Present address: E. R. Scott, School of Integrative Biology, University of Illinois, Urbana, IL 61801, USA.
Behavioural, ecological and genetic evidence confirm the occurrence of host-associated differentiation in goldenrod gall-midges
Article first published online: 23 FEB 2009
© 2009 The Authors. Journal Compilation © 2009 European Society For Evolutionary Biology
Journal of Evolutionary Biology
Volume 22, Issue 4, pages 729–739, April 2009
How to Cite
DORCHIN, N., SCOTT, E. R., CLARKIN, C. E., LUONGO, M. P., JORDAN, S. and ABRAHAMSON, W. G. (2009), Behavioural, ecological and genetic evidence confirm the occurrence of host-associated differentiation in goldenrod gall-midges. Journal of Evolutionary Biology, 22: 729–739. doi: 10.1111/j.1420-9101.2009.01696.x
- Issue published online: 13 MAR 2009
- Article first published online: 23 FEB 2009
- Received 30 June 2008; accepted 30 December 2008
- assortative mating;
- enemy-reduced space;
- host races;
- reproductive isolation;
- Top of page
- Materials and methods
Host-associated differentiation (HAD) is considered a step towards ecological speciation and an important mechanism promoting diversification in phytophagous insects. Although the number of documented cases of HAD is increasing, these still represent only a small fraction of species and feeding guilds among phytophagous insects, and most reports are based on a single type of evidence. Here we employ a comprehensive approach to present behavioural, morphological, ecological and genetic evidence for the occurrence of HAD in the gall midge Dasineura folliculi (Diptera: Cecidomyiidae) on two sympatric species of goldenrods (Solidago rugosa and S. gigantea). Controlled experiments revealed assortative mating and strong oviposition fidelity for the natal-host species. Analysis of mitochondrial DNA showed an amount of genetic divergence between the two host-associated populations compatible with cryptic species rather than host races. Lower levels of within-host genetic divergence, gall development and natural-enemy attack in the S. gigantea population suggest this is the derived host.
- Top of page
- Materials and methods
One of the mechanisms thought to promote speciation in phytophagous insects is shifts to new hosts that lead to the establishment of new species via an intermediate step of host-race formation (Bush, 1969; Jaenike, 1981; Drès & Mallet, 2002). The role of host adaptation in speciation, especially in sympatry, has been one of the most debated topics among evolutionary ecologists over the last four decades (reviews in Tauber & Tauber, 1989; Via, 2001; Drès & Mallet, 2002), since Bush (1969) first argued the occurrence of sympatric host-race formation in the fruit fly Rhagoletis pomonella. Populations that experience host-associated differentiation (HAD – speciation that is driven by the adaptation to hosts) can be categorized as constituting a heterogeneous panmictic species, host-races or established sister species based on the amount of gene flow between them (Drès & Mallet, 2002), with FST estimates for host-races, for example, typically ranging from 0.01 to 0.21 (e.g. McPheron et al., 1988; Via, 1999; Nason et al., 2002; Blair et al., 2005; Stireman et al., 2005). In order for HAD to occur, the host-associated populations must exhibit certain premating and post-mating mechanisms that promote reproductive isolation and thus lead to genetic divergence.
Possible premating mechanisms include assortative mating, wherein individuals show preference for mating within the same host-associated population (Berlocher & Feder, 2002; Rundle & Nosil, 2005), and phenological differences exist in emergence or activity times between host-associated populations (e.g. Bush, 1969; Smith, 1988; Craig et al., 1993; Nason et al., 2002; Thomas et al., 2003). A major post-mating isolating mechanism is host fidelity by ovipositing females, which has been documented in many case studies of phytophagous insects (summary in Drès & Mallet, 2002). Once a shift to a new host has occurred, escape from natural enemies may be a factor that offsets the cost of lower fitness resulting from physiological maladaptations to this host (Price et al., 1980; Jeffries & Lawton, 1984; Jaenike, 1990).
Despite the accumulation in recent years of empirical evidence for HAD in herbivorous insects, the data currently available represent only a tiny fraction of potential cases within this extremely speciose and diverse group. Most documented case studies are based on a single type of evidence (usually genetic), and many involve man-made rather than natural scenarios [e.g. the fruit fly R. pomonella on hawthorn and apple (Bush, 1969), the European corn borer Ostrinia nubilalis on corn and mugwort (Bethenod et al., 2005; Malausa et al., 2005) and the aphid Acyrthosiphon pisum on alfalfa and red clover (Via, 1999; Via et al., 2000)]. Difficulty or reluctance to publish negative results, for cases in which HAD has not been found, further contribute to the difficulty of estimating the prevalence of this process. Many more examples are therefore needed to establish how prevalent HAD is under natural conditions, and under what circumstances it is likely to occur. In providing such empirical evidence it would be most informative to explore HAD in unrelated groups of herbivores that use the same set of hosts (Nason et al., 2002; Blair et al., 2005; Stireman et al., 2005). If HAD has indeed occurred independently in such cases, it would imply that this process constitutes an important source of biodiversity in nature.
An exceptional model system in this context is the rich insect community associated with goldenrods (Solidago spp., Asteraceae) in North America, many members of which are highly specialized herbivores (McEvoy, 1988). Among nine cases of goldenrod-feeding insects in which HAD has been studied to date, four proved to have radiated on the sympatric S. altissima and S. gigantea (Abrahamson & Weis, 1997; Nason et al., 2002; Eubanks et al., 2003; Blair et al., 2005; Stireman et al., 2005). In the present study, we employ a multiple-evidence approach to obtain genetic, behavioural, morphological and ecological indications for HAD in the gall midge Dasineura folliculi (Diptera: Cecidomyiidae), which induces galls on Solidago rugosa and S. gigantea. We rely on multiple rather than on a single type of evidence as we consider it to be a more powerful approach for establishing the existence of HAD, while also inherently providing insight into premating and post-mating mechanisms that lead to and/or maintain HAD in a studied organism (see Via, 2001; Berlocher & Feder, 2002).
Specifically we asked whether: (1) adults exhibit host-related morphological differences, (2) midges prefer mates that originate from the same host (is there assortative mating), (3) females prefer to oviposit on their natal host (is there host fidelity), (4) the host-associated populations are genetically differentiated and (5) one of the hosts offers enemy-reduced space to the population associated with it. We then discuss the significance of parallel host shifts in phylogenetically unrelated organisms as a means for estimating the prevalence of HAD in phytophagous insects.
Materials and methods
- Top of page
- Materials and methods
The gall midge D. folliculi (Diptera: Cecidomyiidae) induces bud galls on the common and sympatric goldenrods S. rugosa and S. gigantea (Dorchin et al., 2007). Adults of this species live for 1–3 days, mate on or under the plants, and lay eggs between leaves that surround the apical buds. The galls do not have defined chambers, and the gregarious larvae develop among widened and thickened leaves that form loose clusters on growing shoot tips, with 5–80 larvae per gall. Larvae mature within 3–4 weeks and then leave the gall to pupate in the soil. Dasineura folliculi is multivoltine, completing at least four generations a year between early May and October (Dorchin et al., 2007).
Collecting and rearing of insects
We collected galls in 13 field sites within a 90-km radius of Lewisburg, Pennsylvania (N40°57′ W76°53′) between late April and October in 2005 and 2006. The collection sites included roadsides, natural areas and old fields that supported large, sympatric populations of S. rugosa and S. gigantea; although in some localities one of the plant species was much more abundant than the other. We collected all the galls we encountered that appeared mature (i.e. containing third-instar larvae) and dissected each of them under a stereomicroscope on the same or on the next day. The numbers of larvae of gall inducers and of all types of natural enemies were recorded for each gall, and mature larvae were transferred into 25-mL plastic cups filled with ProMix BXTM (Premier Horticulture, Dorval, QC, Canada) potting mix to allow pupation. Each soil cup contained all viable larvae originating from a single gall. The cups were individually covered with ventilated caps and kept moist in the laboratory, at room temperature, until the end of emergence (up to 4 weeks). The percentage of galls attacked by different types of natural enemies was compared between the two host plants via likelihood-ratio tests using jmp version 5.1.2 (SAS Institute, Cary, NC, USA).
Emerging adults were preserved in 70% ethyl alcohol for morphological study, and 50–60 individuals from each host-associated population were later mounted on permanent microscope slides in euparal according to the method outlined in Gagné (1989). To detect possible morphological differences between adults from the two host-associated populations, we measured wing length (which reflects body size) and the length of tergite 8 relative to that of tergite 6 of the abdomen in both males and females. The relative lengths of the distal abdominal tergites are one of the few morphological traits that may vary among species in the otherwise rather morphologically uniform adults in the genus Dasineura (Gagné, 1989). The wing and tergite measurements were analysed with a two-way anova (with host plant and sex as main effects). Tergite measurements were analysed after −X −0.5 transformation to better meet the distributional assumptions of anova. Analyses were carried out in jmp version 5.1.2 (SAS Institute).
Adults of both sexes reared from field-collected galls as described above were paired in small glass vials in the four possible host-associated combinations (gigantea female with rugosa male, gigantea female with gigantea male, rugosa female with gigantea male and rugosa female with rugosa male). Because D. folliculi galls yield mostly single sex progeny (Dorchin et al., 2007), all females used in the experiments were known to be virgins. In order for mating to occur, females had to exhibit ‘calling behaviour’, during which they stood still and extended and slightly waved their ovipositor, probably emitting pheromones. Mating never occurred when females were not calling. We conducted the mating experiments in a no-choice design, using a single male and a single female in a vial at a time. Each couple was given 5 min to mate as long as the female was calling. If a female stopped calling during the observation, we stopped the clock and resumed the time measurement once she started calling again. The great majority of mating events occurred within 10–120 s from the time a male was introduced into the vial. Once a female had mated she retracted her ovipositor and did not call or mate again. Overall, we conducted 220 mating trials between individuals reared from the same host (controls), and 139 mating trials between individuals from different hosts (crosses). The individuals used in these trials originated from at least 50 different galls collected at different localities and times. Mating frequencies were compared with likelihood-ratio tests using jmp version 5.1.2 (SAS Institute).
Oviposition-choice and performance experiments
We tested female oviposition preference in a choice experiment in which females were offered both host plants in a greenhouse setting. We introduced individual females immediately after they had mated with males from the same host plant into ca. 0.5-L mesh cages covering 20-cm standard pots that each contained one S. rugosa and one S. gigantea ramet that were propagated from rhizomes. Rhizomes for these experiments were collected from numerous plants in two field sites and, although we did not verify their genetic identity, each of the host species must have been represented by at least six different genets. This design assumed that variation among the plant species would be of greater significance to the gall midges than variation between individual plants within the same species. The ramets were 4–10 weeks old at the time of the experiment and were matched for age and height in each pot. Females remained in the cage until death (1–3 days later), after which the cage was removed and the ramets were inspected for eggs and then daily for gall development. Like other gall midges, D. folliculi adults are very short-lived [males live for up to 24 h, females for up to 3 days (Dorchin et al., 2007)], and therefore females were expected to engage in host searching and oviposition soon after they had mated. Because Dasinuera females do not insert their eggs into the plant tissues, the eggs are visible on the shoot tips following oviposition. Overall, we tested 137 females reared from S. gigantea and 87 females reared from S. rugosa. The differences in host choice (i.e. presence of eggs on the shoot tip) and progeny performance (i.e. gall induction and development of midges to maturity) between females from the two host-associated populations were compared via goodness of fit tests using jmp version 5.1.2 (SAS Institute).
We extracted genomic DNA from 35 whole adult midges with the DNeasy Blood & Tissue extraction kit (QIAGEN Inc. Valencia, CA, USA) and PCR-amplified a fragment of ca. 650 bp of the mitochondrial cytochrome oxidase subunit I (COI) for individuals from each host-associated population across our collecting localities. PCR reactions contained 1–2 μL genomic DNA, 2.5 μL (10 mm) 10× PCR Buffer, 2 μL (10 mm) DNTP solution, 2.5 μL (1 mm) MgCl2, 2.5 μL (10 μm) of forward and reverse primers, 0.2 μL (5 U μL−1) AmpliTaq Gold polymerase (Applied Biosystems, Foster City, CA, USA), and dH2O to 25 μL. The primers we used were LCO1490 (F) and HCO2198 (R) (Folmer et al., 1994). The PCR conditions consisted of an initial 10-min denaturation at 95 °C, 35 cycles of 95 °C for 30 s, 50 °C for 1 min, 72 °C for 1 min and a final 72 °C elongation phase of 4 min. PCR products were purified using the MinElute PCR Purification Kit (QIAGEN Inc.). Sequencing was conducted on an automated ABI Hitachi 3730XL DNA Analyzer (Applied Biosystems) at the Pennsylvania State University nucleic acid facility. Sequences were initially aligned using CodonCode Aligner version 1.5.2 (CondonCode Corporation, Dedham, MA, USA). GenBank accession numbers are given in Appendix 1 (http://www.ncbi.nlm.nih.gov/).
We carried out a maximum-likelihood (ML) phylogenetic analysis of 21 unique ingroup and four unique outgroup haplotypes. First, we identified an appropriate model of molecular evolution using paup* (Swofford, 1998) and the methods of Frati et al. (1997) and Buckley et al. (2002). Both likelihood-ratio tests and the Akaike information criterion supported the GTR+ Γ model, and it was used in all phylogenetic analyses. A heuristic search was then carried out using 10 random addition sequence replicates and Tree Bisection-Reconnection (TBR) branch swapping. ML model parameters were initially estimated using a neighbour-joining topology and fixed for the first heuristic search. Model parameters were then re-estimated using the resulting ML tree, and fixed for a second heuristic search to confirm the results. In order to assess nodal support, we performed 1000 bootstrap pseudoreplicates using model parameters estimated from the ML tree, one addition sequence replicate, retaining one tree per replicate and TBR branch swapping.
The mean genetic distance for the ingroup sequences, mean genetic distances within each of the host-plant groups and net mean genetic distances between the groups were calculated using mega 4.0 (Tamura et al., 2007) and the Tamura-Nei + Γ model, with alpha fixed at 0.25 based on estimates from the phylogenetic analysis (Tamura & Nei, 1993). mega does not include the GTR model, so we chose a model of similar complexity. Standard errors of these values were calculated using 1000 bootstrap pseudoreplicates. We then calculated nucleotide and haplotype (gene) diversity using Arlequin 2.0 (Schneider et al., 2000). Finally, we used Arlequin 2.0 to infer the population genetic structure with an analysis of molecular variance (amova). Here, we defined two populations corresponding to the host-associated populations. We used the Tamura and Nei model with a gamma correction, alpha = 0.25 (Tamura & Nei, 1993) and significance tests based on 1023 permutations.
- Top of page
- Materials and methods
We found significant morphological differences between adult gall midges reared from S. rugosa and S. gigantea (Table 1). Based on wing-length measurements, individuals reared from S. rugosa were larger than individuals from S. gigantea, (F1,106 = 16.0, P = 0.0001), and males were larger than females regardless of host and locality (F1,106 = 135.7, P < 0.0001). The interactions between host and sex and between host and locality were not significant (P = 0.28 and 0.33 respectively). The ratio in length of tergite 8 to tergite 6 of the abdomen was allometrically higher in both males and females from S. rugosa than in those from S. gigantea (females: F1,44 = 19.3, P < 0.001; males: F1,31 = 7.4, P = 0.01), hence male and female post-abdomens are longer in the population associated with S. rugosa. Because segment 8 of the female abdomen constitutes the basal part of the ovipositor, this observation means that females associated with S. rugosa have relatively longer ovipositors.
|S. rugosa||S. gigantea|
|Male||Wing length||2.77 ± 0.03 (N = 31)||2.58 ± 0.04 (N = 28)|
|Tergite ratio||0.47 (N = 14)||0.41 (N = 19)|
|Female||Wing length||2.29 ± 0.04 (N = 27)||2.19 ± 0.04 (N = 24)|
|Tergite ratio||1.70 (N = 25)||1.47 (N = 21)|
The percentage of mating events within the same host-associated populations was significantly higher than between host-associated populations (Table 2), indicating assortative mating in D. folliculi based on host association. Because individuals reared from S. gigantea mated more readily than those from S. rugosa, we also found significant differences in mating percentages between the control groups (i.e. within host-plant species) (χ2 = 14.78, P = 0.0001). No difference was found between the two combinations that represented crosses between individuals from different hosts.
|gigantea♀ × gigantea♂||129||87.6a|
|rugosa♀ × rugosa♂||91||65.9b|
|gigantea♀ × rugosa♂||71||31.0c|
|rugosa♀ × gigantea♂||68||38.2c|
Oviposition-choice and offspring survival experiments
Forty-five of 87 (52%) mated females reared from S. rugosa (‘rugosa females’) and 42 of 137 (31%) mated females reared from S. gigantea (‘gigantea females’) oviposited on potted plants. All of the rugosa females that oviposited and 38 of the 42 gigantea females (90%) chose their natal host species for oviposition, thus indicating significant preference of females for oviposition on their natal host (χ2 = 71.7, P < 0.0001, Fig. 1).
Larvae had higher performance on S. rugosa than on S. gigantea (χ2 = 51.1, P < 0.0001, Fig. 1). Oviposition by all 45 (100%) rugosa females resulted in gall development, whereas only 10 of the 38 (26%) oviposition events among gigantea females resulted in gall development, one of which was on an S. rugosa plant. All but four of the galls produced adults (93%).
Gall-midge populations associated with S. rugosa and S. gigantea were genetically divergent. We identified 22 unique ingroup haplotypes from 35 individuals, and the alignment of 646 bp was unambiguous and contained no indels. The overall mean mtDNA-corrected genetic distance was 0.148 ± 0.024 and the net mean genetic distance between the S. gigantea and S. rugosa groups was 0.188 ± 0.041. ML phylogenetic analysis found genetic differentiation between the host-associated populations with strong bootstrap support (Fig. 2). Within-host mtDNA distances showed a higher level of variation in the population associated with S. rugosa than in the population associated with S. gigantea [0.020 ± 0.003 (N = 14) and 0.015 ± 0.004 (N = 21) respectively]. Similarly, both haplotype (gene) and nucleotide diversity were higher among S. rugosa individuals (0.96 ± 0.036 and 0.042 ± 0.022 respectively) than among S. gigantea individuals (0.89 ± 0.057 and 0.0129 ± 0.007). The amova showed that 78.9% of the genetic variation was between host-associated populations. The corresponding FST value of 0.789 was found to be highly significant (P < 0.0001), suggesting low levels of gene flow between host-associated populations, although increased sample sizes would be needed to confirm this.
Occurrence of natural enemies
Natural enemies found in D. folliculi galls included parasitic wasps from the families Pteromalidae and Torymidae, the inquilinous gall midge Macrolabis americana and larvae of an unidentified lepidopteran that fed on the gall tissue (Dorchin et al., 2007). Each of these natural enemies can have a devastating effect on larvae of the gall inducer. For instance, a single parasitoid larva usually killed many of the gall-midge larvae, and if more than one parasitoid was present, the majority of the gall midges were killed. Larvae of the inquiline M. americana were often found in large numbers in a gall, where they caused indirect mortality of the gall inducer, most probably due to competition for gall resources. Feeding by a single lepidopteran caterpillar typically consisted of tunnelling through the centre of the gall and physically destroying it. The percentage of galls attacked by each of the types of natural enemies was significantly higher on S. rugosa than on S. gigantea (wasps: χ2 = 9.2, P = 0.0024; inquilines: χ2 = 50.1, P < 0.001; others: χ2 = 22.7, P < 0.001), and overall attack by natural enemies was about three times higher in galls on S. rugosa (χ2 = 60.1, P < 0.001, Fig. 3).
- Top of page
- Materials and methods
Evidence for host-associated differentiation
The data we present provide multiple types of evidence showing D. folliculi has differentiated into two distinct populations on its two hosts, S. rugosa and S. gigantea, which represent cryptic species or well-established host races. This is one of a handful of cases for which a full array of genetic, behavioural, morphological and ecological evidence for HAD is available (e.g. R. pomonella, Eurosta solidaginis, Zeiraphera diniana). Dasineura folliculi exhibited assortative mating and host choice for oviposition (i.e. host fidelity), and mtDNA sequence data showed the host-associated populations are genetically divergent.
Even in the absence of host plants, the gall midges showed preference to mate within the same host-associated population, a phenomenon that was not found in E. solidaginis, for example, where no assortative mating was observed in the absence of hosts (Craig et al., 1993). Difficulties stemming from the tiny size, ephemeral adult longevity and other aspects of the biology of gall midges precluded experiments in the presence of hosts. We assumed that differences in mate choice and host fidelity would characterize the host-associated populations in general despite possible variations among different field sites; hence we pooled all individuals from the same host from all field sites for the analyses. Although this set-up precludes a statistical estimation of inter-population variation, our data suggest very limited variation at least with regard to mate choice (N. Dorchin, unpublished data).
The ML tree based on our mtDNA data placed two individuals that were reared from S. rugosa in the clade otherwise representing the population associated with S. gigantea. The occurrence of these exceptions, together with the trend found in the mating experiments, implies that the host-associated populations of D. folliculi experience a certain amount of gene flow. This amount is considerably lower than that found between host races of E. solidaginis (Itami et al., 1998) and of the gall moth Gnorimoschema gallaesolidaginis (Stireman et al., 2005), and is similar to the amount of gene flow between the sister gall-midge species Rhopalomyia solidaginis and R. capitata (Stireman et al., 2005). However, it is noteworthy that in cases of rapid divergence, an incomplete lineage sorting may occur, thereby sorting of genes into monophyletic trees lags behind a speciation event that is already manifested by adaptive traits (Forister et al., 2008).
Allochronic emergence or activity times of adults from different host-associated populations were proposed as important premating factors that prevent gene flow in some species (e.g. Bush, 1969; Smith, 1988; Craig et al., 1993; Nason et al., 2002; Thomas et al., 2003). However, these do not seem to play a major role in the case of D. folliculi, because we did not find appreciable differences in larval development and adult emergence times between the hosts (Dorchin et al., 2007). Although peaks of gall development and adult emergence on S. rugosa usually lag about a week behind those on S. gigantea (N. Dorchin, personal observation), there is considerable overlap in adult emergence between the populations, which means midges from different hosts could potentially interbreed.
A primary post-mating cause for reproductive isolation in D. folliculi appears to be host choice by ovipositing females. If adults prefer to mate on the same host species in which they developed, then an oviposition choice made by a female predisposes her offspring to choose the same host for mating. Coupled with assortative mating, host fidelity by ovipositing females can result in reproductive isolation leading to genetic differentiation between host-associated populations. In D. folliculi, galls developing in the greenhouse contained much larger numbers of larvae than field-collected galls (Dorchin et al., 2007), suggesting that, under field conditions, females spread their eggs among several plants. However, in our greenhouse experiments, females deposited their eggs in a single plant despite the availability of an individual of the alternate host species. This behaviour indicates that females not only choose their natal host species for oviposition but also specifically avoid non-natal hosts (see Forbes et al., 2005).
A second potential source of post-mating isolation is reduced fitness in hybrids (Craig et al., 1997; Feder & Filchak, 1999; Via et al., 2000). Although hybrids between the S. rugosa and the S. gigantea populations of D. folliculi are viable (N. Dorchin, unpublished data), we currently do not know whether their fitness differs from that of the pure populations.
We found both strong host fidelity and morphological differences between host-associated populations in D. folliculi. The differences in body size between the populations could result from variable environmental factors, but may be genetically based, given that these differences were consistent in time and space. Morphological differences between host-associated populations have been documented in several systems (Wood, 1980; Rossi et al., 1999; Nosil et al., 2002; Pappers et al., 2002; Emelianov et al., 2003; Ikonnen et al., 2003; Tabuchi & Amano, 2003; Tokuda et al., 2004; Blair et al., 2005), and when directly associated with mate attraction (Brown & Cooper, 2006) or host use (Carroll et al., 1997; Diegisser et al., 2003), they contribute to the establishment of genetic differentiation. In D. folliculi, we found that females from S. rugosa are larger and have longer ovipositors than those from S. gigantea. A possible explanation is that a longer ovipositor allows better manoeuvrability and thus better access to the pubescent S. rugosa buds, whereas the smooth surface of S. gigantea plants does not require such an adaptation. Indeed, longer ovipositors among Dasineura species are correlated with less accessible oviposition sites, and when accompanied by a divided tergite 8, as in the case of D. folliculi, they are assumed to permit more flexibility of the post-abdomen (Sylvén & Tastás-Duque, 1993). Despite the overall morphometric differences, there may be some overlap between the host-associated populations, especially among males; hence we are reluctant to use these differences alone for distinguishing between the populations.
Direction of the shift
At least four lines of evidence suggest that the direction of the shift in D. folliculi was from S. rugosa to S. gigantea. First, our genetic data show that the average within-host mtDNA distance and nucleotide and gene diversity are lower in the S. gigantea population than in the S. rugosa population. Lower genetic diversity can be expected in a derived lineage due to founder effects. Second, ‘oviposition mistakes’ were observed only among gigantea females in our greenhouse oviposition trials, whereas all 45 females from S. rugosa oviposited on their natal host. The weaker host fidelity exhibited by gigantea females could result from the fact that the development of host preference and recognition mechanisms occurred more recently in this population. Third, populations shifting to a new host are expected to be less adapted to utilizing it and will therefore have reduced fitness on this host compared with populations on the original host (Prokopy et al., 1988; Jaenike, 1990; Brown et al., 1995). In our greenhouse experiments, only 9 of 38 oviposition events (24%) by S. gigantea females on S. gigantea resulted in successful gall induction and development, compared to a success rate of 100% among S. rugosa females on S. rugosa. In cases where oviposition by S. gigantea females did not lead to gall formation, larvae hatched from the eggs and fed on the buds for several days, but died before moulting into second instars. This observation suggests that S. gigantea gall midges are less adapted to their host than those from S. rugosa.
A fourth type of evidence is the level of natural-enemy attack on galls of the two host-associated populations. Enemy-reduced space on a derived host has been suggested as a factor facilitating host shifts or maintaining isolation between conspecific populations in several systems (e.g. Abrahamson et al., 1994; Brown et al., 1995; Feder, 1995; Thomas et al., 2003; Blair et al., 2005). Our results show that for D. folliculi, S. gigantea offers enemy-reduced space with regard to the two main enemy taxa we documented (parasitic wasps and gall-midge inquilines). Although the percentage of galls affected by parasitic wasps and by inquilines varied among study sites (N. Dorchin & G. Lee, unpublished) and generations (Dorchin et al., 2007), it was almost always higher on S. rugosa regardless of the relative local abundance of the two hosts.
Mitochondrial DNA analyses showed that S. gigantea was the derived host in three other case studies of HAD among goldenrod insects, namely, E. solidaginis, G. gallaesolidaginis and R. solidaginis + R. capitata, all of which shifted to S. gigantea from S. altissima (Waring et al., 1990; Brown et al., 1996; Stireman et al., 2005). In all three cases the gigantea population exhibited less genetic variation than the altissima population. This parallel direction of the shift may reflect the fact that S. gigantea is simply less abundant than the other two species due to its narrower ecological requirements (Abrahamson et al., 2005). Alternatively, it is possible that S. gigantea is a younger species, or that it has expanded its distribution into habitats where S. altissima and S. rugosa were already established, thus offering an underused niche to the insects that were associated with these hosts (see Rice, 1987; Berlocher & Feder, 2002). Given the complex and yet unresolved classification of the Solidago species in this group (Zhang, 1996), and the lack of knowledge on their historic distribution in North America (Marks, 1983), these explanations remain speculative. However, it is noteworthy that a shift to S. gigantea has now been documented in four different systems that show varying amounts of gene flow.
All three cases of goldenrod gall-inducing insects mentioned above, as well as the host races of the beetle Mordellistena convicta (Mordellidae) that attack E. solidaginis galls (Blair et al., 2005), involved populations only from S. altissima and S. gigantea, although each of these insects is also found occasionally on S. rugosa. Several other highly specialized herbivores that are shared by the three largely sympatric host plants include the gall-inducing fruit fly Procecidochares atra (Tephritidae) and at least five additional species of gall midges (N. Dorchin, unpublished). These observations imply some involvement of S. rugosa in host shifts and speciation processes in these insects. It appears that despite their presumed distant phylogenetic relations (Zhang, 1996), S. rugosa and the S. altissima species complex (subsection Triplinerviae of Semple & Cook, 2006) are at least physiologically similar with respect to the conditions they offer for the development of gall inducers.
- Top of page
- Materials and methods
Dasineura folliculi joins three species of gall inducers and one inquiline species from goldenrods that show HAD, thus further establishing the goldenrod insect fauna as a model system for the study of ecological diversification in phytophagous insects. Gall inducers in this community provide an excellent system for investigating possible factors that promote HAD, which is particularly true for gall midges, being by far the most numerous and biologically diverse group among goldenrod gallers. Until further molecular study, we cannot say whether the host-associated populations of D. folliculi represent well-established host races or whether they are cryptic species. The high genetic differentiation is indicative of cryptic species, but interspecific copulation, production of hybrids and possible oviposition mistakes suggest a lower degree of speciation. Major reproductive isolation factors found in this study are assortative mating and host fidelity by ovipositing females. We also found that S. gigantea provided enemy-reduced space to the gall midges, thus possibly counterbalancing their poor performance on this host.
- Top of page
- Materials and methods
We thank Alan Snyder, Brian Lucas, Catherine Blair, Anna Latimer, Patti Scarff and Jeff Williams for help with field and laboratory work, Michael J. Wise for assistance in statistical analyses, and Catherine Blair and Michael J. Wise for discussions and comments on earlier versions of the manuscript. This study was supported by Bucknell’s David Burpee Plant Genetics Chair endowment and by National Science Foundation Grant DEB-0343633 to W.G.A. and Jason T. Irwin.
- Top of page
- Materials and methods
- 1997. Evolutionary Ecology across Three Trophic Levels: Goldenrods, Gallmakers, and Natural Enemies. Princeton University Press, Princeton, NJ. &
- 1994. Gallmaker speciation: an assessment of the roles of host-plant characters and phenology, gallmaker competition, and natural enemies. In: Gall-forming Insects (P.Price, W.Mattson & Y.Baranchikov, eds), pp. 208–222. USDA Forest Service, North Central Experiment Station, St Paul, MN General Technical Report NC-174. , , , , , , , , &
- 2005. Ecological divergence among five co-occurring species of old-field goldenrods. Plant Ecol. 177: 43–56. , , &
- 2002. Sympatric speciation in phytophagous insects: moving beyond controversy? Annu. Rev. Entomol. 47: 773–815. &
- 2005. Genetic isolation between two sympatric host plant races of the European corn borer, Ostrinia nubilalis Hübner. II: assortative mating and host-plant preferences for oviposition. Heredity 94: 264–270. , , , , , &
- 2005. Cryptic speciation and host-race formation in a purportedly generalist tumbling flower beetle. Evolution 59: 304–316. , , &
- 2006. Evolution of wing pigmentation patterns in a tephritid gallmaker: divergence and hybridization. In: Galling Arthropods and their Associates (K.Ozaki, J.Yukawa, T.Ohgushi & P.W.Price, eds), pp. 253–263. Springer-Verlag, Tokyo, Japan. &
- 1995. The role of natural-enemy escape in a gallmaker host-plant shift. Oecologia 104: 52–60. , , &
- 1996. Mitochondrial DNA phylogeography of host races of the goldenrod ball gallmaker, Eurosta solidaginis (Diptera: Tephritidae). Evolution 50: 777–786. , &
- 2002. Combined data, Bayesian phylogenetics, and the origin of the New Zealand cicada genera. Syst. Biol. 51: 4–18. , , &
- 1969. Sympatric host race formation and speciation in frugivorous flies of the genus Rhagoletis (Diptera: Tephritidae). Evolution 23: 237–251.
- 1997. Genetic differentiation of fitness-associated traits among rapidly evolving populations of the soapberry bug. Evolution 51: 1182–1188. , &
- 1993. Behavioral evidence for host-race formation in Eurosta solidaginis. Evolution 47: 1696–1710. , &
- 1997. Hybridization studies on the host races of Eurosta solidaginis: implications for sympatric speciation. Evolution 51: 1552–1560. , &
- 2003. Genetic and morphological differentiation in Tephritis bardanae (Diptera: Tephritidae): evidence for host-race formation. J. Evol. Biol. 17: 83–93. , , &
- 2007. Taxonomy, life history, and population sex ratios of North American Dasineura (Diptera: Cecidomyiidae) on goldenrods (Asteraceae). Ann. Entomol. Soc. Am. 100: 539–548. , , , &
- 2002. Host races in plant-feeding insects and their importance in sympatric speciation. Philos. Trans. R. Soc. Lond. B Biol. Sci. 357: 471–492. &
- 2003. Host choice promotes reproductive isolation between host races of the larch budmoth Zeiraphera diniana. J. Evol. Biol. 16: 208–218. , , &
- 2003. One host shift leads to another? Evidence of host-race formation in a predaceous gall-boring beetle. Evolution 57: 168–172. , &
- 1995. The effects of parasitoids on sympatric host races of Rhagoletis pomonella (Diptera: Tephritidae). Ecology 76: 801–813.
- 1999. It’s about time: the evidence for host plant-mediated selection in the apple maggot fly, Rhagoletis pomonella, and its implications for fitness trade-offs in phytophagous insects. Entomol. Exp. Appl. 91: 211–225. &
- 1994. DNA primers for amplification of mitochondrial cytochrome c oxidase subunit I from diverse metazoan invertebrates. Mol. Mar. Biol. Biotechnol. 3: 294–299. , , , &
- 2005. Habitat avoidance: overlooking an aspect of host-specific mating and sympatric speciation? Evolution 59: 1552–1559. , &
- 2008. Considering evolutionary processes in the use of single-locus genetic data for conservation, with examples from the Lepidoptera. J. Insect Conserv. 12: 37–51. , , , &
- 1997. Evolution of the mitochondrial cytochrome oxidase II gene in Collembola. J. Mol. Evol. 44: 145–158. , , &
- 1989. The Plant Feeding Gall Midges of North America. Cornell University Press, Ithaca, NY.
- 2003. Evidence for host race formation in the leaf beetle Galerucella lineola. Entomol. Exp. Appl. 108: 179–185. , , &
- 1998. Factors affecting gene flow between the host races of Eurosta solidaginis. In: Genetic Structure and Local Adaptation in Natural Insect Populations: Effect of Ecology, Life History, and Behavior (S.Mopper & S.Y.Strauss, eds), pp. 375–407. Chapman & Hall, New York. , &
- 1981. Criteria for ascertaining the existence of host races. Am. Nat. 117: 830–834.
- 1990. Host specialization in phytophagous insects. Annu. Rev. Ecol. Syst. 21: 243–273.
- 1984. Enemy free space and the structure of ecological communities. Biol. J. Linn. Soc. 23: 269–286. &
- 2005. Assortative mating in sympatric host races of the European corn borer. Science 308: 258–260. , , , , &
- 1983. On the origin of the field plants of the Northeastern United States. Am. Nat. 122: 210–228.
- 1988. The gall insects of goldenrods (Compositae: Solidago) with a revision of the species of Rhopalomyia (Diptera: Cecidomyiidae). MS Thesis. Cornell University, Ithaca, NY.
- 1988. Genetic differences between host races of Rhagoletis pomonella. Nature 336: 64–66. , &
- 2002. Host-associated genetic differentiation in the goldenrod elliptical-gall moth, Gnorimoschema gallaesolidaginis (Lepidoptera: Gelechiidae). Evolution 56: 1475–1488. , &
- 2002. Host-plant adaptation drives the parallel evolution of reproductive isolation. Nature 417: 440–443. , &
- 2002. Genetically based polymorphisms in morphology and life history associated with putative host races of the water lily leaf beetle, Galerucella nymphaeae. Evolution 56: 1610–1621. , , &
- 1980. Interactions among three trophic levels: influence of plants on interactions between insect herbivores and natural enemies. Annu. Rev. Ecol. Syst. 11: 41–65. , , , , &
- 1988. Behavioral evidence for host races in Rhagoletis pomonella flies. Oecologia 76: 138–147. , &
- 1987. Speciation via habitat specialization: the evolution of reproductive isolation as a correlated character. Evol. Ecol. 1: 301–314.
- 1999. Evidence for host-associated races in a gall-forming midge: trade-offs in potential fecundity. Ecol. Entomol. 24: 95–102. , , &
- 2005. Ecological speciation. Ecol. Lett. 8: 336–352. &
- 2000. Arlequin ver. 2.000: A Software for Population Genetics Data Analysis. Genetics and Biometry Laboratory, University of Geneva, Geneva, Switzerland. , &
- 2006. Solidago. In: Flora of North America, Vol. 20 (Flora of North America Editorial Committee, eds), pp. 107–166. Oxford University Press, New York, NY. &
- 1988. Heritable divergence of Rhagoletis pomonella host races by seasonal asynchrony. Nature 336: 66–67.
- 2005. Host-associated genetic differentiation in phytophagous insects: general phenomenon or isolated exceptions? Evidence from a goldenrod-insect community. Evolution 59: 2573–2587. , &
- 1998. paup*. Phylogenetic Analysis Using Parsimony (*and Other Methods), Version 4b10. Sinauer Associates, Sunderland, MA.
- 1993. Adaptive, taxonomic, and phylogenetic aspects of female abdominal features in Oligotrophini (Diptera: Cecidomyiidae), and four new Dasineura species fro the western Palearctic. Zool. Scr. 22: 277–298. &
- 2003. Host-associated differences in emergence pattern, reproductive behavior and life history of Asteralobia sasakii (Monzen) (Diptera: Cecidomyiidae) between populations on Ilex crenaa and I. integra (Aquifoliaceae). Appl. Entomol. Zool. 38: 501–508. &
- 1993. Estimation of the number of nucleotide substitutions in the control region of mitochondrial DNA in humans and chimpanzees. Mol. Biol. Evol. 10: 512–526. &
- 2007. mega4: Molecular Evolutionary Genetics Analysis (mega) software version 4.0. Mol. Biol. Evol. 24: 1596–1599. , , &
- 1989. Sympatric speciation in insects: perception and perspective. In: Speciation and its Consequences (D.Otte & J.A.Endler, eds), pp. 307–344. Sinauer Associates, Sunderland, MA. &
- 2003. Genetic isolation between two sympatric host-plant races of the European corn borer, Ostrinia nubilalis Hübner. I. sex pheromone, moth emergence timing, and parasitism. Evolution 57: 261–273. , , , &
- 2004. Inter- and intraspecific comparisons between Asteralobia gall midges (Diptera: Cecidomyiidae) causing axillary bud galls on Ilex species (Aquifoliaceae): species identification, host range, and mode of speciation. Ann. Entomol. Soc. Am. 97: 957–970. , , &
- 1999. Reproductive isolation between sympatric races of pea aphids. I. Gene flow restriction and habitat choice. Evolution 53: 1446–1457.
- 2001. Sympatric speciation in animals: the ugly duckling grows up. Trends Ecol. Evol. 16: 381–390.
- 2000. Reproductive isolation between divergent races of pea aphids on two hosts. II. Selection against migrants and hybrids in the parental environments. Evolution 54: 1626–1637. , &
- 1990. Genetic differentiation in the gallformer Eurosta solidaginis (Diptera: Tephritidae) along host plant lines. Evolution 44: 1648–1655. , &
- 1980. Divergence in the Enchenopa binotata Say complex (Homoptera: Membracidae) effected by host plant adaptation. Evolution 34: 147–160.
- 1996. A molecular biosystematic study on North American Solidago and related genera (Asteraceae: Astereae) based on chloroplast DNA RFLP analysis. PhD Dissertation. University of Waterloo, ON, Canada.
- Top of page
- Materials and methods
Samples used for mtDNA analysis with host plants, collecting sites, dates, and GenBank accession numbers. All specimens were collected in Pennsylvania, USA unless otherwise noted.
|Sample name||Host plant||Site||Date||Accession no.|
|gig2541||Solidago gigantea||Turbot||1 July 2006||EU375690|
|gig2542||Solidago gigantea||Turbot||1 July 2006||EU375683|
|gigLB137||Solidago gigantea||Stein Ln. Lewisburg||13 July 2005||EU375666|
|gigCr||Solidago gigantea||Pottsgrove||9 August 2005||EU375667|
|gigFR1||Solidago gigantea||Furnace Rd, Lewisburg||30 May 2006||EU375682|
|gigFR2||Solidago gigantea||Furnace Rd, Lewisburg||30 May 2006||EU375677|
|gigH1||Solidago gigantea||Hughesville||8 June 2006||EU375688|
|gigH2||Solidago gigantea||Hughesville||8 June 2006||EU375689|
|gigLair1||Solidago gigantea||Lairdsville||22 June 2005||EU375668|
|gigLair2||Solidago gigantea||Lairdsville||22 June 2005||EU375669|
|gigM1||Solidago gigantea||Montour Preserve||1 July 2006||EU375695|
|gigM2||Solidago gigantea||Montour Preserve||1 July 2006||EU375696|
|gigMa||Solidago gigantea||Mauses Creek||19 July 2005||EU375681|
|gigNA216||Solidago gigantea||Bucknell University Chillisquaque Creek Natural Area||21 June 2005||EU375665|
|gigNA67||Solidago gigantea||Bucknell University Chillisquaque Creek Natural Area||26 July 2005||EU375679|
|gigNA98||Solidago gigantea||Bucknell University Chillisquaque Creek Natural Area||9 September 2005||EU375684|
|gigSt8||Solidago gigantea||Stein Ln, Lewisburg||8 August 2005||EU375664|
|gigV1||Solidago gigantea||Mauses Creek||10 July 2006||EU375691|
|gigV2||Solidago gigantea||Mauses Creek||10 July 2006||EU375678|
|rugLB1||Solidago rugosa||Lewisburg||20 June 2005||EU375670|
|rugLB3||Solidago rugosa||Lewisburg||20 June 2205||EU375674|
|rugH1||Solidago rugosa||Hughesville||8 June 2006||EU375699|
|rugH2||Solidago rugosa||Hughesville||8 June 2006||EU375700|
|rugLair2||Solidago rugosa||Lairdsville||22 June 2005||EU375685|
|rugLair226||Solidago rugosa||Lairdsville||22 June 2005||EU375686|
|rugM1||Solidago rugosa||Montour Preserve||4 June 2006||EU375675|
|rugM2||Solidago rugosa||Montour Preserve||4 June 2006||EU375692|
|rugNA1||Solidago rugosa||Bucknell University Chillisquaque Creek Natural Area||27 June 2005||EU375671|
|rugNA197||Solidago rugosa||Bucknell University Chillisquaque Creek Natural Area||19 July 2005||EU375680|
|rugNA2||Solidago rugosa||Bucknell University Chillisquaque Creek Natural Area||27 June 2005||EU375672|
|rugNA3||Solidago rugosa||Bucknell University Chillisquaque Creek Natural Area||27 June 2005||EU375673|
|rugRick||Solidago rugosa||Ricketts Glen||29 June 2006||EU375701|
|rugSN1||Solidago rugosa||Selinsgrove||14 June 2006||EU375698|
|rugSN2||Solidago rugosa||Selinsgrove||14 June 2006||EU375702|
|rugV||Solidago rugosa||Mauses Creek||17 July 2006||EU375693|
|Dasineura carbonaria||Euthamia graminifolia||Bucknell University Chillisquaque Creek Natural Area||24 May 2006||EU375703|
|Dasineura n.sp.1(Syd)||Leptospermum laevigatum||Sydney, NSW, Australia||21 August 2004||EU375687|
|Dasineura n.sp.2(Vic)||Leptospermum laevigatum||Pearcedale, Victoria, Australia||23 September 2004||EU375694|
|Mayetiola destructor||Triticum||Mishmar Hanegev, Israel||23 February 2005||EU375697|