Report on fish by-catch on exploratory fishing in Divisions 58.4.1 and 58.4.2

<p>In this paper, we report on fish by-catch during exploratory fishing activities undertaken in Divisions 58.4.1 and 58.4.2 during the 2012 to 2018 seasons. Fish by-catch comprised 14 species or groups of species. In 2018, by-catch represented 12% of the total catch (43 tonnes over 307 tonnes of <em>Dissostichus </em>spp.) in Divisions 58.4.1 and 58.4.2. 98% of the biomass was represented by 2 families: Macrouridae and Channichthyidae. The others most common by-catch species or families were <em>Muraenolepis </em>spp., <em>Antimora rostrata</em> and Artedidraconidae. <em>Raja </em>and <em>Bathyraja </em>were rarely caught. Species composition varied between research blocks except for <em>Macrourus </em>spp. which dominated by-catch composition everywhere. The ratio by-catch to target catch was higher in the eastern part of Division 58.4.1. The ratio <em>Macrourus </em>to <em>Dissostichus</em> increased in every research blocks during the study period, particularly in the eastern blocks where <em>Macrourus </em>CPUE also increased in the last 3 years. <em>Macrourus </em>CPUE was 23kg/1000 hooks in average with relatively little spatial variations except in the east were it could reach 180kg/1000 hooks locally. As found in others areas of the Convention, <em>Macrourus </em>CPUE was 2 times higher for autolines than Spanish lines and trotlines, and it peaked at depths between 900 and 1300m. CPUE distribution of the others by-catch species were much lower and highly heterogeneous in space. <em>Macrourus </em>catch was dominated by females in all research blocks and length frequency distribution did not reveal any temporal changes within research blocks. Length at 50% maturity was estimated for the two most common species of Macrourus (<em>Macrourus whiton</em>i and <em>Macrourus caml</em>) and for <em>Chionobathyscus dewitti</em>. 95% of the <em>Muraenolepis </em>sp. and <em>Antimora rostrata</em> were females in every research blocks with spatial variations in total length frequency distribution.</p>