Journal of Biogeography (J. Biogeogr.) (2015) ORIGINAL Integration of global fossil and modern ARTICLE biodiversity data reveals dynamism and stasis in ant macroecological patterns Benoit Guenard1,2*, Vincent Perrichot3 and Evan P. Economo1,4 1Okinawa Institute of Science and Technology ABSTRACT Graduate University, Onna-son, Okinawa Aim We investigate the dynamics of ant biodiversity patterns and community 904-0495, Japan, 2School of Biological structure from the Eocene until the present. Our goal is to empirically test Sciences, The University of Hong Kong, Pok Fu Lam Road, Hong Kong SAR, China, hypotheses regarding the similarities in composition and structure of fossil and 3CNRS UMR 6118 Geosciences, Universitede modern ant faunas to understand the historical processes influencing modern Rennes 1, Rennes Cedex 35042, France, global biodiversity patterns. 4 Department of Ecology and Evolutionary Location Global. Biology, University of Michigan, Ann Arbor MI 48104, USA Methods We integrated two recently developed databases of the geographical distributions of fossil and modern ant genera and analysed the evolution of diversity and composition since the Eocene, c. 55 Ma. We assembled a third new database on community structure of ants, and used it to compare community structure of fossil assemblages with modern communities in different bioregions. Results The analyses of generic composition and community structure were congruent, supporting a strong affinity between the Western Palaearctic ant fauna and modern Indomalayan and Australasian assemblages, and of a wide- spread Holarctic ant palaeofauna, and affinity between fossil Caribbean and modern Neotropical faunas. In addition, neither generic composition nor com- munity structure of fossil assemblages showed evidence of taphonomic bias towards arboreal taxa. Main Conclusions The aggregated fossil record reveals the dynamic nature of macroecological patterns in ant biodiversity during the Cenozoic, with conti- nental-scale generic extinctions common and important in shaping modern ant assemblages. Our results suggest major compositional changes for the Western Palaearctic bioregion and the Caribbean bioregions. The ant palaeofauna of the Western Palaearctic bioregion was then very similar in composition and struc- ture to the one observed now in the Indomalayan bioregion, supporting the notion that modern-day ‘tropical’ biomes were historically much more wide- spread, while temperate biotas are more recently evolved. Our results under- score the importance of a palaeogeographical perspective in understanding modern-day macroecological patterns. *Correspondence: Benoit Guenard, School of Keywords Biological Sciences, The University of Hong Cenozoic, community structure, extinctions, Formicidae, fossils, generic Kong, Pok Fu Lam Road, Hong Kong SAR, diversity, global diversity patterns, historical biogeography, neofauna, palaeo- China. E-mail: [email protected] fauna. efforts have been bolstered by new informatic tools allowing INTRODUCTION for aggregations of vast amounts of data. As noted by Biologists have long strived to understand the macroecologi- others (Manchester et al., 2009; Jones & Safi, 2011), the cal patterns such as the latitudinal diversity gradient and large-scale biodiversity patterns we observe today are an the distribution of Earth’s biomes. In recent years, these outcome of a long and dynamic evolutionary history driven ª 2015 John Wiley & Sons Ltd http://wileyonlinelibrary.com/journal/jbi 1 doi:10.1111/jbi.12614 B. Guenard et al. by climatological, geological and ecological shifts. The towards certain habitats or strata. We first review large-scale aggregation and synthesis of fossil data is critical for resolv- ant biogeography to establish some context and hypotheses ing the temporal dimension of biodiversity dynamics and for our analysis. understand the processes driving patterns. While data syntheses have advanced rapidly for vertebrates Background and hypotheses and plants, invertebrates lag behind despite constituting the large majority of the global diversity. Among invertebrates, Recent work on ant evolution including molecular phyloge- ants are one of the most ecologically dominant and well- netic analyses (Brady et al., 2006; Moreau et al., 2006; Mor- studied groups. Understanding the mechanisms and the eau & Bell, 2013), diversification hypotheses (Wilson & pathways by which this group achieved such ecological and Holldobler,€ 2005; Moreau et al., 2006; Perrichot et al., evolutionary success represents a major endeavour. Recently, 2008b), and new description of ancient fossils nearly 100 Ma a global database on the distribution of modern ant genera old (Nel et al., 2004; Perrichot et al., 2008a; Barden & Gri- was compiled revealing large-scale patterns in richness and maldi, 2012), indicate that all the genus-level taxa known composition (Guenard et al., 2012). Work on fossil ants has from the first 50–60 Myr following their first appearance are mainly focused on taxonomic analysis of individual speci- now extinct. The rise of the modern ant fauna is a more mens (e.g. Baroni Urbani, 1980; Perrichot et al., 2008a; recent phenomenon occurring probably during the early Dlussky & Wedmann, 2012). Only a few synthetic analyses Cenozoic and continuing until the present (Brady et al., exist on particular deposits with a rich ant palaeocommunity, 2006; Moreau et al., 2006; LaPolla et al., 2013). such as Baltic and Dominican ambers (Wheeler, 1915; Wil- son, 1985; Dlussky & Rasnitsyn, 2009; Dlussky & Putyatina, The Fossil West Palaearctic–Modern Indomalayan- 2014). Recently, a world catalogue of fossil ants was com- Australasian Affinity piled (Perrichot, 2014), and the updated global fossil record was reviewed (LaPolla et al., 2013). However, our macro- One of the pioneering -and still most impressive- studies on scopic view of ant biodiversity, across all genera and across fossil ants to-date was conducted by Wheeler (1915) on 9527 the globe, remains firmly rooted in the present. ant specimens from Eocene (37–42 Ma) Baltic amber. Here, for the first time, we assemble a comprehensive Wheeler noted then that modern American and African ele- global database of ant genera known both from fossil and ments are nearly completely missing from Baltic amber, and modern species, providing relevant historical context to the he described this ant palaeofauna as a ‘mixture of what at patterns observed in extant generic distributions. We evaluate the present day we are able to recognize as at least four dif- the extent to which historical assemblages of genera resemble ferent faunas, the Palaearctic, the Indian, the Malayan and modern faunas, both in the same and in distant geographical the Australian, with a little more than ⅓ of the genera and regions. We use these data on generic composition to evalu- nearly ½ of the species Palaearctic and the reminder belong- ate biogeographical hypotheses developed to explain modern ing to Indomalayan and Australian types’ (Wheeler, 1915, p. ant composition patterns. 12). Wheeler’s observations and conclusions followed those While generic composition varies in informative ways of Emery (1892), who previously noticed the strong affinity across major bioregions, the distribution of diversity at of much smaller sample of Sicilian amber ant fossils with the higher taxonomic levels also varies in important ways that modern Indo-Australian fauna. However, these authors ver- are not captured by the presence–absence of genera. For bally described patterns but did not test them quantitatively. example, compared with faunas in other continents, both Furthermore, as those original studies, subsequent work has local and regional ant faunas of Australia are much more enriched our understanding of both the fossil and modern dominated by species of the subfamily Dolichoderinae than ant faunal distributions, with studies on new deposits avail- other regions. These broad differences in community organi- able (Europe: Dlussky & Rasnitsyn, 2009; Dlussky & Wed- zation also help us interpret the organization of historical mann, 2012; Dominican amber: Wilson, 1985; Baroni ant assemblages. Likewise fossil records may be biased Urbani, 1995; Chiapas: Solorzano Kraemer, 2007), and towards ants living in certain habitats (forest, open, etc.) and improvements in taxonomy of both modern and fossil taxa in different ecological strata (arboreal, epigaeic, hypogaeic) (Ward, 2007). The aggregation of the most recent data for as noted for instance by Dlussky et al. (2009) on the domi- both the fossil and modern ant distributions (Guenard et al., nance of arboreal ant individuals in the Oligocene Bembridge 2010; Perrichot, 2014) allow now for more rigorous compar- Marls deposits. ison of the affinities between those different assemblages. To compare community organization of fossil and modern assemblages, we compiled a global database of modern ant The East Palaearctic–East Nearctic Disjunction community structure from all major biomes and habitat types, as characterized by the distribution of species richness For over 150 years, plant ecologists have noticed strong simi- across subfamilies. Using this data set, we evaluated whether larities between plant communities found in East Asia and the organization of fossil communities resembles any mod- East North America, both sharing similar genera, while those ern-day faunas, and if there is evidence of a taphonomic bias were absent from Europe or West America (Wen, 1999). 2 Journal of Biogeography ª 2015 John Wiley &