Journal: Food Hydrocolloids, vol. 15, p. 521–532, 2001
International Standard Numbers:
Open Access: none
The causes of variation in the particle size distributions of a hot-processed fall-fat dressing (system I) and a cold-processed full-fat mayonnaise (system II) have been studied using a newly developed modification of classical multivariate analysis of variance (MANOVA). The MANOVA analysis uses scores obtained from principal component analysis as input parameters. The modification also involves rules for selecting the number of components to be used, as well as a routine for testing each effect inherent in the experimental designs used. The results of the MANOVA analysis are compared with those obtained using ordinary ANOVA on the average volume, the volume mode diameter and the number mode diameter. The MANOVA gives a more consistent picture of the causes of variation in the particle size distribution than was obtained by choosing a single response such as average volume. In addition, more of the variance is frequently explained by MANOVA. For both systems, emulsifier concentration was identified as the most important variable affecting the volume frequency distributions. This most important effect, as judged from the MANOVA, was not always identified if a single response was used. The hot-processed system had large non-dispersible particles (aggregates) which seemed largely influenced by pH and protein levels. The number mode diameter was influenced by salt level; low salt concentrations gave more small particles than did high salt concentrations. The results were supplemented with microscopic investigations. Only a very small effect of heating procedure was noticed. For the cold processed system, the oil type also affected the particle size distribution. (C) 2001 Elsevier Science Ltd. All rights reserved.