More than fifty published on demersal trawl codend selectivity studies have been carried out in the Mediterranean waters of Turkey since the mid-1980s. In this study, in an attempt to enhance the size selectivity of three species in the Aegean Sea trawl fisheries, three 90 o turned meshes codens were investigated. Experiments were carried out in the Kuşadası Bay of the Eastern Mediterranean, between 19 January and 29 March 2015 with commercial stern trawler. The meshes of tested codends were turned 90 o , two of them have 44 mm mesh size with 300 and 150 meshes around codend circumferences (44T300 and 44T150). The other has 40 mm mesh size with 165 meshes around codend circumferences (40T165). Individual hauls and mean selectivity parameters were estimated by using the CC2000 and EC-Modeller softwares, respectively. The mean L50 values of 44T300, 44T150 and 40T165 were 17.5 ± 0.2 mm, 18.6 ± 0.1 mm and 17.1 ± 0.2 mm carapace length for shrimp, 16.4 ± 0.4 cm, 17.1 ± 0.4 cm and 14.8 ± 0.2 cm total length for horse mackerel, respectively. The pooled L50 values of, 12.8 ± 0.5 cm, 13.2 ± 0.6 cm and 12.1 ± 0.1 cm total length for hake. While there are no differences between L50 values of 40T165 and 44T300 for deepwater rose shrimp (p>0.05), there are significant differences for horse mackerel (p<0.05). The likelihood ratio test results showed that there were no significant differences among all codends for European hake. In conclusion, any effort such as using square mesh, turned mesh codends or halved the number of meshes in the codend to release immatures of some species, is probably to cause the loss of marketable sizes of other species in the Mediterranean. This economical loss of revenues is the essential concern of fishermen that prevents them using alternative codend designs, though such designs are technically applicable. Consequently, the losses of income caused by using T90 mesh codend and halving the number of meshes around codend circumference should also be investigated in future studies.