In light of recent alarming trends in human population growth, climate change, and other environmental modifications, a “Warning to humanity” manifesto was published in BioScience in 2017. This call reiterated most of the ideas originally expressed by the Union of Concerned Scientists in 1992, including the fear that we are “pushing Earth's ecosystems beyond their capacities to support the web of life.” As subterranean biologists, we take this opportunity to emphasize the global importance and the conservation challenges associated with subterranean ecosystems. They likely represent the most widespread nonmarine environments on Earth, but specialized subterranean organisms remain among the least documented and studied. Largely overlooked in conservation policies, subterranean habitats play a critical role in the function of the web of life and provide important ecosystem services. We highlight the main threats to subterranean ecosystems and propose a set of effective actions to protect this globally important natural heritage.

International audienceAimThree broad mechanisms have been proposed to explain geographic variation in species range size: habitat area/heterogeneity, climate seasonality and long-term climate variability. However, it has proved difficult to disentangle their relative role, particularly as temperature seasonality often covaries with the amplitude of long-term temperature oscillations. Here, we shed new light onto this debate by providing the first continental-scale analysis of range size and beta diversity in groundwater habitats, where taxa are not exposed to latitudinal variation in temperature seasonality.LocationEurope.MethodsWe compiled and mapped occurrence data for 1570 groundwater crustacean species. Generalized regression models were used to test for latitudinal variation in geographic range size and to assess the relative role of the three broad mechanisms in shaping present-day patterns of range size. We partitioned beta diversity into its spatial turnover and nestedness components and analysed their latitudinal variation across Europe.ResultsMedian range size increases with latitude above 43 degrees N and the range size of individual species is positively correlated to latitude, even after accounting for phylogenetic effects. Long-term temperature variability accounted for a substantially higher variation in median range size of groundwater crustaceans across Europe than precipitation seasonality and habitat heterogeneity, including aquifer area, elevation range, climatic rarity and productive energy. Spatial turnover contributes significantly more to beta diversity in southern regions characterized by stable historic climates than it does in northern Europe.Main conclusionsOur findings add support to the historic climate hypothesis which suggests that patterns of increasing range size and decreasing species turnover at higher latitudes in the Palaearctic region are primarily driven by long-term temperature oscillations rather than by climatic seasonality and the availability and heterogeneity of habitats

Summary 1. The spatial patterns of groundwater biodiversity in Europe remain poorly known, yet their knowledge is essential to understand local variation in groundwater assemblages and to develop sound conservation policies. We explore here the broad‐scale distribution of groundwater biodiversity across Europe, focussing on obligate subterranean species. 2. We compiled published distributional data of obligate subterranean aquatic taxa for six European countries (Belgium, France, Italy, Portugal, Slovenia and Spain), and conducted a detailed biological survey of six regions (one in Belgium, two in France, one in Italy, one in Slovenia and one in Spain). Based on this data set, we mapped spatial patterns of biodiversity in Europe on a cell grid with 0.2 × 0.2 ° resolution. 3. As of mid‐2006, the total number of described stygobiotic species in the six countries was 930 and the total number of genera with at least one described stygobiotic species was 191. The total number of sampling sites where at least one stygobiont had been collected was 4709, distributed in 1228 of the 4668 grid cells covering the study area. 4. Groundwater stygobiotic biodiversity was dominated by Crustacea with 757 species in 122 genera. Insects were represented by only two species of a single genus of dytiscid beetles restricted to south‐eastern France. 5. The geographic distribution of stygobionts was extremely heterogeneous. Stygobionts were recorded in 26% of the 4668 grid cells and only 33 cells had more than 20 stygobiotic species. These 33 ‘hot‐cells’ of groundwater species richness clustered in seven hotspots. 6. Endemicity was very high, with 43% of the total number of stygobiotic species restricted to a single cell, i.e. <500 km2. 7. Hotspots defined by rarity, number of genera, number of genera with only one species known in Europe, or number of monospecific genera differed markedly in ranking from those based on species richness. However, a core of four hotspots emerged in all cases: one stretching across Slovenia and northeastern Italy, one in the French Pyrenees, one in the Cévennes in southern France and one in the Rhine River valley in northeastern France. 8. Unevenness in stygobiont distribution cannot be explained solely by unevenness in sampling effort. This is indicated in particular by the fact that our comprehensive sampling survey roughly matched the level of taxonomic richness of the studied regions based on previously published information. 9. With sampling effort continuing, a twofold or higher increase in species richness can be expected in several Mediterranean areas, with a potential to discover up to 50% more new species than are currently known in the region.

Ecologists increasingly rely on molecular delimitation methods (MMs) to identify species boundaries, thereby potentially increasing the number of putative species because of the presence of morphologically cryptic species. It has been argued that cryptic species could challenge our understanding of what determine large‐scale biodiversity patterns which have traditionally been documented from morphology alone. Here, we used morphology and three MMs to derive four different sets of putative species among the European groundwater crustaceans. Then, we used regression models to compare the relative importance of spatial heterogeneity, productivity and historical climates, in shaping species richness and range size patterns across sets of putative species. We tested three predictions. First, MMs would yield many more putative species than morphology because groundwater is a constraining environment allowing little morphological changes. Second, for species richness, MMs would increase the importance of spatial heterogeneity because cryptic species are more likely along physical barriers separating ecologically similar regions than along resource gradients promoting ecologically‐based divergent selection. Third, for range size, MMs would increase the importance of historical climates because of reduced and asymmetrical fragmentation of large morphological species ranges at northern latitudes. MMs yielded twice more putative species than morphology and decreased by 10‐fold the average species range size. Yet, MMs strengthened the mid‐latitude ridge of high species richness and the Rapoport effect of increasing range size at higher latitudes. Species richness predictors did not vary between morphology and MMs but the latter increased the proportion of variance in range size explained by historical climates. These findings demonstrate that our knowledge of groundwater biodiversity determinants is robust to overlooked cryptic species because the latter are homogeneously distributed along environmental gradients. Yet, our findings call for incorporating multiple species delimitation methods into the analysis of large‐scale biodiversity patterns across a range of taxa and ecosystems.

Five decades ago, a landmark paper in Science titled The Cave Environment heralded caves as ideal natural experimental laboratories in which to develop and address general questions in geology, ecology, biogeography, and evolutionary biology. Although the ‘caves as laboratory’ paradigm has since been advocated by subterranean biologists, there are few examples of studies that successfully translated their results into general principles. The contemporary era of big data, modelling tools, and revolutionary advances in genetics and (meta)genomics provides an opportunity to revisit unresolved questions and challenges, as well as examine promising new avenues of research in subterranean biology. Accordingly, we have developed a roadmap to guide future research endeavours in subterranean biology by adapting a well‐established methodology of ‘horizon scanning’ to identify the highest priority research questions across six subject areas. Based on the expert opinion of 30 scientists from around the globe with complementary expertise and of different academic ages, we assembled an initial list of 258 fundamental questions concentrating on macroecology and microbial ecology, adaptation, evolution, and conservation. Subsequently, through online surveys, 130 subterranean biologists with various backgrounds assisted us in reducing our list to 50 top‐priority questions. These research questions are broad in scope and ready to be addressed in the next decade. We believe this exercise will stimulate research towards a deeper understanding of subterranean biology and foster hypothesis‐driven studies likely to resonate broadly from the traditional boundaries of this field.

SUMMARYObligatory cave species exhibit dramatic trait modifications such as eye reduction, loss of pigmentation and an increase in touch receptors. As molecular studies of cave adaptation have largely concentrated on vertebrate models, it is not yet possible to probe for genetic universalities underlying cave adaptation. We have therefore begun to study the strongly cave-adapted small carrion beetle Ptomaphagus hirtus. For over 100 years, this flightless signature inhabitant of Mammoth Cave, the world's largest known cave system, has been considered blind despite the presence of residual lens structures. By deep sequencing of the adult head transcriptome, we discovered the transcripts of all core members of the phototransduction protein machinery. Combined with the absence of transcripts of select structural photoreceptor and eye pigmentation genes, these data suggest a reduced but functional visual system in P. hirtus. This conclusion was corroborated by a negative phototactic response of P. hirtus in light/dark choice tests. We further detected the expression of the complete circadian clock gene network in P. hirtus, raising the possibility of a role of light sensation in the regulation of oscillating processes. We speculate that P. hirtus is representative of a large number of animal species with highly reduced but persisting visual capacities in the twilight zone of the subterranean realm. These can now be studied on a broad comparative scale given the efficiency of transcript discovery by next-generation sequencing.

We studied species richness patterns of obligate subterranean (troglobiotic) beetles in the Dinaric karst of the western Balkans, using five grid sizes with cells of 80 × 80, 40 × 40, 20 × 20, 10 × 10, and 5 × 5 km. The same two hotspots could be recognized at all scales, although details differed. Differences in sampling intensity were not sufficient to explain these patterns. Correlations between number of species and number of sampled localities increased with increasing cell size. Additional species are expected to be found in the region, as indicated by jackknife 1, jackknife 2, Chao2, bootstrap, and incidence‐based coverage (ICE) species richness estimators. All estimates increased with increasing cell size, except Chao2, with the lowest prediction at the intermediate 20 × 20 km cell size. Jackknife 2 and ICE gave highest estimates and jackknife 1 and bootstrap the lowest. Jackknife 1 and bootstrap estimates changed least with cell size, while the number of single cell species increased. In the highly endemic subterranean fauna with many rare species, bootstrap may be most appropriate to consider. Positive autocorrelation of species numbers was highest at 20 × 20 km scale, so we used this cell size for further analyses. At this scale we added 137 localities with less positional accuracy to 1572 previously considered, and increased 254 troglobiotic species considered to 276. Previously discovered hotspots and their positions did not change, except for a new species‐rich cell which appeared in the south‐eastern region. There are two centres of troglobiotic species richness in the Dinaric karst. The one in the north‐west exhibited high species richness of Trechinae (Carabidae), while in the south‐east, the Leptodirinae (Cholevidae) were much more diverse. These centres of species richness should serve as the starting point for establishing a conservation network of important subterranean areas in Dinaric karst.

International audienceThe recognition of multi-causality and spatial non-stationarity in the determinants of large-scale biodiversity patterns requires to consider the role of multiple mechanisms, their interactions, and how these mechanisms vary in strength relative to each other across geographical space. Here, we challenge the view that historical climate stability primarily drives European patterns of groundwater crustacean diversity by testing also the role of spatial heterogeneity and productive energy. First, we predicted that the three mechanisms would be equally important at continental scale when analyzed separately, but that the importance of historical climate variability would weaken in joint analyses due to co-variation with the two other mechanisms. Second, we predicted that the three mechanisms would exhibit predictable latitudinal changes in their relative strength. To test these predictions, we selected predictors representing each mechanism and analyzed separately and jointly their effects and interactions using global regression models. We further mapped the independent and overlapping effects of mechanisms across Europe using partial geographically weighted regressions. When analyzed separately, the three mechanisms explained the same amount of variation in species richness, but in the joint analysis, the influence of historical climate stability became hidden in the variation shared with the other mechanisms. Topographic heterogeneity interacted synergistically with actual evapotranspiration and habitat heterogeneity on species richness. Spatial non-stationarity in the independent and overlapping effects of the three mechanisms was the most plausible expla- nation for the hump-shaped latitudinal pattern of crustacean species richness. Productive energy and spatial heterogeneity were important predictors at mid and southern latitudes, whereas historical climate stability overlapped with the two other mechanisms in northern Europe and productive energy in southern Europe. Multi-causality and spatial non-stationarity provide a broader perspective of groundwater biodiversity determinants that revives the importance of spatial heterogeneity and the strong dependence of subterranean communities on food supply from the surface