Search

Summary Propolis has various biological activities closely related to the composition which varies according to environmental factors and also to the method of production. This study was aimed at determining whether or not HR-NMR and multivariate statistical analysis were able to classify propolis according to the method used to harvest it. Sixty propolis samples were analysed in all. The ethanolic propolis extracts were initially analysed for quantification of the main bioactive substances, balsams and waxes. The 1H NMR and heteronuclear multiple bond correlation spectra were then acquired. Spectral data were analysed by the application of multivariate statistical techniques (Factor Analysis and General Discriminant Analysis). The best results were obtained using the 1H NMR which furnishes a sufficiently effective model by analysing the spectral region between 4.50 and 13.00 ppm (predictive capacity: 96.7%).