The use of relatively small hydraulic tow dredge for sampling hard clams (Mercenaria mercenaria) is discussed in conjunction with a sequential sampling plan. The sequential sampling plan, in which the number of replications was not predetermined, was employed to categorize hard clam densities into high, medium and low classes. Its application resulted in a 70% saving in sampling effort compared with a comparable fixed sampling plant when the objective was the classification of density rather than the estimation of a density parameter.