Synapses at larval neuromuscular junctions of the flies Drosophila melanogaster and Sarcophaga bullata are not distributed randomly. They have been studied in serial electron micrographs of two identified axons (axons 1 and 2) that innervate ventral longitudinal muscles 6 and 7 of the larval body wall. The following fly larvae were examined: axon 1--wild-type Sarcophaga and Drosophila and Drosophila mutants dunce(m14) and fasII(e76), a hypomorphic allele of the fasciclin II gene; and axon 2--drosophila wild-type, dunce(m14), and fasII(e76). These lines were selected to provide a wide range of nerve terminal phenotypes in which to study the distribution and spacing of synapses. Each terminal varicosity is applied closely to the underlying subsynaptic reticulum of the muscle fiber and has 15-40 synapses. Each synapse usually bears one or more active zones, characterized by dense bodies that are T-shaped in cross section; they are located at the presumed sites of transmitter release. The distribution of synapses was characterized from the center-to-center distance of each synapse to its nearest neighbor. The mean spacing between nearest-neighbor pairs ranged from 0.84 microm to 1.05 microm for axon 1, showing no significant difference regardless of genotype. The corresponding values for axon 2, 0.58 microm to 0.75 microm, were also statistically indistinguishable from one another in terminals of different genotype but differed significantly from the values for axon 1. Thus, the functional class of the axon provides a clear prediction of the spacing of its synapses, suggesting that spacing may be determined by the functional properties of transmission at the two types of terminals. Individual dense bodies were situated mostly at least 0.4 microm away from one another, suggesting that an interaction between neighboring active zones could prevent their final positions from being located more closely.