Data collected in predominantly nocturnal stratocumulus, as observed during the DYOCMS-II experiment are used to estimate cloud-top entrainment velocities. Estimates are based on two distinct methods, one which evaluates entrainment as the difference between the large-scale subsidence velocity and the local growth of the boundary layer, and another which evaluates entrainment rates from tracer budgets across the top of the cloud layer. The latter, tracer, method is based on four independent tracers: ozone, temperature, water vapor and dimethyl sulfide (DMS). The use of DMS in this respect was a novel aspect of DYCOMS-II. Because DMS only has a source at the surface, and has a lifetime of 2-5 days, it is particularly well suited to such measurements. Entrainment velocity estimates from these data are compared to predictions from a suite of recently proposed entrainment parameterizations. Preliminary results, including some from cloud regimes which are unstable according to the Randall-Deardorff entrainment instability criterion, suggest that most parameterizations greatly over-estimate the amount of mixing at cloud top.