Following the recommendation of WG-FSA in 2014 the Secretariat has undertaken an analysis of by-catch in CCAMLR longline fisheries. The numerical target catch ratio (the number of target fish as a proportion of the total number of fish caught), using haul by haul data from the Ross Sea toothfish fishery from 2008–2014, was used as a single metric to examine the amount of by-catch reported by vessels. This analysis indicated that, as expected, differences in this index arose due to gear and location of fishing. However, there was also evidence of distinct differences in the mean (and distribution) in the index from two groups of Members flagged vessels, that together contribute 84% of the haul by haul data. These differences do not appear to be explained by differences in gear or location and are also apparent in the data reported by observers. However, the way in which observer sampling of by-catch is reported requires clarification. In order to further develop the analysis of by-catch it is essential to consider what additional factors need to be included in that analysis. This consideration should also review whether there have been any misinterpretations of the data, or the assumptions underlying the analysis, that may have led to errors in the results and subsequent interpretation.