The densities of total and pathogenic Vibrio parahaemolyticus in 671 samples of molluscan shellfish harvested in 1999 and 2000 from 14 sites in seven Gulf and Atlantic coast states were determined at 2-week intervals over a period of 12 to 16 months in each state. Changes in V. parahaemolyticus densities in shellfish between harvest and sample analysis were minimized with time and temperature controls. Densities were measured by direct plating techniques, and gene probes were used for identification. Total and pathogenic V. parahaemolyticus organisms were identified with probes for the thermolabile direct hemolysin (tlh) gene and the thermostable direct hemolysin (tdh) gene, respectively. An enrichment procedure involving 25 g of shellfish was also used for the recovery of pathogenic V. parahaemolyticus. The densities of V. parahaemolyticus in shellfish from all harvest sites were positively correlated with water temperature. Shellfish from the Gulf Coast typically had higher densities of V. parahaemolyticus than did shellfish harvested from the North Atlantic or mid-Atlantic coast. Vibrio parahaemolyticus counts exceeded 1,000 CFU/g for only 5% of all samples. Pathogenic (tdh+) V. parahaemolyticus was detected in approximately 6% of all samples by both procedures, and 61.5% of populations in the positive samples from the direct plating procedure were at the lower limit of detection (10 CFU/g). The frequency of detection of pathogenic V. parahaemolyticus was significantly related to water temperature and to the density of total V. parahaemolyticus. The failure to detect pathogenic V. parahaemolyticus in shellfish more frequently was attributed to the low numbers and uneven distribution of the organism.