Ocean deoxygenation and warming have been shown to pose a growing threat to the health of marine organisms and ecosystems. Yet, the potential for acclimation and adaptation remains poorly understood. The aim of this study was to evaluate the effects of transgenerational exposure to reduced oxygen availability and elevated seawater temperature on the chemosensory-dependent mating mechanisms of male amphipods Gammarus locusta. Three subsequent generations were exposed to four experimental treatments for 30 days: i) present-day scenario, ii) warming; iii) deoxygenation; and iv) warming+deoxygenation. After exposure, the number of individuals that reached adulthood was gauged, and adult males from F₀ and F₁ were subjected to behavioral trials to assess their capacity of long-distance female cue detection through quantification of response time, first direction of movement, activity rate, and proportion of time spent in female scent cues. Ocean-warming-induced mortality and reduced oxygen availability had adverse effects on each of the investigated behavioral traits, which were amplified when combined with elevated temperature. Still, when compared to F₀, the F₁ generation demonstrated more adaptability (i.e., higher activity rate and preference for female odors) to the combination of the two stressors, suggesting positive carry-over effects. Nevertheless, full recovery to control levels was not observed. Altogether, this study indicates that future scenarios of ocean deoxygenation and warming have the potential to disrupt chemosensory-dependent mate detection in amphipods, but also suggests possible behavioral adaptations. We call for greater research efforts on long-term impacts of ocean change on the behavioral and physiological processes of benthic coastal communities.

Our study aimed to investigate the transgenerational effects of increased temperature and lower oxygen levels on male gammarids' response to female scent signals. For this purpose, we conducted a binary-choice experiment using a total of 30 males from the F₀ generation and 26 males from the F₁ generation. These trials involved direct observations, during which we recorded four key traits of male individuals: response time (time to initiate movement towards an association zone), proportion of first choice (initial movement towards the female-scented association zone), cumulative time spent in each association zone, and activity rate (proportion of time spent actively moving). 

Data analysis of behavioral outputs was performed through generalized linear mixed models (GLMM) using the negative binomial (number of individuals, response time), binomial (first choice), and Gaussian (cumulative time & activity rate) residual distributions. All models used temperature, oxygen, and generation as categorical fixed factors, and replicate ID as random factors. Transgenerational effects of the treatments on these variables were considered whenever an interacting effect was observed between Generation and Treatment. Full models, with all possible interactions, were tested using the function ‘glmmTMB’ from the package ‘glmmTMB’ and the function 'Anova' from the package 'car' in R, version 3.4.3. Post-hoc multiple comparisons were performed using the package ‘emmeans’ and applying Tukey corrections. Model assumptions and performance were validated using the package “performance”. Data exploration was conducted using the HighstatLibV10 R library from Highland Statistics.