In order to assess the role of input-target interactions in the development of olfactory cortex, the primary afferent fibers from the olfactory bulb to the superficial part of layer I of the cortex (layer Ia) were removed in developing and mature rats. After survival periods that vary from a few days to 2-6 months, changes were assessed in (1) the radial thickness of layer I, (2) the laminar distribution of intracortical associational fibers, which normally terminate in a deep part of layer I (layer Ib), and (3) the distribution of glia in layer I. The findings indicate that the lamination of fibers within layer I is not intrinsically prespecified, but gradually becomes "set" during the first month after birth. If the fibers from the olfactory bulb are removed, the dendrites of cortical cells are capable of accepting inputs from other fiber systems, depending on the maturational state of the dendrites and the ingrowing axons. Development of the abnormal inputs is associated with relatively normal dendritic growth, whereas lack of adequate input results in dendritic atrophy. Thus, after neonatal bulb ablation, the intracortical fibers occupy both superficial and deep parts of layer I, and a normal synaptic density is established throughout the layer. Layer I also develops to nearly its normal adult thickness, although the high density of glia that normally characterizes layer Ia is not apparent. With bulb ablation at progressively older ages (from postnatal day (P-) 3 to 21), the cortical associational fibers show progressively less extension into the denervated layer Ia. Layer I continues to grow, but not to the same extent as after P-1 ablations. In these experiments the glia distribution resembles the pattern present at the time of denervation. After adult olfactory bulb ablation, the long intracortical fibers extend very little into layer Ia, which undergoes pronounced shrinkage and becomes filled with a high concentration of glia. However, partial reinnervation of layer Ia is accomplished by the proliferation of a normally sparse native fiber system, which has been identified only with the Timm method. These results are interpreted as evidence that the normal development of lamination of afferent fibers to the olfactory cortex depends on axodendritic interaction during development.