1. This study was designed to determine whether the duration and pattern of prior insulin exposure modulate the symptomatic and counterregulatory responses to hypoglycaemia. 2. Ten healthy non-obese subjects (five males/five females age 25 +/- 1 years, mean +/- SEM) were made hypoglycaemic in three ways: (i) a hyperinsulinaemic (60 m-units min-1 m-2; plasma insulin concentration 95 m-units/I) clamp, with 1 h of euglycaemia, blood glucose level 4.5 mmol/l, followed by 30 min of hypoglycaemia, at a stable glucose nadir of 2.0 mmol/l (i.e. euglycaemic then hypoglycaemic clamp: E + HC); (ii) an identical hypoglycaemic clamp without preceding hyperinsulinaemic euglycaemia (i.e. a hypoglycaemic clamp: HC); (iii) insulin infusion only, discontinued at a blood glucose level of 3.0 mmol/l (II). Blood glucose level reached the same nadir as on E + HC and HC, and did not fall further. At the glucose nadir, and 15 and 30 min after, the plasma insulin concentration was 23, 7 and 4 m-units/l, respectively, on the II visit. 3. At the glucose nadir, plasma glucagon level, plasma adrenaline level, sweating rate, heart rate, blood pressure, and overall and individual symptom scores (using visual analogue scales) were the same on E + HC, HC and II. 4. There were no significant differences in neurohormonal response between E+HC and HC, but more subjects felt hypoglycaemic on E + HC on arrival at the glucose nadir (P < 0.05).(ABSTRACT TRUNCATED AT 250 WORDS)