CCAMLR has agreed that there is a requirement for systematic observation from the krill fishery, however, current levels of observer coverage are not distributed evenly across years, regions or vessels. The number of observer days as a proportion of the number of days fished varied between c10% (Subareas 48.1 and 48.2) and c30% (in Subarea 48.3). 13 of the 16 vessels that have fished for krill in 48.3 have carried observers at some time, compared to 6 of 15 in 48.1 and 4 of 15 in 48.2. An analysis of existing observer data submitted to the Secretariat indicates that measuring 200 krill from 5 hauls should produce a sufficiently precise length frequency distribution and would free up time for other priority tasks of observers. Suggestions are provided for a revision of the Scientific Observer Manual for consideration by the Working Group.