The firing responses of neurons to current injection determine input-output relationships and influence the operation of neuronal circuits. Commonly evoked patterns include firing a single spike and firing repetitively. We therefore examined the membrane currents underlying these patterns in two types of spinal neuron in young Xenopus laevis tadpoles. Neurons were characterised in vivo using the whole-cell voltage-clamp technique, and categorised using a predictive classification algorithm based on measurements of input resistance, resting membrane potential and soma diameter. Tadpoles were initially anaesthetised with MS-222 (1mg/ml), and immobilised with α-bungarotoxin (10µm) during experiments. Experiments and husbandry were in accordance with the UK Home Office Animals (Scientific Procedures) Act 1986 and received local ethical approval. Single-firing primary-sensory Rohon-Beard (RB) neurons expressed a tetrodotoxin-sensitive transient sodium current at a significantly higher density than repetitive-firing dorsolateral (DL) neurons (RB, 2.5pA/µm2 v. DL, 0.8pA/µm2, n = 4 and 3 respectively, p = < 0.001, independent-sample t-test). High-voltage-activated calcium currents in RB and DL neurons were similar in current density and voltage-dependence (RB, 1.0pA/µm2 v. DL, 0.6pA/µm2, n = 8 and 7 respectively, independent-sample t-test). There was no evidence of persistent sodium currents, low-voltage-activated calcium currents, or hyperpolarisation-activated currents. Macroscopic potassium currents were different in RB and DL neurons, and tail current analysis revealed separable fast (IKf) and slow (IKs) components. These components differed in proportion: IKf mediated 23 ± 10% of the total current in RB neurons and 80 ± 4% in DL neurons (n = 13 and 8 respectively, p = 0.002). IKf was selectively blocked by 4-aminopyridine (4-AP, 2mM) and IKs was selectively blocked by tetraethylammonium (TEA, 10mM). Sequential current and voltage clamp recordings in individual neurons suggested IKs determined if neurons fired repetitively and IKf regulated firing frequencies. To test these predictions we constructed Hodgkin-Huxley-based models of currents, fitting the time-constants and steady-state conductance parameters to the biological data. We incorporated these modelled currents into single compartment model neurons to investigate the effects of changing their relative proportions on the firing pattern. This integrative approach has revealed the role of potassium currents in determining firing patterns.