Relationships between bacon grade schedules and carcass characteristics

A computer model was used to compare alternative bacon grade schedules in terms of the carcass composition, joint proportions and simple meat quality characteristics (muscle pH and light reflectance) of pigs in different grades. The data used were for 350 side-dissected carcasses from years 6 to 8 of the Meat and Livestock Commission's Commercial Pig Evaluation. Analyses were carried out separately for restricted-fed and ad libitum-fed pigs. Overall means and standard deviations for carcass weight and lean percentage were 64·4 ± 3·19 and 51·5 ± 3·98 (restricted) and 65·1 ± 3·57 and 49·9 ± 4·09 (ad libitum). Alternative schedules were compared against a base schedule, typical of commercial practice in Great Britain and involving the following requirements for the principal grade: a maximum P(2) measurement (taken laterally over the M. longissimus) of 15 mm, maximum mid-line fat measurements (shoulder/loin) of 42 22 mm and a minimum side length of 770 mm. Removal of the shoulder/loin and length requirements increased the proportion of pigs in the principal grade by seven percentage units with little effect on the mean carcass characteristics of pigs in the grade. An extreme tightening of the maximum shoulder/loin measurements to 35 15 mm removed half the pigs from the principal grade and increased the mean carcass lean content and the lean content of individual joints for pigs in the grade by one to two percentage units. Introduction of a minimum muscle thickness measurement of 55 mm had a similar effect. In both cases the effect was partially mediated through a general reduction in fatness and could be achieved by tightening the maximum P(2) fat thickness. None of the changes in grade schedules examined materially altered the mean meat quality characteristics of pigs in the principal grade. The analysis illustrates the complexities involved in using grade thresholds based on a number of correlated measurements. The use of a multiple regression approach, in which carcass lean content and commercial value are estimated from measurements known to contribute significantly to the regression, would simplify the situation considerably.