Qualitative and quantitative analyses of terminal arborizations of biocytin-labeled axon terminals were carried out in the cat primary visual cortex (V1). Extracellular iontophoretic injections of 5% biocytin were made into V1 of five adult cats. The animals were perfused 24-48 h after the injections. Labeled-axon fragments were considered to comprise two presumptive groups, according to the qualitative features, thickness, bouton features and appearance of terminal arbors. Forty axon fragments (20 for each presumptive group) were digitized using a microscope with motorized stage and a z-encoder, attached to a microcomputer. The densities of boutons, branching points and axon segments per mm of axon as well as axon segment length were used for comparison of the two groups. The two qualitative groups were confirmed to contain two axon types (I and II), according to cluster analysis of characteristics of the 40 axons in our sample. Forward stepwise discriminant analysis retained two variables as predictors of group membership: axonal length and bouton density. Parametric and non-parametric tests were employed for statistical comparisons (significance at P < 0.01). Type II axon fragments showed the greatest densities of boutons, axonal segments and branching points and the smallest values of length of segments ( P < 0.01). Both the qualitative and quantitative differences found for both types of axons suggest that they belong to different functional classes of neurons, namely spiny (type I) and smooth neurons (type II). Computer-assisted morphometric analysis of individual axon fragments seems to be a suitable approach with which different axon types can be objectively distinguished from each other.