According to the Host-tracking Hypothesis, species of higher trophic levels with a close relationship to their hosts, such as parasites or parasitoids, are expected to show spatio-temporal phylogeographic patterns similar to those of their host. Alternatively, with ecological sorting, a subset of the local species pools might shift to a related host species, thereby disengaging common phylogeographic patterns. Here, we compare the phylogeographic structures of the cynipid rose gall wasp Diplolepis rosae across Europe and of two of its most common parasitoids, the wasps Orthopelma mediator and Glyphomerus stigma, by analysing the sequences of two gene fragments (COI and ITS 2). The phylogeographic structures of the three species associated with roses were incongruent. D. rosae had the lowest genetic diversity with one major clade, O. mediator showed the classical phylogeographic structure for Europe with one eastern and one western clade, and G. stigma had the highest diversity but no geographical structuring. This discordance of geographical patterns may be explained by 1) the dispersal propensity of adult parasitoids or 2) the parasitoids having the ability to switch to another host, while the primary host becomes rare or is even not available. Furthermore there was no indication that phylogenetic patterns were affected by Wolbachia infections. Our results document that communities of closely interacting species may be the result of idiosyncratic biogeographic histories.
Citation: Kohnen A, Richter I, Brandl R (2012) No Concordant Phylogeographies of the Rose Gall Wasp Diplolepis rosae (Hymenoptera, Cynipidae) and Two Associated Parasitoids across Europe. PLoS ONE 7(10): e47156. https://doi.org/10.1371/journal.pone.0047156
Editor: Hans Henrik Bruun, University Copenhagen, Denmark
Received: July 22, 2011; Accepted: September 12, 2012; Published: October 11, 2012
Copyright: © Kohnen et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Funding: This work was supported by the Deutsche Forschungsgemeinschaft (DFG; http://www.dfg.de/index.jsp) within the Priority Programme SPP 1127 (Adaptive Radiation – Origin of Biological Diversity) and by the FAZIT-Stiftung Gemeinnützige Verlagsgesellschaft mbH, Germany (http://www.fazit-stiftung.de). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests: A.K. received a scholarship of the FAZIT-Stiftung for one year during her PhD. The only interest of this founder is to help poor students without salary. The only condition of the scholarship was to finish the PhD-work. This does not alter the authors’ adherence to all the PLOS ONE policies on sharing data and materials.
Plants, phytophagous insects and associated parasitoids together comprise more than 50% of the macroscopic species on Earth  and are therefore an important part of the world’s biodiversity. Based on the hypothesis of Ehrlich and Raven , ecologists long believed that co-evolution and co-speciation between plants and phytophagous insects contributed to this bewildering diversity of insects. Co-evolution requires stable communities of interacting species that may undergo parallel diversification . According to the Host-tracking Hypothesis, pairs of interacting species follow the shifts of the distributional ranges of the host species with either concordance in timing (Contemporary Host-tracking Hypothesis) ,  or with a temporal delay (Delayed Host-tracking Hypothesis) , . Both forms of the hypothesis predict concordant phylogeographic structures for closely interacting species, such as symbionts, mutualists and host-parasite systems –.
Despite these theoretical predictions the evidence for a clear concordance in the phylogeographic structure and timing of diversification of interacting species is rare , , . Even related interacting species sharing the same habitat and history can differ in their genetic structure , . Therefore, during recent years, the paradigm of interpreting the genetic structure of interacting species changed from co-evolution and co-speciation to a more individualistic view , . Models of community assembly understand insect communities with different phylogeographic patterns as their being subsets of regional species pools that depend on the local abiotic conditions , . One example for a community of closely interacting species are plants, gall-inducers and their associated communities. With 1300 described species, the gall wasps (Cynipidae) are one of the largest families of gall-inducing insects . The induced plant galls provide nutritive tissue and a protected habitat for the gall-inducing wasp as well as for a complex community of inquilines, parasitoids and hyperparasitoids. Because of their intimate relationship and their high specialisation, a close co-distribution and co-evolution for these insect groups could be expected.
Here, we focus on the cynipid rose gall wasp system. The univoltine rose gall wasp Diplolepis rosae L. (Hym., Cynipidae) induces conspicuous, multi-chambered galls on Rosa species from several sections . The galls form the basis of a complex community of parasitoids, hyperparasitoids and inquilines , . Two of the most frequent parasitoid species are the endoparasitoid Orthoplema mediator Thunb. (Hym. Ichneumonidae) and the ectoparasitoid Glyphomerus stigma Fabr. (Hym. Torymidae) . These three insect species live closely together in the same habitat, have similar life cycles, and are directly or indirectly dependent on the distribution and history of dog roses. O. mediator has been observed as a parasitoid of three other Diplolepis species, namely D. spinosissimae (Giraud), D. eglanteriae (Hartig) and D. mayri (Schlechtendal). These and most other Diplolepis species introduce much smaller galls than D. rosae, usually with just one chamber for one larva. Likewise, G. stigma attacks not only D. rosae, but also the inquiline species and other parasitoid species within the D. rosae galls. G. stigma has been found to attack other Diplolepis species in Canada, but not in Europe. Nevertheless, it is hypothesised that D. rosae is the main host species of both O. mediator and G. stigma . Switches between different host species however would allow the parasitoid species to colonise areas where D. rosae is rare or even absent.
Stille  suggested for D. rosae and another rose gall wasp species (D. mayri) two re-colonization routes into Sweden. One group was supposed to have come from the South and the other from Finland. Using allozymes he found for D. rosae very low genetic differentiation in southern Sweden  and higher diversity in Europe. According to his findings we expect a classical European differentiation pattern with higher diversity in the South and the differentiation of several lineages originated in glacial refugia using different re-colonization routes –.
Furthermore the reproductive strategy of one species might be important for their current geographical differentiation because of differences in dispersal and consecutive settlement abilities . Parthenogenetic reproduction could reduce gene flow between populations  and might therefore lead to a higher genetic differentiation between populations as well as to a lower genetic variability within populations . Parthenogenetic reproduction can be induced by the intracellular bacteria Wolbachia . The type of Wolbachia and the infection rate of the three chosen insect species differ. The gall inducer D. rosae is infected with Wolbachia type I, which induces parthenogenesis and has a recent infection in central Germany of more than 99% , . In the study of Schilthuizen and Stouthamer , O. mediator was infected with Wolbachia type II, but only at one of three sampling sites. The reproductive impact on O. mediator is currently not known. Infections of G. stigma with Wolbachia have not been observed .
Here we tested two alternative scenarios for the biogeographic histories of Diplolepis and their parasitoids. First, all three insect species have co-evolved, had undergone similar population dynamics and are influenced by their associated life cycles in similar ways. Assuming a tight co-distribution of gall wasps and parasitoids in space and time the same basic patterns of genetic differentiation should be found in gall wasps as well as parasitoid species. Second, the possibility to switch between host species disengaged the close connection of the distributions of both the gall wasp and associated chalcid parasitoids . Therefore the same basic patterns for all three species would not be expected. The parasitoid species may have survived or expanded in regions where D. rosae did not occur. This should have lead to spatial discordance in distributional patterns of genetic diversity. Therefore we would expect different phylogeographic patterns for the parasitoids and their host species. We tested these scenarios using sequences of variable regions in the DNA of the wasps and the parasitoids. One mitochondrial marker and one nuclear locus were chosen , .
A. Geographical distribution of D. rosae samples. Colour codes correspond with clades B. Bayesian 50% majority-rule consensus tree for the COI data of D. rosae. The first number at each node indicates posterior probability values; the second and third numbers indicate bootstrap support of the corresponding clades of neighbour-joining and maximum likelihood trees. With the taxon labels the country and number of sampling site are given. Wolbachia infection is indicated with +, absence with -.
Materials and Methods
Galls of D. rosae were collected in Europe during 2006 and 2007. All sampling sites (not privately owned or protected) and collectors (no specific permits were required because no protected or endangered species were involved) are named in the Table S1. Each gall was kept in separate plastic pots covered with gauze. Inhabitants of the galls were allowed to exit the galls until July and subsequently identified morphologically and stored in 90% ethanol. Only one individual of each species per gall was used for genetic analyses to avoid sampling of closely related siblings or identical samples owing to parthenogenesis in D. rosae.
Total DNA of the emerged gall inhabitants was extracted using spin columns (DNeasy tissue kit, Qiagen, Hilden, Germany). We amplified and sequenced two DNA fragments: ITS 2, with a length of ca. 700 bp; and the mitochondrial COI, with a length of ca. 650 bp.
The ITS 2 fragments of D. rosae and G. stigma were amplified with the primers forward ITS5.8F (5′-GTC CAC GGA TAC AAT TCC CGG ACC-3′)  and reverse ITS 4 (5′-TCC TCC GCT TAT TGA TAT GC-3′) . Amplifications consisted of an initial denaturation at 95°C for 2 min, then 30 cycles with denaturation at 95°C for 30 s, annealing at 55°C for 1 min and extension at 72°C for 1 min. These steps were followed by extension at 72°C for 10 min.
The ITS 2 fragment of O. mediator was amplified with the primers ITS 2 F (5′-GGG TCG ATG AAG AAC GCA GC-3′) and ITS 2 R (5′-ATA TGC TTA AAT TCA GCG GG-3′) . Amplification consisted of 35 cycles with an annealing temperature of 51°C for 1 min.
The COI fragment from all three insect species was amplified with the primers forward LCO (5′-GGT CAA CAA ATC ATA AAG ATA TTG G-3′) and reverse HCO (5′-TAA ACT TCA GGG TGA CCA AAA AAT CA-3′) . Amplification consisted of an initial denaturation at 95°C for 5 min, then 35 cycles with denaturation at 95°C for 30 s, annealing for 45 s at 45°C for D. rosae and at 40°C for O. mediator and G. stigma, and extension at 72°C for 1 min. These steps were followed by extension at 72°C for 10 min.
All reactions were carried out in 20 µl reaction volume containing 2–5 µl of template DNA, 10 mM Tris-HCl (pH 8.3), 50 mM KCl, 1.5 mM MgCl2, 80 µM dNTP, 10 µM of each primer and 1 unit of Taq DNA polymerase (New England Biolabs, Frankfurt a. Main, Germany). ITS 2 fragments of O. mediator were extracted from an agarose gel using the QIAquick gel extraction kit (Qiagen, Hilden, Germany). All other PCR products were purified using a Qiagen MinElute PCR purification kit. Fragments were sequenced directly by Sequencing Laboratories Göttingen GmbH, Germany.
A. Neighbour-joining tree of ITS data. Numbers at nodes indicate bootstrap support. B. Bayesian 50% majority-rule consensus tree for COI data of O. mediator. The first number at each node indicates posterior probability values; the second and third numbers indicate bootstrap support of the corresponding clades of neighbour-joining and maximum likelihood trees. With the taxon labels the country and number of sampling site are given, see Fig. 3. Wolbachia infection is indicated with +, absence with -. The western clade is coloured in blue, the eastern in red. Individuals with conflicting assignment are marked with *.
Haplotypes were assigned based on the COI and ITS 2 sequences. Blue = western clade, red = eastern clade, grey = haplotype found only in Italy, half blue/half red = individuals assigned to eastern clade based on ITS 2 sequences and to the western clade based on COI sequences.
Infection of Wasps by Wolbachia sp
The presence or absence of Wolbachia was tested with specific primer pairs amplifying ca. 600 bp of the wsp gene using primers forward wsp 81F (5′-TGG TCC AAT AAG TGA TGA AGA AAC-3′) and reverse wsp 691R (5′-AAA AAT TAA ACG CTA CTC CA-3′) . Whether the absence of a PCR product was caused either by the absence of Wolbachia or by a reaction failure was checked with a control primer pair. Control primers for D. rosae were forward Dr06-F (5′-CTC ATC TCT TCT TCT TAT CTC AG-3′) and reverse Dr06-R (5′-CCC AGG AGA GCA GAG G-3′) , and control primers for O. mediator and G. stigma were LCO and HCO . All PCR reactions had both a positive control (known infected individual) and a negative control (water). Amplifications consisted of initial denaturation at 94°C for 3 min, then 35 cycles with denaturation at 94°C for 1 min, annealing at 50°C (D. rosae) or 42°C (O. mediator and G. stigma) for 1 min and extension at 72°C for 1 min. These steps were followed by extension at 72°C for 5 min.
Sequences were manually edited and aligned using BioEdit version 7.0.9 . Haplotype frequencies and diversities were estimated using the program DnaSP version 4.10 . The hierarchical likelihood ratio tests and the Akaike information criterion implemented in MODELTEST 3.7  were used to select appropriate models of sequence evolution including outgroup individuals (for D. rosae, two individuals of D. fructuum, GenBank JN252403, JN252404; for the COI sequences of O. mediator, one individual of Pimpla aequalis from GenBank AF146681; and for COI sequences of G. stigma, one individual of Nassonia vitripennis from GenBank EU746551).
The selected models (Table 1) were implemented to calculate corrected distances and to construct phylogeographic relationships by using the neighbour-joining algorithm as well as using maximum likelihood and Bayesian methods. For the latter we used the program MrBayes version 3.1.2 . Four chains per run and two independent runs were used. A run length of 1 million generations and a sample frequency of every 1000 generations were preset with a burnin period of 50000 generations. All analyses were checked for convergence with the program Tracer v1.5 . For the sequences of O. mediator both gene fragments were analysed separately and combined. For neighbour joining and maximum-likelihood analyses (using a heuristic search algorithm) data were supplied to the programs MEGA version 4  and PAUP* version 4.0b10 . To quantify the reliability of the nodes, 1,000 bootstrap replicates were used for these two approaches. For COI sequences of O. mediator, we used analysis of molecular variance (AMOVA) Arlequin version 3.11 .
The D. rosae mtDNA sequences of COI (GenBank JN252324-JN252338) had many more variable sites than the genomic ITS 2 sequences (GenBank JN252386-JN252391) (Table 1). For both loci, we found one frequent haplotype and several haplotypes that occurred only once (Table S1). Individuals from the same sampling site carried the same haplotype, with three exceptions.
For COI sequences, Modeltest selected the model of sequence evolution of Hasegawa, Kishino and Yano  with among site rate variation. In all topologies one major clade was found which was common across Central and Northern Europe and was also recorded in Spain (Fig. 1). Three subclades were recorded only in Southern Europe (Fig. 1A).
As for D. rosae, COI sequences of O. mediator (GenBank JN252339-JN252361) were much more variable than ITS 2 sequences (GenBank JN252392-JN252399); the ITS 2 sequences had two positions at which one or two base pairs were deleted (Table 1). The most frequent haplotype of COI was found in 14 individuals (n = 56). For ITS 2, Modeltest selected the simplest model of sequence divergence, the Jukes-Cantor model , and for COI, Modeltest selected the K81 model with unequal base frequencies .
All topologies with all three methods and both markers separately or combined divided O. mediator into two major clades (Fig. 2, Fig. S1). One clade was found in Western Europe, distributed from Spain to France, the UK, and Denmark, up to Sweden (western clade). The other clade was found in Macedonia, Central Europe, and Lithuania up to Finland (eastern clade, Fig. 3). For three individuals, the two loci showed incongruent clade affiliation (Fig. 2, 3, S1); the ITS 2 sequences assigned all three individuals to the eastern clade, but the COI sequences assigned them to the western clade. In the combined analyses the three individuals were assigned to the western clade. A minor, third clade only distantly related to the two main clades and containing only one individual (from Italy) was identified by the COI sequences and the combined dataset (Fig. 3, S1). Both western and eastern clades were supported by high bootstrap values and showed low genetic distances within (0.45 and 0.58%, respectively) and higher distances between (5.2%) each other. An AMOVA attributed only 6% variation within clades (SS = 53, df = 53), but 94% between clades (SS = 422, df = 1). The pair-wise FST value between clades was 0.94 (p<0.0001). The haplotype diversity H in the eastern and western clades was 0.66 and 0.91, respectively.
The ITS 2 sequences of G. stigma individuals (GenBank JN252400-JN252402) were almost identical, with only one variable position in four individuals from Italy, Macedonia, Ukraine and the Czech Republic. The individual from Ukraine also had a deletion of two base pairs. In contrast, the COI sequences of G. stigma individuals (GenBank JN252362-JN252385) were highly variable (Table 1). Most haplotypes were found only once, and four haplotypes were found twice; modeltest selected the Hasegawa, Kishino and Yano model with among site rate variation . In all topologies,haplotypes were clustered together with high support, but no consistent clades are supported between haplotypes (Fig. 4).
Wolbachia infection was common in the gall wasp D. rosae (91%; n = 79, Fig. 1). We found no geographic pattern in the infection rate. The parasitoid O. mediator showed a much lower infection by Wolbachia; 10% of the sampled individuals were infected (n = 56). Populations with infected individuals were distributed throughout Europe but not in Scandinavia (Fig. 2). The parasitoid G. stigma (n = 28) was not infected with Wolbachia bacteria. The overall infection across species differed significantly (chi-square 27.6, df = 2, p<0.001).
The phylogeographical structure of D. rosae and two of its most common parasitoids, O. mediator and G. stigma, showed little phylogeographic congruence across Europe. Of the two analysed sequences just one, the mitochondrial COI fragment, was variable. The ITS 2 region showed no or little variation in G. stigma and D. rosae. Therefore our results and conclusions are mainly based on the mitochondrial fragment. Due to gene duplication, loss, lineage sorting and horizontal transfer gene trees need not necessarily reflect species trees , . Therefore in a further study our results should be verified with data from more genes.
The most common host plants of D. rosae are dog roses of the section Caninae. Dog roses comprise more than 150 species that hybridise and differ in several characters. Little is known about their biogeographic history except that this section originated by hybridisation events presumably during the last ice ages –. By using almost all lineages of dog roses as a host, the gall wasp is able to colonise the entire palaearctic region. Compared to its two parasitoid species, D. rosae showed low haplotype diversity and low genetic distances between haplotypes. A similar low genetic diversity was found by Stille  with allozymes in Swedish populations. In contrast D. mayri shows higher diversities and a division in two subpopulations in Sweden, which were also geographically divided pointing to two different re-colonisation routes after the last ice ages . The distribution of the common haplotype of D. rosae across Europe points to ongoing high gene flow between sampling sites . The higher haplotype richness in Southern Europe is consistent with the expectation of Pleistocene refugia in that area – and a re-colonisation of the regions north of the Alps by the common haplotype.
In contrast to the gall wasp, the parasitoid species showed much higher haplotype diversities. We found no clear geographical structure for G. stigma for either locus, which points either to high rates of mixing and exchange between populations (and therefore also to good dispersal abilities) or to a large effective population size with incomplete lineage sorting of these specific genes. The pronounced divergence of O. mediator into two lineages, an eastern and a western clade, is a classical pattern found in many plant and animal species , . This split is thought to be the result of survival during the Pleistocene in isolated refugia one located in Iberia and a second one somewhere in the east maybe in the Balkans. In a suture zone between both clades in France and Germany, where both clades live in sympatry, we found three “hybrid” individuals with contrasting mitochondrial and nuclear haplotypes.
Some populations of O. mediator of both geographical lineages were infected with Wolbachia. These bacteria are normally vertically transmitted from the mother to offspring with the cytoplasm of eggs. Some phylogenetic studies however have shown that phylogenies of insect species and associated bacteria are incongruent, which suggests that horizontal transmission must have taken place frequently , . Schilthuizen and Stouthamer  found no horizontal transmission between D. rosae gall wasps and parasitoids, because host and parasitoids are infected with different types or strains of Wolbachia. We found no congruent geographical pattern in the Wolbachia infection of host and parasitoids leading to the conclusion that the infection by Wolbachia had little influence on the phylogeographic structure of hosts and parasitoids. In a recent study, Nicholls and colleagues  have found congruence of lineage divergence as well as temporal patterns within and across trophic levels of two parasitoid species of the genus Megastigmus, both assoociated with oak gall wasps. But this is one rare example and less than a general pattern within oak gall wasps and their parasitoids. Several oak gall wasps have shown broadly compatible geographical distributions and genetic structures with an eastern origin and a re-colonization through central Europe , –, but not always on the same timescales. In more recent times oak gall wasps restricted to Quercus section Cerris expanded their ranges northwards because of human mediated plantings of their host trees , . No such change in the distribution and such strict restriction to one host plant section is known for D. rosae.
The different phylogeographical patterns of the three interacting insect species studied here may be explained by host switching of the parasitoids when the primary host was not available. Indeed, O. mediator and G.stigma have been observed as parasitoids of other Diplolepis species. Since parasitoids are bound to their host species only during the larval stage, further dispersal and host switching is possible at later stages. Dispersal abilities and the ensuing colonisation success also influence population structures . Dispersal and host switching thereby disengages the close association in the geographic dynamics of the gall wasp and its parasitoids.
Bayesian 50% majority rule consensus tree of COI and ITS 2 sequences combined of Orthopelma mediator. The number at each node indicates posterior probability values. With the taxon labels the country and number of sampling site are given, see Fig. 3. The western clade is coloured in blue, the eastern in red.
We thank all collectors of bedeguar galls used in this study, Sandra Schneider, Antje Schmidt and Konstanze Bandmann for help in sorting and identifying the species and Lars Opgenoorth for help with the Bayesian analyses. We are especially thankful for helpful comments on the manuscript by Karsten Schönrogge and Graham Stone.
Conceived and designed the experiments: AK RB. Performed the experiments: AK IR. Analyzed the data: AK. Wrote the paper: AK RB.
- 1. Strong DR, Lawton JH, Southwood R (1984) Insects on plants. Common patterns and mechanisms. Oxford: Blackwell Scientific Publications. 315 p.
- 2. Ehrlich PR, Raven H (1964) Butterflies and plants: a study in co-evolution. Evolution 18: 586–608.
- 3. Schluter D (2000) The Ecology of Adaptive Radiation. New York: Oxford University Press. 288 p.
- 4. Becerra JX (2003) Synchronous coadaptation in an ancient case of herbivory. Proc Natl Acad Sci U S A 100: 12804–12807.
- 5. Wheat CW, Vogel H, Wittstock U (2007) The genetic basis of a plant-insect coevolutionary key innovation. Proc Natl Acad Sci U S A 104: 20427–20431.
- 6. Hayward A, Stone GN (2006) Comparative phylogeography across two trophic levels: the oak gall wasp Andricus kollari and its chalcid parasitoid Megastigmus stigmatizans. Mol Ecol 15: 479–489.
- 7. Stone GN, Lohse K, Nicholls JA, Fuentes-Utrilla P, Sinclair F, et al. (2012) Reconstructing community assembly in time and space reveals enemy escape in a western palearctic insect community. Current Biology 22: 1–6.
- 8. Funk DJ, Helbling L, Wernegreen JJ, Moran NA (2000) Intraspecific phylogenetic congruence among multiple symbiont genomes. Proc R Soc Lond B Biol Sci 267: 2517–2521.
- 9. LaJeunesse TC, Bhagooli R, Hidaka M, DeVantier L, Done T, et al. (2004) Closely related Symbiodinium spp. differ in relative dominance in coral reef host communities across environmental, latitudinal and biogeographic gradients. Mar Ecol Prog Ser 284: 147–161.
- 10. Nieberding C, Morand S, Libois R, Michaux JR (2004) A parasite reveals cryptic phylogeographic history of its host. Proc R Soc Lond B Biol Sci 271: 2559–2568.
- 11. Crandall ED, Jones ME, Munoz MM, Akinronbi B, Erdmann MV, et al. (2008) Comparative phylogeography of two seastars and their ectosymbionts within the Coral Triangle. Mol Ecol 17: 5276–5290.
- 12. Johannesen J, Seitz A (2003) Comparative population genetic structures of the fruit fly Urophora cardui and its primary parasitoid Eurytoma robusta. Entomol Exp Appl 108: 149–157.
- 13. Brändle M, Knoll S, Eber S, Stadler J, Brandl R (2005) Flies on thistles: support for synchronous speciation? Biol J Linn Soc 84: 775–783.
- 14. Dawson MN, Louie KD, Barlow M, Jacobs DK, Swift CC (2002) Comparative phylogeography of sympatric sister species, Clevelandia ios and Eucyclogobius newberryi (Teleostei, Gobiidae), across the California Transition Zone. Mol Ecol 11: 1065–1075.
- 15. Brandl R, Mann W, Sprinzl M (1992) Estimation of the monocot-dicot age through tRNA sequences from chloroplast. Proc R Soc Lond B Biol Sci 249: 13–17.
- 16. Schoonhoven LM, Jermy T, von Loon JJA (1998) Insect-plant biology. From physiology to evolution. London: Chapman and Hall. 409 p.
- 17. Gotelli NJ, McCabe DJ (2002) Species co-occurrence: A meta-analysis of J. M. Diamond’s assembly rules model. Ecology 83: 2091–2096.
- 18. Chase JM (2003) Community assembly: when should history matter? Oecologia 136: 489–498.
- 19. Liljeblad J, Ronquist F (1998) A phylogenetic analysis of higher-level gall wasp relationships (Hymenoptera: Cynipidae). Syst Entomol 23: 229–252.
- 20. Schröder D (1967) Diplolepis ( = Rhodites) rosae (L.) (Hym.: Cynipidae) and a review of its parasite complex in Europe. Techn Bull Commonw Inst Biol Contr 9: 93–131.
- 21. Blair KG (1944) A note on the economy of the rose bedeguar gall, Rhodites rosae, L. Proc South Lond Ent Nat Hist Soc. 1943–44: 55–59.
- 22. Redfern M, Askew RR (1992) Plant galls. The Richmond Publishing Co. Ltd. Slough. 104 p.
- 23. Stille B (1984) The effect of hostplant and parasitoids on the reproductive success of the parthenogenetic gall wasp Diplolepis rosae (Hymenoptera, Cynipidae). Oecologia 63: 364–369.
- 24. Randolph S (2005) The natural history of the rose bedeguar gall and its insect community. British Plant Gall Society. 92 p.
- 25. Stille B (1985a) Host plant specificity and allozyme variation in the parthenogenetic gall wasp Diplolepis mayri and its relatedness to D. rosae (Hymenoptera: Cynipidae). Entomol Gener 10: 87–96.
- 26. Stille B (1985b) Population genetics of the parthenogenetic gall wasp Diplolepis rosae (Hymenoptera, Cynipidae). Genetica 67: 145–151.
- 27. Hewitt GM (1999) Post-glacial re-colonization of European biota. Biol J Linn Soc 68: 87–112.
- 28. Lunt DH, Ibrahim KM, Hewitt GM (1998) MtDNA phylogeography and postglacial patterns of subdivision in the meadow grasshopper Chorthippus parallelus. Heredity 80: 633–641.
- 29. Stauffer C, Lakatos F, Hewitt GM (1999) Phylogeography and postglacial colonization routes of Ips typographus L-(Coleoptera, Scolytidae). Mol Ecol 8: 763–773.
- 30. Santucci F, Emerson BC, Hewitt GM (1998) Mitochondrial DNA phylogeography of European hedgehogs. Mol Ecol 7: 1163–1172.
- 31. Taberlet P, Fumagalli L, Wust-Saucy AG, Cosson JF (1998) Comparative phylogeography and postglacial colonization routes in Europe. Mol Ecol 7: 453–464.
- 32. Werren JH, Baldo L, Clark ME (2008) Wolbachia: master manipulators of invertebrate biology. Nat Rev Microbiol 6: 741–751.
- 33. Plantard O, Rasplus JY, Mondor G, Le Clainche I, Solignac M (1998) Wolbachia-induced thelytoky in the rose gallwasp Diplolepis spinosissimae (Giraud) (Hymenoptera : Cynipidae), and its consequences on the genetic structure of its host. Proc R Soc London B Biol Sci 265: 1075–1080.
- 34. Stouthamer R, Breeuwer JAJ, Hurst GDD (1999) Wolbachia pipientis: Microbial manipulator of arthropod reproduction. Annu Rev Microbiol 53: 71–102.
- 35. Schilthuizen M, Stouthamer R (1998) Distribution of Wolbachia among the guild associated with the parthenogenetic gall wasp Diplolepis rosae. Heredity 81: 270–274.
- 36. Kohnen A, Wissemann V, Brandl R (2011) No host-associated differentiation in the gall wasp Diplolepis rosae (Hymenoptera: Cynipidae) on three dog rose species. Biol J Linn Soc 102: 369–377.
- 37. Hurst GDD, Jiggins FM (2005) Problems with mitochondrial DNA as a marker in population, phylogeographic and phylogenetic studies: the effects of inherited symbionts. Proc R Soc London B Biol Sci 272: 1525–1534.
- 38. Knowles LL, Carstens BC (2007) Delimiting species without monophyletic gene trees. Syst Biol 56: 887–895.
- 39. Rokas A, Nylander JAA, Ronquist F, Stone GN (2002) A maximum-likelihood analysis of eight phylogenetic markers in gallwasps (Hymenoptera : Cynipidae): Implications for insect phylogenetic studies. Mol Phylogenet Evol 22: 206–219.
- 40. White TJ, Bruns T, Lee S, Taylor J (1990) Amplification and direct sequencing of fungal ribosomal RNA genes for phylogenetics. PCR protocols: A Guide to Methods and Applications. Academic Press, Inc. 315–322.
- 41. Wagener B, Reineke A, Lohr B, Zebitz CPW (2006) Phylogenetic study of Diadegma species (Hymenoptera: Ichneumonidae) inferred from analysis of mitochondrial and nuclear DNA sequences. Biol Contr 37: 131–140.
- 42. Folmer O, Black M, Hoeh W, Lutz R, Vrijenhoek R (1994) DNA primers for amplification of mitochondrial cytochrome c oxidase subunti I from diverse metazoan invertebrates. Mol Mar Biol Biotechnol 3: 294–299.
- 43. Braig HR, Zhou WG, Dobson SL, O’Neill SL (1998) Cloning and characterization of a gene encoding the major surface protein of the bacterial endosymbiont Wolbachia pipientis. J Bacteriol 180: 2373–2378.
- 44. Hall TA (1999) BioEdit: a user-friendly biological sequence alignment editor and analysis program for Windows 95/98/NT. Nucleic Acids Symp Ser 41: 95–98.
- 45. Rozas J, Sanchez-DelBarrio JC, Messeguer X, Rozas R (2003) DnaSP, DNA polymorphism analyses by the coalescent and other methods. Bioinform 19: 2496–2497.
- 46. Posada D, Crandall KA (1998) MODELTEST: testing the model of DNA substitution. Bioinform 14: 817–818.
- 47. Ronquist F, Huelsenbeck JP (2003) MrBayes 3: Bayesian phylogenetic inference under mixed models. Bioinform 19: 1572–1774.
- 48. Rambaut A, Drummond AJ (2003) Tracer version 1.5 (computer program). Available: http://tree.bio.ed.ac.uk/software/tracer.
- 49. Kumar S, Tamura K, Jakobsen IB, Nei M (2001) MEGA2: molecular evolutionary genetics analysis software. Bioinform 17: 1244–1245.
- 50. Swofford DL (2002) Paup: Phylogenetic Analysis Using Parsimony (and Other Methods) 4.0 Beta. Sunderland MA: Sinauer Associates. 128 p.
- 51. Excoffier L, Laval G, Schneider S (2005) Arlequin (version 3.0): An integrated software package for population genetics data analysis. Evol Bioinform: 47–50.
- 52. Hasegawa M, Kishino H, Yano T (1985) Dating of the Human-Ape Splitting by a Molecular Clock of Mitochondrial DNA. J Mol Evol 22: 160–174.
- 53. Jukes T, Cantor C (1969) Evolution of protein molecules. In: Munro, editor. Mammalian Protein Metabolism.New York: Academic Press. 21–132.
- 54. Kimura M (1981) Estimation of evolutionary distances between homologous nucleotide sequences. Proc Natl Acad Sci U S A 78: 454–458.
- 55. Rosenberg NA, Nordborg M (2002) Genealogical trees, coalescent theory and the analysis of genetic polymorphisms. Nature Reviews 3: 380–390.
- 56. Lohse K, Sharanowski B, Stone GN (2010) Quantifying the Pleistocene history of the oak gall parasitoid Cecidostiba fungosa unsing twenty intron loci. Evolution 64: 2664–2681.
- 57. Wissemann V (2002) Molecular evidence for allopolyploid origin of the Rosa canina-complex (Rosaceae, Rosoideae). J Appl Bot 76: 176–178.
- 58. Ritz CM, Schmuths H, Wissemann V (2005) Evolution by reticulation: European dogroses originated by multiple hybridization across the genus Rosa. J Hered 96: 4–14.
- 59. Dingler H (1907) Versuch einer Erklärung gewisser Erscheinungen in der Ausbildung und Verbreitung der wilden Rosen. Mitteilungen des Naturwissenschaftlichen Vereins zu Aschaffenburg 6: 1–38.
- 60. Stouthamer R, Breeuwer JAJ, Luck RF, Werren JH (1993) Molecular-identification of microorganisms associated with parthenogenesis. Nature 361: 66–68.
- 61. Schilthuizen M, Stouthamer R (1997) Horizontal transmission of parthenogenesis-inducing microbes in Trichogramma wasps. Proc R Soc London B Biol Sci 264: 361–366.
- 62. Nicholls JA, Preuss S, Hayward A, Melika G, Csoka G, et al. (2010) Concordant phylogeography and cryptic speciation in two Western Palaearctic oak gall parasitoid species complexes. Mol Ecol 19: 592–609.
- 63. Rokas A, Atkinson RJ, Webster L, Csoka G, Stone GN (2003) Out of Anatolia: longitudinal gradients in genetic diversity support an eastern origin for a circum-Mediterranean oak gallwasp Andricus quercustozae. Mol Ecol 12: 2153–2174.
- 64. Stone G, Atkinson R, Rokas A, Csoka G, Nieves-Aldrey JL (2001) Differential success in northwards range expansion between ecotypes of the marble gallwasp Andricus kollari: a tale of two lifecycles. Mol Ecol 10: 761–778.
- 65. Rokas A, Atkinson RJ, Brown GS, West SA, Stone GN (2001) Understanding patterns of genetic diversity in the oak gallwasp Biorhiza pallida: demographic history or a Wolbachia selective sweep? Heredity 87: 294–304.
- 66. Stone GN, Challis RJ, Atkinson RJ, Csoka G, Hayward A, et al. (2007) The phylogeographical clade trade: tracing the impact of human-mediated dispersal on the colonization of northern Europe by the oak gallwasp Andricus kollari. Mol Ecol 16: 2768–2781.