Dye release experiments were performed together with microstructure profiling to compare the two methods of estimating diapycnal diffusivity during summer and fall stratification on the continental shelf south of New England. The experiments were done in 1996 and 1997 as part of the Coastal Mixing and Optics Experiment. During the 100 hours or so of the experiments the area of the dye patches grew from less than 1 km2 to more than 50 km2 <Sundermeyer and Ledwell, 2001>. Diapycnal diffusivities inferred from dye dispersion range from 10-6 to 10-5 m2/s at buoyancy frequencies from 9 to 28 cycles/hour. Diffusivities estimated from the dye and those estimated from dissipation rates in the companion paper by Oakey and Greenan <2004> agree closely in most cases. Estimates of diffusivities from towed conductivity microstructure measurements made during the cruises by Duda and Rehmann <2002> and Rehmann and Duda <2000> are fairly consistent with the dye diffusivities. The dye diffusivities would be predicted well by an empirical formula involving shear and stratification statistics developed by MacKinnon and Gregg <2003> from profiling microstructure measurements obtained at the same site in August 1996. All of the measurements support the general conclusion that the diffusivity, averaged over several days, is seldom greater than 10-5 m2/s in the stratified waters at the site, and usually not much greater than 10-6 m2/s. Severe storms, such as a hurricane that passed over the CMO site in 1996, can dramatically increase the mixing at the site, however.