*2.5. Method Validation Procedure*

The method was validated according to the Codex guidelines [32] and the parameters included linearity, specificity, accuracy, precision, limit of quantification (LOQ), limit of detection (LOD), stock solution stability and sample stability. For the calibration curve, linearity was evaluable by a seven-point detection matrix-matched calibration curve. For gelsemine and koumine, the concentrations were 5, 20, 50, 100, 200, 500 and 1000 ng/g; for humantenmine, the concentrations were 20, 40, 100, 200, 400, 800 and 1000 ng/g. The standard curve was constructed with an added concentration (X) as the abscissa and peak area (Y) as the ordinate.

Recovery was determined by four concentration added levels of QC samples, including QCLL (1 × LOQ), QCL (3 × LOQ), QCM (6 × LOQ) and QCH (30 × LOQ) (*n* = 6) detection. Precision is measured by intraday precision and interday precision, usually expressed as RSD values. Intraday and interday precision is the result of the repeated analysis of honey QC samples at four supplemental concentrations (*n* = 6) for one day (intraday precision) and three consecutive days (intraday precision).

The LOD and LOQ were calculated by preparing a sample of honey at the lowest concentration and calculating the S/N 3 and 10 times. The method specificity was evaluated by comparing multiple blank matrix samples with blank matrix spiked samples (blank honey matrix supplemented with gelsemine, koumine and humantenmine).

The standard stock solution stability was measured by the standard deviation (RSD) of four different time nodes, which were measuring the stock solution at −20 ◦C for 0 d, 7 d, 30 d and 60 d (*n* = 6). The working solution was diluted from the standard reserve solution to near the lower limit and upper limit of quantification, respectively.

The sample stability was measured by the standard deviation (RSD) of four honey samples at different concentration levels, which were freeze–thawed three times at room temperature (25 ◦C) for 24 h and frozen for 30 d at −20 ◦C (*n* = 6).
