The shape of the Mauthner neuron (M-neuron) and the distribution of its afferent synapses were studied between days 2 and 6 after fertilization in the zebrafish Brachydanio rerio. This interval is just after the outgrowth of M-dendrites begins, and during this time the M-cell acquires its definitive shape. The M-cell has two large invariant dendrites: The lateral dendrite terminates in the sensory neuropil of the acoustico-lateral area, and th ventral dendrite terminates in the neuropil of the motor tegmentum. Fine dendrites are present, and mostly arise from three regions; from the terminus of each major dendrite and from the ventral surface of the perikaryon. The number and position of fine dendrites within each of these sets is variable, even among animals from a single isogenic clone. M-cells with improper numbers or positions of large dendrites were never encountered, even early in development. This suggests that their outgrowth is a highly directed process. Large numbers of afferent synapses are formed on the M-cell during the time of dendrite outgrowth. By day 6 there is a mosaic pattern of morphologically distinctive terminals that is similar to the pattern of the adult goldfish M-cell. Identified categories of terminals include (1) myelinated club endings, on the distal part of the lateral dendrite, (2) boutons, on the dendrites and perikaryon, (3) unmyelinated club endings, on the dorsomedial portion of the perikaryon adjacent to the axon cap, and (4) spiral fiber terminals within the axon cap. The nonrandom nature of the input may be ascertained by observing the distribution of electrotonic or gap junctions on the cell surface. These are frequently encountered on the initial segment of the axon (spiral fiber terminals), ventral dendrite and ventral perikaryon (boutons), and distal lateral dendrite (myelinated club endings). Gap junctions are only rarely observed on the dorsal surface of the cell, although this region, like others of the cell, receives large numbers of chemical synaptic contacts. This pattern is similar at all stages studied, which suggests that no large rearrangements in synaptic contacts occur during this developmental period. We discuss these observations in relation to the hypothesis that patterned dendritic growth of the M-cell is directed by synaptic interactions with the afferents.