In order to estimate the abundance of scallops it is necessary to estimate the efficiency of the experimental fishing gear (scallop dredge with rings of 50 mm diameter, 2 m wide and a toothed bar with 30 teeth). The age-size structure of catches does not usually correspond to that of the actual population because of selectivity of the gear. Several previous studies on the Saint-Brieuc scallop stock (English Channel, France) have used a ratio estimator per age group: an assumption of increase in efficiency vs. scallop size is taken into account. The evaluation of the gear efficiency used for this stock since 1986 has been carried out by diving according to a random sampling plan. Factors such as sedimentary heterogeneity, tide, current, and depth are neglected. There are drawbacks as regards to the definition of dredge efficiency by this method: (1) a ratio estimator is biased when the number of samples is small such as the number of diving units in this study; and (2) the definition of the estimator using age rather than scallop size limits the period of assessment validity: intra-annual extrapolations are not valid. These two difficulties are resolved when dredge efficiency is defined against scallop size. A logistical curve is used and requires three parameters: (1) maximum asymptotical efficiency (e(max)); (2) parameter describing the selection deviation (a) and; (3) size corresponding to 50% of e(max) (L-50). The matrix of variances-covariances of three parameters is also obtained.