Seventy years after the discovery that steroid metabolites measured in urine or faeces reflect circulating hormone levels, non‐invasive measurements of glucocorticoids and their derivative from various biological matrices have become one of the most essential tools in the functional ecologist's repertoire (Palme, 2019; Sheriff, Dantzer, Delehanty, Palme, & Boonstra, 2011). Glucocorticoids (cortisol and corticosterone) are vertebrate metabolic hormones involved in energy maintenance and regulation whose secretion increases drastically in response to various physiological and psychological stressors. Although the release of glucocorticoids in response to a stressor aimed at maintaining or restoring homoeostasis and provides direct fitness benefits in the short‐term, chronically elevated glucocorticoid levels can affect key body functions such as immunity, growth or reproduction, even if the intensity of these effects seems to vary from one species to another (Boonstra, 2013).

Given the duality of glucocorticoid effects on animal performance parameters, describing the patterns of glucocorticoids in natural populations became a major element for characterizing the role of stress as an ecological driving force and a key player in evolution. In this context, the vast majority of studies found in the literature aim at determining the environmental (extrinsic) and individual (intrinsic) drivers of glucocorticoids levels. In contrast, very few studies have sought to establish causal relationships between level of glucocorticoids, animal performance, fitness and how these translate into population dynamics (Beehner & Bergman, 2017; Breuner, Patterson, & Hahn, 2008).

Functional links between the causes and consequences of stress on demography were established in an elegant study by Dulude‐de Broin and his collaborators in the current issue of Functional Ecology (Dulude‐de Broin, Hamel, Mastromonaco, & Côté, 2020). Using the mountain goat Oreamnos americanus of the Canadian Rockies as study model, the authors validated for the first time in a wild population of large mammals the ‘predation‐stress hypothesis’ which predict that repeated exposure to predators would affect the fitness components and demographic trajectories of their prey through deleterious effects of chronic stress. To achieve their goal, the authors have combined an impressive array of environmental, physiological and demographic data on mountain goats collected over 23 years in a multi‐predator system, raising the bar high in the field of stress ecology. Exposure to predators, quantified as the probability of encountering one of the predator species at the study site, was determined by intense monitoring in the field and was used as an explanatory variable to model patterns of glucocorticoids levels. Other environmental factors, such as the availability of resources (measured as faecal crude protein peak) and population size, were also quantified each year and taken into account in the analyses. Faeces and hair samples were collected each summer during captures and were used to model annual estimates of glucocorticoid levels in the population while controlling for potential individual effects such as age, sex and body mass and seasonal variations. To link the effects of fluctuating glucocorticoids to population dynamics, the authors quantified reproductive success as the proportion of reproductive females each year in the population. Finally, the causal relationship between the risk of predation, the levels of glucocorticoids at the population level and the reproductive success were established by performing path analyses.

The results are unequivocal: increasing the risk of encountering predators (grizzly and black bears, wolves, cougars, coyotes and wolverines) had a direct positive effect on faecal glucocorticoids metabolites (FGM) which, in turn, negatively affected the proportion of reproductive females in the population. In short, a 30% drop in the proportion of reproductive females was observed in the years when the level of FGM is high compared to the years when the level of FGM is low, with a resounding effect on demography. Although the data do not allow determining whether the stress‐induced increase in glucocorticoids primarily affects female or male fertility—and this constitutes the next challenge for future studies—the current finding supports the idea that the physiological anti‐predator responses bear energy costs that impact reproduction with measurable impact on demography.

These findings are of great value to community ecology as they identify, for the first time in a large mammal, a physiological mechanism by which predator–prey interactions other than predation may shape communities. In many taxa, the impact of those interactions (also referred as non‐consumptive effects of predation) on the demography of their prey are predominant and even greater than the direct effect of predation (Preisser, Bolnick, & Benard, 2005). In large mammals, the non‐consumptive effects of predation are classically attributed to behavioural anti‐predator responses such as habitat shift or temporal avoidance (Say‐Sallaz, Chamaillé‐Jammes, Fritz, & Valeix, 2019). By establishing causal relationships between predation risk, glucocorticoids levels and reproductive outputs, the findings of Dulude‐de Broin et al. (2020) on mountain goats provide an alternative scenario to that proposed for other ungulates for which the predation stress hypothesis is typically not valid (e.g. Creel, Winnie, & Christianson, 2009). Hence, Dulude‐de Broin's observations suggest that the nature of the anti‐predatory response and the associated cost depend on the ecological context. The predation‐stress hypothesis would apply when the presence of predators cannot be predicted and/or proactively mitigated by behavioural responses, and these conditions are met in the case of mountain goats (Dulude‐de Broin et al., 2020).

Although quantifying the effect of the risk of predation on the reproduction and survival of prey in wild populations of large mammals poses serious challenges (Say‐Sallaz et al., 2019), the article by Dulude‐de Broin highlights the importance of considering both physiological and behavioural responses to predation risks for a better understanding of the dynamics of predator–prey interactions. To this end, extrapolating the stress response from proxies of glucocorticoids levels appear to be relevant to link the physiology to fitness components even if there are limitations. In particular, the large unexplained residual variability that emerges from the models, including when many confounding factors are taken into account (Dulude‐de Broin et al., 2020), suggest that glucocorticoids levels may include multiple components, potentially other than stress, that obscure the signal. In the current context where science is bogged down in a replication crisis (Baker, 2016) which does not spare the field of ecology and evolution (Fraser, Parker, Nakagawa, Barnett, & Fidler, 2018), using several complementary approaches to quantify the response stress—as was done in Dulude‐de Broin's study and in other recent studies in stress ecology (e.g. Boudreau et al., 2019)—may soon become a methodological requirement. On a larger scale, there is a need to focus on other physiological aspects such as the metabolic (energy) and somatic costs (e.g. oxidative stress, immune function) associated with the risk of predation in large mammals. This ‘strategic use of multiple approaches to answer a question’ qualified as ‘triangulation’ by Munafò and Smith (2018) is necessary for a global understanding of the physiological cost of predation risk and their cascading effects on prey population dynamics.

发现尿液或粪便中的类固醇代谢产物反映了循环激素水平后的七十年，无创性地测定糖皮质激素及其衍生自各种生物基质的衍生物已成为功能生态学家库中最重要的工具之一（Palme，2019 ; Sheriff ，Dantzer，Delehanty，Palme和Boonstra，2011年）。糖皮质激素（皮质醇和皮质酮）是参与能量维持和调节的脊椎动物代谢激素，其分泌会随着各种生理和心理压力而急剧增加。尽管糖皮质激素的释放是为了应对应激反应，旨在维持或恢复稳态，并在短期内提供直接的健身益处，但长期升高的糖皮质激素水平仍会影响人体关键功能，例如免疫力，生长或繁殖，即使这些物质的强度影响似乎从一个物种到另一个物种（Boonstra，2013）。

考虑到糖皮质激素对动物性能参数的双重影响，描述自然种群中糖皮质激素的模式已成为表征压力作为生态驱动力和进化关键角色的重要要素。在这种情况下，文献中发现的绝大多数研究都旨在确定糖皮质激素水平的环境（外部）和个体（内部）驱动因素。相比之下，很少有研究试图建立糖皮质激素水平，动物性能，健康状况以及它们如何转化为种群动态之间的因果关系（Beehner＆Bergman，2017 ; Breuner，Patterson，＆Hahn，2008）。

Dulude-de Broin及其合作者在当前的功能生态学杂志（Dulude-de Broin，Hamel，Mastromonaco和Côté，2020年）进行的一项优雅研究中，建立了人口统计学压力因果关系之间的功能联系。使用美洲山羊Oreamnos作者以加拿大洛矶山脉为研究模型，首次在大型哺乳动物的野生种群中验证了“捕食压力假说”，该假说假设反复接触捕食者会通过有害作用影响猎物的适应性成分和人口统计学轨迹慢性压力。为了实现其目标，作者在多捕食者系统中结合了超过23年收集的关于山羊的环境，生理和人口统计数据，令人印象深刻，从而提高了压力生态学领域的门槛。捕食者的暴露，被量化为在研究地点遇到一种捕食者物种的概率，是通过现场的严格监测确定的，并用作解释糖皮质激素水平模式的解释变量。其他环境因素 如资源的可用性（以粪便粗蛋白峰值测量）和人口规模，也每年进行量化，并在分析中加​​以考虑。每年夏天在捕获过程中收集粪便和头发样本，并用于模拟人群中糖皮质激素水平的年度估算，同时控制潜在的个体影响，例如年龄，性别和体重以及季节性变化。为了将糖皮质激素波动的影响与种群动态联系起来，作者将生殖成功量化为人口中每年育种雌性的比例。最后，通过进行路径分析，确定了捕食风险，人群中糖皮质激素水平与生殖成功之间的因果关系。每年也进行量化，并在分析中加​​以考虑。每年夏天在捕获过程中收集粪便和头发样本，并用于模拟人群中糖皮质激素水平的年度估算，同时控制潜在的个体影响，例如年龄，性别和体重以及季节性变化。为了将糖皮质激素波动的影响与种群动态联系起来，作者将生殖成功量化为人口中每年育种雌性的比例。最后，通过进行路径分析，确定了捕食风险，种群中糖皮质激素水平和生殖成功之间的因果关系。每年也进行量化，并在分析中加​​以考虑。每年夏天在捕获过程中收集粪便和头发样本，并将其用于模拟人群中糖皮质激素水平的年度估算，同时控制潜在的个体影响，例如年龄，性别和体重以及季节性变化。为了将糖皮质激素波动的影响与种群动态联系起来，作者将生殖成功量化为人口中每年育种雌性的比例。最后，通过进行路径分析，确定了捕食风险，种群中糖皮质激素水平和生殖成功之间的因果关系。每年夏天在捕获过程中收集粪便和头发样本，并将其用于模拟人群中糖皮质激素水平的年度估算，同时控制潜在的个体影响，例如年龄，性别和体重以及季节性变化。为了将糖皮质激素波动的影响与种群动态联系起来，作者将生殖成功量化为人口中每年育种雌性的比例。最后，通过进行路径分析，确定了捕食风险，人群中糖皮质激素水平与生殖成功之间的因果关系。每年夏天在捕获过程中收集粪便和头发样本，并将其用于模拟人群中糖皮质激素水平的年度估算，同时控制潜在的个体影响，例如年龄，性别和体重以及季节性变化。为了将糖皮质激素波动的影响与人口动态联系起来，作者将生殖成功量化为人口中每年繁殖女性的比例。最后，通过进行路径分析，确定了捕食风险，人群中糖皮质激素水平与生殖成功之间的因果关系。为了将糖皮质激素波动的影响与人口动态联系起来，作者将生殖成功量化为人口中每年繁殖女性的比例。最后，通过进行路径分析，确定了捕食风险，种群中糖皮质激素水平和生殖成功之间的因果关系。为了将糖皮质激素波动的影响与种群动态联系起来，作者将生殖成功量化为人口中每年育种雌性的比例。最后，通过进行路径分析，确定了捕食风险，人群中糖皮质激素水平与生殖成功之间的因果关系。

结果是明确的：增加与天敌（灰熊和黑熊，狼，美洲狮，土狼和金刚狼）接触的风险，对粪便糖皮质激素代谢物（FGM）有直接的积极影响，反过来又对粪便中糖皮质激素的代谢产生负面影响。人口。简而言之，与FGM水平低的年份相比，FGM水平高的年份中生殖女性的比例下降了30％，对人口统计学产生了巨大的影响。

这些发现对于群落生态学非常有价值，因为它们首次在大型哺乳动物中发现了一种生理机制，通过该机制，掠食者与猎物之间的相互作用（除了掠食之外）可以塑造群落。在许多生物分类中，这些相互作用（也称为掠食的非消费效应）对其猎物人口的影响是主要的，甚至大于掠食的直接影响（Preisser，Bolnick和Benard，2005年）。在大型哺乳动物中，捕食的非消费影响通常归因于行为上的反捕食者反应，例如栖息地转移或暂时避开（Say-Sallaz，Chamaillé-Jammes，Fritz和Valeix，2019年））。通过建立捕食风险，糖皮质激素水平和生殖输出之间的因果关系，Dulude-de Broin等人的发现。（2020）提出了一种针对其他有蹄类动物提出的替代方案，对于其他有蹄类动物而言，捕食者应激假设通常无效（例如Creel，Winnie和Christianson，2009年）。因此，Dulude-de Broin的观察表明，反掠夺性反应的性质和相关成本取决于生态环境。当不能通过行为反应来预测和/或主动缓解掠食者的存在时，就可以采用掠夺-压力假说，并且在山羊条件下可以满足这些条件（Dulude-de Broin等，2020）。

尽管量化捕食风险对大型哺乳动物野生种群中猎物繁殖和存活的影响提出了严峻挑战（Say‐Sallaz等人，2019年），但Dulude-de Broin的文章强调了同时考虑两种生理因素的重要性以及对捕食风险的行为响应，以便更好地了解捕食者与猎物互动的动态。为此，即使存在局限性，从糖皮质激素水平的代理推断应激反应似乎也与将生理学与适应性成分联系起来有关。特别是从模型中出现了巨大的无法解释的剩余变异性，包括在考虑了许多混杂因素时（Dulude-de Broin等人，2020年），提示糖皮质激素水平可能包含多种成分（可能不是压力）而使信号模糊。在当前的科学陷入复制危机的困境中（Baker，2016），它没有利用生态和进化领域（Fraser，Parker，Nakagawa，Barnett和Fidler，2018），使用几种互补的方法来量化应对压力-在Dulude-de Broin的研究和其他近期关于压力生态学的研究中所做的工作（例如Boudreau等，2019）-可能很快成为方法论上的要求。在更大的规模上，需要关注其他生理方面，例如与大型哺乳动物捕食风险相关的代谢（能量）和躯体成本（例如氧化应激，免疫功能）。Munafò和Smith（2018）将这种``策略性地使用多种方法来回答一个问题''定性为``三角剖分''对于全球理解捕食风险的生理成本及其对猎物种群动态的级联影响是必要的。