We document the distribution of the common toad Bufo bufo and the spined toad B. spinosus at their contact zone across France with data from a mitochondrial DNA RFLP assay, complementing similar work including nuclear markers in the northwest and southeast of France and in Italy. We also reconstruct geographical clines across the species’ contact zone in central France. Bufo bufo is found in the north-eastern half of France. Bufo spinosus is found in the south-western complement. The contact zone they form runs from the Atlantic coast near Caen, France, to the Mediterranean coast near Savona, Italy, and has a length of over 900 km. In central France B. bufo and B. spinosus engage in a hybrid zone with a unimodal genetic signature. Hybrid zone width is ca. 10 km at mitochondrial DNA and averages at 61 km for four nuclear loci. The hybrid zone is distinctly asymmetric with a signature of B. spinosus in B. bufo and not the other way round. We attribute this observation to B. bufo moving southwards at the expense of B. spinosus, with introgression in the direction of the advancing species. We noted substantial geographic variation in characters for species identification. Morphological species identification performs well in France, but breaks down in Italy. Mitochondrial DNA is inconclusive in south-eastern France and Italy. The nuclear genetic markers perform consistently well but have not yet been applied to the zone in full. Possible, but surely heterogeneous ecological correlates for the position of the hybrid zone are mountains and rivers.
Hybrid zones have long been regarded as natural laboratories for the study of speciation (Hewitt, 1988). Inspired by the review of Hewitt (2000), many studies have further documented the location of secondary contact zones in a wide range of hybridizing taxa across Europe (for reviews see Taberlet et al., 1998; Habel et al., 2010; Schmitt and Varga, 2012). This has revealed general patterns of concordance in the location of so-called suture zones, especially among taxa surviving the Pleistocene Ice Ages in different southern European peninsulas, that later expanded out of their refugia to colonize previously unsuitable areas in central and northern Europe. The addition of data from new model systems builds on previously compiled comparative data and helps to address questions related to the timing and the spatial and ecological context of hybrid zone formation.
One emerging model system is formed by the common toad (Bufo bufo) and the spined toad (B. spinosus), which are genetically deeply differentiated yet morphologically similar species that meet up in France (Recuero et al., 2012; Arntzen et al., 2013a). They are not sister species. Once their evolutionary independence had become apparent, a search for morphological features that would allow species identification in the field pointed to adult body size, the positioning of the parotoids and the size and shape of the metatarsus tubercle as species diagnostic characters (Arntzen et al., 2013b). Based on these morphological features, which were consistent with species assignment based on molecular data, we broadly delineated their contact zone in France, from near Caen in the northwest to Lyon in the southeast. The contact zone encompasses great topographic (and thus ecological) heterogeneity, potentially conditioning the initial location, dynamics and outcome of the species contact. Therefore, a finer-scale delineation of the zone can reveal the role of topographic and ecological factors in shaping the contact between the two species.
We have previously shown that B. bufo and B. spinosus do hybridize (Trujillo et al., 2017) and that, in the northwest of France, they do so in a narrow zone (Arntzen et al., 2016). Cline analyses located the centre of the contact zone at the northern slope of the ‘Collines de Normandie’ (Arntzen et al., 2016). In this region, mtDNA is largely co-distributed with nuclear markers and with morphology in defining species borders, and thus mtDNA genotyping can possibly be used as a fast, cost-effective approach to accurately delineate the contact zone in the remainder of the area where the two species meet. Here we use new and published molecular data to document the precise position of the toad hybrid zone from the Atlantic to the Mediterranean coast. We discuss the role of major rivers and topographic features in shaping species boundaries and their interactions. This information can in turn be used to subsequently select and define replicated transects on which to perform more detailed analyses of the hybrid zone, using genomic data in a cline analysis framework. Additionally, the mapping of the contact zone allows fine-scale delineation of the ranges of the two species, which is a basic step required to complete global or regional assessments on their conservation status.
Materials and methods
DNA was extracted from 1428 tissue samples in 140 populations (, range 1-40). For a visualisation of the sampling effort see
Common toad populations studied for mitochondrial DNA species affiliation, with coordinates and sample sizes. Populations that make up the transect in central France are marked with a #. Populations additionally studied are listed at the bottom. For a map with localities plotted see
We studied the B. bufo-B. spinosus species transition in central France in a transect from the southwest to the northeast under a 45-degree angle. The transect was directed through populations B137 and B102 and positioned roughly perpendicular to the mitochondrial DNA contact zone (see fig. 1). Initially 15 populations were analysed for two mtDNA markers (cytochrome-b and 16S) and four nuclear markers (BDNF, POMC, RAG1 and RPL3). Results indicated that there is no significant tail at the B. spinosus side of the transect for either of the genetic markers. Conversely, substantial tails were observed at the B. bufo side for all nuclear markers. We therefore extended the transect northwards with populations from Belgium (B472-476) and the Netherlands (B171, B172-173 and B359). Altogether 19 populations were studied for mtDNA SNPs with an average sample size of (range 6-28) and with 0.5% of missing data. The same populations were analysed for the four nuclear SNPs with an average sample size of (range 2-28), with 2.1% missing data. Analysis followed the procedures described in Arntzen et al. (2016). The molecular data that describe the species transition were analysed with the geographical cline fitting software HZAR (Derryberry et al., 2014) as described in Arntzen et al. (2017a).
The inferred mtDNA haplotype identities (table 1) are plotted per population as B. bufo, B. spinosus or both species in fig. 1. For ease of interpretation the results are presented in Dirichlet cells for which the spatial extrapolation does not exceed ca. 40 km. The B. bufo-B. spinosus contact zone runs across France from near the mouth of the Seine river at the Atlantic coast till near Grenoble at the French Alps. Over those parts of France where the sampling is dense, the width of the area with both species rarely exceeds 50 km. The cytochrome-b sequence for one RFLP outlier data point indicating B. spinosus in the north was actually equivalent to the B. bufo sequence JN647248, whereas three outlier data points indicating B. bufo in the south were equivalent to the B. spinosus sequence JN647327. It thus appears that these outliers correspond to identification errors caused by homoplasious nucleotide substitutions.
In the southeast of France our sampling is thin, the mutual species distribution appears a mosaic and contact zone position and width are difficult to assess. However, data for this region and adjacent Italy have recently become available (Arntzen et al., 2017a) and are reproduced in fig. 1B. From the same source we plot the position of the B. bufo-B. spinosus contact zones at either side of the Alps, as determined from morphology and nuclear markers.
The species transition in central France is described by a steep cline in mitochondrial DNA and smoother clines for nuclear loci (fig. 2, the supporting data are in table 2). The central point for mtDNA is adjacent to population B156 at 46.044 N, 2.498 E, as this is the only population in the transect with mtDNA haplotypes of both species. Hence, the maximum width of the species transition is determined by ‘pure’ populations at either side. These are population B155 with 8 ‘spinosus’ haplotypes and B135 with 8 ‘bufo’ haplotypes. The distance between the populations is 29.4 km, which is at par with bounds of the 95% confidence intervals for the cline estimates at 0.1 km and 17.4 km. At ca. 61 km the nuclear clines are discordant, significantly wider than the mitochondrial cline (table 3). The cline centres are on average 21.1 km displaced towards the B. bufo side of the transect, significantly (not coincident) for BDNF and POMC and not so (coincident) for RAG1 and RPL3.
Molecular genetic data used for cline fitting over a transect in central France. Fs is the frequency of haplotypes and alleles typical for Bufo spinosus. N is the sample size.
We mapped B. bufo and B. spinosus from mtDNA over a large part of France to document their mutual range border and to complete the picture from the Atlantic to the Mediterranean. Species identification from morphology works well in the British Isles (Jersey) and in France, but not in Italy (table 4). As is to be expected, individuals of hybrid origin have intermediate phenotypes (Arntzen et al., 2016) and cannot (and should not) be allocated to either parental species. The mitochondrial and nuclear clines line up in north-western France and in central France, suggesting that the western part of the two-species range border is adequately documented. However, at southern latitudes the species-specific signatures are markedly dissimilar for nuclear versus mitochondrial DNA.
Parameter estimates for the maximum-likelihood geographical clines shown in fig. 2. Position of the clines is relative to the one for mitochondrial DNA and cline width is 1/maximum slope. Distances are in km. Confidence intervals in parentheses are based upon the 2log-likelihood unit support limits. δ and τ are the shape parameters for tail fitting at the right side. The estimated frequencies at either end of the cline are unity for Bufo spinosus and zero for B. bufo. For a visualization of the nuclear clines with their confidence intervals and population data see
Overview on the performance of morphological and molecular characters in use for the identification of Bufo bufo and B. spinosus in western Europe. Character states are diagnostic (plus), ambiguous (plus-minus), or not diagnostic (minus).
General features of the species contact zone are the absence of large areas of B. bufo-B. spinosus sympatry, its long, narrow and more or less linear aspect, with an overall geographical pattern similar to that of other species pairs, like Triturus cristatus and T. marmoratus (Arntzen and Wallis, 1991; Wielstra et al., 2014), or – not in France but more to the north and the east – Natrix grass snakes (Kindler et al., 2017). Some topographical correlates of the zone’s position can tentatively be proposed. The zone starts at the mouth of the Seine and in central France its location appears to coincide with the rivers Loir, Loire and Cher. In north-western France the zone runs along the ‘Collines de Normandie’ low mountain range, and in Italy the zone is located at the northern slopes of the Ligurian Alps. However, the positioning of the contact zone does not seem affected by the Plateau Central and runs straight across the Rhône river into the Alps. Here the position of the hybrid zone partially coincides with the Isère river.
For B. bufo the upper limits in altitudinal range are around 2400 m a.s.l. (Sinsch et al., 2009) and because toad density declines with altitude, the French and Italian ranges of B. bufo are effectively separated by the Alps. These areas represent different intraspecific lineages (Arntzen et al., 2017a). West of the Alps the mutual range border with B. spinosus is situated at ca. 45.0 N and east of the Alps at ca. 44.5 N (fig. 1). In the four transects studied to date, B. bufo and B. spinosus do exchange genetic material, i.e., they hybridize. At an average width of ca. 61 km at nuclear loci, the hybrid zone in central France appears wider than at the Atlantic side (53 km; Arntzen et al., 2016) and the Mediterranean side of France (9 km) and Italy (20 km; Arntzen et al., 2017a). Important features of this hybrid zone are its unimodal structure and the asymmetry in patterns of marker introgression, with B. spinosus haplotypes in the B. bufo side of the contact but not the other way round. This has been attributed to the existence of a moving hybrid zone, with B. bufo moving southwards at the expense of B. spinosus, resulting in introgression in the direction of the advancing species (Arntzen et al., 2017a).
If there are no ecological associations with hybrid zone location this would suggest that ecological constellations influence the mutual toad distribution differently across the species contact. Given the extensive interspecific hybridization observed over the four transects, in different ecological settings, the B. bufo-B. spinosus contact classifies as a heterogeneous hybrid zone (sensu Espregueira-Themudo et al., 2012). The European common toad hybrid zone offers good opportunity for genome-wide analyses in search of both consistently differentiated genomic regions across transects (candidate ‘barrier genes’) as well as locally adapted variants. The system is a promising model for the study of the genetic architecture of species differences. Particular assets are that the species are widespread, abundant and pond-breeding and that they have low dispersal.
Our results can be directly used to delineate species ranges, which is challenging in cases where morphology alone can lead to confusion or there is simply no available information to draw the line (e.g. Sillero et al., 2014). This is the case for these two species, which have an extinction risk status that is currently under re-assessment by the IUCN. While both species currently inhabit relatively large areas and are not considered to be in decline (Crochet et al., 2004; Arntzen et al., 2017b), population declines have also been reported (Petrovan and Schmidt, 2016). One initial step in this assessment will involve delimiting their respective ranges, for which we expect our results will be helpful.
We thank Brandon Ballengée, Mathieu Denoël, Willem Meilink, Rob Veen and Annie Zuiderwijk for assistance with sampling. IMS acknowledges the receipt of a Naturalis ‘Temminck’ visitors grant.
ArntzenJ.W.TrujilloT.ButôtR.VrielingK.SchaapO.Gutiérrez-RodríguezJ.Martínez-SolanoI. (2016): Concordant morphological and molecular clines in a contact zone of the common and spined toad (Bufo bufo and B. spinosus) in the northwest of France. Front. Zool. 13: 52.
RecueroE.CanestrelliD.VörösJ.SzabóK.PoyarkovN.A.ArntzenJ.W.Crnobrnja-IsailovicJ.KidovA.A.CogălniceanuD.CaputoF.P.NascettiG.Martínez-SolanoI. (2012): Multilocus species tree analyses resolve the radiation of the widespread Bufo bufo species group (Anura, Bufonidae). Mol. Phylogenet. Evol. 62: 71-86.
SinschU.SchneiderH.TarkhnishviliD.N. (2009): Bufo bufo Superspezies – Erdkröten-Artenkreis – taxon bufo (Linnaeus, 1758) – Erdkröte – taxon gredosicola L. Müller und Hellmich, 1935 – Gredoserdkröte – taxon spinosus Daudin, 1803 – Riesenerdkröte – taxon verrucosissimus (Pallas, 1811) – Kolchische Erdkröte. In: Handbuch der Reptilien und Amphibien Europas. Froschlurche II, p. 191-335. GrossenbacherK., Ed., Aula Verlag, Wiebelsheim, Germany.
Associate Editor: Uwe Fritz.