.e., 8,970 and 8,560 respectively). Layer II has a significantly higher S/N ratio than layer III (i.e., 8,060 and 4,330, respectively). Our results show that the size of synaptic terminals and contacts is not correlated with the different connectivity patterns of the distinct collicular layers. However, the density of synapses as well as the synapse-to-neuron ratios show a certain degree of laminar differentiation. In particular the superficial visual zone appears to be inhomogeneous in this respect, since layers I and II have a significantly higher density of synapses and higher S/N ratios than layer III. The deeper collicular zone is more homogeneously organized with synaptic densities similar to that of layer III and gradually increasing synapse-to-neuron ratios from layer IV to layer VII.