The number, dendritic morphology, and retinal distribution of displaced ganglion cells were studied in two anuran species, Xenopus laevis and Bufo marinus. Horseradish peroxidase or cobaltic lysine complex was applied to the cut end of the optic nerve, and the size, shape, and retinal position of retrogradely filled ganglion cells displaced into the inner nuclear layer were determined in retinal wholemount and sectioned material. Approximately 1% of ganglion cells in Xenopus and 0.1% in Bufo were found to be displaced. In both species, many of the previously described orthotopic ganglion cell types (Straznicky & Straznicky, 1988; Straznicky et al., 1990) were present among displaced ganglion cells. In Xenopus more displaced ganglion cells were found in the retinal periphery than in the retinal center, and they formed 3 or 4 distinct bands around the optic nerve head. In Bufo the incidence of displaced ganglion cells was higher along the visual streak than in the dorsal and ventral peripheral retina. These results indicate that the distribution of displaced ganglion cells approximates the retinal distribution of orthotopic ganglion cells. One of the likely mechanisms to account for this developmental paradox may be that the formation of the inner plexiform layer, adjacent to the ciliary margin, acts as a mechanical barrier by preventing the entry of some of the late developing ganglion cells into the ganglion cell layer.

In albinos, the retinofugal projections to the ipsilateral side of the brain are reduced (e.g., see Guillery, 1969; LaVail et al., 1978; Lund, 1965). Although all ganglion cell types are affected, in mice the displaced ganglion cell population is the main target of the albino mutation (Dräger & Olsen, 1980). Here we tested whether this preferential effect on displaced ganglion cells is a general consequence of the melanin reduction or a pleiotropic effect unique to the albino locus, retrogradely tracing tracing retinal ganglion cells in normal C57BL/6J mice and in several non-allelic hypopigmentation mutants on the same background: albino (C57BL/6J-c2J), beige (C57BL/6J-bg), pale ear (C57BL/6J-ep), ruby-eye/haze (C57BL/6J-ru-2hz), and pearl (C57BL/6J-pe). All mutants have lower overall cell counts in the ipsilateral projection, but the displaced population is disproportionately affected: the albinos contain 42% of the normal number of displaced ganglion cells, and the other mutants have an average 57% of normal counts.

The reduction in uncrossed retinofugal projections in albinos affects the inputs to the lateral geniculate nucleus and the superior colliculus, but not to the suprachiasmatic nucleus (Dräger, 1974). To address the question in which way the susceptible uncrossed projections differ from the nonsusceptible one, we compared ganglion cells backfilled from the suprachiasmatic nucleus to ganglion cells backfilled from the optic tract at geniculate level. Whereas the uncrossed optic tract projection originates from the binocular region in the ventro-temporal retina and contains a high fraction of large and displaced ganglion cells (Dräger & Olsen, 1980), both the crossed and uncrossed inputs to the suprachiasmatic nucleus originate from the entire retina with a relative preference for the lower nasal region that corresponds to part of the monocular visual field; all ganglion cells projecting to the suprachiasmatic nucleus are of medium size, and they are located in the ganglion cell layer.

These observations allow the following conclusions: (1) All genetic mutants which cause a reduction in ocular melanin, regardless of the molecular or cell-biological mechanism underlying the pigment reduction, result in decreased uncrossed projections; this confirms previous reports (LaVail et al., 1978, Sanderson et al., 1974). (2) The decrease affects only projections involved in binocular vision. (3) In mice, the ganglion cells displaced to the inner nuclear layer, and hence located closer to the retinal pigment epithelium, are disproportionately affected by the melanin reductions. These observations may provide cues to the spatio-temporal mechanism of the melanin action in the embryonic visual system.