Reference-standard and calibration-model dependence of AlCl3-reactive flavonoid estimates in yemeni Malva verticillata L. leaf extract

Abstract

Aluminum chloride colorimetry is widely used for preliminary flavonoid estimation, but results depend on reference standard, calibration model, and unresolved sample optical background. We re-analysed archived data from a completed undergraduate thesis on a methanolic leaf extract of authenticated Yemeni Malva verticillata L. In one within-run calibration, rutin gave A = 11.7113 C + 0.0012 (R 2 = 0.999992) and showed statistically detectable lack of fit (F(2,8) = 12.756, p = 0.00325), although the maximum mean residual was only 0.001709 absorbance unit. At 0.1000 mg/mL extract, the exploratory apparent AlCl 3 -reactive standard-equivalent estimate under reagent-blank conditions was 152.8 mg RE/g dry extract (95% conditional calibration-only inverse-prediction interval, 151.0–154.6). Quercetin yielded 145.6 mg QE/g with the complete-range model (54.2–229.6) and 174.1 mg QE/g with a post hoc lower-range sensitivity model (166.3–181.8); both are 95% conditional calibration-only inverse-prediction intervals. Only the 0.1000 mg/mL extract response was within both standard-response ranges. No sample-matched blank was available. These intervals quantify within-run calibration uncertainty only and exclude matrix, recovery, preparation, volumetric, between-run, and post hoc model-selection uncertainty. The values are not validated total-flavonoid or compound-specific concentrations.