Chordotonal organs of Drosophila embryos have become models for studies of developmental biology and molecular genetics due to their consistent segmental placement and mutability. Our first goal was to find the origin and anatomical correlate of the blood-nerve barrier of this PNS proprioreceptor in wild type embryos. The concept of a blood-nerve barrier for the PNS of the Drosophila embryo is new, and the present data are the first in this regard. A second goal was to reveal the ultrastructure of these four-celled stretch receptors, focusing particularly on the 'core' of this organ: the bipolar neuron enclosed by a scolopale cell. These latter data have resulted in a graphic reconstruction of the chordotonal organ which reveals how the four consistent cells fit together. At Stage 13 we first observed a clearly recognizable scolopale cell with an enclosed neuron. Surprisingly, an operative blood-nerve barrier, comprised of occlusive pleated-sheet septate junctions, exists at this relatively early stage. A blood-brain barrier is not yet functioning in the CNS during this same stage, as the perineurium is not present until Stage 17. Cross-sectional views of a more mature chordotonal organ show that the neuron's inner segment has a 'tongue-in-groove' formation which fits the dendrite into the scolopale cell. Other newly discovered fine structural features are: hemidesmosomes linking individual scolopale rod bundles to the inner dendrite, and a cap cell matrix bonding with the tip of the ciliary dendrite. Functional aspects of these findings are discussed.