Thermoregulatory shivering is initiated if an individual’s body temperature (product of core (TC) and surface (TSk) temperatures) drops below their shivering threshold. Shivering thresholds may be affected by a number of factors including age, ethnicity, body mass index (BMI) and cold acclimatisation. It is not known if variation exists within a young, lean and non-cold acclimatised population. We studied shivering, TC and TSk during cold stress in this population.Twenty healthy subjects (10 male) lay supine between a water perfused blanket and mattress. Baseline characteristics: gender, age, BMI, ethnicity, cold acclimatisation and predisposition to shivering, were recorded. Subjects were cooled for 45 min by perfusing 4°C water through the mattress and blanket. TSk was recorded from the left forearm, index finger and thigh. TC was recorded using an ingestible pill telemetry system. Shivering was measured with surface EMG recordings from the left trapezius, biceps brachii, erector spinae and rectus femoris muscles.All EMG data were rectified and smoothed (τ = 20 s) prior to analysis. The onset of shivering was defined as the time relative to the start of cooling at which EMG amplitude exceeded twice the mean baseline level. All values are expressed as means ± SEM. Baseline characteristics were analysed using multiple logistic regression and Chi-Square tests. Experimental data were analysed with one-way ANOVA.Analysis revealed three discrete clusters of shivering onset latency. We characterised these clusters as: early shiverers (ES) (378.1 ± 87.5 s), late shiverers (LS) (2961.3 ± 137.5 s) and non-shiverers (NS). The onset latencies were different between ES and LS (p < 0.001). Mean TSk at the onset of shivering was higher in ES than LS (p < 0.001) but Tc was not different (p = 0.18). The rate of temperature change during cooling did not differ between the 3 groups. At 2 time points: mean ES onset and mean LS onset, TC and TSk were not different between the groups (ES: TC; p = 0.610, TSk; p = 0.458, LS: TC; p = 0.356, TSk; p = 0.943). The only baseline characteristic that predicted shivering onset latency was the subjective predisposition to shivering (p = 0.007).These data show that discrete patterns of shivering response exist within a young, lean and non-cold acclimatised population. TSk at the onset of shivering differed between the groups and therefore discrete TSk shivering thresholds may exist within this population. It is likely that these patterns are longstanding as individuals were able to predict their predisposition to shivering. The physiological basis of these discrete patterns remains to be established. An enhanced understanding will have direct clinical relevance to obstetric anaesthesia where a proportion of healthy patients shiver following neuraxial blockade.