110000 years of Quaternary beetle diversity change

Biodiversity and Conservation 12: 2077–2089, 2003.  2003 Kluwer Academic Publishers. Printed in the Netherlands. 110 000 years of Quaternary beetle diversity change P. PONEL 1, *, J. ORGEAS 1 , M.J. SAMWAYS 2 , V. ANDRIEU-PONEL 1 , J.-L. DE BEAULIEU 1 , M. REILLE 1 , P. ROCHE 1 and T. TATONI 1 1 ´ ˆ , Institut Mediterraneen ´ ´ d’ Ecologie et de Faculte´ des Sciences et Techniques de St Jerome ´ ´ , case 451, UMR CNRS 6116, F-13397 Marseille cedex 20, France; 2 Invertebrate Paleoecologie Conservation Research Centre, School of Botany and Zoology, University of Natal, South Africa; * Author for correspondence (e-mail: philippe.ponel@ univ.u-3 mrs.fr; fax: 133 -4 -91 -288668) Received 11 February 2002; accepted in revised form 30 October 2002 Key words: Climate change, Climate–beetle–vegetation relationships, Coleoptera, Fossil beetle diversity, Quaternary Abstract. Our first aim was to document the effects of palaeotemperatures and vegetation changes on beetle assemblages, and secondly to determine the extent to which surrogacy analysis at the family taxonomic level reveals patterns evident from lower taxa analysis. The sedimentary sequence sampled on the experimental site of ‘La Grande Pile’ (Vosges, France) covers the whole of the last climatic cycle. Beetle fragments were extracted from 39 coring samples and identified to the lowest possible taxonomic level. A total of 3092 beetle specimens belonging to 394 taxa were identified, more than half to species level. Carabidae, Staphylinidae and Curculionidae families together represented 40% of the overall taxa richness. Beetle taxa richness and assemblage composition varied markedly over time. Average summer temperatures clearly play a major role in diversity patterns, as temperature was positively correlated with taxon richness. Nevertheless, the warmest and the coldest periods were not the richest and the poorest, respectively, and the most humid period did not correspond to maximum beetle richness. Beetle assemblages are likely to fluctuate in response to other factors such as plant diversity and vegetation structure. Steppe-like vegetation did not reduce species richness while dense, homogenous and closed forests did. Family patterns mirrored those observed at the lower taxa level. This makes the family level a convincing alternative to lower taxonomic level analyses by representing a faithful picture of changing beetle diversity over a long period of time. Finally, evolution of beetle diversity over the Quaternary represents a convincing model for evaluating the effect of close and wide past climate changes, and for assisting in management of present-day biodiversity as part of the current anthropogenic global climate change. Introduction Studies on changes in patterns of biodiversity over geological time to the present day have tended to focus either on the very long and ancient periods or the very short and recent period. Changes in biodiversity over a geological time scale, and usually concerning extinct species, are referred to as evolutionary or deep time (Rosenzweig 1995). Such studies are open to conjecture, as dead organism preservation through time is strongly affected by taphonomical and fossilization factors (Wilson 1988). Conversely, patterns of present-day or very recent biodiversity, particularly those concerning living species, are referred to as ecological time. Following on from that, recent biodiversity studies then consider either different 2078 spatial scales or community dynamics in various ecosystems (Rosenzweig 1995). Against this background, very little is known about biodiversity patterns for Quaternary time, which Bennett (1990) considered as an intermediate time scale between the evolutionary time and ecological time. Such a paucity of information has tended to obscure the relative importance of such a time scale, particularly as this historical perspective helps us to understand the organization processes of present-day biodiversity (Birks and Birks 1980; Barbault 1995). During the Quaternary, astronomical cycles generated alternation of glacial and interglacial periods that were repeated every 100 000 years. Each glacial period induced a north–south migration of vegetation belts over several thousands of kilometres (Bell and Walker 1992), and a complete reorganization of plant communities (Reille 1990). Impact of climate cycles also drastically affected animal communities (Lowe and Walker 1997), and among these were beetles for which there is a good record (Elias 1994). Glacial–interglacial alternations greatly affected the geographical distributions of beetles (Coope 1994). Among these range shifts were a southward migration of arctic and eastern-Asiatic species during glacial periods and a northward migration of Mediterranean and south-European species during interglacial periods and temperate transitions (Coope 1990; Buckland and Coope 1991). The climatic oscillations inducing population mixing are likely to preclude evolutionary changes within the beetle assemblage, because gene fluxes are likely to homogenize the populations (Coope 1978; Fatuyama 1987). This has meant that evolution, disappearance or appearance of species during the Upper Pleistocene, has been extremely rare in central Europe (Coope et al. 1961; Coope 1978). It has been suggested that phylogenetic stability is likely to induce physiological stability (Buckland and Coope 1991), and that ecological requirements of most species are likely to be stable over several hundreds of thousands of years. This means that we can reasonably postulate that there is some similarity between ecological requirements of Quaternary species and those of their present-day counterparts. The study of the ‘nearly fossil Coleoptera’ (Coope 1991) can therefore provide a key link between palaeontology and neontology, which need considerable reconciliation (Bell 2000). Beetles better lend themselves to Quaternary studies than many other animals, as they have a hard tegument that lasts long in the anoxic conditions of sediments. Also, they are often abundant in soft sediments, and they can be identified to species level in most cases (Elias 1994). Beetle analyses at family level may represent an interesting complement to that at the species level. Family may be taken as an alternative to reveal the ‘big picture’ of the changes in diversity patterns over a long period of time. The site of ‘La Grande Pile’ (Vosges, France) allowed us to extract a 20-m-deep continuous sedimentary profile, ranging from the end of the penultimate glacial period to the end of the last glacial period. This sequence covers almost the whole of the last climatic cycle. This site has been a reference for many detailed pollen analyses that have enabled reconstruction of the vegetation pattern succession (Woillard 1975, 1978; Woillard and Moock 1982; de Beaulieu and Reille 1992), as well as the palaeotemperatures (Ponel 1995a). 2079 This paper, as a development of a previous palaeoentomological study (Ponel 1995a), analyses the impact of alternating glacial–interglacial stages and consecutive vegetation changes on beetle assemblages during the penultimate climatic cycle. This is the first time that insect biodiversity patterns have been analysed over a continuous period of 110 000 years, from 130 000 to 20 000 before present (BP). A sound understanding of the past consequences of palaeoclimate changes on biodiversity can assist in management of present-day biodiversity, given the current anthropogenic global climate change (Coope 1994). Accordingly, our specific aims here are to (i) document the effects of palaeotemperatures and consecutive vegetation changes on beetle species richness, (ii) examine changes in beetle assemblages, and (iii) determine the extent to which surrogacy analysis at the family taxonomic level reveals patterns evident from species-level analysis. Material and methods Study site The study site (478449 N, 68309140 E) has been described in detail elsewhere (Woillard 1975; de Beaulieu and Reille 1992; Ponel 1995a), so only a very brief description of the area is given here. Located on the Vosges massif piedmont on an interfluvial plateau at 325 m altitude (Ponel 1995a), the ‘Grande Pile’ depression ¨ remained untouched by the extension of the Wurmian ice sheets. The chronostratigraphy was established by Woillard and Moock (1982), then reviewed by de Beaulieu and Reille (1992). The surrounding forest is densely covered by oaks. The central peat bog is covered by Molinia coerulea, Drosera rotundifolia, Vaccinium microcarpum, Eriophorum vaginatum, Menyanthes trifoliata, Potentilla palustris, various Carex spp. and many Sphagnum spp. The bog is currently recolonized by forest tree species such as Betula, Populus, Frangula alnus, and Salix spp., as well as Quercus robur (Woillard 1975). Coring technique and sample treatment The sedimentary sequence yielded 39 samples of 8–9 kg each. After disaggregating sample sediment in water, the insect fragments were concentrated by kerosene flotation (Coope 1986). The identification of fragments was made by direct comparison with a reference collection of extant species. Counting beetle fragments (number of heads, thorax, left and right elytra, etc.) allowed the minimal number of entire specimens to be estimated in each assemblage. Beetle fragments are particularly resistant through time and well preserved in sediments because of their hard chitinous teguments. However, taphonomical processes may induce underestimation of some taxa, and may introduce a bias in fragment or specimen counting. Big species are more subject to fragmentation than smaller ones, and are consequently highly underestimated. Fragment identification is often extremely difficult, particularly as modern identification keys are designed for complete specimens. The 2080 reliability of identification is thus related to the quality of fragment preservation, but nevertheless it is often possible to identify fragments to species or genus level. In the present paper, ‘taxa’ will be referred to the lowest possible identification level (e.g. species, group of species, genus, even sub-family). The ‘satisfactory’ average level of identification to species level may reasonably be considered to be 50% of the overall taxa number. Despite such considerable difficulty, this method is still the only one reliable enough for assessing response of Quaternary beetle assemblages to environmental changes. Climatic data and vegetation history Climatic data were obtained using the ‘mutual climatic range’ (MCR) method (Atkinson et al. 1986) that allows reliable palaeotemperatures to be reconstructed from beetle assemblages (Ponel 1994, 1995a). This method computes current climatic data of the area occupied by taxa of each near-fossil assemblage, as it is assumed that a given species today has the same climatic requirements as it had during the Quaternary (Coope 1978). Moreover, the MCR method depicts the gross thermal climate in which suitable microclimates will occur (Coope and Elias 2000). Non-phytophagous beetles were exclusively considered for climatic reconstruction, as their distribution is relatively independent of vegetation structure and composition (Magagula and Samways 2001) and is determined principally by prey availability and climatic conditions (Samways et al. 1999). In Figure 1 the palaeotemperature curve is based on the median values calculated from the range of estimated summer temperatures drawn from Ponel’s original data (1995a). Vegetation history and major botanical events have been reconstructed from pollen analyses (de Beaulieu and Reille 1992), and may be summarized as follows: • The end of Riss (more than 130 000 years BP, oceanic isotopic stage (OIS) 6) was a glacial period with open landscapes of Poaceae and heliophytes such as Artemisia, open pine stands, and isolated Betula, Salix or Hippophae trees. The vegetational community was comparable to a modern-day steppe. • The Eemian period (130 000–110 000 years BP, OIS 5e) has been divided in two sub-periods. (i) An anathermic period (Eemian I), warmer than today by about 2 8C (Cheddadi et al. 1998; Rioual et al. 2001), characterized at the beginning by the settlement of a pioneer vegetation with Juniperus and Betula, and then followed by a boreal forest with Pinus and Betula. This forest was then replaced by a deciduous forest with Ulmus, Quercus, Corylus, Taxus and Carpinus. (ii) A catathermic period (Eemian II) that corresponds to declining temperatures. This sub-period is characterized by the regional disappearance of mixed deciduous forests and by the expansion of coniferous forests with first Abies and then Picea. The end of the interglacial is marked by the settlement of an open boreal forest with Pinus and Betula. The presence during the Abies phase of the beetle Platypus oxyurus, an insect living in southern Europe (south of France, Corsica, Calabria, Greece), indicates that this period was relatively warm and much wetter than today (Ponel, unpublished data). ´ • Melisey I (110 000–104 000 years BP, OIS 5d) corresponds to a very cold 2081 Figure 1. Average beetle richness per kilogram of sediment (histogram) and MCR-reconstructed average temperatures (fourth order polynomial fitted curve). The time frame represents the approximate dates that have been constructed from various dating methods (de Beaulieu and Reille 1992; Ponel 1995a). • • • • episode characterized at region scale by a tundra with Poaceae, steppe species and Betula. St Germain I (104 000–93 000 years BP, OIS 5c) was a temperate period characterized by the settlement of a boreal forest with Juniperus, Betula and Pinus. Unlike the Eemian, Picea is present from the beginning of the interstadial, indicating colder climatic conditions. This stage was followed by the settlement of a Quercus forest and by a short climatic regression marked by a transitory expansion of boreal taxa and the disappearance of mesophilous trees. Lastly, Quercus and Corylus settled again, associated first to Carpinus and then to Pinus and Picea at the end of the St Germain I. ´ Melisey II (93 000–84 000 years BP, OIS 5b) was a glacial phase with climate ´ and vegetation comparable to Melisey I. St Germain II (84 000–72 000 years BP, OIS 5a) was a temperate period dominated first by a boreal forest with Betula, followed by a mixed deciduous forest with Quercus and Carpinus, and then by a Picea, Betula and Pinus forest. Unlike St Germain I, no climatic reversal occurred. ¨ Wurmian Pleniglacial (72 000–15 000 years BP, OIS 4, 3, 2) was a glacial, cold and arid period, when steppe vegetation with an abundance of Poaceae, Artemisia, and heliophytes developed, and when the only recorded arboreal taxa were Juniperus, Betula and Pinus. The cold maximum of this glacial period was recorded at the end of the Pleniglacial, around 18 000 years BP. 2082 Data analysis As samples varied greatly in weight, bias in sampling effort was overcome by converting ‘taxa’ (which includes different levels of taxonomy, i.e. species, genus, etc.) richness and total abundance of beetles into numbers per kg of sediment. Taxon richness is an approximate measure of actual species richness. However, it does not exactly mirror species richness, as many beetle fragments can only be identified to higher taxonomic levels. Linear regressions were used to determine the effect of palaeotemperatures on average taxa richness and abundance (per kg of sediment). Unlike traditional palaeoentomological studies, beetle assemblages were documented with Correspondance Analyses (COA, ADE4 software; Chessel 1995) performed on average abundances at taxa and family levels. The value of Correspondence Analysis in analyses of fossil beetle assemblages was previously stressed by Cong and Ashworth (1997), but this method is used here for the first time on a long sedimentary sequence. Testing reliability of the family level as a surrogate of taxa level was carried out on straight abundances using correlation coefficient and Mantel test (ADE4 software; Chessel 1995) for species composition. The Mantel test creates, for both taxa and family levels, two association matrices (Bray–Curtis distances) and compares them by testing permutation probabilities (2000 permutations). Results The fauna A total of 3092 beetle specimens belonging to 394 taxa were identified, more than half of them to species level. The fauna was dominated by three families: Carabidae (54 taxa), Staphylinidae (53 taxa) and Curculionidae (51 taxa), which together represented 40% of the overall taxon richness (Ponel 1995a). Beetle richness varied markedly along the sequence (Figure 1). Five peaks of higher taxon richness were recorded. The first one occurred during the Riss–Eemian transition, the second during the second phase of Eemian, the third during St ¨ Germain I, the fourth during the transition of St Germain II – Wurmian Pleniglacial, and the last at the beginning of the upper Pleniglacial (25 000–20 000 years BP). Abundance patterns mirrored those of taxon richness, as there was a high correlation coefficient between number of taxa and abundance of beetles in each sample (r 5 0.707; P , 0.0005; n 5 39). Near species and family richness were highly correlated (r 5 0.82; n 5 39; P , 0.0005), such that here family richness was a reasonable surrogate of near species richness. 2083 Beetle response to climate Taxon richness Average temperatures, calculated from summer temperature amplitudes of the MCR approach, were positively correlated to taxon richness (r 5 0.33; P 5 0.043; n 5 39). Surprisingly, the warmest period (Eemian I) was characterized by a very low taxon richness (Figure 1), comparable to that observed for the coldest period ¨ (end of the Wurmian Pleniglacial). Relative proportions of taxon number from different habitat types changed over time (Figure 2), with ground-active and standing-water beetle richness increasing ¨ along the sequence, as climate cooled. The very cold period (end of the Wurmian Pleniglacial) was characterized almost exclusively by taxa from these two habitats. Conversely, rheophilous taxa and those of the other habitats almost disappeared at this time. Tree-associated taxa (both deciduous and conifer) occurred mostly during the first half of the sequence, when the climate was distinctly warmer. Beetle assemblage composition Ordination of taxa-level abundance (Figure 3) reveals a marked gradient of palaeotemperatures running along the first axis (eigenvalue 0.75). Beetle assemblage progressively changed as very cold-adapted periglacial conditions gave way Figure 2. Relative taxon richness for habitat types of beetles. The different habitats are: ground (ground), paludal (palud), rheophilous (rheoph), tree crown (tree), deciduous forest (decid), coniferous forest (conif), and grassland (herb). The group ‘other’ includes coprophagous, necrophagous, nonspecialist and undetermined species. 2084 Figure 3. COA ordination based on taxa-level abundances. Samples from the sequence are represented on the scatterplot. Groups are made according to information provided from pollen data. to those where warm deciduous or temperate mixed forests dominated. Axis 2 (eigenvalue 0.5) separated samples corresponding to deciduous forests from those of coniferous or mixed (beech–fir) forests. Changes in beetle assemblages at family level closely mirrored those at taxa level (Mantel test, P , 0.001). As many taxa were recorded (n 5 394), the scatterplot (Figure 4) was worth reinterpreting by replacing each taxon by the habitat type to which it belonged. Consistently, taxa related to deciduous vegetation were recorded in sites where deciduous and mixed forests occurred. Conversely, and surprisingly, coniferrelated taxa were present in warm or temperate mixed forests and were absent from cold conifer forests. Ground-active and standing-water taxa mainly occurred in conifer and mixed forests, whereas they were absent from deciduous vegetation, and were widely spread across the temperature gradient. Taxa associated with tree crowns occurred in warm and temperate forests, whether deciduous or mixed. The other habitats did not show any clear trends, with the taxa being widely dispersed on the scatterplot. Discussion Palaeoclimates, palaeoenvironments and beetle diversity Beetle species richness varied markedly over time, with the succession of warm, 2085 Figure 4. COA ordination based on taxa-level abundances. Taxa are labelled on the scatterplot by the habitat type to which they belong. transitional and cold periods. The climatic conditions favouring particular levels of beetle species richness are open to conjecture. However, average summer temperatures clearly play a major role in diversity patterns, as temperature was positively correlated with taxon richness. Given that the various species have 2086 specific optimal temperatures, climatic changes are likely to cause beetle species richness and assemblage patterns to vary considerably. Surprisingly, despite positive correlations between temperatures and beetle taxon richness, the warmest period (i.e. Eemian I, for which temperatures are believed to be higher than today) and major climatic transitions did not have the richest fauna and had in some cases even among the lowest numbers. This finding supports that of Walkling and Coope ¨ (1996) at Grobern, Germany, on beetle assemblages spanning the Eemian / Early Weichselian transition. Conversely, very cold periods were not necessarily unfavourable to high species richness, as cold-resistant taxa replace warm-climate ones. This suggests that past biodiversity patterns do not respond exclusively to temperature or changes in temperatures, as may also be the case today (Samways et al. 1999). The results here suggest that beetle assemblages in the past have been influenced by factors such as decreasing plant diversity and grass cover in dense forests. Furthermore, there may also be uncontrolled bias in the results. Assessing driving forces of beetle assemblage patterns is indeed subject to both reliability of identification and the representativeness of ancient beetle assemblages compared with the past beetle communities truly living at that time (Kenward 1975, 1976). Some of the specimens collected in each sample have come from a remote source and therefore may not belong to the local, resident entomofauna. The appearance of tourist elements is believed to be particularly important at sites subject to strong winds or water currents (Ponel and Richoux 1997). No information is available here to assess the extent to which this situation may be clouding the overall ecological picture given here. The appearance of steppe-like vegetation did not reduce species richness. Even today, open vegetation is usually inhabited by a high number of species (Ponel 1995b). However, during warm periods of the first half of the sequence, the land was covered by very dense, homogenous and closed forests (de Beaulieu and Reille (1992) for the Grande Pile site; Pons et al. (1992) for other Eemian sequences at various sites in France) that diminished beetle species richness, as the herbaceous understorey cover became reduced. This hypothesis is supported by the recurrent drops in richness at times of deciduous or mixed forests. Accordingly, vegetation structure and dynamics play a role in beetle assemblage patterning in addition to the simple influence of temperature. Nevertheless, very low temperatures (6–11 8C maximal summer temperatures) represent the main driving force ¨ for beetle species richness, especially during the coldest period of the late Wurm. Vegetation structure at this time appeared to be of lesser significance than these very low temperatures. Interestingly, the most humid period (Eemian II) did not correspond to maximum beetle species richness. This again suggests a complex interrelationship between environmental condition, vegetation type and its associated beetle assemblages. Beetle assemblages Beetle ecological guilds changed greatly in response to changes in palaeotempera- 2087 tures, with the changes occurring progressively along this gradient. General climate cooling caused rheophilous and tree-associated taxa to become less numerous. This is probably because of the cold, which caused the change from arboreal forests to tundra-like vegetation, and water environments from flowing to stagnant or frozen. Such vegetation changes have been validated by pollen analyses (de Beaulieu and Reille 1992). Quaternary beetle assemblages are therefore reliable indirect indicators of past environmental changes (Coope 1975). Beetle assemblages of steppe-type vegetation appeared to be the only outside functional group to be associated with cold evergreen forests. This may be explained by the coexistence in mosaics of conifer forests and steppes in the cold times. Periods that were supposedly inhabited by conifers did not necessarily support only conifer-inhabiting species. Indeed, conifer beetle species were mainly recorded in mixed forests. This incongruence may be the result of very cold periods where conifers supported only an impoverished tree-crown fauna, with the likelihood of only very few specimens and species being trapped in sediments. Family as a surrogate of species The response of beetle taxa richness to temperature was mirrored by that of family richness. Similarly, multivariate analyses at the family level revealed compositional changes that were evident at the taxon level. Family-level analyses therefore represent a convincing alternative to lower taxonomic level analyses, at least for interpreting beetle assemblage response to past climatic changes. This coarse taxonomic scale matches the relatively coarse temporal scale. However, it is well known that individual beetle species have specific habitat requirements that also reflect their rarity at one place and time (Eyre et al. 2001). So while family-level analyses may present a faithful picture of changing beetle assemblages over long periods of time, specific investigatory questions may require greater taxonomic resolution. References Atkinson T.C., Briffa K.R., Coope G.R., Joachim M.J. and Perry D.W. 1986. Climatic calibration of coleopteran data. In: Berglund B.E. (ed.), Handbook of Holocene Palaeoecology and Palaeohydrology. Wiley, Chichester, UK, pp. 851–858. ´ ´ ´ ` Masson, Paris. Barbault R. 1995. Ecologie generale, Structure et fonctionnement de la biosphere. de Beaulieu J.L. and Reille M. 1992. The last climatic cycle at La Grande Pile (Vosges, France): a new pollen profile. Quaternary Science Review 11: 431–438. Bell M.A. 2000. Bridging the gap between population biology and paleobiology. Evolution 54: 1457–1461. Bell M. and Walker M.J.C. 1992. Late Quaternary Environmental Change: Physical and Human Perspectives. Longman, London. Bennett K.D. 1990. Milankovitch cycles and their effects on species in ecological and evolutionary time. Paleobiology 16: 11–21. Birks H.J.B. and Birks H.H. 1980. Quaternary Palaeoecology. Edward Arnold, London. 2088 Buckland P.C. and Coope G.R. 1991. A Bibliography and Literature Review of Quaternary Entomology. JR Collis Publications, University of Sheffield, Sheffield, UK. Cheddadi R., Mamakova K., Guiot J., de Beaulieu J.L., Reille M., Andrieu V. et al. 1998. Was the climate of the Eemian stable? A quantitative climate reconstruction from seven European pollen records. Palaeogeography, Palaeoclimatology, Palaeoecology 143: 73–85. ` Chessel D. 1995. ADE-4, Ordination sous contrainte. Institut d’analyse des Systemes Biologiques et ´ Socio-economiques, Universite´ de Lyon 1, Lyon, France. Cong S. and Ashworth A.C. 1997. The use of correspondence analysis in the analysis of fossil beetle assemblages. In: Ashworth A.C., Buckland P.C. and Sadler J.P. (eds), Studies in Quaternary Entomology – An Inordinate Fondness for Insects. Quaternary Proceedings No. 5. Wiley, Chichester, UK, pp. 79–82. Coope G.R. 1975. Climatic fluctuations in northwest Europe since the last Interglacial, indicated by fossil assemblages of Coleoptera. In: Wright A.E. and Moseley F. (eds), Ice Ages: Ancient and Modern. Seel House Press, Liverpool Geological Journal Special Issue 6, pp. 153–168. Coope G.R. 1978. Constancy of insect species versus inconstancy of Quaternary environments. In: Mound L.A. and Waloff N. (eds), Diversity of Insect Faunas (Symposia of the Royal Entomological Society of London 9). Blackwell, Oxford, UK, pp. 176–187. Coope G.R. 1986. Coleoptera analysis. In: Berglund B.E. (ed.), Handbook of Holocene Palaeoecology and Palaeohydrology. Wiley, Chichester, UK, pp. 703–713. Coope G.R. 1990. The invasion of Northern Europe during the Pleistocene by Mediterranean species of Coleoptera. In: di Castri F., Hansen A.J. and Debussche M. (eds), Biological Invasions in Europe and the Mediterranean Basin. Kluwer, Dordrecht, The Netherlands, pp. 203–215. Coope G.R. 1991. The study of the ‘nearly fossil’. Antenna 15: 158–163. Coope G.R. 1994. The response of insect faunas to glacial–interglacial climatic fluctuations. Philosophical Transactions of the Royal Society of London B 344: 19–26. Coope G.R. and Elias S.A. 2000. The environment of Upper Palaeolithic (Magdalenian and Azilian) ˆ ´ hunters at Hauterive-Champreveyres, Neuchatel, Switzerland, interpreted from Coleopteran remains. Journal of Quaternary Science 15: 157–175. Coope G.R., Shotton F.W. and Strachan I. 1961. A Late Pleistocene fauna and flora from Upton Warren, Worcestershire. Philosophical Transactions of the Royal Society of London B 244: 379–421. Elias S.A. 1994. Quaternary Insects and Their Environments. Smithsonian Institution Press, Washington, DC. Eyre M.D., Lott D.A. and Luff M.L. 2001. The rove beetles (Coleoptera, Staphylinidae) of exposed riverine sediments in Scotland and northern England: habitat classification and conservation aspects. Journal of Insect Conservation 5: 173–186. Fatuyama D.J. 1987. On the role of species in anagenesis. The American Naturalist 130: 465–473. Kenward H. 1975. Pitfalls in the environmental interpretation of insect death assemblages. Journal of Archaeological Science 2: 85–94. Kenward H. 1976. Reconstructing ancient ecological conditions from insect remains: some problems and an experimental approach. Ecological Entomology 1: 7–17. Lowe J.J. and Walker M.J.C. 1997. Reconstructing Quaternary Environments. Longman, London. Magagula C.N. and Samways M.J. 2001. Maintenance of ladybeetle diversity across a heterogenous African agricultural / savanna land mosaic. Biodiversity and Conservation 10: 209–222. ´ ¨ ´ Ponel P. 1994. Les fluctuations climatiques au Pleniglaciaire wurmien deduites des assemblages ˆ ´ d’Arthropodes fossiles a` La Grande Pile (Haute-Saone, France). Compte-Rendus de l’Academie des Sciences, Paris 319: 845–852. ¨ Ponel P. 1995a. Rissian, Eemian and Wurmian Coleoptera assemblages from La Grande Pile (Vosges, France). Palaeogeography Palaeoclimatology Palaeoecology 114: 1–41. ´ Ponel P. 1995b. Aspects de la biodiversite´ entomologique des contreforts prealpins et des Plans de Canjuers (Var) [Coleoptera]. Faune de Provence 16: 39–50. ´ d’interpretation ´ ´ ` Ponel P. and Richoux P. 1997. Difficultes des assemblages de Coleopteres fossiles quaternaires en milieu d’altitude. Geobios MS 21: 213–219. Pons A., Guiot J., de Beaulieu J.L. and Reille M. 1992. Recent contribution to the climatology of the 2089 last glacial–interglacial cycle based on French pollen sequences. Quaternary Sciences Review 11: 439–448. Reille M. 1990. Leçons de palynologie et d’analyse pollinique. CNRS ed., Paris. Rioual P., Andrieu-Ponel V., Battarbee R.W., de Beaulieu J.L., Cheddadi R., Reille M. et al. 2001. High-resolution record of climate stability in France during the last Interglacial. Nature 413: 293–296. Rosenzweig M.L. 1995. Species Diversity in Space and Time. Cambridge University Press, Cambridge, UK. Samways M.J., Osborn R., Hastings H. and Hattingh V. 1999. Global climatic change and accuracy of prediction of species’ geographical ranges: establishment success of introduced ladybirds (Coccinellidae, Chilocorus spp.) worldwide. Journal of Biogeography 26: 795–812. Walkling A.P. and Coope G.R. 1996. Climatic reconstructions from the Eemian / Early Weichselian ¨ transition in Central Europe based on the coleopteran record from Grobern, Germany. Boreas 25: 145–159. Wilson M.V.H. 1988. Taphonomic processes: information loss and information gain. Geoscience Canada 15: 131–148. ´ ` dans l’Est de la Belgique et des Vosges Woillard G. 1975. Recherches palynologiques sur le Pleistocene lorraines. Travaux du Laboratoire de Palynologie et Phytosociologie, Universite´ Catholique de Louvain. Woillard G. 1978. Grande Pile peat bog: a continuous pollen record for the last 140 000 years. Quaternary Research 9: 1–21. Woillard G. and Moock W. 1982. Carbon dates at Grande Pile: correlation of land and sea chronologies. Science 215: 159–161.