Grouping patterns of 23 individuals.

The Clustering eigenvector method identified 2 principal clusters: social units-1 and -2. Dolphin are identified by their ID code (PHDn, n = 1–23), age class (A =  adult, J =  juvenile and C =  calf), and sex (F =  female, F* =  females with calves, M =  male, UN =  unknown sex) Each dolphin is shown with (1) mean HWIs calculated within (M.in) and (2) between (M.diff.) social units, (3) maximum HWI and (4) minimum HWI, (5) preferred and (6) avoided dolphins (min. HWIs), (7) residence pattern (VF: Very frequent, F: Frequent and R: Rare), (8) occurrence (%) in groups with size lower than the mean group size (GS <5.22), (9) occurrence (%) with trammel nets (TN). Male-male and female-male preferred associations (p. a.) are also shown.