Characterization of honeys by melissopalynology and statistical analysis

We analyzed pollen from 89 honey samples, collected in Leon and Palencia provinces (NW Spain). According to their pollen spectra, 46 were considered monofloral. The most abundant monofloral honeys were Erica types followed by Castanea, Centaurea, Reseda and Helianthus. One hundred and forty-two different pollen types were recorded, belonging to 47 families. Fifty-five of them reached percentages over 3% in at least one sample, while the other 87 types never exceeded 3% in any of the 89 samples. The families that were present in the highest number of samples were Fabaceae, Rosaceae, Cistaceae and Asteraceae. Plant families that had the highest percentages were Fabaceae, Ericaceae, Asteraceae, and Rosaceae. The pollen types that appeared in most samples were Rubus ulmifolius (73 samples), Cytisus scoparius (70) and Mentha aquatica (62); the pollen types that had the highest relative abundance were Erica arborea, Lotus corniculatus, Cytisus scoparius. The pollen types of the Ericaceae family, Jasione montana, and Lavandula latifolia types could be used as indicators of the origin of honeys among five zones in the area studied. The use of cluster and correlation statistical analyses proved useful in characterizing honey samples from a geographical and botanical point of view. The honey samples were divided into four classes according to the data matrix of presence/absence, and into seven classes according to absolute frequencies of pollen types in the samples.