Habitat selection of intertidal caprellid amphipods in a changing scenario

•Habitat selection is main subject to be studied in a global change scenario.•Unambiguous habitat selection experiments are key to avoid confounding preference.•Closely related amphipods can show different selection patterns.•Habitat plasticity of some species may be favouring a taking over its congeners.•Generalizing within genus level is risky, a species-specific approach is mandatory. Habitat selection is a complex process, dependent on numerous fluctuating conditions and key to species coexistence. In a changing global scenario, it will greatly determine the fate of marine organisms and hence is an important subject to be explored. The present study evaluates host specificity of two caprellid amphipod species, Caprella grandimana and Caprella takeuchii, dwelling on a rocky intertidal where the calcifying macroalgae Jania rubens and Ellisolandia elongata show opposite seasonal fluctuation patterns throughout the year. To avoid confounding preference with other factors, the substrate selection experimental design included both multiple choice and non-choice treatments. Macroalgal structure analyses using fractals and interstitial space index were included in the study, as substrate complexity is a main factor driving preference for epifauna. Caprella grandimana actively selected J. rubens, whereas C. takeuchii did not show any preference; both behaviours remaining consistent regardless of the original substrate. Preference for J. rubens is probably owed to its interstitial space and thalli characteristics, as the complexity analysis suggested, since these allow for better refuge against predators and a more suitable surface for grasping. Meanwhile, the plasticity of C. takeuchii seems to favor an ongoing taking over of its congener at the time of the year when J. rubens drops. We highlight the need for rigor when performing substrate selection experiments; the importance of including habitat selection lessons in conservation strategies and modelling studies dealing with global change; and the risk in generalizing results within the family or genus level, which is occasionally inadequate for understanding the functioning of the ecosystem as a whole.