The 2005 assessment was routinely extended, taking account of a further year’s catch, CPUE and catch-at-age data. CPUE shows a continuation of the increase that commenced in 1998. However, sustainable yield estimates are generally less than those for the 2004 assessment, although estimates of current biomass levels relative to K increase. The Reference Case (RC) scenario suggests that a TAC of about 360 MT or less would be appropriate to prevent biomass decline in the future. Other scenarios suggest either higher or lower values than this. If the catch-at-age data are down-weighted, then this 360 MT level for the TAC is increased to 390 MT. On the other hand, the scenario which assumes the 1996+ recruitment to be equal to the average of the previous 10 years is more pessimistic and suggests an appropriate TAC level of only some 300 MT or less to prevent biomass decline. A model (Model 2) which allows for time-varying selectivity is presented and shows promising results. Model 2 is better able to reproduce the recent CPUE trend. Preliminary results for a model that fits to catch-at-length rather than catch-at-age data, using a selectivity-at-length rather than selectivity-at-age function, are presented, but these do not as yet reflect satisfactory fits to the data so that they should not be considered reliable in the context of stock status estimates. These different scenarios reflect very different interpretations of the recent increase in CPUE for the resource. If the catch-at-age data are down-weighted, the model fit essentially ignores them and suggests a recent increase in abundance. However, under either the effort saturation or the time-varying selectivity approaches, spawning biomass is estimated to have decreased further over recent years.