1. In the large single muscle fibres from the barnacle Balanus nubilus, the total fibre Mg concentration was estimated as 15.1 m-mole/kg wet wt., of which about 3-3.5 m-mole/kg wet wt. was extracellular. The diffusible Mg, measured by internal sampling, was 11.5 m-mole/kg wet wt., of which at least half may be complexed to larger diffusible molecules. The free ionized Mg level was estimated as < 5 m-mole/kg wet wt.2. The loss of [(28)Mg]MgCl(2) from both Maia and Balanus muscle fibres following axial micro-injection approximated to first-order kinetics. The maximum rate constant for the loss was 1.51 +/- 0.20 (S.E.) x 10(-5) sec(-1) for Balanus (sixty-seven fibres) and 1.06 +/- 0.46 (S.E.) x 10(-5) sec(-1) for Maia (seven fibres) at 20-25 degrees C.3. The calculated Mg efflux was in the range 6-12 p-mole/cm(2).sec based on this rate constant, assuming isotopic equilibration internally and that the surface area of the fibres approximated to that of a simple cylinder. If account was taken of the area of the cleft system the efflux was reduced by about fifteen times.4. The diffusion coefficient for injected (28)Mg was estimated as 2-3 x 10(-6) cm(2) sec(-1), about half the value in free solution. Injections of 2 M-MgCl(2) or 200 mM-EDTA subsequent to the injection of the isotope caused about a 30% reduction in the tracer efflux.5. External application of salines containing 100 mM-Ca or Mg caused a rapid but reversible inhibition of the magnesium efflux. Similar effects were observed with salines containing 32 mM-Co or Mn chlorides or 1-2 mM-La or Gd chlorides. Polyarginine (200 mug/ml.) had no effect.6. The Mg efflux had a Q(10) of 3-4 over the temperature range of about 5-20 degrees C. It was irreversibly inhibited by the sulphydryl reagent NEM (1 mM), but PCMBS (0.2-2 mM) had no effect. Contractile agents (5 mM caffeine or 200 mM-K salines) and a variety of inhibitors of ion movement or active transport had no appreciable effect on the Mg efflux. Lowering the pH of the saline from 7 to 5 produced a 70% reduction in the efflux which was reversible over short periods of application.7. Replacement of external Na, but not Ca or Mg, with Li, choline or sucrose caused a rapid and partially reversible reduction of the Mg efflux, but increasing the internal Na by micro-injection in zero Na salines had no consistent effect. It is suggested that the extrusion of Mg from these muscle cells is largely dependent upon the inward movement of Na.