Indicators of Bioactivity and Floral Origin of New Zealand Honeys
Goss, C. H. A. (2009). Indicators of Bioactivity and Floral Origin of New Zealand Honeys (Thesis, Doctor of Philosophy (PhD)). The University of Waikato, Hamilton, New Zealand. Retrieved from http://hdl.handle.net/10289/5805
Permanent Research Commons link: http://hdl.handle.net/10289/5805
The hypothesis that NIR might be capable of discriminating one floral source from another was explored. No prior analysis of NIR for New Zealand honeys has been reported. A visual inspection of the NIR spectra of ten New Zealand honey types indicates that beech honeydew honey is significantly different from nectar honeys. Rata honey is the most unique nectar honey with very little variability seen in the NIR spectra compared to other honey types. Both beech honeydew and rata honey can be distinguished from other floral types using Linear Discriminant analysis (LDA) on selected wavelengths. A degree of clustering within other honey types is achieved, however none of these are fully resolved. A Partial Least Squares (PLS) model successfully classified all main New Zealand unifloral honeys with an average correct classification of 93%. 100% of all beech honeydew honeys were correctly classified with close to 100% achieved for rata, kamahi, manuka, rewarewa and clover honeys. Honeys with a clover contribution: tawari, thyme, nodding thistle and vipers‟ bugloss displayed reduced performance in this model with a proportion of samples misclassified as clover honey. These results indicate that the NIR spectra evaluated using a PLS model would be an effective industry classification method for the identification of New Zealand unifloral honeys with the exception of nodding thistle and vipers‟ bugloss honeys. A multi-technique classification model incorporating NIR classification results with conductivity, colour and sugar analysis has been proposed. A series of compounds in manuka honey were examined in respect to UMF activity. The carbohydrate profiles of 38 manuka honeys of varying UMF activity were determined using a combination of HPLC, GC-FID and GC-MS. A method was developed to determine the proportion of nigerose, turanose, maltose and maltulose in reduced and silylated honey using the ratio of m/z 307 to m/z 308 ion responses as determined by GC-MS-SIM. An examination of the glucose and fructose concentrations in manuka honey revealed a moderate correlation between the glucose/fructose ratio and UMF activity. Due to an improvement in chromatographic resolution, the peak assignment of three disaccharides (cellobiose, laminaribiose and gentibiose) differed from that of a previous investigation. Despite the retention time of palatinose being identical to the corresponding peak in honey, an examination of the mass spectra provided strong evidence to suggest that the corresponding honey disaccharide is α-1→2 linked as opposed to β-1→6 linked and that it was therefore unlikely that this peak arose from palatinose. The mono and disaccharide composition of manuka honey was evaluated with respect to the level of UMF activity. Linear Discriminant analysis successfully distinguished between high, moderate and low UMF activity honeys. Glucose was identified as the single most important compound in the discriminant model. The connection between glucose concentration and UMF activity was not unexpected as a significant proportion of UMF activity has been attributed to the presence of methyl glyoxal, a degradation product of glucose. The existences of indicator compounds in honeys from various floral origins were examined. The extractable organic substances of five New Zealand honeys: beech honeydew honey, kamahi, pohutukawa, rata and tawari were determined by GC-MS of methylated extracts. This survey confirmed the results of a previous investigation and established ranges for marker compounds. Due to difficulties in obtaining sufficient certified unifloral honeys, previous studies on these honey types were exploratory only and not published. Statistical analysis of the extractable organic substances showed that each honey contains a unique fingerprint of compounds. Agglomerative clustering successfully separated all honeys into the correct floral group with the exception of two samples. Well separated clusters were produced in the score plot of the first and second Linear Discriminants. 4-Hydroxyphenylacetic acid, salicylic acid, indole-3-acetic acid and an unknown compound (identified by characteristic ions in the mass spectra) were identified as being the most important discriminants, all of which were present in a single floral source.
The University of Waikato
All items in Research Commons are provided for private study and research purposes and are protected by copyright with all rights reserved unless otherwise indicated.
- Higher Degree Theses