BACKGROUND AND OBJECTIVES Minimum T50 and T95 onsets are the medians of time intervals between end of skin and subcutaneous tissue infiltration with local anesthetics and a painful stimulation, which are associated to no pain in 50% and 95% of patients, respectively. This study has evaluated T50 and T95 of subcutaneous 1% lidocaine for lumbar skin puncture. METHODS This was a prospective study with sequential allocation depending on the response of previous patient (Dixon and Massey's up-down method) involving 25 patients submitted to spinal blocks with 25G Quincke needle. Skin and subcutaneous tissue were infiltrated with 3 mL of 1% lidocaine in L3-L4 or L4-L5 interspace with 10 mm 25G needle. Time elapsed between the end of infiltration and 25G Quincke needle puncture was established in logarithmic scale. Initial interval was 30 seconds. Subsequent intervals depended on previous patient's response. If the patient referred pain at 25G Quincke needle puncture, the next patient was allocated to the immediately longer interval; otherwise, the next patient would be allocated to the immediately shorter interval. Dixon and Massey's formulas were applied to calculate T50 and T95. RESULTS Medians (95% confidence limits) of T50 and T95 were 25 seconds (20-31 seconds) and 53 seconds (33-85 seconds), respectively. CONCLUSIONS To obtain lumbar skin anesthesia before lumbar puncture in 95% of patients, one should wait 33 to 85 seconds after the end of subcutaneous infiltration of 3 mL of 1% plain lidocaine.