Reef manta beam (Yards. alfredi) generalised linear mixed models

alfredi and juvenile E. naucrates, (C) M. alfredi and Lutjanus bohar, (D) M. alfredi and Gnathanodon speciosus presence, and (E) M. birostris and Remora remora.

Teenager sharksucker remora (Elizabeth. naucrates).

The highest positions GLMM getting teenager E. naucrates consisted of all predictors (site setting, intercourse, readiness, and you can season). The fresh model (Fig 8) signifies that teenager Age. naucrates were probably to-be expose with Yards. alfredi within giving elements (Or = 2.08), that was 108% likely to be than just on tidy up station (source category). Teenager Age. naucrates was as well as very likely to show up towards teenager Meters. alfredi (Or = dos.37), which have been 137% expected to fully grasp this hitchhiker classification than simply grownups (reference category). Men Yards. alfredi (Otherwise = step one.59) was basically 59% likely to be than just ladies to own juvenile E. naucrates introduce, and teenager Elizabeth. naucrates was basically least more likely present which have M. alfredi inside the SW Monsoon (Or = 0.63); 37% more unlikely than simply inside NE Monsoon (reference category).

Reddish snapper (L. bohar).

The highest ranking GLMM for L. bohar (Fig 8) contained one significant predictor (site function) and suggests that the species were most likely to be present with M. alfredi at cleaning stations (reference category), which was 100% more likely than at feeding areas (OR = 0). There was one model with ?AICc <2 (S1 Table), which was the same as the highest-ranking model with the addition of season, but this predictor was non-significant (p>0.05).

Fantastic trevally (Grams. speciosus).

The highest ranking GLMM for G. speciosus contained two significant predictors (site function and maturity status). Gnathanodon speciosus were most likely to be present with M. alfredi at cleaning stations (reference category), which was 22% more likely than at feeding areas (OR = 0.78) (Fig 8). There were two models with ?AICc <2 (S1 Table). One of these models contained the same predictors as the highest-ranking model with the addition of season, but this predictor was non-significant (p>0.05). The other model contained only the predictor maturity status, which suggested G. speciosus were more likely to be present on juvenile M. alfredi (OR = 1.5), which were 55% more likely to have this hitchhiker species than adults (reference category).