Social Life in the Deep: Variation in Network Structure and Association Patterns of Rough-Toothed Dolphins (Steno bredanensis) Around the Main Hawaiian Islands
Article
Corsi, E, Baird, RW, Mahaffy, SD et al. (2026). Social Life in the Deep: Variation in Network Structure and Association Patterns of Rough-Toothed Dolphins (Steno bredanensis) Around the Main Hawaiian Islands
. MARINE MAMMAL SCIENCE, 42(4), 10.1111/mms.70281
Corsi, E, Baird, RW, Mahaffy, SD et al. (2026). Social Life in the Deep: Variation in Network Structure and Association Patterns of Rough-Toothed Dolphins (Steno bredanensis) Around the Main Hawaiian Islands
. MARINE MAMMAL SCIENCE, 42(4), 10.1111/mms.70281
While coastal delphinid sociality has been extensively investigated across the globe, little is known about the social structure of odontocetes in the open ocean. The rough-toothed dolphin (Steno bredanensis) is a globally distributed oceanic dolphin occurring in tropical and warm-temperate waters. Around the main Hawaiian Islands, two demographically distinct populations have been identified: one off Kauaʻi and Niʻihau, the other off Hawaiʻi Island. We investigated differences in network structure and association patterns between these populations using photo-identification data collected from 2002–2020. Social network analysis revealed that the Kauaʻi–Niʻihau population was more subdivided into distinct social clusters (i.e., had higher modularity: 0.58 vs. 0.47) and less densely interconnected (i.e., lower network density: 0.07 vs. 0.12) than the Hawaiʻi Island population. Mean association strength within clusters was higher off Hawaiʻi Island (0.31) than off Kauaʻi–Niʻihau (0.22). Social connections off Hawaiʻi Island also relied more strongly on a smaller number of individuals that linked clusters together. We suggest that differences in population density and habitat extent may contribute to differences in social network structure. Although low re-sighting rates limited our ability to characterize either population comprehensively, applying identical methods allowed us to identify relative differences in network structure between them.
