TY - RPRT T1 - Otolith shape as a valuable tool to evaluate the stock structure of swordfish (Xiphias gladius) in the Indian Ocean A1 - Mahe,Kelig A1 - Evano,Hugues A1 - Mille,Tiphaine A1 - Bourjea,Jerome AD - Ifremer, laboratoire Ressources Halieutiques, 150 quai Gambetta, 62321 Boulogne sur mer, France AD - Ifremer, Délégation Océan Indien, Le Port, France UR - https://archimer.ifremer.fr/doc/00269/37980/ KW - Xiphias gladius KW - otolith shape KW - Fourier descriptors KW - stock discrimination KW - Indian Ocean N2 - Swordfish (Xiphias gladius) is an oceanic-pelagic species currently fully exploited by several fisheries in the Indian Ocean, with suspicion of overexploitation in the southwest, but without a clear understanding of the real stock structure within this Ocean. Population structure of the Indian stock was studied in the western Indian Ocean using 395 individual samples collected from 2009 to 2014. Sagittal otoliths of the fish have been removed and shape analysis performed on these calcified pieces. Otolith morphometrics data and normalized Elliptical Fourier Descriptors (EFDs) were then extracted automatically by the dedicated imageanalysis system TNPC. Preliminary, side effect was tested by Redundancy analysis (RDA) combined to permutation tests on 91 individual samples and showed no significant differences in the outline shape between the right and left otoliths. Consequently, 395 sagittal otoliths were used to identify stocks among several geographical areas (La Reunion, Mozambique channel, Rodrigues, South Africa, South Malagasy, Sri Lanka and Thailand) within the Indian Ocean. To investigate variations of otolith shape according to 4 explanatory variables, Principal Components Analysis (PCA) was applied to EFDs and a RDA with permutation tests was used on the first 7 PC selected by the broken-stick model. These tests demonstrated no significant effects, neither by sex (p=0.121), sampled year (p=0.725), or total length (p=0.464). Only, geographical area appeared to be significant (p<0.05). Regarding the relationship between the ratio otolith length/otolith width and the total length of fish, size effect was neither significant. Furthermore, Linear Discriminant Analysis (LDA) was performed and overall jackknifed classification success reached 30%. Finally, a clustering analysis has been realised using Ward's hierarchical algorithm, which discriminated 3 different groups; however each group was composed by some individual samples from all geographical areas. In conclusion, all these results did not show a clear geographicalseparation, which corroborate the recent genetic analysis at the Indian Ocean scale while identifying only a single swordfish stock component in this area. Y1 - 2014/09/29 PB - Ifremer ID - 37980 ER -