A PROBE TO MEASURE GR IN LAMB CARCASSES AT CHAIN SPEED

The GR (total tissue thickness at the twelfth rib, 110 mm from the mid line) was measured with the AUS-MEAT sheep probe (ASP) on lamb carcass es at chain speed by an abattoir operator in two experiments, both at the same abattoir. In both experiments, GR was also measured on the ca rcass by an independent operator using a GR knife. For experiment 1, a total of 779 lamb carcasses were measured over four occasions and for experiment 2 data on 607 lamb carcasses were obtained on two occasion s, twelve months after experiment 1. In both experiments, one operator measured the manual GR and a different operator used the ASP. A new v ersion of the ASP was used in the second experiment. In only 30% of ca ses in experiment 1 did the operator probe at the correct rib, and in the majority of cases the thirteenth rib was used as the probing site. In only a small percentage of cases (2%) was the operator more than o ne rib away from the twelfth rib. In 67% of the cases for experiment 1 , the operator probed at the correct site with respect to distance fro m the midline. Models were developed to describe the relationship betw een the manual and ASP GR measurements. In experiment 2, the amount of variation in manual GR explained by ASP measurements was greater than that in experiment 1 (R2 = 0.80 compared with 0.72), and the accuracy of the estimates was significantly increased (+/- 1.54 mm compared wi th 2.31 mm). The better over-all performance of the ASP in the second experiment was indicated by the fact that for 90% of the sample the AS P measurements were within +/-2 mm of the manual GR measurements where as, for experiment 1, the level was less than 70%. Measurement time (d ay), which could be described as an operator 'effect', was identified as an important factor influencing the accuracy of GR estimates but lo cation of the probed site with respect to the GR site was not found im portant. The significance of operator training and monitoring is discu ssed on the basis of the results, as are the implications of the findi ngs for objective purchasing systems.