Introduction
Insect herbivores and their natural enemies, including many parasitoids, constitute a substantial part of global insect diversity and suggest plant–herbivore–parasitoid relationships as one of the major components of terrestrial ecosystems (Hamilton et al., Reference Hamilton, Basset, Benke, Grimbacher, Miller, Novotný, Samuelson, Stork, Weiblen and Yen2010; Hawkins, Reference Hawkins1994; Price, Reference Price2002; Quicke, Reference Quicke1997; Smith et al., Reference Smith, Rodriguez, Whitfield, Deans, Janzen, Hallwachs and Hebert2008; Sugiura, Reference Sugiura2007). Currently, more than half of all known terrestrial species are part of plant–herbivore–parasitoid food webs (Leppänen et al., Reference Leppänen, Altenhofer, Liston and Nyman2013; Novotný et al., Reference Novotný, Miller, Baje, Balagawi, Basset, Cizek, Craft, Dem, RA, Hulcr and Leps2010; Smith et al., Reference Smith, Rodriguez, Whitfield, Deans, Janzen, Hallwachs and Hebert2008). These food webs include both organisms providing ecosystem services such as pollinators and natural enemies of various organisms, as well as economically important pests in agriculture and forestry (González et al., Reference González, Štrobl, Janšta, Hovorka, Kadlec and Knapp2022; Kremen and Chaplin-Kramer, Reference Kremen and Chaplin-Kramer2007; Letourneau et al., Reference Letourneau, Jedlicka, Bothwell and Moreno2009; Martin et al., Reference Martin, Reineking, Seo and Steffan-Dewenter2013). The interactions of these organisms within food webs have led herbivore hosts to evolve a plethora of various defence strategies (Aoyama and Ohshima, Reference Aoyama and Ohshima2019; Brodie and Smatresk, Reference Brodie and Smatresk1990; Stankowich and Campbell, Reference Stankowich and Campbell2016).
A common primary defence strategy of herbivores against enemies such as parasitoids is the creation of physical barriers. Such a strategy is known from a large number of small arthropods, living inside plant material, which, apart from food, serves as protective physical barrier against predators and parasitoids (Aoyama and Ohshima, Reference Aoyama and Ohshima2019). The ability of insects to use plants as their shelter by creating various structures (leaf rolls, galls, mines) is a key factor influencing the parasitoid community (Hering, Reference Hering1951; Fernandes and Price, Reference Fernandes and Price1988; Sinclair and Hughes, Reference Sinclair and Hughes2010; Hrček et al., Reference Hrček, Miller, Whitfield, Shima and Novotny2013). A partially hidden lifestyle (e.g. leaf rolls) as a defence strategy against parasitoids appears to be the least advantageous strategy compared to living on the surface of plants (Hrček et al., Reference Hrček, Miller, Whitfield, Shima and Novotny2013), where individuals are more frequently attacked by predators (Libra et al., Reference Libra, Tulai, Novotný and Hrček2019). Hawkins (Reference Hawkins1994) and Hrček et al. (Reference Hrček, Miller, Whitfield, Shima and Novotny2013) stated that parasitism rate and parasitoid species richness increase with the degree of concealment from exposed feeders through leaf rollers, leaf tiers, and case-bearers; reach maximum values in leaf miners and gallers; and then decrease again in even more concealed borers and root-feeders. In contrast, other studies have shown that exposed hosts can exhibit high parasitism rates (e.g. Konvičková et al., Reference Konvičková, John, Konvička, Rindoš and Hrček2024; Sarfraz et al., Reference Sarfraz, Keddie and Dosdall2005), and that other factors, such as seasonality, may be more important than host concealment (Le Corff et al., Reference Le Corff, Marquis and Whitfield2000). To address this issue, it is necessary to obtain and analyse additional data from various locations around the world.
Despite that, life inside the plant tissue (e.g. leaves) and creating mines is one of the most common life strategies of phytophagous insects (Tooker and Giron, Reference Tooker and Giron2020). Even if mining is not a perfect strategy against parasitoids, mines and specifically their shapes have been suggested as a defence mechanism to avoid or confuse parasitoids (Aoyama and Ohshima, Reference Aoyama and Ohshima2019; Connor and Taverner, Reference Connor and Taverner1997; Needham et al., Reference Needham, Frost and Tothill1928). During larval development, some species of mining butterflies (e.g. Acrocercops transecta Meyrick, 1931 (Lepidoptera: Gracillaridae)) create variously complex and shaped corridors inside plant leaves and increase the area over which the parasitoid must search for its host or provide a space to escape from the parasitoid in case the host is found (Aoyama and Ohshima, Reference Aoyama and Ohshima2019; Ayabe et al., Reference Ayabe, Tuda and Mochizuki2008). As the complexity of mines increases, their level of parasitism decreases. Thus, the complexity of mines is partly a result of the prey–predator relationship between mining species and their parasitoids. For example, the mining moth Phyllonorycter malella (Gerasimov, 1931) (Lepidoptera: Gracillaridae) creates a complex network of mines with unconsumed tissue, which serves as protection (Djemai et al., Reference Djemai, Meyhöfer and Casas2000). The more complex the structure of the mines in the leaf of this species is, the less the caterpillar was parasitised (Aoyama and Ohshima, Reference Aoyama and Ohshima2019). However, the exact role of mining as a defence mechanism against parasitoids is still disputable. Several studies (i.e. Askew, Reference Askew1980; Hawkins, Reference Hawkins1990, Reference Hawkins1994; Hawkins and Lawton, Reference Hawkins and Lawton1987) have shown mining insect species to have more parasitoids than hosts from any other feeding guild (Connor and Taverner, Reference Connor and Taverner1997). However, mining still may be one of the initial strategies to avoid parasitoids. The taxonomic composition of the parasitoid community attacking leaf-mining sawflies (Hymenoptera: Tenthredinidae) has been shown as radically different from that attacking external-feeding sawflies having less species of braconid and ichneumonid (both Hymenoptera: Ichneumonoidea) parasitoids (Pschorn-Walcher and Altenhofer, Reference Pschorn-Walcher and Altenhofer1989). With the decrease in the number of braconid and ichneumonid parasitoids, there was an increase in the number of species from the family Eulophidae (Hymenoptera: Chalcidoidea), which generally parasitise concealed-feeding Lepidoptera, such as leaf miners and leaf rollers (Stireman and Shaw, Reference Stireman, Shaw, Marquis and Koptur2022). This aligns well with observations by Pschorn-Walcher and Altenhofer (Reference Pschorn-Walcher and Altenhofer1989), who suggested that the initial escape from parasitism associated with adopting a concealed-feeding habit, such as leaf mining or rolling, may have provided sufficient impetus to selectively reinforce these feeding strategies. These adaptations likely led to interactions with a distinct and potentially more specialised group of parasitoids, as evidenced by studies like Hrček et al. (Reference Hrček, Miller, Whitfield, Shima and Novotny2013).
Comparing the parasitism rates and the success of different feeding strategies, such as leaf-mining and leaf-rolling, provides a deeper understanding of how these strategies function as defences against parasitoids. Some microlepidopteran species rely exclusively on the mining strategy, while others adopt a combination of mining and rolling during their larval stages (Tooker and Giron, Reference Tooker and Giron2020). This raises the question of how these two strategies are comparable in terms of effectiveness against parasitoid attacks. Therefore, it is essential to evaluate the benefits and limitations of each strategy to better understand their evolutionary significance and ecological implications.
In this study, we analysed the parasitoid community of two microlepidopteran species on hops (Humulus lupulus L. (Rosales: Cannabaceae)), Caloptilia fidella (Reutti, 1853) (Lepidoptera: Gracillaridae) and Cosmopterix zieglerella (Hübner, 1810) (Lepidoptera: Cosmopterigidae), with similar mine complexity but different life strategies through the larval life cycle.
With a wingspan of 9–12 mm, Caloptilia fidella is one of the smallest West-Palaearctic moths in the genus Caloptilia Hübner, 1825. This species is bivoltine, with larvae appearing mainly in July and September, while adults overwinter. The larvae primarily feed on hop but occasionally also on Celtis australis. During development, the larvae first create a triangular leaf mine (1st–3rd instar) before moving to feed externally within a rolled leaf tip or lobe (4th and 5th instars). Pupation occurs in silken cocoons on the underside of the host plant’s leaves (fig. 1A–D; Baugnée and Prins, Reference Baugnée and Prins2010; Laštůvka et al., Reference Laštůvka, Laštůvka, Liška and Šumpich2018; Watson et al., Reference Watson, Eaton and Mcclennon-Warnock2021).
The moth Cosmopterix zieglerella (a wingspan measuring 8–11 mm) is distributed throughout the Palaearctis. This species is monovoltine, with larval activity occurring from July to September. The larvae hibernate within a silken cocoon located in ground detritus. They exhibit monophagy, feeding exclusively on hop. During the initial stages, the larvae construct a narrow, irregular linear mine along the leaf vein, lined with silk and serving as a protective shelter. The larvae of C. zieglerella are green during the early instars of their development. Prior to pupation, the larva acquires a distinctive coloration characterised by red stripes. As development progresses, the mine expands into a diffuse, irregular yellowish-white blotch containing scattered frass (fig. 1E–H; Koster and Sinev, Reference Koster and Sinev2003; Laštůvka et al., Reference Laštůvka, Laštůvka, Liška and Šumpich2018).
Figure 1. Life cycle of two microlepidoptera living on hops. (A–E) Caloptilia fidella: (A) Mining larva within a leaf mine located between the veins at the leaf axil. (B) Leaf mine and a leaf roll with visible silk threads spun by the larva. (C) Silvery silk cocoon with remnants of the larval pupa on the underside of the leaf. (E–H) Cosmopterix zieglerella: (E) Characteristic leaf mine on hop leaves, which later develops into a broader, flattened shape. (F) Young larva. (G) Final instar larva before pupation. (H) Adult moth. (Photo credits: T. Hovorka, K. Holý, and image (H) by Rudolf Bryner).
This study investigates the interactions between two host lepidopteran species C. fidella, C. zieglerella, and their associated parasitoids, focusing on parasitoid species composition, bionomy, and parasitism rates. The aims are to (i) reveal how the contrasting defensive strategies of the two hosts affect their susceptibility to parasitism, (ii) identify key parasitoid species and their life histories, and (iii) assess ecological factors shaping host–parasitoid dynamics.
Material and methods
Sampling and rearing
Sampling was conducted between 2020 and 2022 across 18 sites in the Czech Republic, Slovakia, Hungary, Romania, and Croatia (Supplementary Table 1). Whole hop leaves containing first- or second-generation larvae of C. fidella (1st‒5th instar) were collected, with a total of 50 leaves per site. For C. zieglerella, leaves with visible mines were collected in as many numbers as available per site (always <50 leaves per location) where the species was present. Collected leaves containing mines or rolls with host species were transported to the laboratory in separate ziplock bags for each host and site. Upon arrival, each leaf was carefully inspected and cleaned of any other insect hosts, most commonly (aphids, spider mites, planthopper nymphs, and whiteflies). A sufficiently large section of the leaf containing the mine (C. zieglerella) or the corresponding host stage (C. fidella) was then clipped to allow the host species or its parasitoid to complete its life cycle and placed in sterile plastic petri dish (100 mm diameter, 15 mm height). The petri dishes were stored under controlled conditions (20°C, 75% humidity, and a 16:8 h light:dark cycle) in Trigon Plus ST 2 B SMART climate chambers (TRIGON PLUS Ltd., Czech Republic) for 40 days. Every 2 days until adult insects emerged, the petri dishes were inspected, and the leaves were moistened with a water solution of the fungicide LUNA® (fluopyram 200 g/L and tebuconazole 200 g/L; Bayer AG, Germany) using a laboratory sprayer to prevent mould during rearing. This fungicide has no adverse effects on arthropods (tested by Central Institute for Supervising and Testing in Agriculture, Czech Republic). The specific developmental biology of the host and its parasitoids was recorded. The biology of the parasitoids was determined during their development on the hosts using forceps and dissecting needles to carefully open the leaf shelters, ensuring minimal disruption to their development. Additionally, parasitoids were monitored to determine whether they were koinobionts or idiobionts and classified as either ectoparasitoids or endoparasitoids. Whenever feasible, individual parasitoids were photographed. Direct observation of parasitoid life strategies was supplemented by post-emergence dissection of leaf mines and shelters to analyse remnants of host larvae, pupae, parasitoid cocoons, and the positions of emergence or exit holes. Additionally, the pupae or cocoons of parasitoids were monitored for the potential occurrence of hyperparasitoids. After emergence, the adults of hosts or parasitoids were stored in 96% ethanol at −20°C until DNA isolation.
Genetic analysis
DNA was extracted from specimens preserved in 96% ethanol using the Qiagen DNeasy Blood & Tissue Kit with a modified non-destructive protocol based on Cruaud et al. (Reference Cruaud, Nidelet, Arnal, Weber, Fusu, Gumovsky, Huber, Polaszek and Rasplus2019). This protocol was optimised for high DNA yield from small insect specimens while preserving them for morphological identification. Specimens were incubated with lysis buffer and proteinase K at 56°C for 20 h under gentle vortexing (250 rpm). Post-incubation, lysate was transferred to fresh tubes for DNA purification. Elution was performed in two steps using a pre-warmed elution buffer, yielding a final volume of 100 µL with high DNA concentration. Additionally, for high-throughput processing, some samples were extracted using the Xiril AG X100 Automatic Workstation according to Haas et al. (Reference Haas, Baur, Schweizer, Monje, Moser, Bigalk and Krogmann2021) excluding the semi-destructive step as we used the whole specimens. Target gene fragments included CO1 (mitochondrial DNA): LCO1490 (5′-GGT CAA ATC ATA AAG ATA TTG-3′) and HCO2198 (5′-TAA ACT TCA GGG TGA CCA AAA AAT CA-3′) (Hebert et al., Reference Hebert, Cywinska, Ball and deWaard2003a) and ITS2 (nuclear ribosomal DNA): ITS2-F (5′-TGT GAA CTG CAG GAC ACA TG-3′) and ITS2-R (5′-AAT GCT TAA ATT TAG GGG GTA-3′) (Campbell et al., Reference Campbell, Steffen-Campbell and Werren1994). Polymerase chain reactions (PCRs) were prepared in 25 µL volumes containing FastGene Optima HotStart Ready Mix (12.5 µL), primers (1 µL), PCR clean H2O (8.5 µL), and template DNA (2 µL). Thermal cycling conditions for CO1 included an initial denaturation at 94°C for 2 min, followed by a two-step protocol with annealing temperatures of 45°C for 1 min (for 5 cycles) and 50°C for 1 min (for 35 cycles), extension at 72°C for 1.5 min for every cycle and for ITS2 amplification followed a similar conditions but used a single annealing temperature of 53°C for 45 s (33 cycles). Successful PCR products were purified using ExoSAP-IT™ (1.5 µL of PCR clean H2O, 1.5 µL of ExoSAP-IT, 1.5 µL of PCR product; cycling conditions – 37°C for 15 min and 80°C for 15 min) and sequenced bidirectionally using Sanger sequencing (Eurofins Genomics, Germany, Ebersberg).
Sequences were processed in Geneious Prime 2023.2.1 (https://www.geneious.com/). Initially, sequences were categorised by taxonomy (Chalcidoidea or Ichneumonoidea), host (C. fidella or C. zieglerella), and gene (CO1 or ITS2). Forward and reverse of every sequence were assembled, and BLAST was used to determine the taxonomic identity of individuals. Consensus sequences were generated after successful assembly. Alignments were performed in Geneious using the MAFFT plugin with the E-INS-i strategy for ITS2 and L-INS-i strategy for CO1 (Katoh and Standley, Reference Katoh and Standley2013), followed by manual verification and trimming to a uniform length. For CO1 sequences, translations to amino acids were checked for stop codons using Geneious for pseudogenes or misalignments. A concatenated alignment of both genes and hosts was created for Chalcidoidea, while for Ichneumonoidea, only the CO1 alignment was used due to the lack of high-quality ITS2 sequences. BLAST was used to verify the taxonomic identity of individuals.
Phylogenetic trees were constructed using the maximum likelihood method in RAxML-HPC2 on XSEDE (8.2.12; Stamatakis, Reference Stamatakis2006) via the CIPRES server (Miller et al., Reference Miller, Pfeiffer and Schwartz2010). The GRTCAT model was used with 1,000 bootstrap replicates (Stamatakis, Reference Stamatakis2006). Bootstrap percentages (BP) ≥70% were considered as strong nodal support. Resulting trees for the concatenated Chalcidoidea dataset and the CO1 alignment for Ichneumonoidea were visualised in FigTree v1.4.4 (Rambaut, Reference Rambaut2018). Bayesian analysis (BA) was performed in MrBayes v3.2.7 (Ronquist et al., Reference Ronquist, Teslenko, van der Mark, Ayres, Darling, Höhna, Larget, Liu, Suchard and Huelsenbeck2012), with input files in NEXUS format set two independent runs with 1,000,000 generations each, saving every 1,000th tree. Parameter files were inspected in Tracer (Rambaut, Reference Rambaut2018) and the first 25% of trees were discarded as burn-in. Posterior probabilities (PP) ≥ 0.95 were considered as strong support, PP < 0.90 as weak. Final trees were visualised in FigTree, with graphical edits in iTOL (https://itol.embl.de/; Letunic and Bork, Reference Letunic and Bork2021).
For species delimitation, both gene fragments were used to define species-level entities (OTUs – operational taxonomic units) using two online-available tools. The first method employed was ASAP (Assemble Species by Automatic Partitioning), as described by Puillandre et al. (Reference Puillandre, Brouillet and Achaz2021) and Zhang et al. (Reference Zhang, Sheikh, Ward, Forbes, Prior, Stone, Gates, Egan, Zhang, Davis and Weinersmith2022). This method proposes the delimitation of hypothetical species based on genetic distances calculated between DNA sequences. The default settings in the online ASAP interface (available at https://bioinfo.mnhn.fr/abi/public/asap/) were used for the species delimitation analysis. The second method, Bayesian Poisson Tree Processes (bPTP), delineated species-level entities using maximum likelihood and the Bayesian implementation of the Poisson Tree Processes (PTP) model for species delimitation (available online at https://species.h-its.org/; Zhang et al., Reference Zhang, Kapli, Pavlidis and Stamatakis2013). For the bPTP analysis, the online interface was configured with model settings of 200,000 Markov chain Monte Carlo generations, thinning of 100, and a burn-in of 0.1. ASAP was run on the CO1 and ITS2 sequence alignments in FASTA format without outgroups, while bPTP was run on the phylogenetic trees obtained from these alignments using the maximum likelihood approach. Additionally, genetic distances across sequences containing CO1 and ITS2 gene fragments were calculated in MEGA 11 using the Kimura-2-parameter model (Tamura et al., Reference Tamura, Stecher and Kumar2021). Groups of similar sequences were identified using a 2% barcode threshold (Hebert et al., Reference Hebert, Ratnasingham and De Waard2003a and Reference Heraty and Hawksb).
Morphological identification
To facilitate morphological identification, after lysis every specimen was washed twice in water (15 min each bath) and stored in 80% EtOH. Drying was performed according to a modified protocol (Heraty and Hawks, Reference Heraty and Hawks1998) using hexamethyldisilazane (HMDS). The procedure involved sequential immersion in 90% and 95% EtOH for 30 min each, followed by two washes in 100% EtOH for 15 min each. Specimens were then placed in HMDS for two 30-min intervals. After removing HMDS, specimens were left to dry. Due to the volatile nature of HMDS, the process was conducted under a fume hood with appropriate protective equipment. Once dried, specimens were mounted on cards. Prepared specimens were sorted and identified using a Leica M205C stereomicroscope (Leica Microsystems). Identification was based on available keys for Ichneumonoidea (e.g. Nixon, Reference Nixon1965; Whitfield and Wagner, Reference Whitfield and Wagner1991) and Chalcidoidea (e.g. Nikol’skaya’s, Reference Nikol’skaya1952; Bouček, Reference Bouček1958; Yoshimoto, Reference Yoshimoto1983; Gibson et al., Reference Gibson, Huber, Woolley, Gibson, Huber and Woolley1997).
Statistical analysis
Data were analysed using the freely available statistical program R (version 4.1.1) with the FSA library. A non-parametric Kruskal–Wallis test was conducted to determine whether parasitism rates differed among the mine, leaf roll, and pupae stages, as the data did not meet the assumptions of normal distribution. To further explore these differences, a post hoc Dunn test with Bonferroni correction was applied. Additionally, the Wilcoxon test was used to compare parasitisation rates between the two generations. To estimate mean parasitism rates for each species, we calculated the proportion of parasitised host-infested leaves for each year and then computed the 3-year average. These mean rates are presented alongside parasitoid emergence data in Table 1.
Table 1. Parasitoid emergence from different host stages of Caloptilia fidella, and average annual parasitism of both species
Note: Values for host stages represent the number of parasitoids reared from each stage and their relative frequency. Total parasitism rates were calculated from full host counts. Mean annual parasitism rates for each species (last two rows) were calculated from yearly totals of parasitised and unparasitised leaves. In contrast, emergence data by host stage (top three rows) reflect only the subset of parasitoids successfully reared and identified by host stage, and therefore do not represent stage-specific parasitism rates across the full host population.
Results
Delimitation of parasitoid species
From a total of 162 collected specimens, 139 were successfully sequenced for either CO1 (93 specimens – 67 individuals of Chalcidoidea, 26 of Ichneumonoidea; fragment length ca. 622 bp) or ITS2 (106 Chalcidoidea; fragment length ranging from 326 to 618 bp). For 57 individuals of Chalcidoidea, we sequenced both CO1 and ITS2 fragments (summarised in Supplementary Table 1). Maximum likelihood (RAxML) and BA analyses yielded almost similar topologies for both single locus and concatenated datasets (see figs 2 and S1). Molecular analyses provided significant insights into the identity of parasitoid species associated with Caloptilia fidella and Cosmopterix zieglerella and allowed corroboration with our morphological hypothesis or identified the specimens that could not be determined based on morphology. Specifically, specimens of Chalcidoidea with codes PZ3, RZ23, 37, MV13, Z9, CH3, V1, RB34, Z21, RB26, CH1, CH12, CH4, MV27, 9, 10 showing severe body damages after DNA extraction were assigned to particular clades based on genetic information.
Using morphology, we distinguished 17 species of parasitoids. Within Ichneumonoidea, we identified six species (two species of Ichneumonidae and four of Braconidae). Cosmopterix zieglerella hosted only one species of Braconidae (Microgaster novicia Marshall, 1855), while C. fidella hosted three species of Braconidae and two species of Ichneumonidae. Chalcidoidea were represented by 12 species (all Eulophidae) with 4 species of parasitoids of C. zieglerella and 8 species of C. fidella.
While the morphological analysis identified 17 parasitoid species either on(?) C. fidella or C. zieglerella, molecular species delimitation did not match completely our morphological identification and occasionally varied in relation to the method used in some parasitoid groups. For Ichneumonoidea, based on CO1, ASAP delimited only five putative species, respectively OTUs compared to bPTP or morphology (six OTU species). For Chalcidoidea, based on COI, ASAP identified a total of 10 OTUs and bPTP 12 OTUs, based on ITS2, ASAP delimited 12 OTUs and bPTP 13 OTUs. Results of delimitation methods have been mapped on the maximum likelihood tree calculated using a concatenated dataset in Chalcidoidea (fig. 2) and CO1 dataset in case of Ichneumonoidea (Figures S2–S4).
Figure 2. Phylogenetic tree of Chalcidoidea parasitoids based on RAxML analysis of concatenated COI and ITS2 sequences. Bootstrap values >70 are indicated at branch nodes. Branch colours correspond to different families, while the background shading of parasitoid species represents their respective hosts. Additionally, colours distinguish the host life stages parasitised, recorded parasitoid bionomy, and species delimitation based on genetic distances using the ASAP and bPTP methods.
Both delimitation methods pointed out one probably cryptic species within Chalcidoidea. Specifically, RB24 identified as Sympiesis sericeicornis (Nees, 1834) had no morphological differences from other specimens identified as S. sericeicornis. The genetic distance between these two clades (RB24 and the rest of S. sericeicornis) on CO1 was 8.7% and on ITS2 4.8% and corroborated the hypothesis of the cryptic species. However, we did not decide to describe this specimen (further referred as RB24_Sympiesis cf. sericeicornis as a new species as it requires more intensive research on more specimens, ideally from other host species/areas.
In a few cases, delimitation methods underestimated or overestimated the number of OTUs. ASAP method considered the two (PV28_Dolichogenidea sp. – Braconidae and 28A_Pnigalio sp.1 – Eulophidae), morphologically very well distinguishable species, as the same OTUs as their sister clades (i.e. 2_Cotesia ?ofella and Sympiesis dolichogaster Ashmead, 1888 – in case of CO1, and Pnigalio sp.2 – in case of ITS2). However, the genetic distance on CO1 has been discovered to be over 10% (for PV28_Dolichogenidea sp.) and 8.7% or 4.2% respectively on ITS2 (for 28A_Pnigalio sp.1). Therefore, in this case, we decided to follow the bPTP analysis and recognise into two species. On the other hand, the two specimens of Chrysocharis purpureus Bukovskii, 1938 (CH3 and CH5), although morphologically uniform, were delimited as two separate species with both methods using CO1 with quite large genetic distance (8.7%). However, analysis of the ITS2 dataset and concatenated dataset did not show such divergence and also both delimitation methods presented the specimens as one species. The intraspecific distance of other species of parasitoids did not exceed over 2% neither for COI nor ITS2.
Parasitoids composition
A total of 162 hymenopteran parasitoids were successfully reared from 924 collected leaves containing hosts (C. fidella: 142 individuals of 14 parasitoid species from 774 leaves, overall average parasitisation of 18.34%; C. zieglerella: 20 individuals of 5 parasitoid species from 150 leaves, overall average parasitisation of 13.33%) on 18 collection sites between 2020 and 2022 (Supplementary Table 1). Notably, Elachertus fenestratus (Nees, 1834) was the only species associated with both C. fidella and C. zieglerella. Out of a total of 142 parasitoid individuals associated with C. fidella, 122 emerged from leaf rolls, 12 from mines, and 8 from the pupae of the primary host or primary parasitoid. Of those emerging from leaf rolls, three OTUs were identified as koinobionts, six as idiobionts, and the bionomy of five OTUs was not further detailed. For C. zieglerella, all 20 parasitoids emerged from the host’s mine. Among these OTUs, one was identified as solitary idiobiont ectoparasitoid, three as possible idiobiont ectoparasitoids, and one as solitary koinobiont endoparasitoids (fig. 3).
Figure 3. Overview and associations of observed parasitoids in this study with the studied hosts. The thickness of the lines between species indicates the frequency of parasitoid–host associations, i.e., thicker lines represent more frequent associations. Coloured symbols indicate the developmental stages of the host from which the parasitoids emerged.
Parasitoids associated with Caloptilia fidella
Braconidae were represented by two parasitoid species. Pholetesor circumscriptus (Nees, 1834) emerged as the predominant primary parasitoid, exhibiting koinobiont endoparasitism of leaf-mining larvae and early exophagous stages within a leaf roll (fig. 4A-B). A total of 39 individuals were reared, with a sex ratio favouring females (27 females to 12 males), highlighting the central role of this species in the community. Another braconid parasitised leaf roll, tentatively identified as Cotesia ?ofella exhibiting koinobiont endoparasitism, was represented by a single female.
Figure 4. Parasitoids of the subfamily Microgastrinae reared from hosts and identified to species. (A) Cocoon with polar threads and an opened top, typical for the parasitoid Pholetesor circumscriptus (visible in the bottom left corner is an emergence hole chewed by the parasitoid in the leaf roll of the host Caloptilia fidella). (B) Lateral habitus of P. circumscriptus (female) reared from C. fidella. (C) Characteristic silk cocoon of the parasitoid Microgaster novicia located inside the mine of the host Cosmopterix zieglerella. (D) Lateral habitus of M. novicia (male) reared from the host C. zieglerellA. (Photo credits: T. Hovorka).
Within the Ichneumonidae, Gelis agilis (Fabricius, 1775) and Acrolyta rufocincta (Gravenhorst, 1829) were identified as hyperparasitoids (fig. 5A-B). These species specifically targeted the cocoons of P. circumscriptus, contributing to the intricate parasitisation dynamics. Notably, G. agilis exhibiting solitary idiobiont ectoparasitoidism, emerged from P. circumscriptus cocoons, further emphasising its role in secondary parasitism.
Eulophidae were represented by several species performing with multifaceted ecological roles. Sympiesis dolichogaster (fig. 5C-D) and Sympiesis sericeicornis (fig. 5F) acted mostly as primary idiobiont ectoparasitoids of leafrolls. However, S. dolichogaster also occasionally develops as a hyperparasitoid, attacking prepupae of P. circumscriptus. Solitary idiobiont ectoparasitoid Sympiesis acalle (Walker, 1848) (fig. 5E) and solitary endoparasitoid Chrysocharis purpureus parasitised larvae within mines and leaf rolls (fig. 5G-I), with S. acalle ranking as the second most abundant parasitoid after P. circumscriptus. Another ectoparasitoid, Elachertus fenestratus, primary solitary idiobiont ectoparasitoid, was recorded parasitising late-instar larvae within leaf rolls.
Figure 5. Parasitoids emerged from Caloptilia fidella (CF), Cosmopterix zieglerella (CZI), and Pholetesor circumscriptus (PC). (A) Acrolyta rufocincta (PC), (B) Gelis agilis (PC), (C, D) Sympiesis dolichogaster (PC), (E) S. acalle (CF), (F) S. sericeicornis (CF), (G–I) Pupa, larva, and adult of Chrysocharis purpureus (CF), (J) Egg near paralised caterpillar of CF, (K) Pnigalio sp. (CF, CZI), (L) Larva of Elachertus fenestratus in mine of CZI, (M) E. fenestratus emerged from mine of CZI. (Photo credits: T. Hovorka and K. Holý).
In addition to the above-mentioned species of parasitoids from the family Eulophidae, additional individuals were reared from leaf rolls containing larvae: Baryscapus cf. endemus (female, subfamily Tetrastichinae), Pnigalio sp.3 (female), and Sympiesis sp.1 and Sympiesis cf. sericeicornis (females, subfamily Eulophinae). One individual of Pnigalio sp.3 (male) was reared from a mine (fig. 5K). Precise species-level identification was not possible due to damage to the specimens during either DNA extraction or preparation and CO1 sequences of such OTUs are not presented in GenBank. In the case of the female Pnigalio sp.3, a freely laid egg was observed next to a paralysed larva in leaf roll (fig. 5J). The parasitoid larva fed as a solitary idiobiont ectoparasitoid on the host. The genus Pnigalio has not been previously recorded as a parasitoid of C. fidella larvae, making this a novel host record. The female Baryscapus cf. endemus was reared from a leaf roll without detailed observations of its biology. This represents the first known association of the genus Baryscapus with the host genus Caloptilia. Individuals identified as Sympiesis sp.1 and Sympiesis cf. sericeicornis were reared as primary parasitoids of C. fidella larvae within leaf rolls. Genetically, they differ from previously recorded species of the genus Sympiesis: Sympiesis sp.1 clusters closely with parasitoids of the genus Pnigalio, while Sympiesis cf. sericeicornis is genetically close to species of S. dolichogaster.
Parasitisation of Caloptilia fidella
A total of 774 leaves containing hosts (mines and rolls) were collected, with 534 leaves from the first host generation and 240 from the second host generation. Even if parasitisation rates varied significantly across years (Table 1) and developmental stages, no significant difference in parasitisation rates was observed between the first and second generations (Wilcoxon test, p-value = 0.4801). However, different larval stages exhibited significant variation in parasitoid emergence (Kruskal–Wallis test, p-value: 0.0008573). Out of a total of 142 parasitoids reared from C. fidella, 122 individuals emerged from leaf rolls (relative frequency 86%), 12 from mining larvae (8%), and 8 from pupae (6%). Larvae within leaf rolls were significantly more parasitised compared to pupae and mining larvae (post hoc Dunn test, p-values: 0.007864636 and 0.009008146, respectively). These values reflect only the parasitised individuals and represent relative emergence frequencies rather than actual parasitism rates per stage (see Table 1). Pupation stages exhibited the lowest relative emergence (6%), primarily involving morphologically unidentified Eulophidae (in 74%; four of them were damaged during DNA isolation, and no sequence of those was obtained), Sympiesis dolichogaster (13%), and S. sericeicornis (13%). Mining larvae had the second lowest relative emergence (8%), primarily caused by C. purpureus (in 33%), Sympiesis species (in 22%), Pnigalio sp.3 (in 11%) and three morphologically unidentified OTUs (in 33%; destroyed during DNA isolation). The first and second exophagous larval stages were the most susceptible, with emergence rates reaching 86%, largely attributable to P. circumscriptus (in 32%) and Sympiesis acalle (in 31%).
Parasitoids associated with Cosmopterix zieglerella
Braconidae were represented by Microgaster novicia, which was observed as a solitary, koinobiont endoparasitoid of the mining larvae of C. zieglerella (fig. 4C-D). Only a single female of this species was reared from the samples. The larva of M. novicia emerged from a final (purple) instar larva of C. zieglerella, which exhibited no outward signs of parasitism until shortly before pupation. The parasitoid cocoon, formed within the host’s mine, was distinctly different from that of P. circumscriptus in its absence of polar filaments (fig. 4).
Within Eulophidae, Elachertus fenestratus emerged as the most abundant parasitoid of C. zieglerella (fig. 5L-M). This primary, solitary idiobiont ectoparasitoid was observed in 11 individuals (4 females and 7 males), accounting for a significant proportion of parasitation events (52%). Detailed observations revealed that females of E. fenestratus initially paralysed their host larvae and subsequently consumed them within the mines. In several cases, larval parasitoids were seen actively moving within the mine to avoid light exposure. Other eulophid parasitoids included individuals of Pnigalio sp.1 and Pnigalio sp.2 and an unidentifiable individual of the subfamily Tetrastichinae gen. sp.1. All specimens of this species were poorly preserved, limiting precise identification. In at least one case, Pnigalio sp. was observed laying eggs near a paralysed host larva, suggesting a potential role as a hyperparasitoid. The individual was damaged during DNA isolation, gene amplification failed for both fragments, and precise morphological identification was not possible. Therefore, this individual, along with others of the genus Pnigalio, is not assigned to a particular species.
Parasitisation of Cosmopterix zieglerella
A total of 150 leaves containing hosts were collected. Of this total number of leaves, 20 were parasitised. This host has only one generation per year and does not form leaf rolls during its development. Therefore, all developments took place in the mine. For this reason, tests for differential parasitisation between generations and life stages were not performed for this host species. The mean annual parasitisation rate across the 3 years was 10.6% (Table 1). A total of 20 parasitoids were reared across the 3 years, with solitary emergence observed in all cases. Mining larvae were predominantly parasitised by E. fenestratus (in 52%), which displayed an ectoparasitic strategy. No parasitoid has been reared from host pupae.
Hyperparasitism and novel findings
For C. fidella, hyperparasitism was a significant factor shaping the parasitoid community. P. circumscriptus cocoons were frequently attacked by chalcidoid and ichneumonid hyperparasitoids, with S. dolichogaster exhibiting dual roles as a primary parasitoid of C. fidella and hyperparasitoid of P. circumscriptus. For C. zieglerella, no direct hyperparasitism was observed. However, the role of Pnigalio sp. as potential hyperparasitoids warrants further investigation.
This study uncovered several novel host–parasitoid associations and newly identified S. acalle, S. sericeicornis, and E. fenestratus as parasitoids with an idiobiont life strategy. For C. fidella, these included the first records as a host for P. circumscriptus, S. dolichogaster, and S. acalle. All parasitoids of C. zieglerella were recorded for the first time and therefore represent completely novel findings.
Discussion
In previous studies, the families Eulophidae (Chalcidoidea) and Braconidae (Ichneumonoidea) have been consistently reported as one of the most diverse parasitoid wasp groups associated with leaf mining insects (Salvo et al., Reference Salvo, Valladares and Cagnolo2011). Our study confirms this. Morphological identification of these dark taxa is challenging, and many lineages might represent cryptic species (Hausmann et al., Reference Hausmann, Krogmann, Peters, Rduch and Schmidt2020; Schmidt et al., Reference Schmidt, Schmid-Egger, Morinière, Haszprunar and Hebert2015) that are not possible to distinguish based on morphology only. However, to reveal the complexity of trophic interaction between parasitoids and their hosts requires the most precise species identification for further understanding of life histories and evolutionary consequences between parasitoids and their hosts or host plants. Therefore, besides morphology, we also applied barcoding and species delimitation. Such an integrative approach has been widely used to successfully decipher the identity of various groups of insects without excluding the parasitoids (Budrys et al., Reference Budrys, Orlovskytė, Lazauskaitė and Budrienė2023; Sheikh et al., Reference Sheikh, Ward, Zhang, Davis, Zhang, Egan and Forbes2022; Zhang et al., Reference Zhang, Kapli, Pavlidis and Stamatakis2013, Reference Zhang, Sheikh, Ward, Forbes, Prior, Stone, Gates, Egan, Zhang, Davis and Weinersmith2022).
The two molecular markers used in this study, ITS2 and CO1, provided complementary data for species delimitation. However, ITS2 was particularly effective in Chalcidoidea, offering greater amplification success than CO1. ITS2 is increasingly recognised as a valuable marker for distinguishing closely related parasitoid species, as demonstrated by studies on various parasitoid families, including Microgastrinae (Fagan-Jeffries et al., Reference Fagan-Jeffries, Cooper, Bertozzi, Bradford and Austin2018) and Eulophidae (Perry and Heraty, Reference Perry and Heraty2019). The results of this study further confirm the utility of ITS2 for resolving taxonomic challenges, particularly within morphologically similar taxa.
In our study, both methods (ASAP and bPTP) aligned relatively well with morphological identity of parasitoids with a few exceptions. While in some cases, ASAP underestimated, the number of OTUs identified using morphology (i.e. 28A_Pnigalio sp.1 – on CO1 and ITS2, and 20_Pnigalio sp.3 – on CO1 only), bPTP proved to be more sensitive. It identified a higher number of OTUs than we were able to distinguish based on morphological differences (i.e. RB24_Sympiesis cf. sericeicornis – on CO1 and ITS2, and 37 + Z27_Sympiesis dolichogaster – on ITS2). Underestimation using ASAP might be caused by lack of sufficient number of samples in particular clades (Ahrens et al., Reference Ahrens, Fujisawa, Krammer, Eberle, Fabrizi and Vogler2016; Eberle et al., Reference Eberle, Ahrens, Mayer, Niehuis and Misof2020; Vogel et al., Reference Vogel, Sauren and Peters2024). On the other hand, overestimating of OTUs compared to morphology using bPTP has been corroborated by genetic distances on particular gene fragments (RB24_Sympiesis cf. sericeicornis to the rest of S. sericeicornis – 8.8% on CO1 and 4,8% on ITS2, 37 + Z27_Sympiesis dolichogaster to the rest of S. dolichogaster – 11.5% on ITS2) as the use intraspecific genetic distance on CO1 is set to 2% (Hebert et al., Reference Hebert, Ratnasingham and De Waard2003b). This might point out to cryptic species within particular clades that we, using current taxon sampling and morphology, are not able to resolve.
The performance of bPTP in this study contrasts with findings from other research, such as Sheikh et al. (Reference Sheikh, Ward, Zhang, Davis, Zhang, Egan and Forbes2022) and Zhang et al. (Reference Zhang, Sheikh, Ward, Forbes, Prior, Stone, Gates, Egan, Zhang, Davis and Weinersmith2022), which also applied bPTP and ASAP to identify cryptic species in parasitoid complexes. Both studies reported that bPTP tended to overestimate species richness, likely due to its higher sensitivity to subtle genetic variation. For example, Zhang et al. (Reference Zhang, Sheikh, Ward, Forbes, Prior, Stone, Gates, Egan, Zhang, Davis and Weinersmith2022), in their investigation of cryptic species within the genus Sycophila Walker, 1871 (Hymenoptera: Chalcidoidea), and Sheikh et al. (Reference Sheikh, Ward, Zhang, Davis, Zhang, Egan and Forbes2022), in their study of the Ormyrus labotus Walker, 1843 complex (Hymenoptera: Chalcidoidea), observed discrepancies between the methods, particularly in populations with minor genetic differences or distinct geographic distributions. Both studies relied on the more conservative results from ASAP, suggesting that bPTP might sometimes reflect population-level genetic structuring rather than species-level divergence. The more accurate performance of bPTP in this study can be attributed to the larger genetic distances observed among the parasitoid species associated with C. fidella and C. zieglerella. These larger intraspecific distances likely reduced the chances of overestimation by bPTP. Moreover, the calculated intraspecific and interspecific genetic distances supported species delimitation via bPTP, as evidenced by phylogenetic analyses.
Despite these successes, some limitations were noted. The CO1 dataset was incomplete, lacking sequences for taxa such as Pnigalio sp.2 and Pnigalio sp.3. This discrepancy limited direct comparisons between ITS2 and CO1. For future studies, obtaining datasets with equal taxonomic representation across both genes would allow for a more robust evaluation of the performance of these molecular markers.
The results of this study highlight the importance of integrating molecular methods with a traditional morphological approach. While molecular techniques are powerful, they can sometimes yield overestimated or ambiguous results if used in isolation. Sharkey et al. (Reference Sharkey, Janzen, Hallwachs, Chapman, Smith, Dapkey, Brown, Ratnasingham, Naik, Manjunath and Perez2021) reported cases where molecular methods led to misclassification due to insufficient integration with morphological and ecological evidence. Therefore, an integrative approach combining molecular, morphological, and ecological data is essential for accurate species delimitation (Ahrens et al., (Reference Ahrens, Ahyong, Ballerio, Barclay, Eberle, Espeland, Huber, Mengual, Pacheco, Peters and Rulik2021); Wang et al., Reference Wang, Zhou, Zou, Liu and Peng2024). In our study, molecular methods were used as a complementary tool to confirm morphological identification. Additionally, they proved crucial in cases where certain individuals exhibited slightly different morphological traits, helping us determine that they most likely represent cryptic species that cannot be distinguished morphologically. This was the case, for instance, with Ch. purpureus or an individual showing minor differences, designated as RB24_Sympiesis cf. sericeicornis. Beyond their role in species identification, molecular methods such as ASAP and bPTP have broader applications in ecological and evolutionary research, aiding in biodiversity assessment and species delimitation (Cruaud et al., Reference Cruaud, Underhill, Huguin, Genson, Jabbour-Zahab, Tolley, Rasplus and van Noort2013; Sheikh et al., Reference Sheikh, Ward, Zhang, Davis, Zhang, Egan and Forbes2022; Zhang et al., Reference Zhang, Kapli, Pavlidis and Stamatakis2013). As sequencing technologies continue to advance, molecular approaches will become even more instrumental in uncovering cryptic diversity and elucidating parasitoid evolutionary relationships (Kenyon et al., Reference Kenyon, Buerki, Hansson, Alvarez and Benrey2015).
Parasitoid community of Cosmopterix zieglerella and Caloptilia fidella
Even if larvae of both hosts are monophagous on the same host plant (Humulus lupulus) (Koster and Sinev, Reference Koster and Sinev2003; Laštůvka et al., Reference Laštůvka, Laštůvka, Liška and Šumpich2018; Watson et al., Reference Watson, Eaton and Mcclennon-Warnock2021), Elachertus fenestratus is the only parasitoid species shared between the two host species. Except from families Tortricidae, Gelechiidae, and Coleophoridae, E. fenestratus has been recently referred to as an ectoparasitoid of leaf mining larvae of Phyllonorycter issikii (Kumata) (Lepidoptera: Gracillariidae) from Tilia platyphyllos (Malvales: Malvaceae) (Kosheleva et al., Reference Kosheleva, Belokobylskij and Kirichenko2022). In Japan, Sugiura (Reference Sugiura2011) recorded additional Elachertus species (E. inunctus (Nees, 1834) and another unidentified species) for Caloptilia azaleella (Lepidoptera: Gracillariidae). Similarly, Barry et al. (Reference Barry, Rodriguez-Saona, Polk and Zhang2010) in the USA reported another species of Elachertus as a parasitoid of C. porphyretica. These findings suggest that despite geographical differences, parasitoids of the genus Elachertus are common leaf-miner associates. In general, the genus Elachertus is known as primary, often gregarious, parasitoid of various Lepidoptera larvae (Schauff, Reference Schauff1985), but our observation confirmed E. fenestratus as strictly primary solitary idiobiont ectoparasitoid on both studied hosts either in mines (Cosmopterix zieglerella) or leaf-rolls (Caloptilia fidella). Until now, no parasitoids had been recorded for Cosmopterix zieglerella. This study reveals the first parasitoid associations for this species, identifying five distinct parasitoid–host interactions. Besides Elachertus fenestratus, the other most common parasitoid species included two species of Pnigalio (Pnigalio sp.1 and Pnigalio sp.2 both solitary ectoparasitoids), and one representative of Braconidae, Microgaster novicia, solitary koinobiont endoparasitoid (Fernandez-Triana et al., Reference Fernandez-Triana, Shaw, Boudreault, Beaudin and Broad2020; Shaw, Reference Shaw2023). Other observations of the parasitoid community of Cosmopteryx were made by El-Serwy (Reference El-Serwy2006, Reference El-Serwy2009) in his studies on C. pararufella Riedl, 1976 and C. salahinella Chrétien, 1907, (both Lepidoptera: Cosmopterigidae) where Pnigalio spp. were among the dominant parasitoids. Additionally, Cotesia ruficrus (Haliday, 1834) (Hymenoptera: Braconidae) was frequently recorded in these studies. Ahmad et al. (Reference Ahmad, Ghramh, Khan, Khan, Khan, Elgezouly and Pandey2022) described a novel parasitoid species, Bracon cosmopteryx Ahmad and Pandey, 2022 (Hymenoptera: Braconidae) from C. phaeogastra (Meyrick, 1917) (Lepidoptera: Cosmopterigidae). These findings indicate that parasitoid complexes of Cosmopterix species remain largely unexplored, with only sparse information available for European species of this genus.
The most frequently observed parasitoids of Caloptilia fidella were Pholetesor circumscriptus (Braconidae) and three species of the genus Sympiesis (S. acalle, S. dolichogaster and S. sericeicornis) (Eulophidae). While S. acalle and S. sericeicornis are widely distributed across the whole Holarctis, S. dolichogaster has been mentioned apart from Nearctic and western Palearctic also in Oriental and Australian realm.
Our study confirms P. circumscriptus as a common parasitoid of Caloptilia fidella, aligning with previous findings that species of Pholetesor frequently parasitise Gracillariidae (Liu et al., Reference Liu, He and Chen2016; Mason, Reference Mason1981; Papp, Reference Papp1983; Whitfield, Reference Whitfield2006). In European habitats, Pholetesor spp., particularly those within the circumscriptus species group, are among the most frequent parasitoids of gracillariid leafminers (Ahmad et al., Reference Ahmad, Ghramh and Pandey2020; Whitfield, Reference Whitfield2006), contributing significantly to the natural regulation of these herbivores. These wasps exhibit a koinobiont endoparasitoid strategy, in which the female oviposits into an active leaf-mining larva that continues to develop until the parasitoid reaches its final instar and ultimately kills the host at or near pupation. This developmental strategy is typical of Microgastrinae and allows Pholetesor wasps to exploit their hosts efficiently before emergence (Fernandez-Triana et al., Reference Fernandez-Triana, Shaw, Boudreault, Beaudin and Broad2020). The larvae of Pholetesor construct suspended cocoons within leaf mines or leaf rolls, a behaviour that has been observed in P. circumscriptus and is thought to provide protection against predation or hyperparasitation (Whitfield, Reference Whitfield2006; Whitfield and Wagner, Reference Whitfield and Wagner1991). Some Pholetesor species, such as P. pedias, have also been explored for biological control against agricultural pests (Whitfield, Reference Whitfield2006). Recent molecular studies have revealed substantial cryptic diversity within Pholetesor, with several genetically distinct lineages hidden within morphologically similar species (Fernandez-Triana et al., Reference Fernandez-Triana, Shaw, Boudreault, Beaudin and Broad2020). However, in our study, both molecular delimitation methods confirmed that all analysed specimens belong to P. circumscriptus, with no evidence of cryptic diversity. This suggests that P. circumscriptus represents a genetically homogeneous species, at least within the studied population.While Sympiesis dolichogaster and S. sericeicornis were found in our study predominantly as primary ectoparasitoids mostly of leafrolls, Sympiesis acalle acted as solitary ectoparasitoid of larvae within either mines or leaf rolls. All three species have been observed here as idiobionts; however, the only so far published mention about idiobiont lifestyle came from study of S. sericeicornis on Phyllonorycter comparella (Lepidoptera: Gracillaridae) in Hungary (Szőcs et al., Reference Szőcs, Melika, Thuróczy and Csóka2015). S. scalle has been referred to as a primary ectoparasitoid of larvae or pupae of insects that live in sheltered situations, such as miners and rollers (Maier and Hansson, Reference Maier and Hansson2006). The species develop rather as generalist and is known from several species of Caloptilia and other genera of Gracillariidae, and additionally from many other lepidoptera families such as Gelenichidae, Tischeridae, and Tortricidae (Bouček and Askew, Reference Bouček, Askew, Delucchi and Remaudière1968; Maier and Hansson, Reference Maier and Hansson2006). The two other species have been reported from more distantly related lepidopteran hosts and are considered as generalists, too. The host spectrum of S. dolichogaster either includes concealed larvae of Gelenichidae, Gracillariidae, Tischeridae or exposed larvae of Tortricidae (Bouček and Askew, Reference Bouček, Askew, Delucchi and Remaudière1968, UCD Community 2023). S. sericeicornis is commonly known as primary parasitoid of plethora species of genera Gracillariidae (e.g. Calloptilia, Lithocolletis, Phyllonorycter) and Tischeridae (Bouček and Askew, Reference Bouček, Askew, Delucchi and Remaudière1968; Hansson, Reference Hansson1987).
Besides exhibiting a life style as primary parasitoids, these three generalistic eulophids species have also been observed as facultative hyperparasitoids of various braconids including Pholetesor circumscriptus (Maier and Hansson, Reference Maier and Hansson2006; Hagley, Reference Hagley1985; Maier, Reference Maier1988; Bouček and Askew, Reference Bouček, Askew, Delucchi and Remaudière1968), which corresponds to the results of our study. Sympiesis acalle and S. sericeicornis have also been recorded as hyperparasitoids of a few Eulophidae such as Achrysocharoides or Pnigalio (Askew and Shaw, Reference Askew and Shaw1979; Bouček and Askew, Reference Bouček, Askew, Delucchi and Remaudière1968)
Our findings align with the study by Sugiura (Reference Sugiura2011), who examined parasitoid assemblages of two other Caloptilia species (C. azaleella Brants, 1913 and C. leucothoes Kumata, 1982) that develop on various Rhododendron L. species in Japan. In C. azaleella, which has two generations per year, dominant parasitoids included Apanteles cf. xanthostigma, Pholetesor laetus (Marshall, 1885) (both Hymenoptera: Braconidae), Acrolyta sp. (Ichneumonidae), and Sympiesis dolichogaster (Eulophidae). For C. leucothoes, Sugiura (Reference Sugiura2011) reported the predominance of Achrysocharoides sp. (Hymenoptera: Eulophidae). Additionally, S. dolichogaster and other Sympiesis species were among the most frequently recorded parasitoids of C. azaleella in Mizell and Schiffhauer’s (Reference Mizell and Schiffhauer1991) study conducted in the United States.
The number of parasitoid species recorded for C. azaleella (18 species) in Sugiura (Reference Sugiura2011) is comparable to the 15 species identified for C. fidella in our study, highlighting a similar degree of parasitoid diversity. However, C. leucothoes, which has only one generation per year, had a lower number of recorded parasitoid species (7). A comparable parasitoid composition was observed in Barry et al. (Reference Barry, Rodriguez-Saona, Polk and Zhang2010), who investigated C. porphyretica (Braun, 1923) in the United States. That study identified Pholetesor sp. (Braconidae) and Eulophidae as the most abundant parasitoids. Moreover, Wist and Evenden (Reference Wist and Evenden2013), in their study on C. fraxinella (Ely, 1915) in Canada, noted that the dominant parasitoids were Apanteles polychrosidis Viereck, 1912 (Hymenoptera: Braconidae) and Sympiesis sericeicornis, which was also identified in this study alongside with two unidentified Sympiesis species.
In contrast to the dominant role of Braconidae found in many studies, Wheeler et al. (Reference Wheeler, Dyer, Hight and Wright2017) reported that the parasitoid fauna of C. triadicae Davis, 2013 in the United States was dominated by Goniozus sp. (Hymenoptera: Bethylidae) and Brasema sp. (Hymenoptera: Eupelmidae). However, members of Sympiesis (Eulophidae) were also among the main parasitoids in that study. The differences in parasitoid composition for C. triadicae, an introduced species in the United States, were attributed to its non-native origin and the initial reliance on generalist parasitoids in the new environment. Over time, the parasitoid complex may shift as local parasitoids adapt to this new host, a phenomenon also observed by Grabenweger et al. (Reference Grabenweger, Kehrli, Zweimüller, Augustin, Avtzis, Bacher, Freise, Girardoz, Guichard, Heitland and Lethmayer2010).
Parasitism and host strategies
Leaf mining lifestyles occur in the orders Lepidoptera, Diptera, Coleoptera, and Hymenoptera, and is diverse and abundant in Lepidoptera (Kirichenko et al., Reference Kirichenko, Skvortsova, Petko, Ponomarenko and Lopez-Vaamonde2018; Rott and Godfray, Reference Rott and Godfray2000). Our results indicate that mining is a more effective strategy for reducing parasitism compared to leaf rolling, as C. zieglerella, which exclusively develops as a leaf-miner, and the mining stages of C. fidella exhibited lower parasitism rates than its leaf-rolling stages. Previous studies suggest that mining complexity deters parasitoids, making it harder for them to locate their hosts (Ayabe et al., Reference Ayabe, Tuda and Mochizuki2008). For example, Phyllonorycter malella creates intricate mining networks that reduce parasitism risk by acting as barriers (Djemai et al., Reference Djemai, Meyhöfer and Casas2000). While our study species do not exhibit such extreme modifications, leaf mines (concealed stage) generally provide confined spaces where larvae can actively dislodge parasitoids. However, the relationship between host concealment and parasitism risk remains complex. Kobayashi et al. (Reference Kobayashi, Matsuo, Watanabe, Nagata, Suzuki-Ohno, Kawata and Kato2015) found that leaf-rolling weevils that seal or reinforce their leaf rolls (closed and wrapped rollers = concealed) experience lower parasitism rates than those that leave openings (open rollers = semi-concealed). Similarly, stem-boring (concealed) species exhibited lower parasitism rates than leaf-mining ones, suggesting that physical barriers can enhance protection. Similarly, Hrček et al. (Reference Hrček, Miller, Whitfield, Shima and Novotny2013) found that semi-concealed hosts were more frequently parasitised than exposed hosts, likely due to visual cues associated with leaf damage that parasitoids use to locate their hosts, the active defence of exposed hosts, or competition between parasitoid guilds. In our study, we classify leaf-mining stages as concealed hosts and leaf-rolling stages as semi-concealed hosts.
A key distinction in parasitoid communities is the balance between koinobionts and idiobionts. Koinobionts allow their hosts to continue development after parasitism, whereas idiobionts immediately immobilise and consume their hosts. Hrček et al. (Reference Hrček, Miller, Whitfield, Shima and Novotny2013) found that semi-concealed hosts were primarily parasitised by koinobionts, with Braconidae dominating, and exposed hosts were mostly attacked by koinobiont Tachinidae (Diptera). These two host feeding-mode groups also did not differ in parasitoid community size. In contrast, leaf-mining species are known to be frequently parasitised by Chalcidoidea and Ichneumonidae, with a higher prevalence of idiobionts (Askew and Shaw, Reference Askew and Shaw1979; Hawkins, Reference Hawkins1994). Our study contrasts with Hrček et al. (Reference Hrček, Miller, Whitfield, Shima and Novotny2013), as we recorded a higher proportion of idiobionts in both concealed leaf-mining and semi-concealed leaf-rolling stages, particularly in C. fidella. This might suggest that the koinobiont–idiobiont ratio varies significantly depending on host species, habitat, and parasitoid community composition.
In this context, the semi-concealed stage of C. fidella provides some protection but remains more vulnerable to parasitoids than the fully concealed leaf-mining stage. Larvae in rolls either bend the tip of a leaf or fold its margin, creating a pocket-like shelter that never fully seals, making it more susceptible to parasitoids than species that construct tightly sealed shelters (Kobayashi et al., Reference Kobayashi, Matsuo, Watanabe, Nagata, Suzuki-Ohno, Kawata and Kato2015). This may explain why its leaf-rolling stages had higher parasitism rates than its mining stages, despite the additional structural protection.
Overall, our findings support the hypothesis that semi-concealed hosts experience the highest level of parasitism risk between fully exposed and concealed species as in Hrček et al. (Reference Hrček, Miller, Whitfield, Shima and Novotny2013). However, such results might be influenced by host concealment type, parasitoid searching strategies, and the koinobiont–idiobiont balance. Early gracillariids likely fed on Fabales tree leaves, forming simple blotch mines, while later adaptations, such as keeled or tentiform blotch mines, leaf rolling, and gall formation, emerged as defensive strategies (Aoyama and Ohshima, Reference Aoyama and Ohshima2019; Li et al., Reference Li, St Laurent, Earl, Doorenweerd, Nieukerken, Davis, Johns, Kawakita, Kobayashi, Zwick and Lopez‐Vaamonde2022). However, in contrast to various mining strategies, leaf rolling evolved as the most advanced larval behaviour mode only in a few clades of Gracillariidae (incl. Gracillariinae, subf. of Calloptilia) (Kawahara et al., Reference Kawahara, Plotkin, Ohshima, Lopez-Vaamonde, Houlihan, Breinholt, Kawakita, Xiao, Regier, Davis and Kumata2017; Li et al., Reference Li, St Laurent, Earl, Doorenweerd, Nieukerken, Davis, Johns, Kawakita, Kobayashi, Zwick and Lopez‐Vaamonde2022). Although Li et al. (Reference Li, St Laurent, Earl, Doorenweerd, Nieukerken, Davis, Johns, Kawakita, Kobayashi, Zwick and Lopez‐Vaamonde2022) suggested advanced larval behaviours, such as making keeled or tentiform blotch mines, rolling leaves and making galls, may have accelerated the diversification of Gracillariidae by avoiding larval parasitoids, our results indicate leaf rolling might not be a sufficient strategy to avoid parasitism at least in C. fidella. Further studies with expanding taxonomic sampling across gracillariid hosts could provide further insights into how parasitoid–host interactions shape the evolution of leaf-mining and leaf-rolling behaviours in Gracillariidae.
Conclusion
This study highlights how the level of host concealment may influence parasitoid communities and parasitism rates in two microlepidopteran species on hops. Our findings suggest that C. fidella, which transitions from mining to leaf rolling, tends to experience higher parasitism in its semi-concealed stage than in its mining stage. In contrast, C. zieglerella, which remains a leaf miner throughout its development, exhibited lower parasitism rates. This pattern is consistent with the hypothesis that leaf mining provides greater protection from parasitoids than leaf rolling, at least within the studied system. However, given the limited number of host species examined and the confounding between concealment type and host identity, our ability to generalise these patterns is constrained.
The differences in parasitoid assemblages between the two hosts, particularly the higher representation of idiobiont parasitoids in C. fidella, suggest potential functional differences in host susceptibility linked to concealment strategy. Yet, as concealment level is fully confounded with host species in this study, we caution against drawing firm conclusions about causal relationships.
Beyond ecological patterns, this study underscores the utility of molecular species delimitation in resolving taxonomic challenges in parasitoid communities. The combination of ITS2 and CO1 markers, applied using the bPTP and ASAP methods, proved valuable in identifying cryptic diversity and refining parasitoid taxonomy. These findings support the continued integration of molecular and morphological approaches in host–parasitoid research.
Overall, while this study provides valuable case-specific observations, its conclusions regarding concealment and parasitism should be viewed as preliminary. Broader comparative studies across multiple species within each concealment category are needed to more rigorously test the role of host concealment in shaping parasitoid communities. Nonetheless, our results offer a foundation for future research and highlight several candidate taxa and life-history traits for more extensive ecological and evolutionary investigations.
Supplementary material
The supplementary material for this article can be found at https://doi.org/10.1017/S000748532510031X
Acknowledgements
This project was supported by the Grant Agency of Charles University (GAUK) project no. 375421, Ministry of Culture of the Czech Republic IP DRKVO 2024–2028/5.I.b, 00023272, National Museum of the Czech Republic (TH) and grant PRIMUS/24/SCI/015 (PJ). We are grateful to our colleagues at the Department of Zoology, Faculty of Science, Charles University; the Department of Entomology, National Museum of the Czech Republic; the Department of Entomology, State Museum of Natural History Stuttgart; the Department of Integrated Crop Protection against Pests, Czech Agrifood Research Center (institutional support MZE-RO0423) and to Nela Gloríková, for their valuable collaboration and support. We also thank Rudolf Bryner for providing the photograph of an adult Cosmopterix zieglerella (fig. 1H).
Competing interests
The authors declare that they have no conflicts of interest related to this work. The research was conducted in accordance with institutional and national ethical guidelines.
Open access
