. 1997. Comparison of objective external carcass measurements and subjective conformation scores for prediction of lamb carcass quality. Can. J. Anim. Sci. 77: 217-223. Carcass measurements were collected from 1505 lambs over a 2-yr period by a single technician. The lambs (rams, ewes and wethers) were classified as either meat-type or wool-type, with meat-type subdivided by frame size into light, medium or large. Carcass measurements included carcass length (L), length of hind leg (T), depth of hind leg (H) and circumference of hind leg (G). Agriculture and Agri-Food Canada graders measured tissue depth of all lambs at the GR site (11 cm from carcass midline over the 12th rib) and assigned subjective conformation scores for the regions of the leg, loin and shoulder. Sixty-nine lambs across groups were dissected into primal cuts to determine saleable meat yield. Carcass length and T decreased (P < 0.05) with increasing conformation score, whereas G and H increased (P < 0.05) with increasing conformation score. Wool-breed and light-frame lambs had lower conformation scores (P < 0.01) for the regions of the leg and loin than did medium-and large-frame lambs. Comparing the results of carcass dissection, wool-breed lambs had a lower saleable meat yield (P < 0.05) than large-frame lambs. Saleable meat yield also tended to be lower for light-frame than for medium-and large-frame lambs, although the numbers of medium-and light-frame lambs dissected were low. The carcass measurements used in this study accurately predicted percentage of the high priced leg cuts (R 2 = 0.80, RSD = 0.6%), and were marginally more accurate than subjective conformation scores in predicting saleable yield from dissection (R 2 =0.61, RSD = 1.3%). However the carcass measurements used in this study were not valuable predictors of percentages of shoulder or loin and rack in the primal cuts. These data do not indicate any unjustified bias in the subjective conformation scores currently in use in the Canadian Classification System for lamb, as the breed types and frame sizes assigned lower conformation scores also had lower saleable meat yields. Anim. Sci. 77: 217-223. Des mesures de carcasse ont été prises pendant deux ans par un même technicien sur un total de 1 505 agneaux. Les agneaux (mâles entiers, mâles castrés, agnelles) étaient classés en type à viande ou type à laine, le premier étant fractionné selon le format en petit, moyen ou grand. Les mensurations comprenaient : la longueur de la carcasse (L) et la longueur (T), la profondeur (H) et la circonférence (G) du gigot. Ce sont les services de classement du ministère de l'Agriculture et de l'Agroalimentaire du Canada qui se chargeaient de mesurer la profondeur des tissus aux sites de classement officiels, soit à 11 cm de la ligne dorsale au niveau de la 12 e côte, et d'attribuer des notations de conformation subjectives pour le gigot, la longe et l'épaule. Soixante-neuf agneaux, tous groupes confondus, étaient découpés en grosses pièces pour établir le rendement boucher. La longueur de l...