Folia 4/02-def

Folia Zool. – 51(4): 307–318 (2002) Genetic differentiation and population structure within Spanish
common frogs (Rana temporaria complex; Ranidae, Amphibia)
Michael VEITH1*, Miguel VENCES2, David R. VIEITES3, Sandra NIETO-ROMAN3 and Antonio PALANCA3 1 Zoologisches Institut der Universität Mainz, Saarstraße 21, D-55099 Mainz, Germany; e-mail: michael@oekologie.biologie.uni-mainz.de 2 Zoologisches Forschungsinstitut und Museum Alexander Koenig, Adenauerallee 160, 53113 Bonn, Germany; e-mail: m.vences@t-online.de 3 Laboratorio de Anatomía Animal, Facultade de Ciencias Biolóxicas, Universidade de Vigo, Apdo. 874, 36200 Vigo (PO), Spain; e-mail: apalanca@uvigo.es Received 2 February 2001; Accepted 23 April 2002 A b s t r a c t . Genetic differentiation of Rana temporaria from the Pyrenean and Cantabrianmountains in Spain was studied by means of allozyme electrophoresis. 24 loci were analysed in104 specimens from 15 populations: nine populations from the Pyrenean massif, fivepopulations from the area of the Cantabrian mountain chain (regions of Galicia, Asturias, andBasque Country), and one population from Germany. Three distinct clusters were distinguishedby phenetic analysis: (a) the Pyrenean samples and the single population from the BasqueCountry, (b) the populations from Galicia and Asturias) and (c) the German population.
Ordination (PCA) resulted in one principle component (PC1) that separated Cantabrian fromPyrenean populations, and in a second one (PC2) that separated the single German populationfrom the Iberian ones. PC1 indicated introgression that was corroborated by west-east clines inseveral alleles along the Cantabrian chain. The rather clear separation of the Cantabrian andPyrenean clusters (mean genetic distance 0.121) suggests that two genetically differentsubspecies of R. temporaria may be distinguished in Spain. The absence of fixed allelicdifferences between populations refutes recent hypotheses of the existence of syntopic siblingspecies within R. temporaria in Spain. Biogeographically, the Pyrenean and Cantabrianpopulations possibly originated in two separate colonisation events starting from differentglacial refuges. The strong morphological differentiation of Pyrenean R. temporaria populationsis not paralleled by genetic divergence, and may better be explained by ecological factors suchas climate, altitude and vegetation.
Key words: population genetics, taxonomy, allozymes, Spain, Rana temporaria
Introduction
The systematics of the Iberian brown frogs, subgenus Rana (Rana) according to D u b o i s(1992), has long been discussed. Especially the species affiliation of populations from thePyrenean mountain range has not been satisfactorily studied in the past as shown by therecent discovery and description of a new species, Rana pyrenaica Serra-Cobo, 1993, whichis well differentiated by adult and larval morphology, and ecology (S e r r a - C o b o 1993,S t r i j b o s c h 1996, V e n c e s et al. 1998b).
Currently, four brown frog species are recognised from the Iberian Peninsula. (1) Rana pyrenaica is, as far as is known, restricted to the Pyrenean mountain range (S e r r a -C o b o 1997). (2) Rana iberica Boulenger, 1879, an endemic, brook-dwelling speciesoccurs mainly in north-western Spain and northern Portugal (E s t e b a n 1997a). (3) Ranadalmatina Bonaparte, 1840, a mainly Central European species, occurs in a restricted area in the Basque Country and Navarra (G o s á 1997); records from Catalonia and possibly fromthe French Pyrenees are due to misidentification (see L l o r e n t e et al. 1995, D u b o i s1998). (4) Rana temporaria Linnaeus, 1758, is a species with a vast distribution areaincluding almost all European countries (G r o s s e n b a c h e r 1997); in Spain it isrestricted to a northern stretch largely corresponding to the Pyrenean and Cantabrianmountain ranges (E s t e b a n 1997b).
Geographic variation of Rana temporaria on the Iberian Peninsula and adjacent regions has so far not been sufficiently studied. However, present taxonomy indicates a remarkabledifferentiation. Populations from the Basses Alpes in France (referable to the taxon R. t.
honnorati) differ from German populations in tadpole morphology and mean nuclear DNAcontent (S p e r l i n g et al. 1996). In north-western Spain, populations attributed to thesubspecies R. t. parvipalmata (Seoanne, 1885) have a slightly different advertisement calland a reduced foot webbing (G a l á n 1989, V e n c e s 1992) as compared to Germanpopulations. They are allozymatically well differentiated from other Spanish and fromCentral European populations (A r a n o et al. 1993). The status of the taxon R. t.
canigonensis Boubée, 1833, from Mont Canigou in the French Pyrenees, is unsolved (seeD u b o i s 1983). Recently, P a l a n c a et al. (1995) defined morphotypes of brown frogsfrom the Spanish Pyrenees of Aragon; one of these morphotypes was named Ranaaragonensis Palanca Soler, Rodriguez Vieites et Suárez Martínez, 1995, unintentionallyconstituting a valid taxon description due to the lack of explicit statement that the name justreferred to a morphotype. A lectotype of R. aragonensis was later designated (V e n c e s etal. 1998a), but the status of the taxon remains uncertain.
The by now single study on allozyme variation in Iberian brown frogs (A r a n o et al. 1993) did not comprise Pyrenean R. temporaria populations. These authors were thusunable to clarify the status of the populations inhabiting this massif as well as their relationto R. t. parvipalmata and R. t. temporaria.
The aim of the present study is to test whether there exists more than one R. temporaria- like species in the western Pyrenees. This was (i) assumed as one possible explanation of theobserved morphological divergence among Pyrenean populations (V e n c e s et al. 1998b),and (ii) deduced from the co-existence of two seemingly separated taxa at localities in theAragonese Pyrenees (P a l a n c a et al. 1995). To test this hypothesis we analysed the geneticdifferentiation of the Pyrenean R. temporaria populations by means of allozyme studies.
Material and Methods
Specimens were collected by opportunistic day and night searching. They were sacrificedusing chlorobutanol. Femur muscle tissue and liver was removed from freshly dead specimensand frozen at -80°C for electrophoresis. Specimens were fixed in 5% formaldehyde or 95%ethanol, and stored in 70% ethanol. Vouchers were deposited in the collections of theZoologisches Forschungsinstitut und Museum Alexander Koenig (ZFMK), Bonn, and theMuséum National d’Histoire Naturelle (MNHN), Paris.
Specimens of Rana temporaria were collected at the following localities (Fig. 1) from West to East Iberian Peninsula (Spain), and in Germany. Galicia: (1) Serra da Capelada,province of La Coruña (CAP; 43°44’N/7°56’W; 5 specimens; no vouchers preserved); (2)Serra dos Ancares, province of Lugo, (ANC; 42°50’N/7°00’W; 4 specimens; MNHN1998.136-138, ZFMK 68854); Asturias: (3) Puerto de Somiedo, province of Oviedo (PSO; 43°11’N/6°17’W; 14 specimens; ZFMK 68361-68374); (4) Espina near Salas, Los Porcinos(ESP; 42°24’N/6°19’W; 5 specimens; ZFMK 68402-68405); (5) near Picos de Europa (PIC;43°17’N/4°56’W; 6 specimens; ZFMK 68379-68384); Basque Country: (6) Puerto deAltube, province of Álava (PAL; 42°19’N/2°52’W; 5 specimens; ZFMK 68393-68397);Aragón (Huesca province): (7) between Oza and Aguas Tuertas, (OAT; 42°51’N/0°40’W; 11 specimens; ZFMK 65399-65409); (8) Aguas Tuertas, (AGU; 42°49’N/0°35’W; 8 specimens; ZFMK 65410-65416, 65437); (9) upper Canal Roya valley, (CRO;42°47’N/0°30’W; 8 specimens; ZFMK 65419-65426); (10) Pico de Anayet, (ANA;42°46’N/0°26’W; 3 specimens; ZFMK 65427-65429); (11) between Formigal and Portalet,(FOR; 42°47’N/0°24’E; 2 specimens; ZFMK 65417-65418); (12) Respomuso, Circo dePiedrafita (RES; 42°49’ N/0°17’W; 8 specimens; ZFMK 65430-65436); (13) Ibones de laFacha (FAC; 42°48’N/0°15’W; 15 specimens; ZFMK 68347-68360); (14) Barranco Ordiso,Bujaruelo (BUJ; 42°43’N/0°9’W; 6 specimens; ZFMK 65439-65444); Germany: (15)environments of Bonn, (BON; 50°53’N/7°9’E; 4 specimens; MNHN 1998.135).
Fig. 1. Map showing sample locations. Circles, populations of the Cantabrian cluster; squares, populations of the
Pyrenean cluster.
Samples of Rana iberica (RIBE; Salas, Asturias, Spain; 42°25’N/6°16’W; 6 specimens; ZFMK 68875-78), Rana pyrenaica (RPYR; Zuriza, Aragón, Spain; 42°54’N/0°48’W; 2 specimens; ZFMK 65447-65449) and Rana macrocnemis (RMAC; Tavas, Turkey;37°42’N/29°03’E; 4 specimens; vouchers are preserved in the Musée National d’HistoireNaturelle, Paris, MNHN 2000.660-2000.663) were used for hierarchical outgroup rooting.
Pieces of muscle and liver were homogenised in Pgm buffer (H e b e r t & B e a t o n 1993). Electrophoresis was run on cellulose acetate (CA) plates from Helena Diagnostics,Texas. We used four different buffer systems for separation of allozymes (Table 1): Phosphatebuffer, pH 7.2 (PP 7.2); tris-maleic buffer, pH 7.0 (TM 7.0); tris-citric buffer, pH 7.2 (TC 7.2);tris-glycine buffer, pH 8.5 (TG 8.5). Twenty enzyme systems provided data on 24 presumptivegene loci (Table 1). Allozyme loci and alleles were numbered according to their electrophoreticmobility, either anodal or cathodal, with the fastest being 1 or a, respectively.
Allele frequencies and population genetic variability estimates (mean heterozygosity, average number of polymorphic loci and average number of alleles) were calculated for allsamples using G-STAT (S i g i s m u n d 1997). We tested for syntopic occurrence ofdifferent taxa by calculating deviations of observed genotype frequencies from ideal Hardy-Weinberg proportions χ2-Test; rare alleles were pooled to avoid expected genotypefrequencies below 1.0; G-STAT). Subsequently, we corrected within populations formultiple tests across polymorphic loci (sequential Bonferroni correction as outlined byR i c e 1989). Inbreeding parameters according to W r i g h t ’ s (1951) F-statistics werecalculated with G-STAT for ingroup samples with n ≥ 4 by the procedure described in Table 1. Enzyme systems, enzyme commission (E.C.) number, buffer systems and tissues used in electrophoresis.
NADP-dependent malate dehydrogenase (malic enzyme) dipeptidase with alanine-leucine as substrate tripeptidase with glycine-leucine-leucine as substrate dipeptidase with phenylalanine-proline as substrate W e i r & C o c k e r h a m (1984). Variances of these estimators were obtained byjackknifing populations. Inbreeding estimates deviating ±1.96 standard deviation (SD) fromzero were regarded as significant.
We used N e i ’ s (1972) standard genetic distance to build an UPGMA tree. Both calculations were performed with NTSYS (R o h l f 1990). 1000 bootstrap replicates(F e l s e n s t e i n 1985) were run using the subroutine SEQBOOT as implemented in PHYLIP3.5c (F e l s e n s t e i n 1993). Since UPGMA cluster analysis hardly allows for the detectionof intergraded populations (it forces all populations into a dichotomic branching pattern) anda priori information on intergradation was not available we used a principle componentanalysis (PCA) using the alleles as characters and their frequencies as states in order to detectpotentially intergraded populations relative to pure populations of the detected lineages.
In 24 studied loci, we identified 77 different alleles among the samples (Table 2). In theUPGMA phenogram (Fig. 2) based on N e i ’ s (1972) standard genetic distances as shown inTable 3, the Spanish populations were clearly separated into two geographic clusters. Oneincluded the Galician and Asturian samples (referred to subsequently as the Cantabrian cluster),and a second cluster composed of all Pyrenean samples and the single population from theBasque Country (Pyrenean cluster). The German sample was basal to the Iberian samples.
The PCA detected only two principle components (Eigenvalue > 1) that accounted for ca. 40% of the total variance. PC1 explained 23.6 % of the total variance and discriminatedamong the Spanish samples. The position of the Basque sample was intermediate betweenthe Galicia/Asturias and the Pyrenean samples in a two-dimensional plot of PC1 and PC2(Fig. 3). PC2 explained 16.8 % of the total variance, and clearly separated the Germansample from the Spanish samples (Fig. 3).
Fig. 2. UPGMA phenogram of all samples based on N e i ’ s (1972) standard genetic distances; bootstrap p-values
>50% for 1000 replicates are shown.
Fig. 3. Scatterplot of the first two principal components of genetic variance.
The mean genetic distance of D = 0.121 between the Pyrenean and the Cantabrian cluster (Tables 3 and 4) was not due to fixed allelic differences. The genetic distances to the Germansample were considerably higher (D = 0.161 and 0.197, respectively). Of the three outgroupspecies, R. pyrenaica showed the closest affinities to R. temporaria (D ranged from 0.212–0.379).
In only five populations one out of 6–11 polymorphic loci deviated from Hardy- Weinberg proportions at the 5% level (gpi in OAT, idh1 in PSO, pepB in FAC, pepD1 inRES, and pepD2 in PIC). However, after Bonferroni correction none deviated significantly.
Table 4. Mean ± S D of N e i ’ s (1972) genetic distances between clusters and taxa (minimum and maximum
distances are given in parentheses). The Pyrenean cluster of R. temporaria includes the populations from Euskadi
and Aragon; the Cantabrian cluster includes those from Galicia and Asturias (see Fig. 2).
R.temporaria
R.temporaria
R.temporaria
R.pyrenaica
R.iberica
R. temporaria
0.027 ± 0.011
R. temporaria
0.121 ± 0.032
0.059 ± 0.018
R. temporaria
0.161 ± 0.024
0.197 ± 0.021
R. pyrenaica
0.212 ± 0.013
0.279 ± 0.030
R. iberica
0.451 ± 0.012
0.500 ± 0.035
R. macrocnemis
0.531 ± 0.018
0.623 ± 0.046
Thirty percent of the total genetic variance of all samples was due to within population = 0.284) was distributed among populations (Table 5). In the Cantabrian cluster the situation was similar, whereas samples from thePyrenean cluster were much more homogeneous (only 44% of genetic variance distributedamong populations). The degree of population subdivision was also less pronounced (F 0.109) in the Pyrenean cluster than in the Cantabrian cluster (F Six loci (apk, ck, gldh, gapd, ldh2, tre) were monomorphic among all studied brown frog species (Table 2). Of the remaining loci only mdh was monomorphic among R. temporariasamples. None of the 17 loci polymorphic within R. temporaria was diagnostic for eitherpopulation or geographic cluster. Several alleles that were present in more than one populationof one cluster were absent or almost completely absent in the other: (a) allele a of 6pgd, alleled of 6pgd, allele c of gpi, allele b of me2 and allele c of mpi were characteristic for thePyrenean cluster; (b) allele c of ldh1, allele b of pepD2 and allele b of pgm2 were present in allor almost all populations of the Cantabrian cluster but almost entirely absent in all populationsof the Pyrenean cluster. The mean frequency of the latter three alleles decreased from West toEast along the Cantabrian chain (Fig. 4). In the Galician and Asturian populations, theircombined average frequency ranged from 0.55 to 0.4, followed by a steep decrease betweenthe easternmost Asturias population (PIC) and the single population from the Basque country(PAL). The latter was clearly grouped within the Pyrenean cluster (mean genetic distance toother populations of this cluster: 0.036). However, PAL also showed a low genetic distances toPIC which is the geographically nearest population of the Cantabrian cluster (D = 0.056).
Table 5. W e i r & C o c k e r h a m ’ s (1984) F-statistics averaged over 17 polymorphic loci of all R. temporaria
populations with n ≥ 4, and of subsamples from the Cantabrian and Pyrenean clusters. Standard deviations (SD)
were obtained by jackknifing samples. Populations FOR and ANA were excluded from the analysis due to their
low sample size.
n
F
SD(F )
F
SD(F )
F
SD(F )
Fig. 4. Changes in frequencies of “typical” R. t. parvipalmata alleles (ldh1-c, pepD2-b and pgm-b) from the
easternmost population (CAP) to population OAT in the West; we fitted a logistic regression model (p < 0.05) to
describe the decrease of average allele frequencies from East to West.
Discussion
Our data demonstrate the existence of two geographically separated genetic lineages ofR. temporaria in Iberia. The lack of fixed allelic differences between these clusterscontradicts the results of A r a n o et al. (1993). They found fixed allelic differences betweena R. t. parvipalmata and an Alava-Barcelona-Germany cluster at two loci, icd-2 and lcd-2,which should be homologous to our loci idh1 and ldh1. In our study, both loci showfrequency differences between the Cantabrian and the Pyrenean clusters, but no fixedalternative alleles. This may at least partially be explained by the small size of some samplesof A r a n o et al. (1993; only two specimens in the geographically intermediate Asturias andAlava samples) with a low chance to detect comparatively rare alleles. C l i n a l v a r i a t i o n i n a l l e l e f r e q u e n c i e s The lack of fixed allelic differences between the two clusters may account for either a shortseparation time or for gene flow following secondary contact. Indication for the latter may bethe decrease of mean frequency of R. t. parvipalmata alleles from West to East whichresembles a clinal pattern (discordant among alleles, but evident when averaging allelefrequencies over loci). No such cline exists for typical Pyrenean alleles.
In principle, several scenarios may account for this pattern: (1) an initial polymorphic but homogeneous population broke up into local populations; (2) an initial isolation-by-distance structure produced frequency clines at single loci; afterwards it broke up into localpopulations; (3) two populations evolved divergently in isolation and formed a cline aftersecondary contact; gene flow among local populations subsequently broke up again.
After the break up of a homogeneous polymorphic population into isolated subpopulations (scenario 1) drift would produce a geographically irregular pattern of allele frequencies, beingclinal neither at single loci nor on average. An initial isolation-by-distance pattern withsubsequent isolation of local populations (scenario 2) would well explain the irregularfrequency pattern of single alleles. Again there is no rationale to assume that on average a clinewould result since isolation-by-distance would produce non-parallel clines at different loci. We therefore prefer scenario 3. It well explains (i) clinal variation when averaging loci, since two populations that differentially evolved in isolation would always produce a clinewhen hybridising after secondary contact, and (ii) the irregular frequencies pattern at singleloci due to genetic drift after subsequent isolation. The high inbreeding coefficients (F ) indicate that the degree of isolation among populations is still high, even within geographicalclusters. Whether there exists a transition zone between the two clusters with ongoing geneflow remains open. If gene flow still does occur between clusters, it is likely to take placesomewhere between PIC and PAL.
Large parts of the montane areas of the Pyrenees and of the Cantabrian mountain chain which are currently densely populated by R. temporaria did not constitute suitableamphibian habitats during glaciation periods. The genetic differentiation of Spanish R.
temporaria populations as found in the present study could be explained by a scenario inwhich two separated groups of populations, one in Galicia/Northern Portugal (where refugesof deciduous forests existed during glaciation maxima; see B a r b a d i l l o et al. 1997),and one more to the east, remained in isolation during considerable time. At the end of theglaciations, the two populations, meanwhile genetically differentiated, came into secondarycontact. It may well be that since the Pleistocene such a process of range expansion andretreat may have occurred repeatedly, resulting in a geographical mosaic of allelefrequencies as is discussed for Spanish Salamandra salamandra (A l c o b e n d a s et al.
1994, 1996) and for many organisms in general (e.g., T a b e r l e t et al. 1998). However,the observed pattern in Spanish R. temporaria would not be in conflict with the muchsimpler scenario of gene flow through a single secondary contact and genetic driftsubsequently altering allele frequencies at random in isolation.
Rana pyrenaica, which is morphologically and ecologically rather similar to the (as far asknown allopatrically distributed) R. iberica, showed a much closer genetic affinity to R.
temporaria (Table 4). Its specific differentiation is for the first time corroborated on geneticgrounds, however, a detailed discussion of its relationships to other western palearctic brownfrogs will be given elsewhere.
The total absence of fixed alleles characterising any Spanish population or cluster together with the low genetic distance between the Cantabrian and the Pyrenean cluster (seeV e i t h 1996 for a review of species-specific genetic distances among Europeanamphibians) support A r a n o ’ s et al. (1993) conclusion that the Cantabrian and thePyrenean R. temporaria populations are differentiated at the subspecies level without anyfurther taxonomically relevant substructure. Therefore, the hypotheses of V e n c e s et al.
(1998a, b) that more than one species of the Rana temporaria complex may occur in thePyrenees can be rejected. We also reject the theory of P a l a n c a et al. (1995) whoassumed co-existence of two separated taxa at localities in the Aragonese Pyrenees (Circode Piedrafita), which corresponds to our locality RES. This and all other populations are in Hardy-Weinberg equilibrium, which is not expected when different taxa are pooled into onesample. In addition, ongoing morphological studies of the Circo de Piedrafita population,including more than 1000 specimens, failed to discriminate two separate morphs (M.
V e n c e s , pers. obs. 1998 and 1999). A specific status of the taxon R. aragonensis from theCirco de Piedrafita can thus be excluded.
On the other hand, it remains true that the different Pyrenean populations of Rana temporaria included in this study are markedly heterogeneous in their external morphology.
For example, the population from Valle de Bujaruelo (BUJ) is composed of very largespecimens (males 78–86 mm, females 77–100 mm SVL; N=2/3), whereas only a minorproportion of specimens of other populations reach a similar size (only 12 out of 828 malesfrom Respomuso (RES) reached a SVL >78 mm, and none exceeded 80 mm; V e n c e s etal. 1999). Many specimens from the high-altitude populations (e.g. RES) had a large numberof black dorsal markings (R i o b ó et al. 2000) which were absent in the lower-altitudepopulation PAL. Relative hindlimb length was also variable among populations (M.
V e n c e s pers. obs.), and within the Cantabrian cluster important differences are foundbetween the small, long-legged and poorly webbed CAP specimens and the larger PICspecimens, several of which have relatively short legs and more extensive webbing. Thisvariation may better be explained by ecological parameters. The CAP population lives close to the type locality of R. temporaria parvipalmata (Seoane, 1885). The Cantabrian cluster therefore corresponds to this taxon. In contrast, thePyrenean cluster may correspond to R. t. canigonensis Boubee, 1833. The fact that in thestudy of A r a n o et al. (1993) the easternmost included R. temporaria population from theMontseny massif in Catalonia clustered close to a population from Basque Country makes itlikely that the whole Pyrenees between Catalonia and Basque Country are inhabited bya genetically homogeneous group of populations. Consequently, the taxon aragonensis is tobe seen as junior synonym of R. t. canigonensis. Thanks are due to Pedro Galan R e g a l a d o (La Coruña) for his help in the field, and Alain Dubois (Paris) forimportant advice. Nicolas P a l a n c a (Vigo) helped during collection of specimens in Spanish high-altitudehabitats. Thanks to Dagmar K l e b s c h for laboratory assistance. We are indebted to the regional authorities ofGalicia, Asturias, Euskadi, and Aragon for the permit of collection of specimens.
L I T E R A T U R E
ALCOBENDAS, M., DOPAZO, H. & ALBERCH, P., 1994: Genetic structure and differentiation in Salamandra salamandra populations from the Northern Iberian Peninsula. In: Greven, H. & Thiesmeier, B. (eds), Biologyof Salamandra and Mertensiella. Deutsche Gesellschaft für Herpetologie und Terrarienkunde, Bonn: 7–23. ALCOBENDAS, M., DOPAZO, H. & ALBERCH, P., 1996: Geographic variation in allozymes of populations of Salamandra salamandra (Amphibia: Urodela) exhibiting distinct reproductive modes. J. Evol. Biol., 9: 83–102.
ARANO, B., ESTEBAN, M. & HERRERO, P., 1993: Evolutionary divergence of the Iberian brown frogs. Ann. Sci. Nat., Zool., 14: 49–57.
BARBADILLO, L. J., GARCÍA-PARÍS, M. & SANCHIZ, B., 1997: Orígenes y relaciones evolutivas de la herpetofauna ibérica. In: Pleguezuelos, J. M. (ed.), Distribución y biogeografía de los anfibios y reptiles enEspaña y Portugal. Granada, Universidad de Granada and AHE: 47–100.
DUBOIS, A., 1983: Notes sur les grenouilles brunes (Groupe de Rana temporaria Linné, 1758). II. Les Grenouilles du Mont Canigou (Pyrenees Orientales). Alytes, 2: 19–26.
DUBOIS, A., 1992: Notes sur la classification des Ranidae (Amphibiens Anoures). Bull. mens. Soc. linn. Lyon, DUBOIS, A., 1998: Mapping European amphibians and reptiles: collective inquiry and scientific methodology.
ESTEBAN, R., 1997a: Rana iberica. In: Pleguezuelos, J.M. (ed.), Distribución y biogeografía de los anfibios y reptiles en España y Portugal. Granada, Universidad de Granada and AHE: 161–163. ESTEBAN, R. 1997b: Rana temporaria. In: Pleguezuelos, J.M. (ed.), Distribución y biogeografía de los anfibios y reptiles en España y Portugal. Granada, Universidad de Granada and AHE: 169–171. FELSENSTEIN, J., 1985: Confidence limits on phylogenies: an approach using the bootstrap. Evolution, 39: 783–791.
FELSENSTEIN, J., 1993: PHYLIP, version 3.5. Washington.
GALÁN, P., 1985: Morfología y fenología del tritón palmeado, Triturus helveticus (Razoumowsky, 1789) en el noroeste de la Península Ibérica. Alytes (Spain), 3: 31–50.
GOSÁ, A., 1997: Rana dalmatina. In: Pleguezuelos, J.M. (ed.), Distribución y biogeografía de los anfibios y reptiles en España y Portugal. Granada, Universidad de Granada and AHE: 158–160.
GROSSENBACHER, K., 1997: Rana temporaria Linnaeus, 1758. In: Gasc, J. P., Cabela, A., Crnobrnja-Isailovic, J., Dolmen, D., Grossenbacher, K., Haffner, P., Lescure, J., Martens, H., Martínez-Rica, J. P., Maurin, H.,Oliveira, M. E., Sofianidou, T. S., Veith, M. & Zuiderwijk, A. (eds), Atlas of Amphibians and Reptiles inEurope. Societas Europaea Herpetologica and Muséum National d’Histoire Naturelle (IEGB/SPN), Paris:158–159.
HEBERT, P. D. N. & BEATON, M. J., 1993: Methodologies for allozyme analysis using cellulose acetate electrophoresis. Helena Laboratories, Beaumont.
LLORENTE, G., MONTORI, A., SANTOS, X. & CARRETERO, M. A., 1995: Atlas dels Amfibis i Rèptils de Catalunya i Andorra. Edicions El Brau, Barcelona. NEI, M., 1972: Genetic distance between populations. Amer. Nat., 106: 283–292.
PALANCA SOLER, A., RODRIGUEZ VIEITES, D. & SUÁREZ MARTÍNEZ, M., 1995: Contribución al estudio anatómico del géneo Rana L., 1758 en el alto Aragón. Lucas Mallada (Huesca), 7: 227–247.
RICE, W. R., 1989: Analyzing tables of statistical tests. Evolution, 43: 223–225.
RIOBÓ, A., REY, J., PUENTE, M., MIRAMONTES, C. & VENCES, M., 2000: Ontogenetic increase of black dorsal pattern in Rana temporaria. Brit. Herpetol. Soc. Bull., 70: 1–6.
ROHLF, F J., 1990: NTSYS: Numerical taxonomy and multivariate analysis system, version 1.60. Setauket, NY.
SERRA-COBO, J., 1993: Descripción de una nueva especie europea de rana parda (Amphibia, Anura, Ranidae).
SERRA-COBO, J., 1997: Rana pyrenaica. In: Pleguezuelos, J.M. (ed.), Distribución y biogeografía de los anfibios y reptiles en España y Portugal. Granada, Universidad de Granada and AHE: 167–168. SIGISMUND, H. R., 1997: GSTAT, version 3. München.
SPERLING, P., VENCES, M. & BÖHME, W., 1996: Vorläufige Bemerkungen zum taxonomischen Status von Rana temporaria honnorati Heron-Royer, 1881. Salamandra, 32: 99–112.
STRIJBOSCH, H., 1996: Pas ontdekt: Rana pyrenaica, de Pyreneeën-kikker. Lacerta, 54: 53–55.
TABERLET, P., FUMAGALLI, L., WUST-SAUCY, A.-G. & COSSON, J.-F., 1998: Comparative phylogeography and postglacial colonization routes in Europe. Mol. Ecol., 7: 453–464. VEITH, M., 1996: Molecular markers and species delimitation: examples from the European batrachofauna.
VENCES, M., 1992: Zur Biologie der nordwestspanischen Braunfrösche Rana iberica Boulenger, 1879 und Rana temporaria parvipalmata Seoane, 1885. Salamandra, 28: 61–71.
VENCES, M., PALANCA SOLER, A., RODRIGUEZ VIEITES, D. & NIETO ROMAN, S., 1998a: Designation and description of a lectotype of Rana aragonensis Palanca Soler et al., 1995 (Anura: Ranidae). Herpetozoa,10: 129–134.
VENCES, M., KUPFER, A., LLORENTE, G.A., MONTORI, A. & CARRETERO, M.A., 1998b: Description of the larval stages of the Pyrenean frog, Rana pyrenaica Serra-Cobo, 1993 (Amphibia: Ranidae). Boll. Mus. reg.
Sci. nat. Torino, 15: 1–23.
VENCES, M., PIQUÉ, N., LOPEZ, A., PUENTE, M., MIRAMONTES, C. & RODRIGUEZ, D., 1999: Summer habitat population estimate and body size variation in a high altitude population of Rana temporaria. Amph.-Rept., 20: 440–443.
WEIR, B. S. & COCKERHAM, C. C., 1984: Estimating F-statistics for the analysis of population structure.
Evolution, 38: 1358–1370.
WRIGHT, S., 1951: The genetical structure of populations. Ann. Eugenics, 1: 323–354.

Source: http://fox.ivb.cz/folia/51/4/307-318.pdf