1. Giant axons from the squids Dosidicus gigas, Loligo forbesi and Loligo vulgaris were internally perfused with 550 or 275 mM KF plus sucrose and bathed in artificial sea water containing 45Ca, 28Mg or mixtures of 45Ca-28Mg or 45Ca-22Na. Resting influxes and extra influxes during voltage-clamp pulses were measured by collecting and counting the internal perfusate. 2. For Dosidicus axons in 10 mM-CaCl2 the resting influx of calcium was 0-016 +/- 0-007 p-mole/cm2 sec and a linear function of external concentration. For two experiments in 10 and 84-7 mM-CaCl2, 100 nM tetrodotoxin had no effect. Resting calcium influx in 10 mM-CaCl2 was 0-017 +/- 0-013 p-mole/cm2 sec for Loligo axons. 3. With 55 mM-MgCl2 outside the average resting magnesium influx was 0-124 +/- 0-080 p-mole/cm2 sec for Loligo axons. Discarding one aberrant point the value is 0-105 +/- 0-046 which is not significantly different from the resting calcium influx for Dosidicus fibres in 55 mM-CaCl2, given as 0-094 p-mole/cm2 sec by the regression line shown in Fig. 1. In two experiments 150 nM tetrodotoxin had no effect. 4. With 430 mM-NaCl outside 100 nM tetrodotoxin reduced the average resting influx of sodium in Dosidicus axon from 27-7 +/- 4-5 to 25-1 +/- 6-2 p-mole/cm2 sec and for Loligo fibres in 460 mM-NaCl from 50-5 +/- 4 to 20 +/- 8 p-mole/cm2 sec. 5. Using depolarizing pulses of various durations, the extra calcium influx occurred in two phases. The early phase was eliminated by external application of tetrodotoxin. The results of analysis are consistent with, but do not rigorously demonstrate, the conclusion that the tetrodotoxin sensitive calcium entry is flowing through the normal sodium channels (cf. Baker, Hodgkin & Ridgway, 1971). 6. Measurements of extra influxes using 22Na and 45Ca simultaneously indicate that the time courses of tetrodotoxin sensitive calcium and sodium entry are similar but not necessarily identical. It is very doubtful that any significant calcium entry occurs before the sodium or is involved in the activation of the sodium system. 7. These measurements confirm for Loligo, as previously shown for Dosidicus axons, that the magnitude and time course of the sodium entry during a depolarizing pulse deduced from electrical measurements is the same as that measured with 22Na. 8. Using 28Mg, or mixtures of 45Ca and 28Mg, we observed a single phase of magnesium entry which was insensitive to external tetrodotoxin or internal tetraethyl ammonium. The magnitude of the magnesium influx was considerably greater than the calcium extra entry and large enough to have been detected in the experiments of Meves & Vogel (1973) if it represented current. 9. We suggest the possibility that the calcium and magnesium extra influxes, after external treatment with tetrodotoxin, during a depolarizing pulse, do not contribute to the measured current.