Temperature is an important factor in determining species distribution patterns in ectothermic organisms (Angilleta, 2009). As sea surface temperatures continue to rise, understanding how these changes will affect species distributions demands a broad understanding of organisms' physiological sensitivities to temperature across their native range. There is overwhelming evidence that temperature increases associated with anthropogenic climate change are having widespread ecological consequences on marine species

7Corals are threatened worldwide due to rapidly warming oceans associated with anthropogenic 8 climate change. A growing body of work has explored how corals respond to heat stress, however, 9responses of corals across the full range of their thermal breadth are rarely explored even though 10 winter colds are also known to induce coral stress. Here, we leverage the temperate stony coral, 11Astrangia poculata, which naturally exhibits a facultative symbiosis with Symbiodiniaceae, to 12 explicitly examine how thermal challenges influence coral hosts in isolation from their symbionts. 13Aposymbiotic A. poculata were collected from Woods Hole, MA, the northern range limit for this 14 species. Corals were thermally challenged in two independent common garden experiments (Heat 15 challenge: 31˚C, 10 days; Cold challenge: 6˚C, 16 days) to determine the effects of divergent 16thermal stressors. Behavioural responses to food stimuli were monitored throughout the thermal 17 challenges and genome-wide gene expression profiling (TagSeq) was used to characterize 18 molecular underpinnings of the coral's response to stress in its aposymbiotic state. Behaviourally, 19both thermal challenges induced polyp retraction and colonies failed to respond to food stimuli. 20Surprisingly, seven times as many genes were differentially expressed under cold challenge heat 21 challenge. Despite a greater magnitude of response under cold challenge, significant similarities 22 in gene expression were detected across the two thermal challenge experiments, with many of the 23 most responsive genes having been previously implicated in coral heat stress response. Given that 24 our data were generated from aposymbiotic colonies, we hypothesize that these genes previously 25linked to heat stress are more likely indicative of a generalized stress response. Overall this work 26 highlights the unique insights facultatively symbiotic corals offer for deciphering the stress 27 response of the coral host in the absence of bleaching. 28 29 30

Arboreal embryos of red-eyed treefrogs, Agalychnis callidryas, hatch prematurely in response to hypoxia when flooded and mechanosensory cues (MC) in snake attacks, but hatching later improves tadpole survival. We studied ontogenetic changes in risk assessment and hatching performance of embryos in response to flooding and physical disturbance. We hypothesized that risk assessment decreases as hatchling survival improves and hatching performance increases as embryos develop. Because snakes eat faster than embryos asphyxiate, we hypothesized that embryos decide to hatch sooner and hatch faster in response to MC. We video-recorded individual embryos hatching in response to each cue type, then compared the incidence and timing of a series of events and behaviors from cue onset to complete hatching across ages and stimuli. Latency from cue to hatching decreased developmentally in both contexts and was shorter with MC, but the elements contributing to those changes differed. Hypoxia-assessment involved position changes, which decreased developmentally along with assessment time. MC-assessment occurred more rapidly, without movement, and decreased with age. The first stages of hatching, membrane rupture and head emergence, were surprisingly age-independent but faster with MC, congruent with greater effort under more immediate risk. In contrast, body emergence and compression showed ontogenetic improvement consistent with morphological constraints but no cue effect. Both appropriate timing and effective performance of hatching are necessary for continued development. Different stages of the process vary with development and environmental context, suggesting combinations of adaptive context- and stage-dependent behavior, cue-related constraints on information acquisition, and ontogenetic constraints on elements of performance.

Arboreal embryos of red-eyed treefrogs, Agalychnis callidryas, hatch prematurely in response to hypoxia when flooded and mechanosensory cues (MC) in snake attacks, but hatching later improves tadpole survival. We studied ontogenetic changes in risk assessment and hatching performance of embryos in response to flooding and physical disturbance. We hypothesized that risk assessment decreases as hatchling survival improves and hatching performance increases as embryos develop. Because snakes eat faster than embryos asphyxiate, we hypothesized that embryos decide to hatch sooner and hatch faster in response to MC. We video-recorded individual embryos hatching in response to each cue type, then compared the incidence and timing of a series of events and behaviors from cue onset to complete hatching across ages and stimuli. Latency from cue to hatching decreased developmentally in both contexts and was shorter with MC, but the elements contributing to those changes differed. Hypoxia-assessment involved position changes, which decreased developmentally along with assessment time. MC-assessment was passive, faster, and also decreased with age. The first stages of hatching, membrane rupture and head emergence, were surprisingly age-independent but faster with MC, congruent with greater effort under more immediate risk. In contrast, body emergence and compression showed ontogenetic improvement consistent with morphological constraints but no cue effect. Both appropriate timing and effective performance of hatching are necessary for continued development. Different stages of the process vary with development and environmental context, suggesting combinations of adaptive context- and stage-dependent behavior, cue-related constraints on information acquisition, and ontogenetic constraints on elements of performance.Summary statementDevelopment reduces risk assessment and improves hatching performance by red-eyed treefrog embryos in both flooding and simulated attacks. Predation cues elicit faster decisions and hatching, congruent with more immediate risk.