Restoring and maintaining megafauna in today’s human‐dominated landscapes is a major challenge for conservation. The refugee species concept (Kerley et al., 2011) is helpful in reminding us that species’ habitat use might have changed over time, and that the option space for restoring such species can therefore be wider than appreciated. In this context, Kerley et al. (2020) express concerns that our assessment of potential reintroduction sites for European bison (Perzanowski et al., 2019) might ignore possibly shifted baselines. We welcome their comment and the opportunity to explain how our work relates to the hypothesis that European bison are a refugee species.

Kerley et al. (2020) misinterpret our habitat assessment as based on ‘the paradigm that the European bison is a forest specialist’. Our analyses make use of a large and diverse European bison habitat use dataset (telemetry data from 45 individuals across all free‐ranging herds in Poland). As we have detailed elsewhere (Kuemmerle et al., 2018), these data unequivocally show that European bison (a) prefer mosaic‐type landscapes, (b) today inhabit landscapes with widely varying shares of forests and open areas, and (c) sometimes select for open areas, and sometimes not. This suggests that European bison are fairly generalist as regards landscape composition – a point we have made repeatedly in the past (Kuemmerle et al., 2011; Kuemmerle et al., 2018), and a point consistent with other studies of European bison habitat selection (including those cited in Kerley et al., 2020). In fact, we are not aware of any recent study portraying European bison as strict forest specialists. The habitat use data of Kuemmerle et al. (2018), cited in Kerley et al. (2020) as evidence for European bison selecting for open areas, formed the basis for identifying potential European bison reintroduction sites in Poland and these sites had on average about 20% open areas (Perzanowski et al., 2019). In sum, our assessment neither frames the European bison as a strict forest specialist nor do our data and recommendations preclude potential preferences for open areas.

More generally, Kerley et al. (2020) touch upon the question of whether or not European bison qualify as a refugee species. All recent studies of European bison habitat use we know of suggest that European bison are neither a strict forest specialist nor a strict grassland specialist. For example, European bison have exhibited intermediate diets between true grazers and true browsers, along with high diet plasticity, throughout the Holocene (Bocherens et al., 2015; Hofman‐Kamińska et al., 2019), a pattern also found in contemporary herds (Krasińska et al., 2014). Vegetation reconstructions (Kuemmerle et al., 2012a) and paleozoological data (Benecke, 2005) suggest that European bison have, for at least 8000 years, been associated with landscapes containing a considerable amount of forest alongside open areas – as they are today (Kuemmerle et al., 2011; Kuemmerle et al., 2018). Likewise, life history, behavioral and morphological evidence point to generalist diets and habitat use (see Kerley et al., 2011 for a summary), but do not support the framing of the species as a strict forest or grassland specialist. Thus, current evidence suggests that (a) habitat use and forest association of European bison might not have changed as dramatically during the Holocene as postulated by the refugee hypothesis and (b) many contemporary European bison populations are not ‘forest‐confined’, as suggested by Kerley et al. (2020). This questions notions of a shifting baseline syndrome, as well as the accuracy of classifying European bison as a refugee species. Good practice in conservation science should allow for the possibility that an intriguing hypothesis might still be rejected.

What is the way forward for European bison conservation? We fully agree that European bison conservation should consider a wide range of landscapes as potentially suitable, including more open landscapes, as we have argued in the past (Kuemmerle et al., 2018; Perzanowski et al., 2019). Regions where agriculture has been abandoned are indeed worth examining for their megafauna restoration potential. Yet we believe that the value of these landscapes does not primarily come from their current land‐cover composition, but from the declining human pressure that characterizes some of them. Potential conflicts with agriculture and forestry are today arguably the main barriers for restoring European bison and other megafauna. However, conflict potential is neither generally high or low in forest landscapes nor in landscapes where agricultural abandonment happens. For example, abandoned fields across Poland often occur alongside much larger areas of managed fields, signaling high risk for conflicts between European bison and farmers in such areas. Our vision is one where large metapopulations of European bison once again have a place within their former range. Achieving such a vision will not be served by focusing on establishing European bison herds in specific habitat types, but by identifying networks of sites where overall conflict potential is low – regardless of the specific degree of forest cover (Kuemmerle et al., 2012b; Kuemmerle et al., 2018).

Saving the European bison from the brink of extinction and restoring the species across parts of its historical range is a rare conservation success story. Once extinct in the wild, today more than 4500 wild European bison occur in more than 40 herds across eastern Europe, with steadily increasing populations (Krasińska et al., 2014). Suggesting that these conservation efforts are in fact putting the species at risk (Kerley et al., 2020) is unsupported by scientific evidence, diminishes the efforts of those who have contributed to European bison conservation in the 20th and 21st century, and is detrimental to European bison conservation. We wholeheartedly agree with Kerley et al. (2020) that we must ‘now avail our efforts to save the species with the full capacity of our best available and rigorous science’. We suggest bison researchers should unite in such efforts and leave oversimplifications and polarized debates behind them. Identifying a viable future for European bison, and other megafauna, in Europe will require a forward‐looking perspective that does not seek to recreate the past, and that embraces the full range of restoration options.

在当今人类主导的景观中恢复和维护大型动物是保护工作的主要挑战。难民物种的概念（Kerley等，2011）有助于提醒我们，物种的栖息地使用可能会随着时间而改变，因此恢复此类物种的选择空间可能会比人们理解的要宽。在这种情况下，Kerley等人。（2020年）表达了对我们对欧洲野牛潜在再引入位点的评估（Perzanowski等人，2019年）的担忧，这可能会忽略基线的可能变化。我们欢迎他们的评论，并有机会解释我们的工作如何与欧洲野牛是难民物种这一假说联系起来。

Kerley等。（2020年）将我们的栖息地评估误解为基于“欧洲野牛是森林专家的范式”。我们的分析利用了一个庞大而多样的欧洲野牛栖息地使用数据集（来自波兰所有自由放养牛群的45个人的遥测数据）。正如我们在其他地方详细介绍的那样（Kuemmerle et al。，2018），这些数据明确表明欧洲野牛（a）喜欢镶嵌式景观，（b）今天居住在森林和开放区域所占比例差异很大的景观中，以及（c）有时选择开放区域，有时则不选择。这表明欧洲野牛在景观组成方面相当笼统，这是我们过去反复提到的观点（Kuemmerle等。，2011 ; Kuemmerle et al。，2018），这一观点与欧洲野牛栖息地选择的其他研究一致（包括Kerley et al。，2020中引用的那些）。实际上，我们还没有发现任何将欧洲野牛描绘成严格的森林专家的最新研究。Kuemmerle等人的栖息地利用数据。（2018），在Kerley等人中引用。（2020年）作为欧洲野牛选择开放区域的证据，为确定潜在的欧洲野牛在波兰的再引入地点打下了基础，这些地点平均约有20％的开放区域（Perzanowski等人，2019年））。总而言之，我们的评估既没有将欧洲野牛定为严格的森林专家，也没有根据我们的数据和建议排除对开放区域的潜在偏爱。

更一般地，Kerley等人。（2020）谈到了欧洲野牛是否有资格成为难民物种的问题。我们知道的所有有关欧洲野牛栖息地使用的最新研究都表明，欧洲野牛既不是严格的森林专家，也不是严格的草原专家。例如，在整个全新世中，欧洲野牛在真正的放牧者和真正的浏览者之间表现出中等的饮食，以及高的饮食可塑性（Bocherens等人，2015 ;Hofman-Kamińska等人，2019），这也是当代牛群中的一种模式（Krasińska等，2014）。植被重建（Kuemmerle等，2012a）和古动物学数据（Benecke，2005）表明，至少在8000年以前，欧洲野牛已经与包含大量森林和开阔地带的景观相关联（今天如此）（Kuemmerle等人，2011； Kuemmerle等人，2018）。同样，生活史，行为和形态学证据也表明了人们的饮食习惯和栖息地的利用（有关摘要，请参见Kerley等，2011），但不支持将该树种定为严格的森林或草原专家。因此，目前的证据表明：（a）全新世时期欧洲野牛的栖息地使用和森林联系可能没有像难民假说所假设的那样发生显着变化；（b）许多当代欧洲野牛种群没有被“森林限制”，因为由Kerley等人提出。（2020年）。这质疑基线综合症转移的观念，以及将欧洲野牛归为难民物种的准确性。保护科学的良好实践应该允许一个有趣的假设仍然被拒绝的可能性。

欧洲野牛养护的前进方向是什么？我们完全同意，欧洲野牛养护应考虑到广泛的景观潜力，包括我们以前所说的更开放的景观（Kuemmerle et al。，2018 ; Perzanowski et al。，2019）。放弃农业的地区的大型动物恢复潜力确实值得研究。然而，我们认为，这些景观的价值并非主要来自其目前的土地覆盖构成，而是来自代表其中某些特征的人类压力的下降。今天，与农业和林业的潜在冲突可以说是恢复欧洲野牛和其他大型动物的主要障碍。但是，潜在的冲突一般都不会在森林景观或发生农业遗弃的景观中高或低。例如，波兰各地的废弃土地经常与更大范围的管理土地并存，这表明欧洲野牛与该地区农民之间发生冲突的风险很高。我们的愿景是，欧洲野牛的大型种群再次在其先前范围内占有一席之地。要实现这样的愿景，将不会专注于在特定栖息地类型上建立欧洲野牛群，而是通过识别总体冲突潜力较低的站点网络而不论森林覆盖的具体程度如何（Kuemmerle等，2012b； Kuemmerle等人，2018）。

保护欧洲野牛免于灭绝的边缘，并在其历史范围的各个部分恢复该物种，这是一个罕见的保护成功案例。一旦在野外灭绝，如今在整个东欧的40多个牧群中出现了4500多种野生欧洲野牛，种群数量稳步增加（Krasińska等，2014）。这表明这些保护工作实际上使该物种处于危险之中（Kerley等，2020），没有科学证据支持，减少了在20和21世纪为欧洲野牛保护做出贡献的人们的努力，并且对欧洲野牛养护。我们完全同意Kerley等人的观点。（2020年），我们必须“现在尽我们最大的努力和最严谨的科学能力竭尽全力保护该物种”。我们建议野牛研究人员应团结起来，为此类工作留下过多的简化和两极化的辩论。确定欧洲野牛和其他大型动物的可行的未来，将需要一种前瞻性的观点，即不试图重现过去，而应涵盖所有的修复方案。