Innocuous cold is detected in skin thermoreceptors by activation of the cold and menthol receptor TRPM8. The activation threshold of the cold- and menthol-activated current shifts towards lower temperatures during sustained cooling (cold adaptation), due to an increase in [Ca² ]_i (Reid et al., 2002). Cold adaptation does not occur in outside-out patches (Reid & Flonta, 2002) indicating that it is not intrinsic to the channel, but depends on the cellular environment. To clarify cold adaptation mechanisms in TRPM8, patch clamp recordings were made in cold- and menthol-sensitive rat dorsal root ganglion (DRG) neurones, pre-selected by imaging of [Ca² ]_i (methods described in Reid et al., 2002; rats killed by CO₂ inhalation followed by decapitation), and in rat TRPM8 expressed in HEK293 cells.Cold adaptation of TRPM8 was intact in HEK cells, and therefore does not depend on a mechanism found only in sensory neurones. Cold adaptation was not preserved by keeping the cytoplasm intact: in amphotericin “perforated vesicle”outside-out patches from DRG neurones (Levitan & Kramer, 1990), cold adaptation was absent and thermal threshold was shifted dramatically to lower temperatures (perforated vesicle, 24.1 ± 2.4 °C; amphotericin whole cell in the same neurones before patch excision, 37.6 ± 0.9 °C; n = 6; both in 100 µM (-)-menthol), exactly as we have reported in conventional outside-out patches (Reid & Flonta, 2002). In contrast, in some inside-out patches, cold adaptation was intact (3 of 10 patches from DRG neurones, 3 of 9 patches from HEK cells), and the current in these patches decayed with a time constant of 26-94 s (DRG, 75.4 ± 12.9 s; HEK, 48.9 ± 38.9 s;mean ± SD), similar to that in intact neurones (62 – 69 s; Reid & Flonta, 2001).This indicates that cold sensitivity and cold adaptation of TRPM8 do not depend on an intact cytoplasm, but rather on the local integrity of the membrane around the channel; this is always disrupted in forming an outside-out patch but is often preserved in inside-out patch formation. The simplest mechanism that could account for these observations would be an accessory Ca² sensing protein interacting with TRPM8 in a membrane-delimited manner, and reducing its cold sensitivity on binding Ca² ; the drastic loss of cold sensitivity on outside-out patch formation could result from membrane or cytoskeletal disruption or loss of an accessory protein that sensitises TRPM8. All procedures accord with the principles of UK legislation.