Multivariate statistics as a quick tool to screen perennial forages for forage value.

Perennial forages seeded as straight grasses and legumes or in mixtures make a valuable contribution as feed for the beef cattle industry in Alberta. With new varieties constantly being released, there is need to quickly screen these for their forage value. Principal component analysis (PCA) is a multivariate analysis which can simplify a large dataset by identifying and representing patterns in fewer dimensions or factors. In total, 23 and 20 perennial straight grass and legume varieties respectively, and 15 commercial mixes were seeded as a complete randomized block design on small plots with 4 replicates on May 31, 2021 at Debolt, Alberta. Plots were harvested on July 12 and September 15, 2022 as first and second cut, respectively. The harvested biomass was weighed, and a sub-sample was air-dried to determine forage dry matter (DM). Dried forage samples were submitted for various forage quality analyses. A PCA was conducted on the correlation matrix of the data set of forage DM yield and quality. The KaiserMeyer-Olkin (KMO) measure and Bartlett’s test of sphericity verified the sampling adequacy (KMO = 0.71 and P > 0.001 respectively) for the analysis. The first three principal components (PC) had eigenvalues over Kaiser’s criterion of 1 and in combination explained 81 % of the variance (Table 1). Most of the legumes were greater in crude protein, yielded greater in forage DM, macro minerals (Ca, Mg and P) and micro minerals (Zn, Fe and Cu) compared with grasses and mixtures (Figure 1). Among the legumes, the following alfalfa varieties exhibited high forage value (i.e. combined both forage yield and quality): Peace Alfalfa, PV Ultima Alfalfa, Alfalfa Dalton, Algonquin, AC Yellow Head, Trueman Alfalfa. Noteworthy are the 3 non-alfalfa species (birdsfoot trefoil, sainfoin and cicer milkvetch) which loaded on the opposite quadrant to most of the alfalfa varieties. The grass varieties were more dispersed on the multivariate space of the first two PC thus exhibiting the inherent variability in forage value of the grass collection. A few grass species nonetheless distinguished themselves in terms of greater forage yield and quality and include AC Saltlander, Richard Timothy, Rocky Mountain Fescue and Hybrid Bromegrass. Most of the grasses were high in fiber and loaded primarily opposite to the positive forage value attributes. The mixtures loaded similarly to the grasses with a couple loading close to the legumes but most close to the grasses, revealing their high fiber content. The mixtures which loaded close to the legumes were legume-dominated mixtures such as Alfalfa/Sainfoin Mix, All LegumeMix, HayMix and LegumeMaster Mix which had between 90 to 100% legumes. PCA proved to be a useful tool in distinguishing treatments with high forage value as these corresponded with widely used varieties in Alberta. Tables and Figures 2.pdf [ABSTRACT FROM AUTHOR]