The foveas of nine avian species, initially selected for the presence of a deep fovea and representing a wide range of eye sizes and ecological habits, were studied with quantitative light microscope techniques. Considerable variation was observed in the location and configuration of the avian foveas, although they appeared to be ''convexiclivate'' in shape when compared with the fovea of the rhesus monkey. Comparisons of foveal cell densities (receptor nuclei and ganglion cells) across species showed an increase in the average number of cells/visual degree² with increasing eye size; similarly, an increase occurred in receptor nuclei relative to ganglion cell density. Thus, smaller eyes showed a coarser retinal grain and a lower ''coincidence ratio'' of receptors to ganglion cells than was found in the largest eyes. There appeared to be no relationship between receptor densities/mm² and (a) eye size, (b) depth of foveal clivus, or (c) width of foveal clivus. However, a negative correlation was generally observed between the width of the foveal clivus and eye size. Two foveas were seen in the red-tailed hawk, goshawk, sparrow hawk, and least tern. The central fovea was more differentiated, with greater densities of both receptor nuclei and ganglion cells than was observed in the temporal fovea of the same species. Further conclusions, particularly with respect to potential visual acuity, await quantitative measurements of foveal cone densities across species.

The hippocampal complex (hippocampus and parahippocampalis) is known to play a role in spatial memory in birds and is known to be larger in food-storing versus non-storing birds. In the present study, we investigated the relative volume of the hippocampal complex in four food-storing corvids: gray-breasted jays (Aphelocoma ultramarina), scrub jays (Aphelocoma coerulescens), pinyon jays (Gymnorhinus cyanocephalus), and Clark's nutcrackers (Nucifraga columbiana). The results show that Clark's nutcrackers have a larger hippocampal complex, relative to both body and total brain size, than the other three species. Clark's nutcrackers rely more extensively on stored food in the wild than the other three species. Clark's nutcrackers also perform better during cache recovery and operant tests of spatial memory than scrub jays. Thus, greater hippocampal volume is associated with better performance in laboratory tests of spatial memory and with stronger dependence on food stores in the wild.

The central projection and retinal distribution of displace ganglion cells (DGC's) are described for the pigeon. Discrete, localized injections of horseradish peroxidase (HRP) into the nucleus of the basal optic root (nBOR) complex labeled as many as 4,800 DGC's in the contralateral retina. The greatest densities of DGC's were observed in the more peripheral regions of the middle and inferior temporal regions of the retina, with lowest densities occurring in the inferior nasal, red field, and foveal areas. Large HRP injections of the tectal lobes, which did not include the pretectal, accessory optic (nBOR), hypothalamic, or thalamic visual nuclei, labeled only ganglion cells within the ganglion cells layer. An HRP injection centered within the nucleus lentiformis mesencephali, also including portions of the optic tectum and optic tract, labeled only ganglion cells within the ganglion cell layer of the contralateral retina. DGC's thus appear to be the primary, if not exclusive, source of retinal afferents to the nBOR complex in pigeon. The observed retinal distribution of DGC's indicates that the areas of retina with the greatest density of cells in the receptor layer, inner nuclear layer, and ganglion cell layer are relatively devoid of DGC's. Since the nBOR complex projects directly upon the vestibulocerebellum and oculomotor nuclei, DGC's would thus appear to be involved in neural circuits that mediate oculomotor reflexes and visuomotor behavior.

In the pigeon, the nucleus of the basal optic root, a component of the accessory optic system, projects directly upon the vestibulo-cerebellum. This nucleus receives a prominent projection composed of large-diameter retinal axons, known as the basal optic root. The cells of origin of this tract were identified using horseradish peroxidase (donor:hydrogen-peroxide oxidoreductase, EC 1.11. In all vertebrates, efferent nerve pathways from the retina terminate upon a number of distinct central nervous system structures including the optic tectum, dorsal and ventral thalamus, pretectum, and the accessory optic nuclei. Retinal neurons which give rise to these projections are known collectively as ganglion cells. Dogiel (1), Cajal (2, 3) and others have demonstrated that ganglion cells are a morphologically heterogeneous population and may be distinguished by variations in somatic size, configuration and lamina of dendrite distribution in the inner plexiform layer, as well as by axon caliber. Recently, West and Dowling (4) and West (5) have described as many as 15 morphologically distinct categories of ganglion cells in the ground squirrel retina.The vast majority of ganglion cells are located in the ganglion cell layer proper, immediately above (distal to) the layer of optic nerve axons, but Dogiel (1) also noted the presence of large ganglion cells along the inner margin of the inner nuclear layer. These have subsequently been referred to as the "displaced ganglion cells of Dogiel," and are found in all vertebrate classes (6). They are, however, particularly conspicuous in avian retinae (1). Relatively little information is available concerning their anatomical and physiological characteristics (7-9).Confronted with the rich morphological variety of ganglion cells, many workers have sought to determine which ganglion cell types terminate within each central target. These studies, however, have relied almost exclusively upon the characteristic of cell size without utilizing information concerning the configuration and lamina of ganglion cell dendritic arborizations within the inner plexiform layer (10-13). Furthermore, the analysis has been complicated by the observations that some central areas receive input from several types of ganglion cells (using the criterion of cell size), while other data suggest that a single type of ganglion cell may project to several different central targets (12). The use of horseradish peroxidase (HRP) (donor:hydrogen-peroxide oxidoreductase, EC 1.11.1.7) as a retrograde marker to identify the cells of origin of a particular projection has proven of considerable value in clarifying the problem of ganglion cell-central nervous system relationships. For example, injections of HRP into either the optic tectum or the dorsal thalamus has permitted the identification of cells in the retina that project to these regions (9-13). The HRP method, however, has not revealed sufficient morphological detail to permit unequivocal correlation of ganglion cell type and central target. A second m...

The accessory optic system of Rana pipiens was investigated by autoradiographic, horseradish peroxidase, and Golgi techniques, revealing a complexity of neuroanatomical organization previously unrecognized. Retinal afferents project to the nucleus of the basal optic root (nBOR) via a primary bundle and more diffuse, medial bundle of optic axons, both of which contain large- and small-diameter fibers. At least six types of retinal ganglion cell contribute to the basal optic root (BOR), including giant ganglion cells, two intermediate-sized ganglion cell types, small ganglion cells, and two types of displaced ganglion cell. The major retinal projection is contralateral, but a small, ipsilateral component also exists. Afferents from neurons which are postsynaptic to the thalamic retinal terminal fields also reach nBOR. Four distinct cell types were identified within the terminal field of nBOR: stellate neurons (63%), amacrine cells (19%), elongate neurons (14%), and large ganglionic neurons (4%). Both stellate and amacrine cells appear to be intrinsic neurons, while elongate and ganglionic neurons constitute the efferent neuron population of nBOR. In addition, cells which lie medial to the terminal field, pyriform and commissural neurons, send dendrites into nBOR. Pyriform neurons project to the nucleus of the medial longitudinal fasciculus (nMLF) and cranial nerve nuclei III and IV, while commissural neurons project to the contralateral nBOR. Large reticular neurons of the nMLF also send dendrites into nBOR. Efferent projections from nBOR were observed in the large-celled pretectal nucleus and in nucleus lateralis profundus. A second major projection originates from the peri-nBOR region and is associated with the oculomotor system and with the nMLF. Efferent projections from the nMLF to the vestibular nuclei and to the rostral spinal cord were also observed, as well as projections which reach the brainstem from the large-celled pretectal nucleus, the posterior thalamic and anterior mesencephalic central gray.