Pig producers in the United Kingdom (UK) are paid on the basis of carcass weight and backfat depth at the P2 position (65mm from the edge of the dorsal mid-line, at the level of the last rib). In processing plants, this measurement is assessed using the optical, Ulster or Hennessy probes and in live pigs measurements can be taken using various ultrasonic devices. Since backfat is considered a heritable trait (heritability value between 0.4 and 0.6) (Whittemore, 1993), it is of interest to the pig producer and breeder to measure it on live pigs when selecting for lean meat in replacement gilts and breeding boars. There are several commercially available ultrasonic devices and it is essential assessments made using these devices have strong correlations with backfat measurements taken at the processing plant. The accuracy of these correlations has not however been fully investigated. Indeed, work by Pomar et al (2001) indicated that assessments of fat using an Ultrascan 50 ultrasound system are only moderately accurate. The aim of this study was therefore to assess the accuracy of ultrasonic devices for backfat prediction as compared to the measurements obtained at the processing plant.