1. Gap junction formation was compared in the absence and presence of small peptides containing extracellular loop sequences of gap junction (connexin) proteins by measuring the time taken for pairs of spontaneously beating embryonic chick heart myoballs to synchronize beat rates. Test peptides were derived from connexin 32. Non-homologous peptides were used as controls. Control pairs took 42 +/- 0.5 min (mean +/- S.E.M.; n = 1088) to synchronize. 2. Connexins 32 and 43, but not 26, were detected in gap junction plaques. The density and distribution of connexin immunolabelling varied between myoballs. 3. Peptides containing conserved motifs from extracellular loops 1 and 2 delayed gap junction formation. The steep portion of the dose-response relation lay between 30 and 300 microM peptide. 4. In loop 1, the conserved motifs QPG and SHVR were identified as being involved in junction formation. In loop 2, the conserved SRPTEK motif was important. The ability of peptides containing the SRPTEK motif to interfere with the formation of gap junctions was enhanced by amino acids from the putative membrane-spanning region. 5. Peptides from loop 1 and loop 2 were equivalently effective; there was no synergism between them. 6. The inclusion of conserved cysteines in test peptides did not make them more effective in the competition assay.