To predict in vitro digestibility from some plant characters of orchardgrass plants, multiple regression analyses were done by the selected variables method in the following three ways. A) Estimation of hay digestibility from hay characters. B) Estimation of aftermath digestibility from hay characters. C) Estimation of aftermath digestibility from aftermath characters. Data derived from 758 plants of 8 varieties (n=91-97) were used to this calculation. Seven characters in hay and 8 characters in aftermath were investigated besides in vitro digestibility (Table 1). As the results of multiple regression analyses in each variety, significant multiple correlation coefficients were obtained in 6, 6, and 8 varieties in A, B, and C, respectively. Using 758 plants pooled 8 varieties, numbers of selected variable were 4, 6, and 8 characters, and multiple correlation coefficients were R=0.303, 0.354, and 0.375 in A, B, and C, respectively (Table 2-5). Judging from the value of standard regression coefficients and frequency of selected variables in multiple regression equations (Table 5), it is concluded that growth habit, leaf width and date of ear emergence in hay, and growth habit, degree of diseases, leaf color and vigor in aftermath were important characters affecting in vitro digestibility. Though significant coefficients were obtained through A to C, the values of the coefficients were comparatively low. More detailed investigation for important characters is required for getting effective multiple regression equations to predict in vitro digestibility of orchardgrass plants.
View full abstract