Spatial variation in meristic and morphometric characteristics of sardine Sardinops sagax around the coast of southern Africa

• Published in: African journal of marine science Vol. 41; no. 1; pp. 51 - 60
• Main Authors: Groenewald, G, Moloney, CL, van der Lingen, CD
• Format: Journal Article
• Language: English
• Published: Grahamstown Taylor & Francis 02.01.2019; Taylor & Francis Ltd
• Subjects: Allometry; Body shape; canonical variate analysis; Coastal environments; Condition factor; Fish; Fish populations; management; Marine fishes; Meristic counts; Morphometry; Phenotypic variations; Population structure; Regions; Regression analysis; Sardinops sagax; Shape; Spatial variations; stock discrimination; Stocks; Stomach; Variability; Vertebrae; vertebral count; South Africa; Namibia; Africa
• ISSN: 1814-232X; 1814-2338
• DOI: 10.2989/1814232X.2019.1569161

Spatial variability in phenotypic characteristics within a fish population may be used to infer the existence of multiple stocks, and knowledge of the population structure of exploited species is important for their sustainable management. In this study we investigated geographic variability in meristic (vertebral count) and morphometric (body shape) characteristics of sardine Sardinops sagax from three southern African regions: Namibia, the South African west coast, and the South African south coast. There were significant regional differences in vertebral counts, with Namibian fish tending to have more vertebrae than South African fish, but with no difference between sardine off the South African west and south coasts. Body shape was characterised using morphometric landmarks, and shape changes were investigated using geometric morphometrics, with influences on fish shape assessed using multivariate regressions. To remove the effects of allometry, the 22 shape variables (represented by Procrustes coordinates) were regressed against fish size (represented by centroid size). There was a significant allometric effect, which was removed by using the residuals of these regressions in further analyses. A multivariate multiple regression was applied to the 22 size-corrected shape variables and to three covariates: condition factor, stomach fullness and fat stage. The significant influence of these covariates was removed by using the residuals from this regression in a canonical variate analysis, where corrected shape variables were classified in relation to region, year and sex. The Mahalanobis distances from the canonical variate analysis differed significantly for most combinations of classifier variables, with clear separations among the three regions. Sardinops sagax off Namibia had thicker bodies and more contracted tail regions than South African fish, whereas S. sagax off the South African south coast had thinner bodies and smaller heads than those off the west coast. These results support the hypothesis of multiple sardine stocks around southern Africa.