Organization of nucleus rotundus, a tectofugal thalamic nucleus in turtles. III. The tectorotundal projection

Nucleus rotundus is the primary thalamic recipient of projections from the optic tectum in pond turtles. Although the projection of the retina to the optic tectum is known to be topographically organized, earlier studies suggest that the tectorotundal projection is not topographically organized. Three types of analyses are used in this paper to characterize the organization of the projection of the optic tectum to nucleus rotundus. First, large iontophoretic injections of horseradish peroxidase into the optic tectum anterogradely fill axons with reaction product after the use of a cobalt-enhanced diaminobenzidine procedure. These preparations show that shafts of axons in the tectothalamic tract give rise to thinner, primary collaterals that enter nucleus rotundus from its caudolateral aspect and form sparsely branching arbors within the nucleus. Very thin secondary collaterals branch from these collateral bear terminal collaterals with frequent varicosities. Although the total size of such arbors is unknown, the evidence suggests that each arbor is large in relation to the size of nucleus rotundus. Thus, injection sites restricted to central tectum label axons throughout nucleus rotundus. Second, subtotal lesions of the tectum produce degeneration throughout nucleus rotundus in silver degeneration preparations. Finally, analysis of electron microscopic degeneration material indicates that tectal boutons are distributed along the full lengths of the dendrites of rotundal neurons, but not on their somata. These boutons form asymmetric synaptic junctions and contain round synaptic vesicles. In view of the relatively large size of the dendritic fields of rotundal neurons, these data suggest that the tectorotundal projection is both strongly convergent on individual neurons and strongly divergent from single tectorotundal axons. This type of organization is consistent with physiological evidence that rotundal neurons have receptive fields that cover at least one-half of the contralateral visual field and often include the entire hemifield. It seems unlikely that nucleus rotundus can be involved in neuronal transactions that preserve detailed spatial information, but it may be involved in processing information on other visual parameters such as stimulus velocity or color.