Mateo R, Bosch-Reig F
Department of Analytical Chemistry, University of Valencia, Dr. Moliner 50, E-46100 Burjasot, Valencia, Spain.
J Agric Food Chem. 1998 Feb 16;46(2):393-400. doi: 10.1021/jf970574w.
To ascertain the most discriminant variables for seven types of Spanish commercial unifloral honeys, stepwise discriminant analysis was performed. Fifteen parameters [pH; water content; electrical conductivity; x, y, and L, chromatic coordinates from the CIE-1931 (xyL) color space; fructose; glucose; sucrose; maltose; isomaltose; maltulose; kojibiose; and the fructose/glucose and glucose/water ratios] were considered. The studied honey types were rosemary, citrus, lavender, sunflower, eucalyptus, heather, and forest. The most discriminant variables, as selected by the multivariate program, were electrical conductivity, color (x, y, L), water content, fructose, and sucrose. All sunflower, eucalyptus, and honeydew honey samples and >90% of the samples from the remaining honey types were correctly classified by using the classification functions devised by the program. The overall proportion of accurately arranged samples was 95.7%. Results were validated by the "jackknifed" procedure and showed that electrical conductivity, color, water content, fructose, and sucrose are highly useful parameters to classify unifloral honeys, although microscopical analysis of honey sediment remains the fundamental tool.
为确定七种西班牙商业单花蜂蜜的最具判别力的变量，进行了逐步判别分析。考虑了十五个参数[pH值；水分含量；电导率；CIE - 1931（xyL）颜色空间中的x、y和L色度坐标；果糖；葡萄糖；蔗糖；麦芽糖；异麦芽糖；麦芽酮糖； kojibiose；以及果糖/葡萄糖和葡萄糖/水比率]。所研究的蜂蜜类型有迷迭香、柑橘、薰衣草、向日葵、桉树、石南和森林蜂蜜。多元程序选择的最具判别力的变量是电导率、颜色（x、y、L）、水分含量、果糖和蔗糖。使用该程序设计的分类函数，所有向日葵、桉树和甘露蜂蜜样本以及其余蜂蜜类型中超过90%的样本都被正确分类。准确排列样本的总体比例为95.7%。结果通过“留一法”程序得到验证，结果表明电导率、颜色、水分含量、果糖和蔗糖是用于单花蜂蜜分类的非常有用的参数，尽管蜂蜜沉淀物的显微镜分析仍然是基本工具。
