Fine-scale estimates of diapycnal diffusivity ? are computed from CTD and XCTD data sampled in Drake Passage and in the eastern Pacific sector of the Southern Ocean and are compared against microstructure measurements from the same times and locations. The microstructure data show vertical diffusivities that are one-third to one-fifth as large over the smooth abyssal plain in the southeastern Pacific than they are in Drake Passage, where diffusivities are thought to be enhanced by the flow of the Antarctic Circumpolar Current over rough topography. Fine-scale methods based on vertical strain estimates are successful at capturing the spatial variability between the low-mixing regime in the southeastern Pacific and the high-mixing regime of Drake Passage. Thorpe scale estimates for the same data set fail to capture the differences between Drake Passage and eastern Pacific estimates. XCTD profiles have lower vertical resolution and higher noise levels after filtering than CTD profiles, resulting in XCTD ? estimates that are, on average, an order of magnitude higher than CTD estimates. Overall, microstructure diffusivity estimates are better matched by strain-based estimates than by estimates based on Thorpe scales, and CTD data appear to perform better than XCTD data. However, even the CTD-based strain diffusivity estimates can differ from microstructure diffusivities by nearly an order of magnitude, suggesting that density-based fine-structure methods of estimating mixing from CTD or XCTD data have real limitations in low-stratification regimes such as the Southern Ocean.