The laminar distribution and structure of the supragranular cells projecting from primary auditory cortex (AI) to the second auditory cortex (AII) in the cat were studied with horseradish peroxidase. Injections in AII retrogradely labeled somata in ipsilateral cortical layers I-VI of AI. A bimodal laminar disposition was found, with approximately 40% of the labeled cells in layer III, 25% in layer V, and 10-15% each in layers II, IV, and VI; only a few cells were found in layer I. The labeled cells were scattered in small aggregates between which unlabeled neurons were interspersed. There was some, though not a strict, topographical distribution of the labeled cells according to the locus of the injection in AII. Injections in the caudal part of AII labeled cells in more rostral AI, while rostral AII injections labeled cells in more caudal AI. Ventral AII injections labeled more ventrally located AI cells, while more dorsal AII injections labeled more dorsally situated AI cells. AII injections also labeled cells in other auditory cortex subdivisions, including the posterior ectosylvian, ventroposterior, temporal, and dorsal auditory zone/suprasylvian fringe cortical areas, and in some non-auditory cortical areas. In layers II and III, both pyramidal and non-pyramidal cells were labeled. More pyramidal cells were labeled in layer III than layer II (80% vs. 62%), and the proportion of non-pyramidal cells in layer II was more than twice that in layer IV (27% vs. 12%). The types of labeled cells were distinguished from one another on the basis of size, somatic and dendritic shape, and laminar distribution. The profiles of labeled cells in these experiments were compared to, and correlated with, those in Golgi-impregnated material. In layer II, the classes of corticocortical projecting cells consisted of small and medium-sized pyramidal, bipolar, and multipolar cells. Those in layer III included small, medium-sized, and large pyramidal neurons, and bipolar and multipolar cells. The average somatic area of the labeled cells did not differ significantly from that of the unlabeled cells, and both were about equal in somatic size to neurons accumulating tritiated gamma-aminobutyric acid in layers II and III. These findings suggest that there is convergent, ipsilateral input onto AII from every layer in AI, and from other cortical auditory and non-auditory areas. A morphologically heterogeneous population of cells in AI contributes to these projections. Diversity in the cytological origins of corticocortical projections implies functional differences between layers II and III since the latter also projects commissural, while layer II in the cat, does not.