With the aim of finding methods that could constitute a solid alternative to melissopalynological and physicochemical analyses to determine the botanical origin (floral or honeydew) of honeys, the free amino acid content of 46 honey samples has been determined. The honeys were collected in a small geographic area of approximately 2000 km(2) in central Spain. Twenty-seven honey samples were classified as floral and 19 as honeydew according to their palynological and physicochemical analyses. The resulting data have been subjected to different multivariant analysis techniques. One hundred percent of honey samples have been correctly classified into either the floral or the honeydew groups, according to their content in glutamic acid and tryptophan. It is concluded that free amino acids are good indicators of the botanical origin of honeys, saving time compared with more tedious analyses.

The amino acid composition of 53 honey samples from Spain, consisting of 39 floral, 5 honeydew, and 9 blend honeys, has been determined. Physicochemical characteristics, polyphenolic content, amino acid composition, and estimation of the radical scavenging capacity against the stable free radical DPPH of the honey samples were analyzed. The resulting data have been statistically evaluated. The results showed that pH, acidity, net absorbance, electrical conductivity, and total polyphenolic contents of the honeys showed a strong correlation with the radical scavenging capacity. The correlation between the radical scavenging capacity of honey and amino acid contents was high with 18 of the 20 amino acids detected, with correlation values higher than those obtained for polyphenolic content. These results suggest that the amino acid composition of honey is an indicator of the sample's scavenging capacity.

No abstract

Fourier transform infrared spectroscopy (FT-IR) was used to determine 20 different measurands in honey.The reference values for 144 honey samples of different botanical origin were determined by classical physical and chemical methods. Partial least squares regression was used to develop the calibration models for the measurands studied. They were validated using independent samples and proved satisfying accuracies for the determination of water (R 2 =0.99), glucose (0.94), fructose (0.84), sucrose (0.91), melezitose (0.98) and monosaccharide content (0.82) as well as fructose/glucose ratio (0.98), glucose/water ratio (0.94), electrical conductivity (0.98), pH-value (0.87) and free acidity (0.96). The prediction accuracy for hydroxymethylfurfural, proline and the minor sugars maltose, turanose, erlose, trehalose, isomaltose and kojibiose was rather poor. The results demonstrate that mid-infrared spectrometry is a valuable, rapid and non-destructive tool for the quantitative analysis of the most important measurands in honey.