Determination of the deviation between judges on the evaluation of beef marbling standard number

The purposes of this research were to investigate the macroscopic feature of carcass cross section with a large individual variation for the BMS number independently assigned by 16 judges, and to predict BMS number using image analysis traits. Clear images of carcass cross section between the 6-7 th rib of 100 Japanese Black steers were used to assign the BMS number. The average, mode and median of BMS numbers from 16 judges, and the average of judgment values excluding the higher and lower 3, 2 BMS values from 16 judges, and the highest and lowest BMS values were obtained and their correlation coefficients were calculated. Ratio of marbling area to rib eye area (FATPER), overall coarseness of marbling (O_COARSE), coarseness of the largest marbling particle in rib eye (M_COARSE), coarseness of single marbling particle (S_COARSE), ratio of minor and major axis of rib eye (MM_RATIO), and complexity of rib eye shape (COMP) were calculated by image analysis. The factor of the variation of the BMS judgment was investigated. As extremely high correlation coefficients (0.938-1.000) were found among all variables related to the judgment value of the BMS number, average BMS was adopted as a variable representing the judgment values. M_ COARSE and MM_RATIO significantly influenced the dispersion of the BMS in the analysis of factor on the variation in BMS judgment value (P<0.05). Multiple regression equation with the average BMS assigned by 16 judges as the dependent variable were obtained by the AIC method starting with 10 independent covariates for image analysis traits. Rib eye area, FATPER, O_COARSE and M_COARSE were selected in the multiple regression equation, and the coefficient of determination of the prediction equation was very high (0.895). Dispersion in the judgment of the BMS number might be due to the rough marbling particle in the rib eye or flattened rib eye of the carcass.