Abstract.-Analyses of the population dynamics of blue crab Callinectes sapidus have been complicated by a lack of estimates of the instantaneous natural mortality rate (M). We developed the first direct estimates of M for this species by solving Baranov's catch equation for M given estimates of annual survival rate and exploitation rate. Annual survival rates were estimated from a tagging study on adult female blue crabs in Chesapeake Bay, and female-specific exploitation rates for the same stock were estimated by comparing commercial catches with abundances estimated from a dredge survey. We also used eight published methods Our results indicate that natural mortality of blue crab is higher than previously believed, and we consider M values between 0.7 and 1.1 per year to be reasonable for the exploitable stock in Chesapeake Bay. Remaining uncertainty about M makes it necessary to evaluate a range of estimates in assessment models.The estimation of natural mortality rates is one of the most difficult and most critical elements of many fishery stock assessments. The natural mortality rate is a key determinant of the potential productivity of a stock and thus the amount of exploitation a stock can sustain. In general, assuming that natural mortality and harvest mortality are additive, stocks with higher natural mortality rates are more productive and are able to sustain higher rates of exploitation. Lacking evidence to the contrary, most stock assessments assume that natural mortality is constant through time as well as across the sizes or ages of the exploited animals. Thus, a single estimate of the instantaneous natural mortality rate (M) is presumed to apply to the entire exploitable stock.The values used for M in assessment models can have substantial effects on model results, biological conclusions, and management recommendations. For a simple age-structured model, Clark (1999) found that stock abundance and target harvest rates could be severely overestimated when M was overestimated by as little as 0.1 per year or less, especially when fishing mortality was low (F , 0.3 per year). Similarly, harvest policies for U.S. West Coast groundfish based on a catch-at-age model were sensitive to changes in M of less than 0.05 per year (Williams 2002). Using a length-structured model for red king crab Paralithodes camtschaticus in Bristol Bay, Alaska, Zheng et al. (1997a, 1997b found that stock rebuilding and longterm harvest strategies were highly sensitive to changes in M of 0.2-0.3 per year. These and other results indicate that it is desirable to have precise knowledge about M for assessment purposes.Unfortunately, estimates of M used in stock assessment models are often uncertain, partly because it is difficult and expensive to estimate the parameter. In practice, values of M for use in stock assessments are obtained by two types of methods, which we refer to as direct and indirect. Direct methods involve estimating M from data pertaining solely to the species or stock of interest. Direct methods include field ...