Intracellular movement of vesiculated pigment granules in angelfish melanophores is regulated by a signalling pathway that triggers kinesin and dyneinlike microtubule motor proteins. We have tested the relative importance of intracellular Ca2+ ([Ca2+]i) vs cAMP ([cAMP]i) in the control of such motility by adrenergic agonists, using fluorescence ratio imaging and many ways to artificially stimulate or suppress signals in these pathways. Fura-2 imaging reported a [Ca2+]i elevation accompanying pigment aggregation, but this increase was not essential since movement was not induced with the calcium ionophore, ionomycin, nor was movement blocked when the increases were suppressed by withdrawal of extracellular Ca2+ or loading of intracellular BAPTA. The phosphatase inhibitor, okadaic acid, blocked aggregation and induced dispersion at concentrations that suggested that the protein phosphatase PP-1 or PP-2A was continuously turning phosphate over during intracellular motility. cAMP was monitored dynamically in single living cells by microinjecting cAMP-dependent kinase in which the catalytic and regulatory subunits were labeled with fluorescein and rhodamine respectively (Adams et al., 1991. Nature (Lond.). 349:694-697). Ratio imaging of F1CRhR showed that the alpha 2-adrenergic receptor-mediated aggregation was accompanied by a dose-dependent decrease in [cAMP]i. The decrease in [cAMP]i was both necessary and sufficient for aggregation, since cAMP analogs or microinjected free catalytic subunit of A kinase-blocked aggregation or caused dispersal, whereas the cAMP antagonist RpcAMPs or the microinjection of the specific kinase inhibitor PKI5-24 amide induced aggregation. Our conclusion that cAMP, not calcium, controls bidirectional microtubule dependent motility in melanophores might be relevant to other instances of non-muscle cell motility.