ABSTRACT
Summary Principal component analysis and other multivariate analyses have not been used extensively to analyze forage data. Our objective was the synthesis and summary of 98 digestion trials on forages using a principal-component analysis. The ten variables studied in all trials were dry matter (DM), crude protein (CP), digestible protein (DP), crude fiber (CF), ether extract (EE), nitrogen-free extract (NFE), ash (A), gross energy (GE), total digestible nutrients (TDN), and digestible energy (DE). Acid detergent fiber (ADF), acid detergent lignin (ADL), and cellulose (C) were summarized, in addition, for some forages. When data on thirteen variables in 24 hays were summarized, the first two principal components accounted for 62.9% of the within-group variance. DE, TDN, CP, DP, EE, and A were positively correlated and ADF, CF, ADL, and C were negatively correlated with the first principal component. CP, DP, and ADL were positively correlated and NFE was negatively correlated with the second principal component. When data were plotted according to principal components 1 and 2, separation of grass hays from legume hays and hays made from cool-season and warm-season plants was observed. Two principal components accounted for 58%-77% of the within-group variation in five other groups of hays and silages. Principal-component analysis permitted optimum summarization and clustering of desired information. It also permitted graphical presentation of variates according to the 2 principal components for easy diagnosis and evaluation. One may also identify latent clustering factors in the population.
