1. Sealed pigeon erythrocyte 'ghosts' were prepared containing ATP and the Ca2+-activated photoprotein obelin to investigate the relationship between cyclic AMP formation and internal free Ca2+. 2. The 'ghosts' were characterized by (a) morphology (optical and electron microscopy), (b) composition (haemoglobin, K+, Na+, Mg2+, ATP, obelin), (c) permeability to Ca2 , assessed by obelin luminescence and (d) hormone sensitivity (the effect of f8-adrenergic agonists and antagonists on cyclic AMP formation). 3. The effect of osmolarity at haemolysis and ATP at resealing on these parameters was investigated. 4. Sealed 'ghosts', containing approx. 2% of original haemoglobin, 150mM-K+, 0.5mM-ATP, 103-104 obelin luminescence counts/106 'ghosts', which were relatively impermeable to Ca2+ and in which cyclic AMP formation was stimulated by f,-adrenergic agonists over a concentration range similar to that for intact cells, could be prepared after haemolysis in 6mM-NaCl/3n1M-MgCl2/50mM-Tes, pH7, and resealing for 30min at 37°C in the presence of ATP and 150mM-KCl. 5. The initial rate of adrenaline-stimulated cyclic AMP formation in these 'ghosts' was 30-50 % ofthat in intact cells and was inhibited by the addition of extracellular Ca2+. Addition of Ca2+ to the 'ghosts' resulted in a stimulation of obelin luminescence, indicating an increase in internal free Ca2+ under these conditions. 6. The ionophore A23187 increased the rate of obelin luminescence in the 'ghosts' and also inhibited the adrenaline-stimulated increase in cyclic AMP. 7. The effect of ionophore A23187 on obelin luminescence and on cyclic AMP formation in the 'ghosts' was markedly decreased by sealing EGTA inside the 'ghosts'. 8. It was concluded that cyclic AMP formation inside sealed pigeon erythrocyte 'ghosts' could be inhibited by more than 50% by free Ca2+ concentrations in the range 1-1O0uM.A considerable amount of indirect evidence exists that changes in intracellular Ca2+ play an important part in the action of certain hormones (Berridge, 1976;Hales et al., 1977;Rasmussen & Goodman, 1977). Investigations of the effects of Ca2+ on adenylate cyclase have mostly been carried out on brokencell preparations, in the absence of ionic gradients across the cell membrane, and have often involved the use of high non-physiological concentrations of Ca2+ (Hales et al., 1977). Although the concentration of intracellular free Ca2+ has only been measured in a few large cells such as barnacle muscle (Ashley & Ridgway, 1970) and the squid giant axon (Baker et al., 1971;Baker, 1972), it now seems likely that under physiological conditions the concentration of free Ca2+ in cells does not exceed 10pUM. In intact pigeon erythrocytes it has been reported that adenylate cyclase can be inhibited by concentrations of Ca2+ less than 10puM (Campbell & Siddle, 1976). Further progress in the investigation of interactions between Vol. 176 intracellular Ca2+ and cyclic nucleotides has been hampered by the lack of suitable preparations where it is possible to study directly changes in ...