SELECTIVITY EXPERIMENTS IN THE NE MEDITERRANEAN - THE EFFECT OF TRAWLCODEND MESH SIZE ON SPECIES-DIVERSITY AND DISCARDS

In the present study we compare the overall performance of the 14mm (f rom knot-to-knot) diamond-shaped codend (14D), presently used in Helle nic waters, with those of 20mm square-shaped (20S) and 20mm diamond-sh aped (20D) trawl codends. The experiments took place at 12 stations in the western Aegean Sea and were based on the covered codend method (c over codend: 10 mm from knot-to-knot, 10D). The results of applying mu ltivariate techniques, to the different station/codend combinations su ggest that they form groups which differ from each other in terms of s pecies composition, species diversity, and catch numbers/weights per h our of fishing. In general, the 10D covers of 20D and 20S were separat ed from the 10D covers of 14D, The 20D and 20S codends were generally grouped together and separated from those of 14D and 10D. Finally, the 14D codend either form groups by themselves or group together with th e 10D cover codends. In addition, the results indicated that the 20D a nd 20S codends, when compared with 14D, allow significantly more indiv iduals and number of species to escape through the meshes and are char acterised by considerably higher commerical/non-commercial ratios in t erms of both weight and number. In absolute numbers and weights 14D ca ught a very large quantity of non-commercial species whereas the weigh t of the commercial species caught was among the lowest. The commercia l/non-commercial ratios of 20D were higher than those of 20S and, in a bsolute numbers and weights, 20S caught relatively more non-commercial species than 20D. These facts clearly indicate that 14D is harmful in terms of species diversity for the Hellenic demersal stocks and hence its use should be prohibited. (C) 1997 International Council for the Exploration of the Sea.