Grimace scales have been developed and used for pain assessment in rats, rabbits and horses. Using a grimace scale validated for sheep (Guesgen 2014), 64, Merino lambs were assessed for pain following surgical and tail-docking (marking). Demeanour of the nose, mouth, cheek, eyes and ear posture were scored on a 3 point scale, and values combined to calculate an overall 0 (no pain) to 2 (obvious) result. Lambs were randomly allocated to 4 treatment groups: Sham (S); marking, no pain relief (C); marking, flunixin in feed (4.0mg/kg) at 24 h and 1 h prior to marking (F); and marking, flunixin injected (2.0mg/kg) 1 h prior to marking (I). Lambs were recorded, in their pens (free) and when restrained (held) at 0, 0.5, 6, 24 and 48 h relative to marking. For analysis the scores given for each facial unit were averaged to give an overall score. Ears were not included as they could not be scored with confidence. Data at time 0 h were analysed by Chi-square test and data post-marking were analysed as proportions using a binomial test. Prior to marking, there was a significant association between free vs. held and incorrect pain assessment χ2 (1) = 12.70, P<0.001, with 62% of held lambs and 18% of free of free lambs incorrectly identified. Post-marking, a free vs. held effect (P=0.03) and treatment effect (P<0.001) were observed. Held C lambs were classified as in pain more often than held S and I lambs (P<0.05). Free C lambs were scored as in pain more often than free S, F and I lambs (P<0.05 for all). S lambs were scored as in pain less often than I and F lambs (P<0.05) A lack of different ear postures was found, perhaps due to lambs orienting to the observer. Hollowing of cheeks were infrequently observed, possibly due to lambs having wrinkled, woolly faces. The grimace scale shows potential as a rapid and practical method to assess pain in lambs, but lambs should be observed without interference, as restraint may change aspects of facial expression.