Since the discovery of the functional diversity of Na+ currents in primary sensory neurones, the physiological reasons for their heterogeneity have been the subject of investigation and speculation. This is certainly true for the tetrodotoxin-resistant and kinetically slow NaV1.8 current expressed in small diameter nociceptive neurones. The normal expression pattern is confined to neurones in sensory ganglia, and this channel has been the subject of scrutiny over many years as a possible novel pharmacological target for the treatment of pain. Because of its relatively depolarized activation voltage dependence, one possible functional attribute is that the channel may continue to function in small diameter afferents when depolarized, for example in the face of raised external K+. This possibility was recognized by Jim Elliott and colleagues many years ago, but it is in the study of cold reception that this plausibility has become something close to a certainty. In the presence of large and long-lasting receptor potentials, when other Na+ channel subtypes should be substantially inactivated, NaV1.8 still functions and endows excitability to corneal cold receptors (e.g. Carr et al. 2002). The knock-out of NaV1.8 has recently been shown to suppress nociception under cold conditions (Zimmermann et al. 2007), and also to inhibit menthol-induced enhanced cold responses (Zimmermann et al. 2007). A widely held view concerning gene knock-outs is that the functional consequences can be modified or diluted by alterations in the expression of other genes, and this argument was put forward at the time in order to explain the phenotype of the NaV1.8 knock-out mouse, where the (apparently) fairly mild phenotypic changes accompanying channel loss were in part explained by a compensatory up-regulation of NaV1.7. What has been described now in the paper by Sturzebecher et al. (2010) in this issue of The Journal of Physiology is a cell-autonomous tethered-toxin approach for the investigation of NaV1.8 function, where the expression of NaV1.8 is expected to be normal, but where channel function is suppressed (although not abolished) by the provision of tethered conotoxin MrVIa (Daly et al. 2004) on the outside of the nerve membrane. The utilized transgene encoding the toxin included genomic DNA incorporating modified SCN10a, resulting in expression of the toxin in sensory neurones when the SCN10a promoter is activated. The toxin was engineered as the most distal part of a glycophosphatidylinositol anchored extracellular protein, and the evidence presented indicates that in practice this fusion protein is sufficiently flexible to allow the tethered-toxin to exhibit its pharmacological action and block NaV1.8. The partial toxin block of NaV1.8 brings about a reduction in noxious cold sensitivity and a substantially reduced behavioural response to noxious cold (0°C; Fig. 4C in Sturzebecher et al. 2010; highlighted in the authors’ Supplemental movie, available online). These findings are consistent with the abolition of behavioural responses to cold following the ablation of neurochemically distinguished isolectin-IB4+ve neurones, by endogenous production of diphtheria toxin controlled by the SCN10a promotor (Abrahamsen et al. 2008). A contribution made by Snape et al. (2010) was to show that IB4+ve, small diameter sensory neurones, well known to be endowed with more functional NaV1.8 than similarly sized IB4–ve neurones, retain their excitability with applied depolarization approaching –20 mV, potentials at which IB4–ve neurones were rendered inexcitable. The following conclusions seem almost inescapable. Firstly, nociceptive nerve endings in the cold may be undergoing a resting depolarization to some degree, by loss of K+ channels as suggested by Gordon Reid (Reid & Flonta 2001) (whether or not they contain e.g. TRPM8 menthol receptors or other cold-gated channels), or tetrodotoxin-sensitive Na+ channels become functionally unavailable because of the effect of cold on gating (Zimmermann et al. 2007). Secondly, NaV1.8 allows sensory encoding to occur in depolarized endings in the cold (Carr et al. 2002; Zimmermann et al. 2007), and finally, NaV1.8 is the likely channel to encode cold (Carr et al. 2002; Zimmermann et al. 2007; Sturzebecher et al. 2010). It is our expectation that NaV1.8 will play other additional special roles that have yet to be clearly defined.