A trade-off between latitude and elevation contributes to explain range segregation of broadly distributed cave-dwelling spiders

A fundamental goal in spatial ecology is to understand how the distribution of species varies along latitudinal and elevational gradients. This stems from the understanding that latitude and elevation are primary drivers affecting temperature variations on Earth's surface, in turn, that plays a critical ecological role. These spatial gradients have been primarily documented using highly dispersive surface species—butterflies, birds, and plants—whereas studies on subterranean organisms remain scattered. The orb-web cave spiders Meta bourneti and M. menardi are ubiqui tous inhabitants of European caves whose distributions stretches over a continental distance. They share a similar ecological niche, which should translate into competitive exclusion in co-occurring areas. Therefore, it can be predicted that there should be an effective spatial segregation between the two species along broadscale spatial gradients. Using a dataset of > 3,000 georeferenced records, we show that the two species are primarily segregated along the latitudinal gradient, with M. menardi pro-gressively becoming more frequent and M. bourneti rarer from south to north. In their overlapping range (36.5–53.4° latitude), the two species are secondarily separated along an elevational gradient, with M. menardi occupying, on average, sites at higher elevations than M. bourneti . However, in the northernmost part of its range and in the absence of its competitor, M. menardi inhabits caves at lower elevations. This clear pattern provides a textbook example of the trade-off between latitude and elevation in determining habitat segregation of broadly distributed competing species.


| INTRODUC TI ON
Latitude and elevation are among the most conspicuous broadscale ecological gradients used by macroecologists to explain patterns of species distributions in space and time (Hillebrand, 2004;Stevens, 1989Stevens, , 1992. This stems from the understanding that these are primary drivers affecting temperature variations on Earth's surface (Aigang et al., 2009) and, in turn, that temperature plays a critical ecological role for animal life (Colinet et al., 2015;Polato et al., 2018). There is an intimate relationship between latitude and elevation in that varying these two factors in the same direction (toward higher elevations and the poles or vice versa) generally leads to analogous changes in thermal conditions. Ultimately, this relationship provides a convincing explanation for the disjunct arctic-alpine distributions of several animals and plants (Muster et al., 2009).
The importance of obtaining a nuanced understanding of species distributions along latitudinal and elevational gradients has gained momentum amid the current climatic emergency (Ripple et al., 2019), given that species are rapidly shifting their distribution poleward and toward higher elevations in response to a global temperature increase (Chen et al., 2011;Lenoir et al., 2020). Furthermore, there is a growing appreciation of the utility of latitude and elevation gradients for advancing ecological theory, as these gradients can be used to uncover some of the mechanisms shaping spatial biodiversity patterns (Sanders & Rahbek, 2012).
General theory predicts that segregation along spatial gradients should be particularly visible in highly dispersive species, such as birds (Auer & King, 2014;Burner et al., 2020;Graham et al., 2009), insofar as the effect of antagonistic interactions require the species to be in contact (that is, some sort of spatial proximity). Conversely, the effect of competition is often found to be of limited importance in driving the distribution of poorly dispersive species, such as subterranean ones (Mammola et al., 2020;Zakšek et al., 2019). Lacking effective dispersal, range size of subterranean organisms is often best predicted by historical factors (Zagmajster et al., 2014) and availability of habitats (Christman & Culver, 2001;Culver et al., 2006). Yet, while this reasoning applies to most specialized subterranean species, often ranging over 30 km or less (Bregović et al., 2019), there exist rare exceptions of cave-dwelling species with ranges stretching over continental distances. Often, these outliers are organisms lacking pronounced adaptations to subterranean life and able to effectively disperse through surface habitats (Mammola, 2019), meanwhile showing a high ecological plasticity in terms of habitat breadth (Ficetola et al., 2020).
An example of such widely distributed cave-dwelling organisms are orb weaver spiders in the genus Meta (Tetragnathidae). In general, Meta spiders are large-sized predators (total body length from 3.8 to 16.0 mm; Hesselberg et al., 2019) inhabiting the entrance area of caves and other habitats with similar ecological conditions such as cellars, railway tunnels, hollow trees, and crevices (Hesselberg et al., 2019;Mammola & Isaia, 2014). Two species of Meta occur in Europe, M. bourneti Simon 1922 andM. menardi (Latreille, 1804), both ranging over an enormous geographic expanse when compared to that of most subterranean spiders Mammola & Isaia, 2017). The explanation for this broad distribution has often been searched for in the unusual life cycle of these spiders (Mammola & Isaia, 2014: fig. 6), involving a phase of airborne dispersal (ballooning) outside caves (Smithers & Smith, 1998, Smithers, 2005a; but see Hesselberg et al., 2019 for cautionary arguments). Being ubiquitous in caves and easily identified in the field, these spiders have become "workhorses" for subterranean arachnological research (Hesselberg et al., 2019).
Many aspects of their biology have been investigated in recent years including their diet (Novak et al., 2010;Smithers, 2005b), activity patterns , habitat preferences (Chiavazzo et al., 2015;Lunghi, 2018;Lunghi et al., 2017;Mammola & Isaia, 2014Manenti et al., 2015;Růžička et al., 2013) Given that the two species are of comparable size and occupy a similar niche within subterranean habitats, they should be in direct competition for space and resources (Mammola & Isaia, 2014). In fact, the coexistence between the two species within the same subterranean habitat has never been proved convincingly (see supporting information in Mammola & Isaia, 2017), suggesting that some kind of competitive exclusion dynamics may be in force.
Brignoli ( Given that latitudinal and elevational gradients are essentially thermal seasonality gradients, these differences between the two species are likely to reflect a shift in their climatic niches (Gasparo & Thaler, 2002). In fact, the adaptation to different climatic conditions was invoked as the main mechanism through which M. bourneti and M. menardi avoid direct competition, as empirically observed for Italian (Mammola & Isaia, 2014) and British populations (Mammola, 2017), as well as through the European range of the two species (Mammola & Isaia, 2017). However, some caution should be exercised in interpreting the pattern reported here, given that the broadscale distribution of the two species may also bear the signature of past biogeographic events and a so far undetected evolutionary component (Warren et al., 2014), and even cryptic diversity within the nominal species. Indeed, recent studies emphasized how there is often a high degree of overlooked diversity in broadly distributed subterranean species (Delić et al., 2017;Eme et al., 2018;Esposito et al., 2015;Hedin, 2015). The lack of phylogenetic data across the distribution of these two species prevented us from formally testing species hypotheses in European Meta, as well as to explore evolutionary hypotheses regarding both their distribution and evolution of microclimatic niche. Coupling ecological information with phylogenetic data would be an interesting future development as new DNA sequences of Meta become available. All in all, the segregation pattern reported can thus be taken as a neat, textbook example of the trade-off between latitude and elevation in determining habitat segregation of broadly distributed species occupying similar niches.

ACK N OWLED G EM ENTS
The authors thank Peter Harvey from the British Arachnological Society for making the dataset from the Spider Recording Scheme available. Stefano Mammola was supported by the CAWEB project "Testing macroecological theory using simplified sys-

AUTH O R CO NTR I B UTI O N
EL, TH, and SM contributed data. SM performed analyses and wrote the first draft. All authors critically contributed to the writing by suggestion and additions to the text.

DATA AVA I L A B I L I T Y S TAT E M E N T
The thinned database used to generate the analysis is available in