ESTIMATION OF HAY ENERGY VALUE WITH A REGRESSION EQUATION
Jasna M. A. STEKAR, F. Zagožen, Andrej Golob · Acta agriculturae Slovenica Suplement · 1996
825 random samples of hay, 655 the first cut and 170 aftergrass, were analysed according to Weende analysis. The results were accounted with the statistical package SPSS so that multiple regessions were calculated for the starch equivalent (SE) and net energy for lactation (NEL) for all samples and separately for the first cut and aftergrass respect. The equations were carried out with the stepwise inculding the separate nutrients. All multiple regression coefficients (R) were very higly significant. For the first cut samples, the estimation of SE on the base of all nutrients was very accurate (R=0.9796, s.e.=1.4502). For the practice, it is sufficient to estimate it with the crude fibre (CF) and crude protein (CP) contents only (R=0.9613, s.e.=1.9861). In the case of estimation of NEL it is necessary to take into account all nutrients (R=0.9429, s.e.=0.1822). It is possible to estimate SE well with all nutrients (R=0.95488, s.e.=0.14164) in afftergrass. For NEL CP, ether extract and nitrogen-free extractives (R=0.93487, s.e.=0.14164) are sufficient. When the samples are not defined, SE and NEL should be estimated with all nutrients (R=0.97186, s.e.=1.69134 resp. R=0.95646, s.e.=0.15974).