3.2. Method performance
Detector response D.C. Woollard et al. / Food Chemistry 69 (2000) 201±208 GLC technique has been demonstrated to be less precise than HPLC following replicate analysis of a typical milk powder (mean: 5.35 mg/100 g; RSDR: 17.01%; n: 30).
A range of samples sourced from both New Zealand and the USA, and including the NIST SRM 1846 were
subjected to analysis with alternative LC columns. Comparative data are presented in Table 2, along with
MBA and GLC results. The above data demonstrate that each column was capable of reliable quantitation of pantothenate in milk powder and infant formula, conrming the suitability of the proposed HPLC method for application to the routine analysis of infant formulas. It was also apparent during these later studies and further comparisons with
the Luna column, that there was no signi®cant dierence between HPLC and GLC (tmeas: 0.62 vs tcrit: 2.06, df: 25, P: 0.05), with r of 0.933 and linear regressiony . 1:026x ÿ 0:004. None of the columns was capable of unambiguously