Abstract
Human-mediated secondary contact of recently diverged taxa offers valuable opportunities for studying the evolutionary mechanisms involved in the establishment and maintenance of genetic boundaries between taxa. We used mitochondrial and microsatellite markers to examine a recently introduced population of the spur-thighed tortoise (Testudo graeca) of mixed origin in the Doñana National Park (SW Spain). The earliest records of tortoises in Doñana trace back to the 18th century, but several population reinforcements in the 20th century with animals from Morocco are well-documented. Consequently, different genetic lineages, which represent distinct subspecies, are thought to co-exist there. Our results confirmed the presence of distinct lineages by revealing that tortoises of the subspecies T. g. marokkensis were introduced into a local allochthonous T. g. graeca population. Unexpectedly, T. g. marokkensis haplotypes exclusively appeared in males, and admixture levels were statistically sex-biased toward males. The sex ratio of the population deviated from parity, with males being 2.36-fold more abundant than females. Our results indicated that population reinforcements had a strong effect on the genetic composition of this population and aggravated its sex ratio deviation. We predict that this sex-biased pattern of introgression is ephemeral and advocated to the near loss of T. g. marokkensis haplotypes.
Similar content being viewed by others
Introduction
Species boundaries originate from limited or lacking gene flow between species caused by a set of intrinsic natural barriers. Introgression can be defined as the exchange and stable integration of alleles between genetically distinct populations and across such boundaries1,2,3. The resulting patterns, caused by hybridization and repeated back-crossing, largely depend on hybrid fitness and their ability to backcross with parental species4, 5. It has been widely demonstrated that introgression does not occur homogeneously within populations [e.g. refs 6,7,8]. Differences in mutation rates, selection pressures or demographic factors can cause cytonuclear discordance9 or sex-biased levels of introgression1. Cytonuclear discordance, defined as a significant difference between mitochondrial and nuclear genomes, is expected since the mitochondrial genome is inherited maternally in most species and has a fourfold smaller effective population size than the nuclear genome10, 11, which results in the quicker loss of ancestral polymorphisms over time12. Cytonuclear discordance is favored by mainly male-mediated gene flow, leading to sex-biased introgression. Sex-biased introgression is also often explained by Haldane’s rule, which states that “when in the offspring of two different animal races one sex is absent, rare, or sterile, that sex is the heterozygous (heterogametic) sex”13. Alternatively, differences in dispersal capabilities between males and females, or in selective pressures, can also promote such patterns1.
Abbott et al.2 recently pointed out that introgression could either promote or retard divergence processes, and vastly varies among hybridizing taxa and their divergence stages. In recently diverged groups, genetic boundaries are often subtle and allow gene flow by incomplete reproductive isolation. This is often the case in introductions that cause secondary contacts of recently diverged groups [e.g. refs 14,15,16]. Thus environmental disturbances and introductions of wild or domestic taxa are valuable opportunities to study the evolutionary mechanisms involved in the establishment and maintenance of genetic boundaries between taxa. However, these situations are not desirable from a conservationist point of view. Exchange of genes between recently isolated lineages could result in loss of local adaptations15, or the transmission of genes from domestic to wild populations can even threaten natural processes in ecosystems (e.g. exchange of genes conferring herbivory resistance17).
Here we address introgression patterns by examining the origin and historic management of the spur-thighed tortoise, Testudo graeca Linnaeus, 1758, in southwest Spain, which putatively led to the admixture of relatively recently diverged genetic lineages from North Africa. The spur-thighed tortoise is one of the most widely distributed tortoise species in the world. Its western Palearctic distribution range embraces North Africa, the Middle East, Asia Minor and southeast Europe. Moreover, a few small isolated populations occur at two distant sites in southwest and southeast regions of the Iberian Peninsula, and also on Majorca, Sardinia and Sicily18, 19. In North Africa, a monophyletic group of six mitochondrial lineages shows a west-east vicariant diversification pattern with allo- and parapatric ranges (Fig. 1a)20,21,22. Meanwhile, only one mitochondrial lineage per population is present on the European side of the Western Mediterranean, whose populations generally result from historical introductions from North Africa20, 22,23,24 (Fig. 1a). Among them, the population in southeastern Spain has been identified as the most ancient as it dates back to around 20,000 years ago, but is also derived from North African founders25. Despite its iconic location in the Doñana National Park (southwest Spain; referred to as “Donaña” henceforth), the other population that inhabits the Iberian Peninsula remains genetically unstudied.
The earliest references of tortoises that inhabit Doñana trace back to the 18th century. Historical texts mention expressis verbis “icotecas” and “galapagos” to specifically differentiate between tortoises and turtles26. It has also been well documented that this population was repeatedly reinforced using tortoises from Morocco throughout the 20th century, as illustrated by some early photographs that document the history of this National Park (Fig. 2). In particular, Valverde27 estimated that less than 100 Moroccan spur-thighed tortoises were introduced into the reserve between 1948 and 1954. The single genetic study, which addressed the origin of T. graeca in Doñana, included four samples from this population and was based exclusively on mitochondrial data. It found a shared lineage for Doñana, southeastern Spain, Majorca, east Morocco and northern Algeria23. This lineage is currently identified with the subspecies T. g. graeca 20. Based on this finding, the hypothesis that T. graeca arrived in the Iberian Peninsula before the Strait of Gibraltar opened could be rejected, which agrees with the fact that no fossil record exists28.
By considering the geographical location of the Doñana population, its origin could be either explained by (i) introduction or, alternatively, (ii) by it representing a relic from a formerly continuous range across southern Spain29. Furthermore, as very few individuals from the Doñana population have been studied to date, the effects of recent introductions of tortoises from Morocco could have been overlooked. Apart from T. g. graeca, two other subspecies (T. g. marokkensis and T. g. soussensis; Fig. 1a) occur in Morocco, both with much wider distribution ranges than T. g. graeca. Thus their genetic signatures could be detectable when more tortoises from Doñana are studied.
To elucidate this situation in more detail, the specific aims of the present study were to: (i) explore the genetic impact that the recent species management could have had; (ii) characterize the origin of the spur-thighed tortoises in Doñana. To do so, we analyzed 85 spur-thighed tortoises from Doñana by sequencing a mitochondrial DNA fragment (cyt b) and genotyping them at eleven microsatellite loci. These molecular data were analyzed together with the available cyt b sequences for all the North African lineages (Supplementary Table S1)20, 22, 24, 25, 30 and genotypes for all the populations potentially related with the tortoises from Doñana (southeastern Spain and North Africa; Fig. 1b).
Toward our first aim, we calculated the proportion of the different mitochondrial haplotypes in Doñana and inferred their lineage. Mismatch distributions31 served to identify the signatures of recent population reinforcements. We also calculated individual admixtures by a Bayesian clustering approach32. We estimated the biases between mtDNA and microsatellite data, and assessed sex-biased levels of introgression. Sex-biased introgression was not expected because Testudo graeca is known to have temperature-dependent sex determination, like the majority of chelonians33, 34, and because we used neutral genetic markers.
Toward our second aim, we expected genetic signatures for a recent origin, in particular low genetic variation and weak divergence from the source population, as previously found in other western European T. graeca populations22, 24, 25. We inferred these patterns from population genetic descriptors and Bayesian cluster analyses. Then we calculated mismatch distributions31, 35 and neutrality indices36, 37, which could reflect the signatures of demographic expansions using mitochondrial data. Microsatellites can also provide evidence for departures from stability, including Hardy-Weinberg and linkage disequilibrium, or heterozygosity excess at selectively neutral loci as a result of loss of low-frequency alleles38, 39. Apart from these analyses, we used Bayesian demographic approaches40, 41 to reconstruct the diversification history of the Doñana population prior to the reinforcements with Moroccan tortoises.
Results
The Testudo graeca lineages from Doñana
The 85 cyt b sequences from Doñana corresponded to three different haplotypes (Supplementary Table S1). Two abundant unique haplotypes were recorded for the first time. Both corresponded to the T. g. graeca lineage (B1.35 and B1.36, shared by 57 and 19 of the individuals, respectively; GenBank accession numbers: LT838800 and LT838801, respectively). Another haplotype was present in nine males, which represented the lineage of T. g. marokkensis (B2.1). This haplotype is widely distributed in northern Morocco (Fig. 3). A parsimony network connected all the haplotypes using the 95% criterion. The two unique haplotypes from Doñana differed by two mutational steps, and by one step from the B1.2 haplotype that occurs in North Africa, southeastern Spain and Majorca (Fig. 3 and Supplementary Table S1). A mismatch distribution also supported the co-existence of different lineages in Doñana, as indicated by two markedly different peaks (rg = 0.67; p = 0.99; Fig. 4a). The lesser frequency of the T. g. marokkensis haplotypes supported the notion that this lineage was introduced into Doñana during the twentieth-century population reinforcements.
Subsequent analyses using microsatellites considered T. g. graeca from North Africa (n = 99) and southeastern Spain (n = 102), as well as T. g. marokkensis from Morocco (n = 19). These markers not only corroborated the existence of two different lineages in Doñana, but also the admixture between them. One locus (GP81) proved monomorphic. For the remaining ten, 110 alleles were found, ranging from two (Test76, Test96) to 28 alleles (GmuD51). No significant linkage disequilibrium was revealed for any of the 55 pair-wise locus combinations, and none locus showed at sampling sites significant homozygote excess indicating the presence of possible null alleles. After these initial analyses, we used a cluster number (K) of four in Structure analyses32, with prior group information that corresponded to (1) Doñana, (2) T. g. graeca from southeastern Spain, (3) T. g. graeca from North Africa (Algeria and Morocco) and (4) T. g. marokkensis from North Africa (Morocco). The analysis supported the differentiation of these four groups. The Doñana samples were assigned mostly to a specific cluster. Of the 85 analyzed samples, 14 males showed consistent assignment values over 0.125 to the North African T. g. marokkensis group that were, therefore, identified as having mixed ancestries of T. g. graeca and T. g. marokkensis. These individuals corresponded to eight of the nine tortoises that harbored the mitochondrial haplotypes of T. g. marokkensis. Not one single female with mixed ancestries was found, and no sample showed consistent mixed ancestry with any other group (Fig. 5). The 14 samples with mixed ancestries from Doñana and the T. g. marokkensis samples from Morocco were removed from subsequent analyses to explore the origin of the original T. g. graeca Doñana population (i.e. prior to the population reinforcements).
Genetic differentiation of T. g. graeca in Doñana
Using only the pure T. g. graeca from Doñana, Structure analyses without prior population information revealed that K = 3 was the optimal cluster number, although K = 2 also yielded a high ΔK value (K = 2: ΔK = 171.8; K = 3: ΔK = 240.58; K = 4: ΔK = 4.59; Supplementary Fig. S1). Different runs were consistent and showed moderate levels of admixture among clusters. For K = 2, all the North African individuals were placed in one cluster and all the Spanish tortoises in the second. Higher resolution was achieved for K = 3 with the individuals from the two Iberian populations divided into two clusters (Doñana and southeastern Spain; Fig. 6). In the latter, 93% of the tortoises from Doñana, 74.5% of those from southeastern Spain, and 95% of the tortoises from North Africa were assigned to their respective clusters with membership probabilities of >0.9.
Pair-wise F ST values among the sampling sites were generally significant (Supplementary Table S2). The weakest differentiation (F ST = 0.002) was found between the sites Guetourfa (GU) and Messad (MS) in Algeria, while the highest value (F ST = 0.37) was observed between a site in the northern Molouya Valley (MO) in Morocco and another site in the north of the southeastern Spanish range (Crisoleja; CR). The lowest F ST values (0.079 for both) for the Doñana population were found for two southeastern Spanish sites: south of the Bas Mountain (SB) and Chinas (CH).
Hierarchical AMOVAs supported the two-cluster solution of Structure. A more pronounced differentiation was revealed for K = 2, with 16.74% of global genetic variance between the Iberian and North African T. g. graeca, 7.73% among the sampling sites within clusters, and 11.61% that corresponded to the variance among individuals within sites (F RT = 0.17; F SR = 0.09; F ST = 0.25; F IS = 0.15; F IT = 0.36; all p = 0.001). For K = 3, 12.99% of global variance occurred among the two Iberian and the North African populations, 7.4% among sampling sites, and 12.25% among individuals (F RT = 0.13; F SR = 0.09; F ST = 0.2; F IS = 0.15; F IT = 0.32; all p = 0.001).
Genetic diversity and inference of recent demographic processes for T. g. graeca in Doñana
The T. g. graeca tortoises from Doñana showed lower levels of genetic diversity for cyt b than the North African and southeastern Spanish representatives of the same genetic lineage, but higher levels compared to those from other recently introduced western European populations (Table 1). The mismatch distribution of the T. g. graeca haplotypes did not reveal any signs of recent demographic expansion (rg = 0.67; p = 0.89; Fig. 4b). In the same vein, the Tajima’s D and Fu’s F S values yielded positive values, but neither revealed a significant deviation from zero (D = 1.40; p = 0.92; F S = 3.34; p = 0.95).
The microsatellite data also supported a stable population scenario for Doñana. The allelic diversities between the two Iberian populations were similar, but were slightly lower than for the North African T. g. graeca. The expected heterozygosity levels were similar across the three populations and all the F IS values came close to zero (which indicates a balance between the expected and observed heterozygosities; Table 2). After Bonferroni correction, all the loci corresponded to Hardy-Weinberg equilibrium expectations for Doñana. The heterozygosity tests did not reveal signatures of heterozygosity excess or deficiency for this population (SMM: p = 0.46; TPM: p = 0.95). Consequently, no recent demographic bottlenecks or expansions were inferred for the Doñana population.
Historical divergence and past demographic inference of T. g. graeca in Doñana
The historical order of divergence inferred from Bananas 41 matched the genetic differentiation revealed by Structure (Fig. 7a). The binary rooted tree topology assigned the populations from North Africa to a single group, which identified the populations from Doñana and southeastern Spain as sister populations.
In order to set a date for the origin of T. g. graeca in Doñana, we investigated its demographic history using IMa240. For these analyses, we considered the Doñana (DO) and Chinas (CH) sites as representative for Doñana and southeastern Spain, which were characterized by the lowest pair-wise F ST values. All the effective sample size (ESS) values for the estimated parameters exceeded 100 at the different replicates, which indicates sufficient Markov chain Monte Carlo mixing across the parameter space42. The marginal posterior probability densities of the demographic parameters did not differ among replicates. The split between Doñana and southeastern Spain was estimated to be 947.5 years, with 95% highest posterior density intervals (HPDI) of 200.8–2,599.5. The effective Doñana population size was estimated at 136.3 individuals (95% HPDI: 40.5–323.9) and at 2,816.8 individuals (95% HPDI: 1,048.25–5,322.8) for southeastern Spain. The plots of the marginal posterior probability densities of the demography parameters are shown in Fig. 7b,c.
Tests for sex-biased introgression and cytonuclear discordance of the spur-thighed tortoises from Doñana
The Structure results for the microsatellite data indicated signatures of admixture between T. g. graeca and T. g. marokkensis in 16.6% of our samples. We did not find signatures of cytonuclear discordance as 57.14% of these samples harbored the mitochondrial haplotype of the T. g. marokkensis lineage.
However, and unexpectedly, not a single female from Doñana was identified as having a mixed ancestry according to the microsatellite data, and no female tortoise harbored the allochthonous T. g. marokkensis haplotype. These sex-biased distributions were significant according to the chi-square tests (for mtDNA: χ2 = 8.05; df = 3; p = 0.045; for microsatellites = χ2 = 11.93; df = 3; p = 0.008). Based on our sampling, the sex ratio of the population deviated from parity, and males were 2.36-fold more abundant than females (N females = 25; N males = 59; N subadults = 1; p < 0.001). After removing the 15 males with mixed genotypes or T. g. marokkensis haplotype (14 from Structure and one from mtDNA analyses), the population’s sex ratio remained skewed (p = 0.03).
Discussion
In the present study, we provide evidence for introgression signatures between two recently diverged lineages of spur-thighed tortoises that co-exist in Doñana as a result of introductions and historical management. In the Doñana population we found a major proportion of haplotypes and genotypes that represented T. g. graeca, which undoubtedly constitute the original genetic lineage in Doñana. However, approximately 15% of tortoises had a genetic impact from T. g. marokkensis, the subspecies that was introduced in the past century as part of population reinforcements.
Our results reveal no population expansion or bottleneck signatures, but indicate demographic stability for the original population, but also corroborated a recent origin of the Doñana population. Compared to the southeastern Spanish T. g. graeca, a weak genetic differentiation appears, which suggests this population as the source of the tortoise population in Doñana. The southeastern Spanish T. g. graeca originated, in turn, from North African founders in prehistorical times, and the role that humans could have played as potential dispersers is still a matter of debate20, 25, 43, 44.
According to our results, the populations from Doñana and southeastern Spain are very closely related. The low genetic differentiation between them and their young historical divergence definitively rule out the possibility that both represent vestiges of an ancient wider distribution in the Iberian Peninsula, as previously suggested29. By considering the wider distribution range and the effective population size of the species in southeastern Spain, our results instead suggest that T. g. graeca was introduced into Doñana with founder individuals from southeastern Spain. In Doñana there are several examples of introduced species, including the recent introduction of fish (e.g. Gambusia holbrooki or Cyprinius carpio 45) or game species (e.g. Dama dama 46). Other examples include ancient introductions of chameleons (Chamaeleo chamaeleon) or genets (Genetta genetta), which may be related to Phoenicians or Arabs (some 3,100–500 years ago)47, 48. Further historical and archaeological research is needed to gain a better understanding of these and other historical introductions across the Strait of Gibraltar because based on molecular data analyses alone, minor inaccuracies in calibration could dramatically change date estimates.
The present study also provides firm evidence for the intrinsic mechanisms that cause a male excess in the Doñana population as a result of exogamy. This population’s sex ratio has been traditionally described as significantly skewed toward males, with ratios of up to 3.32:1 adult males:adult females26. In contrast, the majority of the T. graeca populations show balanced sex ratios [e.g. refs 49,50,51], and very few studies have reported deviations from a balanced sex ratio for the genus Testudo. In these cases, sex ratios skewed toward males correlate positively with high population densities and aggressive courtship behavior (T. hermanni in Greece52, 53) or with an over-collection of larger females (T. graeca in Morocco49). In Doñana, highly admixed genotypes occur exclusively in males. However, one male with a weak admixed genotype harbors a T. g. marokkensis haplotype (Fig. 5), which suggests that although admixed females occur, they are rare. Sex-biased introgression is often explained by Haldane’s rule13, which requires heterogametic sex chromosomes. This pattern is not expected in T. graeca because it has no sex chromosomes and temperature-dependent sex determination32, 33. Alternatively, other mechanisms may explain the observed lack of admixed females: first, the impact of exogamy on the sex determination process; second, sex-biased selection in favor or against admixed individuals. In the former, changes in threshold temperatures to determine sex33 could bias the determination of clutches of hybrids to favor males. Previous studies have examined T. graeca reproduction in Doñana and have found highly variable hatching success rates, incubation temperatures and incubation periods between different nesting periods and years54, 55. Part of this variation could be caused by the admixture of both lineages, which would render this population as being particularly interesting for studying the mechanisms that underlie sex determination in T. graeca. The epigenetic mechanisms that govern temperature-dependent sex determination are largely unknown56, 57, and exogamy may affect it. Rarity of admixed females can be explained alternatively by lower female survival due to either genomic constraints (endogenous selection) or their poor adaptation to local environmental conditions (exogenous selection)58. In a similar vein, better fitness of hybrid males compared to pure males, known as the heterosis effect or hybrid vigor59, would increase the genetic signatures of the admixed males in the population. Long-term demographic studies in the Doñana have found that adult survival rates are high, constant and similar between males and females, while recruitment and juvenile survivals are variable. These studies have also observed a negative population trend over the 1980–1985 period, while a positive trend has been found for 1992–199526, 60. Continuous population monitoring, combined with parallel genetic analyses, would help to elucidate whether exogamy impacts adult or early life stages, and if is related to the negative population dynamics reported after population reinforcements.
Thus the observed male-biased sex ratio is a prelude of future and stable cytonuclear discordance in which T. g. graeca haplotypes will be fixed, but T. g. marokkensis alleles will persist across nuclear genomes. Similar patterns for the species have been reported in both eastern Algeria, in the secondary contact zone of T. g. graeca and T. g. nabeulensis 25; and Transcaucasia, in the contact zone of T. g. armeniaca, T. g. buxtoni and T. g. ibera 61, 62. To date, these patterns have been explained by the low dispersal capabilities of females61, 62. However, the present study suggests that exogamy could promote sex ratio deviations, which would play a role in maintaining the genetic integrity of genetic lineages across the species’ range1.
Finally despite its anthropogenic origin, the Doñana population constitutes a valuable empirical system for studying the evolutionary mechanisms that underlie introgression. Our findings also underline the need to carefully manage wild tortoise populations. Unexpected consequences of inadequate introductions or translocations may threaten the viability of wild populations, especially if they have low densities.
Methods
Sampling and available sequences considered
Two field surveys were conducted in autumn 2012 and spring 2013. Two sites were studied in the Doñana National Park, which were 12 km apart. Sixty-five tortoises were studied from “El Puntal” (DO) and four from “Marismillas” (DO2). Adult tortoises were sexed according to external secondary characters, while one subadult was treated as not sexable26. Sixteen additional samples with known sex were provided by the Doñana Biological Station, which were used in a previous study to address multiple paternity and sperm storage in T. graeca 63. In summary, 85 samples from Doñana were analyzed.
In order to represent the probable sources of T. graeca in Doñana, our previously accumulated data set of 396 sequences from the western Mediterranean clade was used. The considered mitochondrial DNA fragment comprised the complete mitochondrial cytochrome b gene and 25 bp of the adjacent tRNA-Thr gene (1164 bp). This data set corresponded to 250 samples from North Africa, 97 from southeastern Spain, 36 from Sardinia, 12 from Majorca and one from Sicily, which were analyzed for previous studies (Fig. 1a; Supplementary Table S1)20, 22, 24, 25, 30. After identifying the lineages present in the Doñana, the Dresden Zoology Museum and the Miguel Hernández University collections provided additional samples that are representative of their potential sources for their genotyping with microsatellites (Fig. 1b). In addition to the Doñana samples, the final data set of genotypes (n = 305) included samples from six southeastern Spanish sites (subspecies T. g. graeca; n = 102) and from 26 sites from Morocco and Algeria (subspecies T. g. graeca and T. g. marokkensis; n = 99 and n = 19, respectively). No samples from Majorca were considered because the origin of this population is very recent and is related directly with Algerian populations22.
Laboratory procedures
Total genomic DNA was isolated from an ethanol-preserved blood samples by the salt-extraction protocol64. Cyt b sequences were amplified by polymerase chain reaction (PCR) and subsequently sequenced. For this purpose, previously described primers (CytbG65, mt-c-For2, mt-f-na3 and mt-E-Rev266) and PCR conditions66 were used. PCR products were purified by precipitation in 1 volume of PCR product (30 μL), 1 volume of 4 M NH4Ac (30 μL) and 12 volumes of EtOH (100%; 360 μL). DNA was pelleted by centrifugation, washed with 70% ethanol and subsequently dissolved in 20 μL of H2O. PCR products were sequenced in an ABI 3130xl sequencer (Applied Biosystems, Foster City, CA, USA).
Tortoises were genotyped using eleven fluorescent-labeled microsatellites43: GmuB08HEX, GmuD16NED, GmuD51PET (designed for Glyptemys muhlenbergii 67), GP55HEX, GP61HEX, GP816-FAM, GP966-FAM (designed for Gopherus polyphemus 68), Test106-FAM, Test216-FAM, Test716-FAM and Test76HEX (designed for Testudo hermanni 69). Microsatellite loci were individually PCR-amplified in a final volume of 20 μl using 1 unit of Taq polymerase (Biotaq, Ecogen, Spain) with the buffer recommended by the supplier and a final concentration of 1.6 mM MgCl2, 0.2 mM of each dNTP (dNTPs Mix, Ecogen, Spain), 0.4 μM of each primer, and approximately 10–40 ng of total DNA. PCR products were analyzed by gel electrophoresis in two separate runs (first run: Test71, Test10, Gp61, Gp81, Gp55; second run: Test76, Test96, Test21, GmuD16, GmuB08, GmuD51) in an ABI 3130xl sequencer and with the GeneMapper software (Applied Biosystems).
Assessing introgression due to recent population reinforcements in Doñana
The mitochondrial cyt b sequences were aligned with ClustalW using the default parameters as implemented in Mega 4.070. The identified new haplotypes were deposited in GenBank and named following the nomenclature of previous works20, 22, 24, 25. The mitochondrial DNA variation of the Doñana population was characterized by constructing a parsimony network because this approach is especially effective when representing phylogenetic relationships within closely related taxa71. Networks were built using Tcs 72. To detect the genetic signatures of introgression in Doñana population, we calculated and plotted a mismatch distribution using cyt b sequences with Dnasp 73. With an infinite sites model, stable populations would be depicted by stochasticity as ragged and multimodal distributions, whereas introgression processes would usually be reflected in bimodal distributions. The raggedness index rg was considered to test the fit of the observed data to the expected distribution74. The impact of the recent population reinforcements in the Doñana population was quantified by calculating the percentages of the individuals that pertained to different lineages. It was assumed that the lineage with a much smaller proportion was that introduced in the last century. To detect any biases in the percentage of introduced haplotypes between sexes, Chi-square tests were used.
Microsatellite loci were examined for null alleles using Micro-Checker 2.2.375. Genotypic linkage disequilibrium between pairs of microsatellite loci was tested by the Markov chain Monte Carlo (MCMC) method implemented in Genepop 4.076. Tortoise ancestry was examined by the Bayesian clustering approach of Structure 2.3.432. K = 4 was fixed by considering four groups: Doñana, southeast Spanish T. g. graeca, North African T. g. graeca and North African T. g. marokkensis. These groups were used as priors for the ‘locprior’ option77. Burn-in was set at 105 and the number of iterations was set at 106. The individuals from Doñana with cluster membership proportions of >0.125 to any of the other groups were treated as having mixed ancestries (corresponding to a single great-grandparent over three generations). To evaluate the convergence of cluster memberships across individuals, 10 replicates of the analysis were run. As mentioned before for the mitochondrial data, Chi-square tests for independence were used to detect any biases in the percentages of haplotypes between sexes.
Genetic diversity, differentiation and potential sources of T. graeca in Doñana prior to population reinforcements
Microsatellite data and the Structure software also served to explore the genetic differences in the Doñana population and its potential source sites in southeastern Spain and North Africa (eastern Morocco and northern Algeria). For this purpose, Structure was run with and without prior population information from K = 1 to K = 25 (total amount of collection sites, plus one) using the admixture model and the ‘correlated allele frequencies’ option. Burn-in was set at 104 and the number of iterations was set at 105. To evaluate convergence and to estimate optimal genetic clustering, 10 replicates were run for each K value. The number of populations that best fitted the data set was defined by the ΔK method78, as implemented in Structure Harvester 79. The assignment of genotypes to clusters was visualized using Distruct 80.
We calculated the haplotype (h) and nucleotide (π) estimates of diversity from the mitochondrial cyt b sequences using Dnasp 73. Genalex 6.5.0181 was used to calculate the descriptive statistics of genetic diversity from the microsatellite data in the genetic groups obtained from clustering: average number of alleles (N A), mean number of effective alleles (A E), average observed heterozygosity (H O), unbiased average expected heterozygosity (uH E) and inbreeding estimates (F IS), and to test for Hardy-Weinberg deviations. This software was also used to calculate the pair-wise F ST values as measures of genetic differentiation among the considered sites, and hierarchical AMOVAs were run to determine how genetic variation was distributed among clusters, sites and individuals. In the calculation of the pair-wise F ST values, only the sampling sites with at least five genotyped individuals were considered.
Historical divergence and the past demographic history of T. graeca in Doñana prior to population reinforcements
The mismatch distributions from the cyt b sequences served to detect the genetic signatures of past demographic expansions in the Doñana population. Mismatch distributions were calculated and plotted with Dnasp, and tested using the raggedness index rg. In the mismatch distributions, expansion processes were expected as unimodal distributions. In addition to mismatch distributions, the F S 36 and D 37 statistics were calculated in Dnasp using the cyt b data. Significantly negative values indicate excess rare alleles, which was expected according to recent demographic expansions. Contrarily, positive values indicate the elimination of rare alleles by recent bottleneck events. These analyses were carried out according to selective neutrality and population equilibrium assumptions, and their significance was assessed with 1000 coalescent simulations.
The demographic and diversification history of the Doñana population was also inferred using microsatellite data. Bottleneck 1.2.02 39 served to test current excess or defect of heterozygosity, which was likely to arise from recent population size reductions or expansions. The step-wise mutation model (SMM) and the two-phase mutation model (TPM) with 78% step-wise mutations (p s ), 22% multistep mutations (p g ), and a variance of 3.1 among multiple steps, as recommended for microsatellite data82, were used. Then the historical divergence of the sampling sites was inferred from the pair-wise analysis in the Bananas software41. This approach allowed us to identify the sister population of the Doñana population. Bananas uses Bayesian inference, and the Adaptive Metropolis Algorithm as an adaptive Markov chain Monte Carlo sampler, to infer Neighbor Joining trees in which branch lengths represent mean diversification times τ (τ = t/N; being t the number of generations and N the population size) among populations. The model is based on a diffusion approach to the transition density of a neutral infinite alleles Wright-Fisher model. The number of iterations was set at 105, burn-in at 104, and thinning at 10. Only those sampling sites with at least three genotyped samples were considered in this analysis. Finally, the split between the Doñana population and its sister population was estimated using the IMa2 software40. Four independent replicates were performed to simultaneously estimate divergence times (t) and effective population sizes (Ne). The Step-wise Mutation Model was assumed and a run length of 2 × 106 steps was computed after a burn-in of 105 steps. The upper bounds of the prior distributions for parameters were set at t = 0.1; qa = 10 (effective ancestral population size); q1 and q2 = 0.1 (effective sizes of the Doñana and southeastern Spain populations, respectively) and m1 and m2 = 0 (the migration rates between both populations). For scaling the demographic parameters to individuals and years, a generation time of 17.72 years was used, which corresponded to the mean estimates for the Doñana population60 and the mutation rates of loci (μ) were between 3 × 10−9 and 2 × 10−7. The mutation rates were calculated from the number of alleles at each locus and the divergence time of the T. g. graeca lineage22. Parameter convergence and chain mixing were assessed by ESS above 100, and the consistency of results across replicated runs.
Ethic statements
Sampling of spur-thighed tortoises in Doñana was conducted under the authorization and following the protocols approved by the Junta de Andalucía (permit nos.: 21/2012 and 2013/15). The sampling protocols were approved by the Ethic Committee of the Miguel Hernández University (DBA-AGC-001-12), in accordance with the approved guidelines.
References
Petit, R. J. & Excoffier, L. Gene flow and species delimitation. Trends Ecol. Evol. 24, 386–393 (2009).
Abbott, R. et al. Hybridization and speciation. J. Evol. Biol. 26, 229–246 (2013).
Harrison, R. G. & Larson, E. L. Hybridization, Introgression, and the Nature of Species Boundaries. J. Hered. 105, 795–809 (2014).
Broyles, S. B. Hybrid bridges to gene flow: a case study in milkweeds (Asclepias). Evolution 56, 1943 (2002).
Borge, T., Lindroos, K., Nádvorník, P., Syvänen, A.-C. & Saetre, G.-P. Amount of introgression in flycatcher hybrid zones reflects regional differences in pre and post-zygotic barriers to gene exchange: Introgression in flycatchers. J. Evol. Biol. 18, 1416–1424 (2005).
Steeves, T. E., Maloney, R. F., Hale, M. L., Tylianakis, J. M. & Gemmell, N. J. Genetic analyses reveal hybridization but no hybrid swarm in one of the world’s rarest birds. Mol. Ecol. 19, 5090–5100 (2010).
Pons, J.-M., Sonsthagen, S., Dove, C. & Crochet, P.-A. Extensive mitochondrial introgression in North American Great Black-backed Gulls (Larus marinus) from the American Herring Gull (Larus smithsonianus) with little nuclear DNA impact. Heredity 112, 226–239 (2014).
Zieliński, P. et al. No evidence for nuclear introgression despite complete mtDNA replacement in the Carpathian newt (Lissotriton montandoni). Mol. Ecol. 22, 1884–1903 (2013).
Toews, D. P. L. & Brelsford, A. The biogeography of mitochondrial and nuclear discordance in animals. Mol. Ecol. 21, 3907–3930 (2012).
Hudson, R. R. & Turelli, M. Stochasticity overrules the ‘three-times rule’: genetic drift, genetic draft, and coalescence times for nuclear loci versus mitochondrial DNA. Evolution 57, 182 (2003).
Zink, R. M. & Barrowclough, G. F. Mitochondrial DNA under siege in avian phylogeography: nuDNA vs. mtDNA in phylogeography. Mol. Ecol. 17, 2107–2121 (2008).
Funk, D. J. & Omland, K. E. Species-Level Paraphyly and Polyphyly: Frequency, Causes, and Consequences, with Insights from Animal Mitochondrial DNA. Annu. Rev. Ecol. Evol. Syst. 34, 397–423 (2003).
Haldane, J. B. S. Sex ratio and unisexual sterility in hybrid animals. J. Genet. 12, 101–109 (1922).
McDonald, D. B., Parchman, T. L., Bower, M. R., Hubert, W. A. & Rahel, F. J. An introduced and a native vertebrate hybridize to form a genetic bridge to a second native species. Proc. Natl. Acad. Sci. 105, 10837–10842 (2008).
Feulner, P. G. D. et al. Introgression and the fate of domesticated genes in a wild mammal population. Mol. Ecol. 22, 4210–4221 (2013).
Goedbloed, D. J. et al. Genome-wide single nucleotide polymorphism analysis reveals recent genetic introgression from domestic pigs into Northwest European wild boar populations. Mol. Ecol. 22, 856–866 (2013).
Yang, X. et al. Transgenes for insect resistance reduce herbivory and enhance fecundity in advanced generations of crop-weed hybrids of rice: Enhanced fecundity in transgenic weedy rice. Evol. Appl. 4, 672–684 (2011).
Iverson, J. B. A Revised Checklist with Distribution Maps of the Turtles of the World (Privately published, 1992).
Buskirk, J. R., Keller, C. & Andreu, A. C. Testudo graeca (Linnaeus, 1758) – Maurische Landschildkröte in Handbuch der Reptilien und Amphibien Europas, Schildkröten (Testudines) I (ed. Fritz, U.) 125–178 (Aula, 2001).
Fritz, U. et al. Mitochondrial phylogeography of Testudo graeca in the Western Mediterranean: Old complex divergence in North Africa and recent arrival in Europe. Amphib.-Reptil. 30, 63–80 (2009).
Anadón, J. D. et al. Individualistic response to past climate changes: niche differentiation promotes diverging Quaternary range dynamics in the subspecies of Testudo graeca. Ecography 38, 956–966 (2015).
Graciá, E. et al. Expansion after expansion: dissecting the phylogeography of the widely distributed spur-thighed tortoise, Testudo graeca (Testudines: Testudinidae). Biol. J. Linn. Soc. (in press), doi:10.1093/biolinnean/blx007.
Alvarez, Y. et al. Mitochondrial DNA haplotyping of Testudo graeca on both continental sides of the Straits of Gibraltar. J. Hered. 91, 39–41 (2000).
Vamberger, M., Corti, C., Stuckas, H. & Fritz, U. Is the imperilled spur-thighed tortoise (Testudo graeca) native in Sardinia? Implications from population genetics and for conservation. Amphib.-Reptil. 32, 9–25 (2011).
Graciá, E. et al. The uncertainty of Late Pleistocene range expansions in the western Mediterranean: a case study of the colonization of south-eastern Spain by the spur-thighed tortoise, Testudo graeca. J. Biogeogr. 40, 323–334 (2013).
Andreu, A. C., Díaz-Paniagua, C. & Keller, C. La tortuga mora (Testudo graeca L.) en Doñana. (Asociación Herpetológica Española, 2000).
Valverde, J. A. Estructura de una comunidad mediterranea de vertebrados terrestres. (Consejo Superior de Investigaciones Cientificas, 1984).
Morales Pérez, J. V. & Serra, A. S. The Quaternary fossil record of the genus Testudo in the Iberian Peninsula. Archaeological implications and diachronic distribution in the western Mediterranean. J. Archaeol. Sci. 36, 1152–1162 (2009).
Báez, J. C., Estrada, A., Torreblanca, D. & Real, R. Predicting the distribution of cryptic species: the case of the spur-thighed tortoise in Andalusia (southern Iberian Peninsula). Biodivers. Conserv. 21, 65–78 (2012).
Fritz, U. et al. Phenotypic plasticity leads to incongruence between morphology-based taxonomy and genetic differentiation in western Palaearctic tortoises (Testudo graeca complex; Testudines, Testudinidae). Amphib.-Reptil. 28, 97–121 (2007).
Rogers, A. R. & Harpending, H. Population growth makes waves in the distribution of pairwise genetic differences. Mol. Biol. Evol. 9, 552–569 (1992).
Pritchard, J. K., Stephens, M. & Donnelly, P. Inference of population structure using multilocus genotype data. Genetics 155, 945–959 (2000).
Pieau, C. Temperature and sex differentiation in embryos of two chelonians, Emys orbicularis L. and Testudo graeca L. in Intersexuality in the Animal Kingdom (ed. Reinboth, R.) 332–339 (Springer, 1975).
Valenzuela, N. & Lance, V. Temperature-dependent sex determination in vertebrates (Smithsonian Books, 2004).
Slatkin, M. & Hudson, R. R. Pairwise comparisons of mitochondrial DNA sequences in stable and exponentially growing populations. Genetics 129, 555–562 (1991).
Fu, Y. X. Statistical tests of neutrality of mutations against population growth, hitchhiking and background selection. Genetics 147, 915–925 (1997).
Tajima, F. The effect of change in population size on DNA polymorphism. Genetics 123, 597–601 (1989).
Wang, J. L., Caballero, A. & Hill, W. G. The effect of linkage disequilibrium and deviation from Hardy-Weinberg proportions on the changes in genetic variance with bottlenecking. Heredity 81, 174–186 (1998).
Cornuet, J. M. & Luikart, G. Description and power analysis of two tests for detecting recent population bottlenecks from allele frequency data. Genetics 144, 2001–2014 (1996).
Hey, J. Isolation with Migration Models for More Than Two Populations. Mol. Biol. Evol. 27, 905–920 (2010).
Siren, J., Marttinen, P. & Corander, J. Reconstructing Population Histories from Single Nucleotide Polymorphism Data. Mol. Biol. Evol. 28, 673–683 (2011).
Hey, J. & Nielsen, R. Multilocus methods for estimating population sizes, migration rates and divergence time, with applications to the divergence of Drosophila pseudoobscura and D. persimilis. Genetics 167, 747–760 (2004).
Salinas, M. et al. Genetic assessment, illegal trafficking and management of the Mediterranean spur-thighed tortoise in Southern Spain and Northern Africa. Conserv. Genet. 12, 1–13 (2011).
Graciá, E. et al. Genetic patterns of a range expansion: The spur-thighed tortoise Testudo graeca graeca in southeastern Spain. Amphib.-Reptil. 32, 49–61 (2011).
Sánchez-Polaina, F. J. & Fernández-Delgado, C. Historia de las introducciones de peces en la reserva natural de la laguna de Zóñar (Córdoba, España). Oxyura 9, 135–140 (1997).
Duque, A. El mito de Doñana (Fundación José Manuel Lara, 2004).
Gaubert, P. et al. Comparative phylogeography of two African carnivorans presumably introduced into Europe: disentangling natural versus human-mediated dispersal across the Strait of Gibraltar. J. Biogeogr. 38, 341–358 (2011).
Paulo, O. S., Pinto, I., Bruford, M. W., Jordan, W. C. & Nichols, R. A. The double origin of Iberian peninsular chameleons. Biol. J. Linn. Soc. 75, 1–7 (2002).
Znari, M., Germano, D. J. & Macé, J.-C. Growth and population structure of the Moorish Tortoise (Testudo graeca graeca) in west-central Morocco: possible effects of over-collecting for the tourist trade. J. Arid Environ. 62, 55–74 (2005).
Rouag, R. et al. Population structure and demography of an Algerian population of the Moorish tortoise, Testudo graeca. Anim. Biol. 57, 267–279 (2007).
Buică, G., Iosif, R. & Cogălniceanu, D. Demography and conservation of an isolated Spur-thighed tortoise Testudo graeca population in Dobrogea (Romania). Ecologia Balkanica 5, 97–106 (2013).
Hailey, A. Adult survival and recruitment and the explanation of an uneven sex ratio in a tortoise population. Can. J. Zool. 68, 547–555 (1990).
Hailey, A. & Willemsen, R. E. Population density and adult sex ratio of the tortoise Testudo hermanni in Greece: evidence for intrinsic population regulation. J. Zool. 251, 325–338 (2000).
Díaz-Paniagua, C., Keller, C. & Andreu, A. C. Hatching success, delay of emergence and hatchling biometry of the spur-thighed tortoise, Testudo graeca, in south-western Spain. J. Zool. 243, 543–553 (1997).
Díaz-Paniagua, C., Andreu, A. C. & Keller, C. Effects of temperature on hatching success in field incubating nests of spur-thighed tortoises. Testudo graeca. Herpetol. J. 16, 249–257 (2006).
Matsumoto, Y. & Crews, D. Molecular mechanisms of temperature-dependent sex determination in the context of ecological developmental biology. Mol. Cell. Endocrinol. 354, 103–110 (2012).
Shen, Z.-G. & Wang, H.-P. Molecular players involved in temperature-dependent sex determination and sex differentiation in Teleost fish. Genet. Sel. Evol. 46, 26 (2014).
Moore, W. S. An evaluation of narrow hybrid zones in vertebrates. Q. Rev. Biol. 52, 263–277 (1977).
Shull, G. H. What is ‘heterosis’? Genetics 33, 439–446 (1948).
Díaz-Paniagua, C., Keller, C. & Andreu, A. C. Long-term demographic fluctuations of the spur-thighed tortoise Testudo graeca in SW Spain. Ecography 24, 707–721 (2001).
Mashkaryan, V. et al. Gene flow among deeply divergent mtDNA lineages of Testudo graeca (Linnaeus, 1758) in Transcaucasia. Amphib.-Reptil. 34, 337–351 (2013).
Mikulíček, P., Jandzik, D., Fritz, U., Schneider, C. & Široký, P. AFLP analysis shows high incongruence between genetic differentiation and morphology-based taxonomy in a widely distributed tortoise. Biol. J. Linn. Soc. 108, 151–160 (2013).
Roques, S., Díaz-Paniagua, C. & Andreu, A. C. Microsatellite markers reveal multiple paternity and sperm storage in the Mediterranean spur-thighed tortoise. Testudo graeca. Can. J. Zool. 82, 153–159 (2004).
Aljanabi, S. M. & Martínez, I. Universal and rapid salt-extraction of high quality genomic DNA for PCR-based techniques. Nucl. Acids Res. 25, 4692–4693 (1997).
Spinks, P. Q., Shaffer, H. B., Iverson, J. B. & McCord, W. P. Phylogenetic hypotheses for the turtle family Geoemydidae. Mol. Phylogenet. Evol. 32, 164–182 (2004).
Praschag, P., Hundsdörfer, A. K. & Fritz, U. Phylogeny and taxonomy of endangered South and South-east Asian freshwater turtles elucidated by mtDNA sequence variation (Testudines: Geoemydidae: Batagur, Callagur, Hardella, Kachuga, Pangshura). Zool. Scr. 36, 429–442 (2007).
King, T. L. & Julian, S. E. Conservation of microsatellite DNA flanking sequence across 13 Emydid genera assayed with novel bog turtle (Glyptemys muhlenbergii) loci. Conserv. Genet. 5, 719–725 (2004).
Schwartz, T. S., Osentoski, M., Lamb, T. & Karl, S. A. Microsatellite loci for the North American tortoises (genus Gopherus) and their applicability to other turtle species. Mol. Ecol. Notes 3, 283–286 (2003).
Forlani, A. et al. Identification and characterization of microsatellite markers in Hermann’s tortoise (Testudo hermanni, Testudinidae). Mol. Ecol. Notes 5, 228–230 (2005).
Tamura, K., Dudley, J., Nei, M. & Kumar, S. MEGA4: Molecular Evolutionary Genetics Analysis (MEGA) Software Version 4.0. Mol. Biol. Evol. 24, 1596–1599 (2007).
Posada, D. & Crandall, K. A. Intraspecific gene genealogies: trees grafting into networks. Trends Ecol. Evol. 16, 37–45 (2001).
Clement, M., Posada, D. & Crandall, K. A. TCS: a computer program to estimate gene genealogies. Mol. Ecol. 9, 1657–1659 (2000).
Librado, P. & Rozas, J. DNASP v5: a software for comprehensive analysis of DNA polymorphism data. Bioinformatics 25, 1451–1452 (2009).
Harpending, H. C. Signature of ancient population growth in a low-resolution mitochondrial DNA mismatch distribution. Hum. Biol. 66, 591–600 (1994).
Van Oosterhout, C., Hutchinson, W. F., Wills, D. P. M. & Shipley, P. MICRO-CHECKER: software for identifying and correcting genotyping errors in microsatellite data. Mol. Ecol. Notes 4, 535–538 (2004).
Rousset, F. GENEPOP’007: a complete re-implementation of the GENEPOP software for Windows and Linux. Mol. Ecol. Resour. 8, 103–106 (2008).
Hubisz, M. J., Falush, D., Stephens, M. & Pritchard, J. K. Inferring weak population structure with the assistance of sample group information. Mol. Ecol. Resour. 9, 1322–1332 (2009).
Evanno, G., Regnaut, S. & Goudet, J. Detecting the number of clusters of individuals using the software STRUCTURE: a simulation study. Mol. Ecol. 14, 2611–2620 (2005).
Earl, D. A. & vonHoldt, B. M. STRUCTURE HARVESTER: a website and program for visualizing STRUCTURE output and implementing the Evanno method. Conserv. Genet. Resour. 4, 359–361 (2012).
Rosenberg, N. A. DISTRUCT: a program for the graphical display of population structure. Mol. Ecol. Notes 4, 137–138 (2003).
Peakall, R. & Smouse, P. E. GENALEX 6: genetic analysis in Excel. Population genetic software for teaching and research. Mol. Ecol. Notes 6, 288–295 (2006).
Piry, S. BOTTLENECK: a computer program for detecting recent reductions in the effective size using allele frequency data. J. Hered. 90, 502–503 (1999).
Acknowledgements
We thank ICTS–RBD (Estación Biológica Doñana), the Monitoring Team on Natural Resources and Processes of the Estación Biológica Doñana, the “Doñana de Almonte” High School and volunteers for their help with fieldwork. Further thanks go to Severine Roques and Carmen Díaz-Paniagua, who provided samples for this study. Toni Sánchez-Zapata, Victoria Jiménez-Franco and Ana Sanz-Aguilar made helpful comments on an earlier manuscript version. Financial support was granted by the Spanish Ministry of Economy, Industry and Competitiveness and the European Regional Development Fund (Projects CGL2012-33536, CGL2015-64144; MINECO/FEDER, UE). E.G. was supported by the Valencian Regional Government with a postdoctoral grant (APOSTD/2015/048). Estación Biológica Doñana provided the photographs of Fig. 2 and the permissions for their publication. We thank the anonymous reviewers for their valuable comments.
Author information
Authors and Affiliations
Contributions
E.G., R.C.R.C., A.C.A., A.G. and F.B. conceived the initial ideas and collected samples; E.G. and U.F. processed samples; E.G. analyzed the data and led the writing. All the authors discussed the findings, and commented on and revised earlier manuscript versions.
Corresponding author
Ethics declarations
Competing Interests
The authors declare that they have no competing interests.
Additional information
Publisher's note: Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Electronic supplementary material
Rights and permissions
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 license, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons license 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 license, visit http://creativecommons.org/licenses/by/4.0/.
About this article
Cite this article
Graciá, E., Rodríguez-Caro, R.C., Andreu, A.C. et al. Human-mediated secondary contact of two tortoise lineages results in sex-biased introgression. Sci Rep 7, 4019 (2017). https://doi.org/10.1038/s41598-017-04208-4
Received:
Accepted:
Published:
DOI: https://doi.org/10.1038/s41598-017-04208-4
This article is cited by
-
Alternative vegetation trajectories through passive habitat rewilding: opposite effects for animal conservation
Landscape Ecology (2024)
-
Colonization routes uncovered in a widely introduced Mediterranean gecko, Tarentola mauritanica
Scientific Reports (2023)
-
Phylogeographic and phenotypic divergence between two subspecies of Testudo graeca (T. g. buxtoni and T. g. zarudnyi) across their contact zone in Iran
Scientific Reports (2022)
-
Mechanical and Acoustic Emission Characteristics of Coal at Temperature Impact
Natural Resources Research (2020)
Comments
By submitting a comment you agree to abide by our Terms and Community Guidelines. If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.