Schmitt et al. Frontiers in Zoology (2021) 18:34 https://doi.org/10.1186/s12983-021-00418-9 RESEARCH Open Access Biogeography of Italy revisited: genetic lineages confirm major phylogeographic patterns and a pre-Pleistocene origin of its biota Thomas Schmitt1,2,3*, Uwe Fritz4, Massimo Delfino5,6, Werner Ulrich7 and Jan Christian Habel8 Abstract Warm-adapted species survived the cold stages of the past glacial-interglacial cycles in southern European peninsulas and recolonized major parts of Central and Northern Europe in the wake of postglacial warming. However, many of the genetic lineages which differentiated within these refugia predate even the Pleistocene. One of these refugia is the Italian Peninsula with its surrounding islands. In this meta-analysis, we compiled phylogeographic patterns of multiple species across this region. We transformed existing geographic information on 78 animal and plant species (with a total of 471 genetic lineages) within 17 predefined areas into presence/ absence matrices. We elaborated three matrices: (i) only old lineages exclusively distinguished by deep splits, (ii) only young lineages distinguished by shallow (i.e. recent) splits, and (iii) presence/absence of the respective species. To infer biogeographic relationships between the predefined areas, we performed bootstrapped neighbour joining cluster analyses on these three matrices. In addition, we reviewed the geological history of Italy to identify causes of the observed biogeographic patterns. We found Sardinia and Corsica to be biogeographically closely linked with each other, and that they diverge strongly from all other regions. Sicily also diverges strongly from all other regions, while the intra-island differentiation was comparatively low. On the Italian mainland, Calabria exhibited the most pronounced biogeographic differentiation, often with several lineages present, resulting from old vicariance events within the region. Furthermore, southern Apulia and the Po Plain with adjoining areas of northern peninsular Italy displayed considerable distinctiveness. Admixture prevailed in the areas between these three regions. The ancient isolation of Sicily, as well as Sardinia plus Corsica, resulted in endemic lineages with only moderate recent exchange with adjacent mainland regions. Pronounced diversification occurs within the Italian Peninsula. The complex tectonic activities, such as shifting (micro)plates, submergence of major parts of peninsular Italy with the genesis of numerous Pliocene islands, in combination with the climatic cycles during the Pleistocene have most likely generated the current biogeographic pattern of species. Immigrations from the Balkan Peninsula into northern Italy partly accounted for the distinctiveness of this region. * Correspondence: thomas.schmitt@senckenberg.de 1 Senckenberg Deutsches Entomologisches Institut, 15374 Müncheberg, Germany 2 Department of Zoology, Institute of Biology, Faculty of Natural Sciences I, Martin Luther University Halle-Wittenberg, 06099 Halle (Saale), Germany Full list of author information is available at the end of the article © The Author(s). 2021 Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated in a credit line to the data. Schmitt et al. Frontiers in Zoology (2021) 18:34 Page 2 of 13 Keywords: Pliocene, Quaternary cycles, Marine transgressions, Plate tectonics, Topography, Barrier, Refugia, Genetic structures, Phylogeography Introduction The geological and geographic history of Italy The Mediterranean Basin is a renowned biodiversity hot- Due to the northwards drift of the African plate and the spot, in which taxa evolved and survived the Pleistocene shrinking of the Tethys Sea, Italy has undergone quite cold phases [1–3]. However, many species have a much complex geological transformations during the Tertiary longer history of permanency and evolution in this re- [10]. In this process, the succession and intertwining of gion, reaching far into the Neogene [4]. Within the several geological phenomena was involved. Among Mediterranean region, Italy has long been acknowledged others, these include the diachronic genesis of the Alps to be an independent Mediterranean sub-centre of dif- and Apennines, various eustatic events which repeatedly ferentiation and glacial persistence, which was later con- caused significant shifts of the coastline, and the rotation firmed by molecular analyses highlighting the Italian of blocks that ultimately led to the genesis of Corsica, Peninsula as one of the major southern European refugia Sardinia, as well as part of Sicily and Calabria. The bio- [5–7]. However, when addressing the fine-scale biogeo- geographically most relevant geological events are there- graphic substructures of the Mediterranean refugia, the fore briefly outlined. geographically larger sub-centres (such as the Maghreb, The Alpine orogeny is an effect of the closure of the Iberia, the Balkan Peninsula and Anatolia) showed pro- Tethys Sea after the collision of Adria (a northern Afri- nounced sub-structuring, the so-called refugia-within- can peninsula) and Europe. It started already in the Mid- refugia [8]. These structures are particularly pronounced dle Cretaceous and is still on-going [11, 12]. It is likely in the southernmost parts of the larger sub-centres, sup- that the Alps had at least partly emerged already in the porting the idea that these were not homogeneous refu- Cretaceous, but the collision and hence their significant gial areas but further sub-structured into an, often large, elevation occurred much later, during the Eocene. It has number of geographically small subunits in which dis- been estimated that from the Oligocene until today, the tinct lineages evolved in geographic separation [8]. How- Alps were uplifted by about 1 mm/year [12], but Noc- ever, only little was known about the fine-scale quet et al. [13] recently calculated that the actual uplift biogeographic structures of Italy and it seemed to be less reaches ~ 2.5 mm/year in the north-western Alps. It is diverse [2]. It was therefore often assumed that the bio- due to erosion processes that the Alps are not now geographic differentiation within Italy, due to its limited much higher. Consequently, the current Alpine land- geographic size, is in general simpler than that of the scapes are mostly the product of the glacial events dur- large sub-centres, for which much more data were avail- ing the last few million years [12]. able by that time [2] (Italy: 301,217 km2 [9]; compared The origin of Sardinia, Corsica, the southernmost tip to 583,254 km2 (Iberia); ca. 500,000 km2 (Balkan Penin- of Calabria and the Monti Peloritani area in north- sula); ca. 750,000 km2 (Anatolia); and the Mediterranean eastern Sicily is linked to the rotation of a block formerly refugium in the Maghreb is at least as large as Anatolia). located in the western Mediterranean sector. According Recent studies on the phylogeography of Italy have to Advokaat et al. [14], the Corso-Sardinian block expe- now revealed a much more complex phylogeographic rienced a counter-clockwise rotation of about 45° already structuring than expected from its limited geographic in the Eocene, followed by a further rotation of about size [3]. Many of these patterns predate the Pleistocene, 50°, mostly during the early Miocene, being much better reflecting divergences correlated with old geological pro- documented than the first [15]. This overall rotation of cesses, like a “museum of the Neogene” in which old about 95° shifted the block from the southern edge of phylogeographic lineages are preserved. Since the review the European continent to its current position and was by Schmitt in 2007 [2], a wealth of new data has become essentially completed by 15 Ma (Fig. 1a). While the rota- available, allowing a comprehensive meta-analysis of the tion of the Corso-Sardinian block is intensively investi- biogeography of Italy with its surrounding islands. gated, the rotation of the precursors of the Below, we examine these data in the context of the com- southernmost tip of Calabria (from the Piana di Sibari to plex geological and geographic history of the region, the strait of Messina) and the Monti Peloritani in north- which is first briefly summarised to make it more eastern Sicily has attracted much less attention. How- understandable. ever, the geology of this area is the result of a mixture of events that span about 60 Ma and is among the most complex of the Italian Peninsula. These two “now south- ern” blocks followed the others, but their rotation Schmitt et al. Frontiers in Zoology (2021) 18:34 Page 3 of 13 Fig. 1 The changeful history of Italy and its surrounding regions from the mid-Miocene to the Pleistocene. (a) Mid-Miocene (17–16 Ma, the white areas are excluded, black lines show the recent situation), (b) the ‘Lago-Mare’ period (i.e., the first Messinian cycle or the lower gypsum mega- sequence), latest Miocene (6.0–5.6 Ma), (c) Pliocene (5.3–2.6 Ma), (d) Last Glacial Maximum (21 ka), similar distributions of land and sea are also assumed for the previous ice ages of the late Pleistocene. Maps based on Bosellini [12] and Vai [17] Schmitt et al. Frontiers in Zoology (2021) 18:34 Page 4 of 13 amplitude was much greater. They rotated along with rise in the Mediterranean of more than ten metres the Corso-Sardinian block until about 15 Ma (Fig. 1a), per day, and it has been estimated that 90 % of the but then kept rotating counter-clockwise during the fol- water was transferred in a short period of time ran- lowing 8 Ma, while the Tyrrhenian Basin opened [12]. ging from a few months to two years [21]. The Calabrian–Peloritanian orogeny is therefore an ar- The palaeogeographic maps of the Pliocene Italian cuate segment of the peri-Mediterranean orogenic Al- Peninsula show a reduced area of dry land compared to pine nappe system [16]. the present (Fig. 1c). A combination of factors caused this, including differences in the height of the mountain The origin of the Apennines is linked to the rotation chains, tectonic activity, and eustatic changes. This, for of the Corso-Sardinian block and the opening of the example, resulted in a repeated transformation of south- Tyrrhenian Sea. About 15 Ma, when the Corso- ern Calabria into a chain of islands. After a moderate Sardinian block reached its final position, the proto- rise estimated to have reached about 20 to 25 m during Apennine chain was only partly formed and exposed the early Pliocene, the sea level fluctuated repeatedly, above sea level. The northern proto-Apennine chain was showing however a general decrease matching a general already in a position close to the one currently occupied decline in global temperatures, and finally reaching the by Tuscany, whereas the southern proto-Apennine chain current level [22]. was located where the Tyrrhenian Sea is located today Because of these sea level oscillations, the extent of (Fig. 1a; [12]). The following counter-clockwise rotation the coastal plains changed considerably throughout of the proto-Apennine chain and its final structuring in time, including the largest plain in Italy: the Po Plain. the chain of today, occurred in the last 8 Ma as a conse- This plain belongs almost entirely to the Apennine quence of the opening of the Tyrrhenian Sea, the youn- foredeep, and extremely thick Pliocene and Quater- gest and among the deepest (about 3600 m) of the nary sediments lie under the present-day surface [23]. basins of the Mediterranean Sea [12]. Along with part of the bottom of the Adriatic Sea, Sea level fluctuations certainly impacted massively the plain is covered by sediment eroded from the the dispersal of terrestrial organisms and, therefore, Alps. Its height ranges from 150 m a.s.l. in its west- gene flow. Three episodes could be particularly rele- ern part to areas below sea level in the east [12]. vant for the understanding of the extant associations: During the Pliocene transgressions it was basically the Messinian Salinity Crisis (MSC; Fig. 1b), the Plio- non-existent, because it was replaced by a relatively cene transgression (Fig. 1c), and the glacial regres- deep gulf when the Adriatic Sea reached the western sions that most recently culminated with the Last edge of the plain, but later on, during the LGM (from Glacial Maximum (LGM; Fig. 1d). The best studied, 26.5 ka to about 20 ka; [24]), when the sea level but still not fully understood event is the MSC, which dropped by about 130 m [22], it stretched in a south- occurred between 5.97 and 5.33 Ma ago [18]. Major easterly direction, reaching the latitude of Ancona or parts of the Mediterranean Sea quickly dried out, even southwards in the direction of the Gargano leaving an arid landscape similar to intermontane de- Promontory (Fig. 1d; [12, 25]), i.e., opening a broad serts of California and Nevada [12, 19], interspersed connection to the western Balkan Peninsula. during stage 3 with non-marine palaeobiotopes, i.e., a The connections of the islands to the mainland are brackish lake system inhabited by ostracods, molluscs difficult to assess using a geologic basis that goes be- and dinocysts, the so-called ‘Lago-Mare’ (Fig. 1b; see yond the evaluation of sea level variations. Therefore, Andreetto et al. [20], and references therein). How- the detection of biogeographic affinities is usually in- ever, new evidence from the Italian fossil record indi- volved in the study of these connections. As summa- cates that fish assemblages were dominated by taxa rized by Masini et al. [26], the Italian fossil record characterized by a strict marine affinity, thereby sug- provides information about three areas that are or gesting that, at least around the Italian Peninsula, the were islands in the past: the Gargano paleo- Mediterranean was filled by marine waters with a archipelago, the Sardinia–Maritime Tuscany paleo- normal salinity [20]. In any case, it is likely that the bioprovince, and the Sicilian insular complex. During paleo-environments and the paleogeography of some the late Miocene–early Pliocene, Gargano was part of portions of the circum-Mediterranean area were sig- an archipelago whose faunal assemblages had a poly- nificantly modified during the MSC. This period phasic origin due to different dispersal events. Worth ended at about 5.33 Ma with the Zanclean flood, the noting is the presence of the same endemic taxa in opening of the Gibraltar Strait, and the rapid refilling Gargano and Scontrone (Abruzzi), a slightly older lo- of the Mediterranean Basin with Atlantic waters. Ac- cality that is currently more than 110 km west of the cording to some interpretations of the geological evi- fossiliferous localities of Gargano. The Apulian dence, this may have involved peak rates of sea level Schmitt et al. Frontiers in Zoology (2021) 18:34 Page 5 of 13 platform was connected to the Apennine chain during the early Pleistocene. The late Miocene Tusco-Sardinian palaeo-bioprovince, including at least northern Sardinia and southern Tus- cany, was unequivocally characterized by the presence of the same endemic mammalian taxa, indicating geo- graphic proximity, or some kind of connection (see Abbazzi et al. [27], with references therein). Since the mid-Pleistocene, populations of various mammals were isolated as Sardinia became an island, and these subse- quently differentiated into endemic taxa. Their assem- blages displayed considerable stability, in contrast to the changes that occurred on the mainland [26]. Towards a comprehensive understanding of the phylogeography of Italy To comprehensively understand the phylogeographic structure of Italy (including Corsica), we analysed 90 phylogeographic studies based on different molecular markers (i.e., mt, cp, and nuDNA sequences; allozyme polymorphisms; microsatellites; others) and various taxa (including 54 vertebrate, 12 invertebrate, and 12 plant species). Note that these species are mostly terrestrial, and although some semi-aquatic taxa are also included, available data is still insufficient to draw clear conclu- sions on the biogeography of the freshwater systems of Italy. Based on these data, we identify the geographic lo- cations of the centres of differentiation (i.e., the refugial Fig. 2 Overview of 17 predefined areas across Italy: A: Po Plain; B: Ligurian and northern Tyrrhenian area; C: central Adriatic, D: central areas where populations were isolated and consequently Tyrrhenian area, E: Tavoliere and Gargano (and surrounding areas); F: diversified into genetic lineages) and the paleo- southern Tyrrhenian area, G: Murge (and surrounding areas); H: geographic events which were responsible. We dissected Salentine peninsula; I: Sila; K: Serre; L: Aspromonte; M: northwestern Italy into 17 predefined geographic areas, according to Sicily; N: northeastern Sicily; O: southern Sicily; P: southern Sardinia; well established geological (e.g. microplates, mountain R: northern Sardinia; S: Corsica. Predefined areas use mountain ranges as divides, but their pattern is also more fine-grained in ranges) and landscape structures (e.g. [28–31]). Against regions that should be more relevant for the diversification of a the background of the geological history of the Pliocene species into different genetic lineages due to their geologic history and Pleistocene range shifts, we ask the following three (see text) questions: 1. Do the islands of Sardinia plus Corsica as well as biogeographically closely linked to Sardinia [34, 35], they Sicily correspond to biogeographical sub-regions in were both included in the present study (Fig. 2). Accord- comparison to the Italian Peninsula? ing to orography and established geographic provinces 2. Do the biogeographic patterns revealed by our (e.g. [28–31]), the following 17 areas were distinguished analyses correlate with geological microplates and (from north to south): A: Po Plain; B: Ligurian and islands, and are they compatible with what is northern Tyrrhenian area; C: central Adriatic, D: central known about the geological history of the Neogene? Tyrrhenian area, E: Tavoliere and Gargano (and sur- 3. Do sub-refugia exist within the Italian peninsular rounding areas); F: southern Tyrrhenian area, G: Murge refugium? And if so, where are these located? (and surrounding areas); H: Salentine peninsula; I: Sila; K: Serre; L: Aspromonte; M: north-western Sicily; N: Materials and methods north-eastern Sicily; O: southern Sicily; P: southern Sar- Study region dinia; R: northern Sardinia; and S: Corsica. These areas We considered all major parts of Italy, but excluded the are of different sizes, reflecting the different levels of Alpine areas, because these represent a biogeographic geographic complexity found throughout Italy. Their entity very different from the rest of Italy [32, 33]. Since resolution into smaller units is higher in regions that, Corsica (politically part of France) is known to be due to their geological history (see above), are expected Schmitt et al. Frontiers in Zoology (2021) 18:34 Page 6 of 13 to be centres of diversification of species into different distribution in Italy so that either the lineage of an ad- genetic lineages. The large islands Sicily and Sardinia joining area is present, or another still unknown lineage. were also divided into three and two areas, respectively, This, in particular, affected the islands Sardinia and Si- to assess possible intra-island differentiation; Corsica cily: if genetic information was not available for each of was not subdivided, due to its smaller size. However, size their, respectively, two and three areas, the area(s) with- was not the sole reason for doing so. As Sicily was com- out data was/were considered ambiguous, although the posed of several parts during the Earth’s history (see presence of this lineage on the entire island is likely. above), this subdivision makes phylogeographic signals Areas with unsupported presence of a genetic lineage resulting from this detectable; additionally, the western entered the statistical analyses as absence to avoid artifi- parts have the closest links to North Africa. On the cially increased biogeographic distributions and de- other hand, Corsica is a relatively uniform block of acid creased discrimination power. Note that two or more bedrock forming one high mountain area surrounded by lineages can co-occur within one single area. In cases in limited areas at lower altitudes. The landscape structures which one such lineages only marginally entered an area, and geology of Sardinia, like Sicily, are much more com- this lineage was also recorded as present. plex, so that splitting it in two parts (i.e. northern and In addition, we assigned lineages to two age categories southern) seemed desirable. in cases where genetic structuring within a species showed a clearly hierarchical age structure and thus en- abled an unequivocal distinction: Biogeographic data We evaluated 127 articles (published up to the end of (i) Deeply divergent lineages (considering exclusively February 2021) addressing biogeography and population the deepest level of diversification in the data set genetics across major parts of peninsular Italy (including and thus reflecting more ancient processes, i.e., the Po Plain) and/or the three major adjacent islands representing old lineages), resulting in 213 units, (Corsica, Sardinia, and Sicily). Ninety of these articles and. contained data suitable for our analysis, i.e., data for sev- (ii) younger genetic lineages (corresponding only to the eral of the predefined areas, based on different molecular shallow genetic divergences and hence reflecting marker systems (i.e., mitochondrial (56), nuclear (10) more recent processes; note that young genetic and plastid (11) DNA sequences, microsatellites (8), allo- lineages are often nested in the old ones, i.e., one zyme polymorphisms (6), AFLPs (1), ISSR (1), chromo- old genetic lineage can comprise two or more some number polymorphisms (1); note that one genetic young lineages), resulting in 258 units. data set has sometimes been derived from more than one article, and that more than one data set may have If no such hierarchical structure was detected, no dis- been presented within one article. An overview of all tinction into old and young lineages was performed. In studies is provided in Appendix S1. We further assigned these cases, the derived lineages were included in the all studies to one of the following taxonomic categories: analysis of the old genetic lineages. As molecular clocks vertebrates (54 species), invertebrates (12), and plants are missing for many of the analysed datasets and the (12). In total, we analysed phylogeographic data sets of heterogeneity of the existing ones did not allow the im- 78 species representing 471 different genetic lineages plementation of a posteriori assignments of molecular from a total of 238 molecular markers, across their Ital- clocks to all of them, this differentiation into old and ian range. young lineages merely reflects the hierarchical genetic For each species, we assessed whether each genetic structuring within the respective taxon. Therefore, these lineage is present or absent in each of our 17 predefined two sets of genetic lineages cannot be unequivocally areas. In many cases, genetic information was not avail- assigned to two absolute time windows. It is only a rela- able for the entire distribution range in Italy and Cor- tive ranking. Nevertheless, the biogeographic structuring sica. If so, for an individual species, areas without revealed by analysis of the set of old lineages, separated genetic data were either considered ambiguous or by deep genetic splits, should be more ancient than the assigned to the lineage occurring in the surrounding structures revealed by the analyses of the geographic dis- areas, if this was unequivocally possible (i.e., the respect- tribution of the young lineages, distinguished only by ive lineage was present in all adjoining areas). Ambiguity shallow splits. Consequently, no clear dating of the was assigned to a lineage within an area without a record structures was possible in our analyses. However, in (1) if the adjoining areas had different genetic lineages cases where phylogeographic analyses have utilised mo- (or even ambiguities) and a decision could not be made lecular clocks, diversification events separating old line- whether the respective lineage is present or not, or (2) ages have mostly been assigned to the Pliocene, whereas the area without record is at the margin of the species Pleistocene age was often assigned to the diversification Schmitt et al. Frontiers in Zoology (2021) 18:34 Page 7 of 13 events involving young lineages. As a third matrix, we 1000 bootstrap samples. Note that our data structure also used a presence/absence data set at the level of spe- based on presences and absences did not allow for the cies. Additionally, we performed our analyses separately application of maximum likelihood methods to construct for vertebrates and plants; the amount of data for inver- the lineage split trees. Parsimony analysis performed tebrates was insufficient for such analyses. with the present presence–absence data set returned highly inconsistent results. Numbers of lineages per area Statistics and the respective dissimilarity matrices are presented in For the two divergence levels (i.e., old and young genetic Appendix S2. lineages) and the presence/absence data of species, we We further calculated the pattern of spatial co- constructed three lineage × area presence-absence matri- occurrence of each species at each level based on ces. Therefore, our study is basically a phenetic ap- Sørensen dissimilarities J using the Turnover software proach, similar to a co-occurrence analysis in [36]. As the Sørensen score depends on the total num- community ecology [36]. Consequently, our approach is bers of occurrences and areas, direct comparisons of raw not molecular in a narrow sense. For each of these three scores might be misleading. Therefore, we applied a null matrices, we calculated differently rooted neighbour model approach and compared observed scores with joining cluster trees (unrooted, rooted in the Po Plain, those obtained from 1000 equiprobable reshufflings of and rooted in southern Sicily) based on the Sørensen focal lineages across the 17 areas. We used equiprobable index of compositional dissimilarity using Past 4.01 [37]. reshuffling because there is no a priori reason to expect Rooting in the Po Plain and Sicily was used to better that certain lineages cannot occur in any one of the Ital- infer the possible latitudinal gradients in the biogeo- ian areas. From these null models, we calculated for each graphic patterns. Given that Sicily and southern Italy species i standardised effect sizes SES i ¼ σJ ii ; where ΔJi is served as major glacial refuges, we examined whether the difference between observed and expected Ji, and σi the existence of such refuges might be supported by re- is the standard deviation of the null model distribution. spective latitudinal gradients of postglacial colonisation. Under the assumption that this distribution is approxi- In phenetic analyses, neighbour joining is still the best mately normal, SES values > |2.00| indicate statistical approach if the data structure requires the use of a dis- significance at the two-sided 5 % error level. similarity metric, as in the present case. This approach is least sensitive to the specific underlying evolutionary processes and robust with respect to deviations from Results branching additivity [38]. We tested the reliability of the Neighbour joining cluster analyses (Figs. 3 and 4) de- tree topology using comparisons from the matrix with tected, with significant statistical corroboration, major Fig. 3 Neighbour joining cluster analyses (Sørensen dissimilarities) based on the occurrence of 471 genetic lineages. We used deeply divergent (i.e. old) genetic lineages (a–c; level I including 213 lineages) and such distinguished by shallow differentiations (i.e., referring to a younger time frame) (d–f, level II representing 258 lineages) for 78 species among 17 areas (A–S) as defined in Fig. 2. Trees in a and d are unrooted; in b and e rooted in the Po Plain; in c and f rooted in southern Sicily. Numbers show the percentage of bootstrapped trees (1,000 replicates) corroborating the focal split. Colours link areas consistently clustering together. The dark red square is highlighting the closest link of Sicily (M, N, O) Schmitt et al. Frontiers in Zoology (2021) 18:34 Page 8 of 13 Fig. 4 Neighbour joining cluster analyses (Sørensen dissimilarities) based on the occurrence of 78 species within 17 areas (A–S) as defined in Fig. 2. Trees are unrooted (a), rooted in the Po Plain (b), and in southern Sicily (c). Numbers show the percentage of bootstrapped trees (1,000 replicates) corroborating the focal split. Colours link sections consistently clustering together. The dark red square is highlighting the closest link of Sicily (M, N, O), which consistently is Sardinia and Corsica (P, R, S). The order of areas best matches the latitudinal mainland gradient biogeographic structuring with segregation of several Tyrrhenian area B and the central Adriatic area C than congruent geographic groups across the different genetic these from each other (Fig. 3d-f). lineages and taxa. In all analyses, Sardinia and Corsica Analyses restricted to vertebrates yield a pattern simi- form one biogeographic unit, but with each island dis- lar to that obtained for all taxa, but the Po Plain is more playing pronounced differentiation. If the old lineages isolated in the former (Appendix S3 Fig. A1). At both are considered only, these two islands appear to be levels, vertebrates of Corsica and Sardinia cluster closely loosely connected to northern Italy (Fig. 3a-c, level I). Si- to Sicily (Fig. A1). A separate analysis for the plant taxa cily (M, N, O) represents a well-supported region of its (older lineages only) yielded a strong differentiation be- own with little internal structure in all analyses. At the tween Sicily and Calabria on the one hand and the rest old lineage level, Sicily is consistently most closely linked of Italy on the other (Fig. A2). Furthermore, Corsica to Calabria, but with low statistical support. Mainland clusters with northern Italy, while Sardinia shows closer Italy is divided into three regions: northern Italy (A–C, affinities to Apulia (Fig. A2). i.e., to the latitude of Rome) and two southern regions, The biogeographic pattern revealed for species occur- i.e., Calabria (I, K, L) and Apulia (G, H). For the north- rences resembles that of the younger genetic lineages ern Italy group, the Po Plain is somewhat more differen- (Fig. 4). An unrooted neighbour joining tree supports tiated from the two adjoining peninsular areas B and C the similarity of Sardinia and Corsica (Fig. 4a), with Si- than these among each other, although located on differ- cily clustering closest but with remarkable distinction. ent sides of the Apennine chain (Fig. 3). Apulia does not Despite the geographic proximity of southern Calabria cluster close to Sicily (Fig. 3, level I). The south-central and Sicily, our data do not reveal a consistent clustering Italian areas D, E and F do not consistently cluster with of both regions and thus no close biogeographic cohe- the regions of northern or southern Italy, but the south- sion. Northern Italy (A–C) is supported as a distinct ern Tyrrhenian area (F) mostly clusters close to Apulia. unit. Rooting of the occurrence tree in the Po Plain re- Rooting the tree in southern Sicily improves the discrim- sults in a clear north–south gradient of species occur- inatory power of the tree and separates this island from rences (Fig. 4b). The strongest discriminatory power has the rest of Italy (Fig. 3c), with Sardinia and Corsica again the tree rooted in southern Sicily (Fig. 4c) clearly indi- recovered as a single distinct cluster close to northern cating the distinctiveness of Sardinia and Corsica, Sicily, Italy. as well as of southern Calabria (K, L). Also, Apulia (G, H) and northern Italy with the Po Plain (A–C) appear to be distinct biogeographic units. Mantel correlations of In turn, the younger lineages return consistent links of Sørensen dissimilarities based on distribution patterns of Sardinia and Corsica with Sicily (Fig. 3d-f). Nevertheless, genetic lineages and species recover high correlations for Sicily as well as Sardinia with Corsica again appear as the deeply divergent lineages (r = 0.82, permutation P < distinct clusters. Analyses of the younger lineages also 0.0001) and a weaker correlation for the younger line- retain Apulia and Calabria as distinct genetic clusters ages (r = 0.57, P < 0.0001). (Fig. 3d-f). Again, the southern Tyrrhenian area F clus- In our species-specific approach, we look at the pat- ters close to Apulia. The Po Plain remains somewhat tern of co-occurrence of all species (Fig. 5). Fifty-two of more distinguished from the Ligurian and northern the 77 species with data at the old divergence level (i.e. 67.5 %) have positive Sørensen dissimilarities across the Schmitt et al. Frontiers in Zoology (2021) 18:34 Page 9 of 13 with the geological history of Italy and its surrounding islands in the Neogene. Below, we discuss in detail this relationship between genetic structures and the geo- logical, orographic and environmental history. The biogeographic uniqueness of Sardinia and Corsica is easily explained by their geological history [39]. Both islands are part of the Tyrrhenian microplate and de- tached from the eastern Iberian Peninsula via a counter- clockwise rotation of their microplate [14], allowing for a long-standing isolation that explains the high amount of old endemic lineages and the high degree of endemic species [40]. The high cohesiveness between the two islands is not only explained by their shared long-term history, but also by repeated connections during periods of glacial low sea levels (Fig. 1d-f, [39, 41]). For the older lineages, a signal for a relationship between Sardinia plus Fig. 5 Standardized effect sizes SES of the Sørensen index of Corsica and northern peninsular Italy was detected. This dissimilarity in spatial co-occurrence between genetic lineages of all reflects the geographic proximity of these two regions, species studied in dependence of the species-specific number of but is also the result of relatively recent (most likely lineages. Red dots refer to the deeply divergent (i.e. old) lineages, Pleistocene) immigrations from the mainland, facilitated blue to younger lineages. Permutation significance of the blue regression line: *: P = 0.03) by its connection to Corsica and Sardinia via the Tuscan Archipelago during glacial periods of low sea levels [42]. This has been shown, for instance, for one lineage of the 17 regions, while 25 of the 39 species for which informa- Italian wall lizard Podarcis sicula [43, 44] and the hedge- tion on younger lineages is available (64.1 %) have posi- hog Erinaceus europaeus [45]. However, care is required tive SES scores (Fig. 5), indicating a tendency towards in the interpretation of some cases, because human in- spatial lineage segregation. In the younger lineages, troductions to these islands have also been recorded, spatial segregation decreases with increasing numbers of such as the European pond turtle Emys orbicularis [46, species occurrences (Fig. 5). 47]. Also, the old links between Sardinia plus Corsica and the nearby mainland might have been overwritten by the probably more frequent and more recent links Discussion with Sicily. The latter loose biogeographic cohesiveness As an answer to our first research question about the is only reflected by the shallowly divergent genetic line- biogeographic importance of the three large islands ad- ages, indicating recent exchange between both, e.g. as joining peninsular Italy, our meta-analysis strongly sug- shown for the Italian wall lizard Podarcis sicula [43, 44]. gests that Sardinia plus Corsica as well as Sicily Furthermore, the biogeographic cohesiveness between represent the two most divergent biogeographic units. Corsica and Tuscany is more pronounced than that of Both have some similarity when only shallow genetic di- Sardinia and Tuscany, which is easily explained by the vergences are considered but not for the lineages distin- geographic location of Sardinia, which could exchange guished by deep divergences (i.e., reflecting the old level with the mainland via Corsica [48]. However, numerous of differentiation). Mainland Italy also shows a strong studies have emphasised that Sardinian endemics are biogeographic structuring. Southern Calabria (K, L), frequently young, and clearly of European origin [40, Apulia (G, H) as well as the northern half of peninsular 49–56], hereby underlining the importance of a pathway Italy (to the latitude of Rome, areas B and C, including via Corsica, or even direct immigration from the Italian the Po Plain, A) consistently cluster together and corres- mainland. pond to three distinct biogeographic units. Northern Ca- The strong phylogeographic isolation of Sicily appears labria (I) is part of the Calabrian region but is at first glance unexpected when its present geographic distinguished remarkably from southern Calabria (K, L) location is considered, separated from Calabria only by in all analyses. The remaining areas in southern central the 3-km-wide Strait of Messina (Fig. 2). However, prior and southern Italy (i.e., D, E, F) constitute a transition to the Pleistocene, the island was more isolated than zone between the two southern Italian regions on the today (Fig. 1a, c). During the Messinian Salinity Crisis one hand and northern Italy on the other. Consequently, (Miocene-Pliocene boundary), Sicily was connected with we confirm the hypotheses implicit in our second initial Calabria and Tunisia [25, 57], allowing an influx of Afri- question, that the biogeographic structures correlate can biota. In addition, low glacial sea levels repeatedly Schmitt et al. Frontiers in Zoology (2021) 18:34 Page 10 of 13 exposed large areas between Sicily and Tunisia (Fig. 1d), Calabria still has a strong influence on the faunal com- which served as stepping stones to facilitate later ex- position of mainland Italy. If so, northern Calabria change [58]. Due to this geological history, Sicily has a would represent the source (initial leading edge) of the high level of biogeographic distinctiveness compared to northwards range expansions during the late Pliocene or mainland Italy, exemplified by several island endemics, the Pleistocene interglacials, whereas the two southern such as the Sicilian pond turtle Emys trinacris [47]), and areas formed the rear edge and did not contribute to the Sicily is often more similar to the geographically more northwards range expansions (cf. [62]). distant Tunisia than to nearby Calabria [4]. The similar- The history and role of the Salentine Peninsula (H) is ities between Sicily and Calabria are relatively low for considerably different. This region was a single island older and younger genetic lineages, underlining that the during the Pliocene and before (Fig. 1a–c). Later, the exchange between these regions was for long periods ra- Salentine Peninsula was linked to mainland Italy via the ther restricted, even after the final formation of Calabria emerging Murge region (G). This is in line with the hy- at a time when they were close geographic neighbours. pothesis that just one major pre-Pleistocene centre of Furthermore, because Sicily is not orographically divided faunal differentiation exists in this region, and that it be- into several parts, a strong internal biogeographic struc- came a major source of dispersal, influencing major ture as seen in many parts of peninsular Italy is absent, parts of Central Italy. However, our results based on all apart from some moderate differentiations, mostly in lineages also reveal southern and central Apulia (G, H) aquatic and semi-aquatic species linked to different as genetically more distinct than more northern regions, water catchments (e.g. Emys trinacris [59]). indicating more northern centres of differentiation dur- Answering our third starting question, we found that ing the Pleistocene, e.g., in the area of the paleo-island the genetic structure of the biota of peninsular Italy is Gargano [26, 63], see below. Nevertheless, some of the highly complex and in accordance with the refugia- distinctiveness of south-eastern Italy might also be due within-refugia model of Gómez & Lunt [8], i.e., several to Pleistocene immigration from the Balkan Peninsula geographically distinct micro-refuges exist within the (i.e. trans-Ionian dispersal), facilitated by the reduced classical refugial centres [5]. At least some of the line- size of the Adriatic Sea, or even a Quaternary land ages from such micro-refuges have originated prior to bridge (Fig. 1d; [64]). Some of these colonising organ- the Pleistocene [60]. Such lineages have survived in dif- isms later became extinct in Italy, e.g. the toad Pelobates ferent glacial micro-refugia but did not necessarily ori- syriacus [64], while others still remain, such as several ginate there. Perhaps they only retreated to these areas species of darkling beetles (Tenebrionidae) [65], the secondarily, during glacial cold phases, as might be the moth species Erannis ankeraria [66] and Charanyca case in the Italian newt Lissotriton italicus [61] and the apfelbecki, as well as the mountain butterfly Coenonym- Italian wall lizard Podarcis sicula [44]. During the Plio- pha rhodopensis [67]. cene, most of southern Italy was divided into different Our results indicate that the diversification of much of islands (due to higher sea levels), matching roughly the the extant flora and fauna of peninsular Italy probably areas of Calabria, the Salentine Peninsula (the “heel” of predates the Pleistocene and that main centres of differ- southern Apulia), and Monte Gargano (the “spur” of entiation were southern Calabria and southern Apulia. Italy). Major parts of the central and northern Italian However, a number of younger differentiation centres Peninsula were submerged (Fig. 1c, [3]), explaining why might have been located north of that, especially in the the most divergent, i.e. oldest, genetic lineages of many Gargano area (E), the southern Tyrrhenian area (F) and species that are today widespread in mainland Italy are northern Calabria (I). It is likely that these more north- detected in southern Italy. ern centres shaped the range expansions and retractions Calabria comprised three separate islands during the during the Pleistocene climatic cycles, simultaneously re- Pliocene (Fig. 1c) and before (today represented by the inforcing the genetic uniqueness of the more southern three areas I, K and L). The most ancient genetic line- Italian areas. ages of peninsular Italy are often found in this region, The populations from the Po Plain and the two adjoin- with one of the most striking examples being the Italian ing peninsular areas B and C form a single faunal unit. newt Lissotriton italicus [61]. Consequently, the entire This applies to older and younger genetic lineages as Calabrian region is genetically distinct from the well as to species composition. Importantly, these areas remaining Italian areas, reflecting the long-lasting isola- (i.e. A–C) are more strongly distinguished from southern tion of this region [3, 12, 14]. However, only the two Italy than would be expected from their location and the southern areas of Calabria (K, L) form a well- young geological age of the Po Plain, that originated in distinguished biogeographic region, whereas northern the Pleistocene (Fig. 1c, d; [12]). This can partly be ex- Calabria (I) represents a link to the more northern areas. plained by Pleistocene survival of cold-resistant taxa in This could indicate that the orographic division of Schmitt et al. Frontiers in Zoology (2021) 18:34 Page 11 of 13 situ. Repeated immigration from the Balkan Peninsula situ, is most likely responsible for the unexpectedly also contributed to the observed distinctiveness. high biogeographic distinctiveness of this region. The Apennines, dividing peninsular Italy longitudinally from north to south, apparently had only a minor influ- ence on the biogeographic structure of the peninsula. Supplementary Information Thus, pairs of areas located at the same latitude but on The online version contains supplementary material available at https://doi. org/10.1186/s12983-021-00418-9. different sides of this mountain chain (i.e., B and C; D and E) are always nearest neighbours in all cladograms Additional file 1: Appendix S1. Overview of all studies analysed with based on the geographic distribution of older and youn- main phylogeographic patterns. ger lineages. Most of the few cases of differentiated line- Additional file 2: Appendix S2. (a) Number of genetic lineages within ages east and west of the Apennines refer to semi- 17 Italian biogeographic areas used for the present study. (b) Geographic distribution of all lineages over the 17 Italian biogeographic areas. (c) aquatic species, such as the European pond turtle Emys Sørensen dissimilarity matrix between 17 Italian biogeographic regions orbicularis [46] or the barred grass snake Natrix helve- based on 213 and 258 genetic lineages, respectively. tica [68]. For such taxa, the watershed formed by the Additional file 3: Appendix S3. (A1) Unrooted neighbour joining Apennine chain seems to impede dispersal, fostering cluster analyses (Sørensen similarities) of vertebrate (Fig. A1) and plant species (Fig. A2). segregation. A similar east-west divergence (with some exceptions) is also known for the terrestrial Italian wall Acknowledgements lizard Podarcis siculus [43, 44]. In general, however, the We thank Stefano Dominici (Florence, Italy), who kindly provided relevant Apennines do not constitute an important biogeographic literature and discussed the Pliocene transgressions, Katja Kramp (Berlin, divide. Germany) providing figure 2 and Andrew Liston (Müncheberg, Germany) correcting our English. Three anonymous referees contributed valuable comments on a draft version of our article. WU acknowledges an Conclusions internal NCU grant. The present study confirms on the meta-scale what Authors’ contributions had emerged in individual studies of smaller species´ UF, TS and JCH screened the existing literature, TS and JCH extracted the assemblages: The Italian Peninsula and the neighbour- phylogeographic information from the collected literature, WU conducted ing large islands Corsica, Sardinia, and Sicily display the statistical analyses. TS, JCH, WU and UF interpreted the data, TS, JCH, MD, WU and UF contributed in writing this article. All authors approved the final pronounced phylogeographic diversification patterns. submittal. For many species, the geographic density of different lineages, and sometimes the age of the splits separat- Funding ing lineages, are remarkably high in our study region. No funding was received to conduct this study. Open Access funding enabled and organized by Projekt DEAL. These characteristics are often more pronounced than the geographic complexity and ages observed for the Availability of data and materials Iberian and the Balkan peninsulas, and hence contra- Articles from which data were used in our analyses are compiled in an electronic appendix. dict the previous assumption that a relatively simple structure and shallow level of differentiation exists in Declarations Italy, due to its small size [2]. On the contrary, we Ethics approval and consent to participate conclude that the complex and diverse geological his- not applicable. tory of Italy, including strong tectonic activity, shift- ing (micro)plates, the submergence of major parts of Consent for publication Not applicable. the current peninsula, the subsequent formation of numerous islands, as well as the Pleistocene climatic Competing interests cycles (Fig. 1) have fostered the evolution of the The authors declare that they have no competing interests. highly complex biogeographic patterns observed today Author details and an unexpected richness in biodiversity. The long- 1 Senckenberg Deutsches Entomologisches Institut, 15374 Müncheberg, lasting isolation of Sicily as well as Sardinia plus Cor- Germany. 2Department of Zoology, Institute of Biology, Faculty of Natural sica has produced old endemic biota, which only re- Sciences I, Martin Luther University Halle-Wittenberg, 06099 Halle (Saale), Germany. 3Entomology and Biogeography, Institute of Biochemistry and cently (i.e., during the Pleistocene) have moderately Biology, Faculty of Science, University of Potsdam, 14476 Potsdam, Germany. interacted with the neighbouring mainland regions. 4 Museum of Zoology (Museum für Tierkunde), Senckenberg Natural History The vanished Pliocene islands in the Calabrian and Collections Dresden, 01109 Dresden, Germany. 5Dipartimento di Scienze della Terra, Università di Torino, Via Valperga Caluso 35, 10125 Torino, Italy. southern Apulian areas are the origin of other en- 6 Institut Català de Paleontologia Miquel Crusafont, Universitat Autònoma de demic biota, which only recently expanded north- Barcelona, Edifici ICTAICP, Carrer de les Columnes s/n, Campus de la UAB, wards, with further differentiations. Continuous 08193 Cerdanyola del Vallès, Barcelona, Spain. 7Department of Ecology and Biogeography, Nicolaus Copernicus University, 87-100 Toruń, Poland. immigration from the Balkan Peninsula to northern 8 Evolutionary Zoology, Department of Biosciences, University of Salzburg, Italy, combined with subsequent ice age survival in 5020 Salzburg, Austria. Schmitt et al. Frontiers in Zoology (2021) 18:34 Page 12 of 13 Received: 19 November 2020 Accepted: 6 June 2021 24. Clark PU, Dyke AS, Shakun JD, Carlson AE, Clark J, Wohlfarth B, et al. The Last Glacial Maximum Science. 2009;325:710–4. 25. Thiede J. A glacial Mediterranean. Nature. 1978;276:680–3. 26. Masini F, Petruso D, Bonfiglio L, Mangano G. Origination and extinction References patterns of mammals in three central Western Mediterranean islands from 1. Myers N, Mittermeier RA, Mittermeier CG, da Fonseca GAB, Kent J. the Late Miocene to Quaternary. Quat Int. 2008;182:63–79. Biodiversity hotspots for conservation priorities. Nature. 2000;403:853–8. 27. Abbazzi A, Delfino M, Gallai G, Trebini L, Rook L. New data on the vertebrate 2. Schmitt T. Molecular biogeography of Europe: Pleistocene cycles and assemblage of Fiume Santo (North-West Sardinia Italy) and overview on the Late postglacial trends. Front Zool. 2007;4:11. Miocene Tusco-Sardinian palaeobioprovince. Palaeontology. 2008;52:251–425. 3. Hewitt GM. Mediterranean peninsulas: the evolution of hotspots. In: Zachos 28. Console R, Di Giovambattista R, Favali P, Presgrave BW, Smriglio G. FE, Habel JC, editors. Biodiversity Hotspots: Distribution and protection of Seismicity of the Adriatic microplate. Tectonophysics. 1993;218:343–54. conservation priority areas. Dordrecht: Springer; 2011. pp. 123–47. 29. Critelli S, Le Pera E. Post-Oligocene Sediment-Dispersal Systems and Unroofing 4. Husemann M, Schmitt T, Zachos FE, Ulrich W, Habel JC. Palaearctic History of the Calabrian Microplate, Italy. Int Geol Rev. 1998;40:609–37. biogeography revisited: evidence for the existence of a North African 30. Oggiano G, Funedda A, Carmignani L, Pasci S. The Sardinia-Corsica refugium for Western Palaearctic biota. J Biogeogr. 2014;41:81–94. microplate and its role in the Northern Apennine Geodynamics: new 5. De Lattin G. Beiträge zur Zoogeographie des Mittelmeergebietes. Verh dt insights from the Tertiary intraplate strike-slip tectonics of Sardinia. Boll Soc Zool Ges. 1949;143–51. Geol Ital. 2009;128:527–39. 6. Taberlet P, Fumagalli L, Wust-Saucy A, Cosson J. Comparative 31. Carmignani L, Conti P, Cornamusini G, Pirro A. Geological map of Tuscany phylogeography and postglacial colonization routes in Europe. Mol Ecol. (Italy). J Maps. 2013;9:487–97. 1998;7:453–64. 32. Bologna MA, Mazzotti S. Biogeography. In: Bernini F, Doria G, Razzetti E, 7. Hewitt GM. Post-glacial re-colonization of European biota. Biol J Linn Soc. editors. Atlas of the Italian Amphibians and Reptiles. Firenze: Societas 1999;68:87–112. Herpetologica Italica, Edizioni Polistampa; 2006. pp. 654–77. 8. Gómez A, Lunt DH. Refugia within refugia: Patterns of phylogeographic 33. Menchetti M, Talavera G, Cini A, Salvati V, Dincă V, Platania L, et al. Two concordance in the Iberian Peninsula. In: Weiss S, Ferrand N, editors. ways to be endemic. Alps and Apennines are different functional refugia Phylogeography of Southern European Refugia. Dordrecht: Springer; 2007. during climatic cycles. Mol Ecol. 2021;30:https://doi.org/10.1111/mec.15795. pp. 155–88. 34. De Lattin G. Grundriß der Zoogeographie. Jena: Gustav Fischer; 1967. 9. Doglioni C, Flores G. An introduction to the Italian geology. Potenza: 35. Lanza B, Corti C. Evolution of knowledge on the Italian herpetofauna during Lamisco; 1997. the 20th century. Boll Mus civ St nat Verona. 1993;20:373–436. 10. Berra F, Angiolini L. The evolution of the Tethys region throughout the 36. Ulrich W, Gotelli NJ. Pattern detection in Null Model Analysis. Oikos. 2013; Phanerozoic: A brief tectonic reconstruction. In: Marlow L, Kendall C, and L. 122:2–18. Yose L, editors. Petroleum systems of the Tethyan region. AAPG Memoir 37. Hammer Ø, Harper DAT, Ryan PD. PAST: Paleontological StatisticsSoftware 106; 2014. p. 1–27. Package for Education and Data Analysis. Palaeontologia Electronica; 2001. 11. Rosenbaum G, Lister GS, Duboz C. Relative motions of Africa, Iberia 38. Mihaescu R, Levy D, Pachter L. Why neighbor-joining works. Algorithmica. and Europe during the Alpine Orogeny. Tectonophysics. 2002;359: 2009;54:1–24. 117–29. 39. Ketmaier V, Caccone A. Twenty years of molecular biogeography in the 12. Bosellini A. Storia geologica d’Italia: gli ultimi 200 milioni di anni. Bologna: west Mediterranean islands of Corsica and Sardinia: Lessons learnt and Zanichelli; 2005. future prospects. In: Silva-Opps, editor. Current Progress in Biological 13. Nocquet J-M, Sue C, Walpersdorf A, Tran T, Lenôtre N, Vernant P, et al. Research. Rijeka: InTech; 2013. pp. 71–94. Present-day uplift of the western Alps. Sci Rep. 2016;6:28404. 40. Grill A, Casula P, Lecis R, Menken S. Endemism in Sardinia. In: Weiss S, 14. Advokaat EL, van Hinsbergen DJJ, Maffione M, Langereis CG, Vissers Ferrand N, editors. Phylogeography of Southern European Refugia. RLM, Cherchi A, et al. Eocene rotation of Sardinia, and the Dordrecht: Springer; 2007. pp. 273–96. paleogeography of the western Mediterranean region. Earth Planetary 41. Strumia F. Evidence of recent cladogenesis in the Chrysididae fauna of Sci Lett. 2014;401:183–95. Sardinia and Corsica islands (Hymenoptera, Chrysididae). Atti della Società 15. Gattacceca J, Deino AL, Rizzo R, Jones B, Henry B, Beaudoin B, et al. Toscana di Scienze Naturali. Memorie Serie B. 2012;119:75–7. Miocene rotation of Sardinia: new paleomagnetic and geochronological 42. Fattorini S. Influence of Recent Geography and Paleogeography on the constraints and geodynamic implications. Earth Planet Sci Lett. 2007;258: Structure of Reptile Communities in a Land-Bridge Archipelago. J Herpetol. 359–77. 2010;44:242–52. 16. Cirrincione R, Fazio E, Fiannacca P, Ortolano G, Pezzino A, Punturo R. The 43. Podnar M, Mayer W, Tvrtković N. Phylogeography of the Italian wall lizard, Calabria-Peloritani Orogen, a composite terrane in Central Mediterranean; its Podarcis sicula, as revealed by mitochondrial DNA sequences. Mol Ecol. overall architecture and geodynamic significance for a pre-Alpine scenario 2005;14:575–88. around the Tethyan basin. Period Mineral. 2015;84:701–49. 44. Senczuk G, Colangelo P, De Simone E, Aloise G, Castiglia R. A combination 17. Vai GB. Over half a century of Messinian salinity crisis. Bol Geol Min. 2016; of long term fragmentation and glacial persistence drove the evolutionary 127:615–32. history of the Italian wall lizard Podarcis siculus. BMC Evol Biol. 2017;17:6. 18. Vasiliev I, Mezger EM, Lugli S, Reichart G-J, Manzi V, Roveri M. How dry was 45. Seddon JM, Santucci F, Reeve NJ, Hewitt GM. DNA footprints of European the Mediterranean during the Messinian salinity crisis? Palaeogeogr hedgehogs, Erinaceus europaeus and E. concolor. Pleistocene refugia, Palaeoclimatol Palaeoecol. 2017;471:120–33. postglacial expansion and colonization routes. Mol Ecol. 2001;10:2187–98. 19. Hsü KJ. The Mediterranean was a desert: A voyage of the Glomar 46. Lenk P, Fritz U, Joger U, Wink M. Mitochondrial phylogeography of the European Challenger. New Jersey: Princeton University Press; 1983. pond turtle, Emys orbicularis (Linnaeus 1758). Mol Ecol. 1999;8:1911–22. 20. Andreetto F, Aloisi G, Raad F, Heida H, Flecker R, Agiadi K, et al. Freshening 47. Vamberger M, Stuckas H, Sacco F, D’Angelo S, Arculeo M, Cheylan M, et al. of the Mediterranean Salt Giant: controversies and certainties around the Differences in gene flow in a twofold secondary contact zone of pond terminal (Upper Gypsum and Lago-Mare) phases of the Messinian Salinity turtles in southern Italy (Testudines: Emydidae: Emys orbicularis galloitalica, E. Crisis. Earth-Sci Rev. 2021;216:103577. o. hellenica, E. trinacris). Zool Scr. 2015;44:233–49. 21. Garcia-Castellanos D, Estrada F, Jiménez-Munt I, Gorini C, Fernàndez M, 48. Dapporto L, Schmitt T, Vila R, Scalercio S, Biermann H, Dincă V, Vergés J, et al. Catastrophic flood of the Mediterranean after the Messinian et al. Phylogenetic island disequilibrium: evidence for ongoing salinity crisis. Nature. 2009;462:778–81. long-term population dynamics in two Mediterranean butterflies: 22. Miller KG, Mountain GS, Wright JD, Browning JV. A 180-million-year record Phylogenetic disequilibrium in Mediterranean butterflies. J Biogeogr. of sea level and ice volume variations from continental margin and deep- 2011;38:854–67. sea isotopic records. Oceanography. 2011;24:40–53. 49. Caccone A, Milinkovitch MC, Sbordoni V, Powell JR. Molecular 23. Castiglioni GB. Geomorphology of the Po plain. Fourth international biogeography: using the Corsica-Sardinia microplate disjunction to calibrate conference on geomorphology, Italy 1997. Suppl Geogr Fis Dinam Quat. mitochondrial rDNA evolutionary rates in mountain newts (Euproctus). J Evol 1999;II T.3:7–20. Biol. 1994;7:227–45. Schmitt et al. Frontiers in Zoology (2021) 18:34 Page 13 of 13 50. Marchi A, Addis G, Hermosa VE, Crnjar R. Genetic divergence and evolution of Polyommatus coridon gennargenti (Lepidoptera: Lycaenidae) in Sardinia. Heredity. 1996;77:16–22. 51. Lecis R, Norris K. Population genetic diversity of the endemic Sardinian newt Euproctus platycephalus: implications for conservation. Biol Conserv. 2004;119:263–70. 52. Grill A, Schtickzelle N, Cleary DFR, Nève G, Menken SBJ. Ecological differentiation between the Sardinian endemic Maniola nurag and the pan- European M. jurtina. Biol J Linn Soc. 2006;89:561–74. 53. Grill A, Raijmann LEL, Van Ginkel W, Gkioka E, Menken SBJ. Genetic differentiation and natural hybridization between the Sardinian endemic Maniola nurag and the European Maniola jurtina. J Evol Biol. 2007;20:1255–70. 54. Kreuzinger AJ, Fiedler K, Letsch H, Grill A. Tracing the radiation of Maniola (Nymphalidae) butterflies: new insights from phylogeography hint at one single incompletely differentiated species complex. Ecol Evol. 2015;5:46–58. 55. Todisco V, Fiedler K, Puliga E, Dincă V, Vodă R, Grill A. The origin of the Sardinian Blue, Pseudophilotes barbagiae (Lepidoptera, Lycaenidae): out-of- Europe or out-of-African. Genome. 2015;58:288. 56. Todisco V, Grill A, Fiedler K, Gottsberger B, Dincă V, Vodă R, et al. Molecular phylogeny of the Palaearctic butterfly genus Pseudophilotes (Lepidoptera: Lycaenidae) with focus on the Sardinian endemic P. barbagiae. BMC Zool. 2018;3:4. 57. Giraudi C. The Apennine glaciations in Italy. In: Ehlers J, Gibbard PL, editors. Quaternary glaciations–extent and chronology. Part I: Europe. Amsterdam: Elsevier Science; 2004. pp. 215–24. 58. Heiser M, Dapporto L, Schmitt T. Coupling impoverishment analysis and partitioning of beta diversity allows a comprehensive description of Odonata biogeography in the Western Mediterranean. Org Divers Evol. 2014;14:203–14. 59. Vecchioni L, Marrone F, Arculeo M, Fritz U, Vamberger M. Stand out from the Crowd: Small-Scale Genetic Structuring in the Endemic Sicilian Pond Turtle. Diversity. 2018;12:343. 60. Schultze N, Spitzweg C, Corti C, Delaugerre M, Di Nicola MR, Geniez P, et al. Mitochondrial ghost lineages blur phylogeography and taxonomy of Natrix helvetica and N. natrix in Italy and Corsica. Zool Scr. 2020;49:395–411. 61. Canestrelli D, Sacco F, Nascetti G. On glacial refugia, genetic diversity, and microevolutionary processes: deep phylogeographical structure in the endemic newt Lissotriton italicus. Biol J Linn Soc. 2012;105:42–55. 62. Hampe A, Petit RJ. Conserving biodiversity under climate change: the rear edge matters: Rear edges and climate change. Ecol Lett. 2005;8:461–7. 63. De Giuli C, Masini F, Torre D, Boddi V. Endemism and bio-chronological reconstructions: the Gargano case history. Boll Soc Paleontol Ital. 1986;25: 267–76. 64. Blain HA, Delfino M, Berto C, Arzarello M. First record of Pelobates syriacus (Anura, Amphibia) in the Early Pleistocene of Italy. Palaeobiodivers Palaeoenviron. 2016;96:111–24. 65. Fattorini S. Tenebrionid beetle distributional patterns in Italy: multiple colonisation trajectories in a biogeographical crossroad. Insect Conserv Divers. 2014;7:144–60. 66. Nahirić A, Beshkov S. First report of Desertobia ankeraria (Staudinger, 1861) (Lepidoptera: Geometridae) in Serbia. Acta entomol serbica. 2016;21:143–6. 67. Louy D, Habel JC, Abadjiev S, Schmitt T. Genetic legacy from past panmixia: High genetic variability and low differentiation in disjunct populations of the Eastern Large Heath butterfly. Biol J Linn Soc. 2013;110:281–90. 68. Kindler C, Fritz U. Phylogeography and taxonomy of the barred grass snake (Natrix helvetica), with a discussion of the subspecies category in zoology. Vertebr Zool. 2018;269–81. Publisher’s Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.