Hymenopteran parasitoids reared from European gall midges (Diptera, Cecidomyiidae)

Abstract We report the results of investigations 2010 through 2023 of hymenopteran parasitoids associated with gall midges in Europe. A total of 242 collections of gall midges were made, from each of which one to several parasitoid species emerged, resulting in ca. 200 recorded parasitoid species and 267 host-parasitoid interaction records. The parasitoid families involved were Eulophidae (63 species), Platygastridae (56 species), Torymidae (34 species), Pteromalidae (31 species), Ceraphronidae (5 species), Eupelmidae (4 species), Eurytomidae (2 species) and Encyrtidae (1 species). As many as 159 interactions are reported for the first time, significantly enlarging our knowledge of gall midge – parasitoid interactions on the species level. Even more interesting, 51 host records are for parasitoid species for which no host was previously known. Similarly, 28 species of gall midge are reported as host to named parasitoids for the first time. Additionally, 91 parasitoid records were the first for the country in question. Differences between the rearing methods applied and their suitability for recording species with contrasting life histories, are discussed.


Introduction
The gall midges (Diptera, Cecidomyiidae) are arguably the largest and most diverse insect family and, simultaneously, one of the most incompletely known in terms of both basic species cataloguing and knowledge of species' biology (e.g.Espírito-Santo and Fernandes 2007, Hebert et al. 2016, Chimeno et al. 2022, Srivathsan et al. 2023).The families of parasitoid Hymenoptera associated with gall midges remain similarly very poorly known.Srivathsan et al. (2023) identified 20 insect families that account for more than half of Malaise-trapped insect species richness at sites across the globe.Cecidomyiidae and two families of parasitoid Hymenoptera frequently associated with them, i.e.Platygastridae and Eulophidae, figured prominently in the list.The large proportion of undescribed species, "dark taxa", highlights the shortfall in current biodiversity knowledge.Another equally important knowledge gap is the ignorance of species' trophic interactions, something that has been coined the "Eltonian shortfall" (Hortal et al. 2015).
The Cecidomyiidae are small delicate flies (Tokuda and Yukawa 2021).The majority of species are phytophagous and display a wide range of host-plant relationships and lifehistory strategies (Dorchin et al. 2019).Many species induce galls on the host plants, but others inhabit leaf sheaths, grass spikelets or dense inflorescences without causing obvious malformation of the plant parts.Many species are univoltine, fewer are bivoltine or multivoltine.In most temperate species, larvae leave their host plant to hibernate and later pupate in the soil.Pupation in situ, for example, in galls, seems to have arisen several times within the clade.All of these life-history characteristics have been shown to influence the size and family composition of their associated parasitoid assemblages (Hawkins and Gagné 1989).
The study of gall midge parasitoids is almost as old as studies of the gall midges themselves.For example, J.-J.Kieffer (1857Kieffer ( -1925) ) described several hundred species of Cecidomyiidae and Platygastridae and some Chalcidoidea.H. F. Barnes (1902Barnes ( -1960) ) and his students made numerous biological studies on gall midges and their parasitoids, in particular, species associated with crops and ornamental plants.More recently, Skuhravá and Thuróczy (2007) reported 39 species of Chalcidoidea and Platygastroidea reared from galls of 50 species of gall midges, which had been collected between 1955 and 1996 in the Czech Republic and elsewhere in central and southern Europe.Similar studies have been published by Tudor and Neacşu (1983), Askew and Harris (2007), Buhl and Jørgensen (2010), Jennings (2021) and others.Nevertheless, the knowledge of gall midge -parasitoid relationships remains incomplete at best and is mainly based on rearing adult parasitoids directly from midge-induced galls.
The Chalcidoidea parasitoids of gall midges are mainly found in the families Eulophidae, Pteromalidae (incl.Pirenidae and Systasidae) and Torymidae, with fewer species of Braconidae, Encyrtidae and Eurytomidae being involved (Yukawa et al. 2021).Another major group of gall midge parasitoids is Platygastridae (Platygastroidea), while fewer Ceraphronidae (Ceraphronoidea) have been recorded.This suite of hymenopteran parasitoids exhibits a wide and complex array of life-history characteristics.One helpful simplification is the division of parasitoids into koinobionts and idiobionts (sensu Askew and Shaw (1986)).Koinobionts let their hosts continue normal development and behaviour, such as the midge larvae leaving the galls to continue life and to pupate elsewhere.Koinobionts are usually endoparasitoids and expected to be more narrowly hostspecialised, due to the prolonged physiological intimacy with a living and metabolising host (Hawkins 1994).Idiobionts, in contrast, immediately kill or permanently paralyse their hosts.They are typically ectoparasitoids and they sometimes show association with particular host plants, attacking phylogenetically diverse assemblages of phytophagous insects using that particular host plant (Hawkins 1994).
Here, we report primary data on the parasitoid Hymenoptera that, during the years 2010 -2023, we have reared from larvae of Cecidomyiidae in Europe using a suite of methods.The methodology is discussed in relation to the life-history characteristics of both gall midge hosts and of their koinobiont and idiobiont parasitoids.

Materials and methods
Gall midge larvae of a suite of species were collected from their natural feeding site, be it plant galls, plant inflorescences or rust (Pucciniales) sori on plants.The majority of collections were made in Denmark, several from Sweden and Poland and few from other European countries (Spain, Hungary, Romania and Lithuania).Some collections were specifically bred out to obtain adult parasitoid wasps.Other collections were made with the aim of rearing adult gall midges for taxonomic work and studying their life cycles, the parasitoids being obtained as a bycatch.Thus, the collection strategy may be considered opportunistic, dictated by the species of gall midges that we have happened to encounter.
Adults emerging directly from galls or other plants parts inhabited by gall midge larvae, sometimes from galls kept over winter; 3.
Adults emerging from mature larvae or pupae extracted from galls and transferred individually to gelatine capsules; 4.
Adults emerging from soil in pots, to which gall midge larvae had been transferred earlier the same year (4A) or in the preceding year (4B).Parasitoid emergence is often simultaneous with adult gall midge emergence, but is sometimes delayed (even up to a year);

5.
No rearing attempted, or rearing failed; parasitoid reported at the genus level if possible.
Method 4 was applied to gall midge species which hibernate and pupate in the soil.Midge larvae were either transferred to clean soil in pots using a small paintbrush or, in some cases, the inhabited plant parts were left on the soil surface in similar pots for a few days (then removed to avoid mould), allowing the larvae to move to the soil of their own accord.
Often, a combination of the two ways was used, first an appreciable number of larvae were transferred by hand to ensure a minimum number were secured, and then additional inhabited plant parts were added.Larvae of endoparasitic, koinobiont Hymenoptera would be transferred with the midge larvae, while ectoparasitic larvae destined to pupate in situ may potentially have been discarded with the plant parts.Pots with inhabited soil were kept overwinter outdoors in Kårup Skov, Denmark under ambient conditions of temperature and precipitation until early the following spring, then taken indoors to allow adult insects to emerge from pots in mesh bags or spacious perforated plastic bags.
The identification work was based on available keys, original diagnoses, revisions and comparision with paratypes and other specimens kept in the personal collections of RRA and PNB.With regard to the many Platygastridae originally described by Walker, we based our work on the revision and re-description of species by Vlug (1985).We report all parasitoid records that can be considered new to the given gall midge host, even in the many cases where the parasitoid identification is incomplete and requires more work, including rearing more material, undertaking taxonomic revisions and describing new species.That choice is motivated by our focus on the "Eltonian shortfall".

Results
A total of 242 collections was made, representing 109 Cecidomyiidae species, some of which belong to species that are yet to be formally described (designated an interim name to avoid ambiguity).
Clearly, these observations may represent nothing more than chance absence of parasitoids in single gall midge populations.

Discussion
Biotic interactions between species is an essential component of biodiversity (Gaüzère et al. 2022), but this element remains elusive as long as basic natural history knowledge on host-parasitoid relationships is incomplete or absent.Our results clearly demonstrate how incomplete our knowledge of gall midge -parasitoid interactions is.They also show that  Hymenopteran parasitoids reared from European gall midges (Diptera, Cecidomyiidae) parasitoid specimens may be obtained in reasonable quantities using simple methods.These could easily be applied in school teaching and citizen science.The identification work, in contrast, requires highly specialised knowledge and remains a severe bottleneck.
With few exceptions, our host gall midge collection had low levels of replication, with most species being represented by a single collection.Thus, we expect that more species of parasitoid per gall midge species -and, conversely, more hosts per hymenopteran parasitoid -would have resulted, had our sampling allowed collection of larger quantities of hosts from more geographically dispersed sites and -in particular for gall midge species with more than one generation per year -more dispersed over the year and also between years.Previous analyses have shown plant galling Diptera, including Cecidomyiidae, to be generally associated with a low number of parasitoids per host species (Hawkins 1994).However, detailed studies of parasitoid complexes associated with individual gall midge species or groups of species using a single host plant, have demonstrated that the parasitoid communities typically consist of a dozen or more species (e.g.Tscharntke et al. (1991), Roskam (2013)).Better coverage per host species would also establish primary host-parasitoid relationships with greater certainty.
Previous analyses of the parasitoid communities associated with gall midges have found idiobionts to dominate over koinobionts by a factor two to four (Hawkins 1994, table 4.1).
Our results seem to suggest a more even balance between koinobionts and idiobionts.For example, Platygastridae, which attack eggs and young larvae of gall midges (Askew 1975), are exclusively endoparasitic and koinobiont and stand out as one of the two most speciose parasitoid families, only outnumbered by Eulophidae.At the opposite end of the continuum is Torymidae subfamily Toryminae, in which the majority of species are idiobiont ectoparasitoids of insect inhabitants of plant galls, mainly attacking third instar larvae or pupae of gall midges.Pteromalidae, Eulophidae and Eurytomidae appear to take a position intermediate between these two extremes (Yukawa et al. 2021).Torymidae were mainly obtained by our rearing method 3, i.e. extraction of mature parasitoid larvae or pupae by dissecting galls and rearing adults in gelatine capsules.In contrast, Platygastridae were mainly obtained through our method 4, i.e. breeding adult parasitoids from soil, to which host larvae had been transferred earlier.These patterns were to be expected from the life history of the parasitoids and their gall midge hosts.For Eulophidae and Pteromalidae, the two methods were about equally productive.These results suggest that more complete knowledge of parasitoid faunas of Cecidomyiidae are best obtained by a combination of rearing methods.The advantage of the first mentioned method is that it is targeted and establishes host-parasitoid relationships with greater certainty, while its disadvantage is that endoparasitoids are overlooked.Our method 2, i.e. rearing parasitoids directly from galls, the method used by far the most frequently in the past, involves the risk of obtaining parasitoids associated not with gall midges, but with other insects inhabiting the galled plant parts.Additionally, this method targets gall midge species that pupate in the galls.The main advantage of method 4 is that it is suitable for obtaining the mainly koinobiont parasitoids of gall midges leaving the galls as larvae to pupate in the soil, which is the most common condition amongst temperate Cecidomyiidae.
If midge larvae are transferred to soil individually, rather than by placing inhabited plant parts on the soil surface for larvae to move voluntarily, this method is as accurate in establishing host-parasitoid relationships as is method 3. The fact that most of the first host records for parasitoid species and first parasitoid records for gall midge hosts were obtained using method 4, i.e. transfer of larvae to soil, is because this method has rarely been applied in the past.It suggests that this method is indispensable for obtaining primary life history data for koinobiont parasitoid species and for filling the knowledge gap of gall midge-parasitoid interaction networks.

name Gall midge host name Method Collection date Emg. month Coll. C. Int. nov. Biogeo. nov.
Table, the degree of certainty in identification to species level is indicated as follows: 1. Identified to the genus level (in one case family level), but not identified further, for example, "Aprostocetus sp."; 2. Positively placed in a named species group, for example, "Acerotella sp.(evanescens group)"; 3. Morphologically close to a named taxon, but probably a separate species, for example, "Synopeas sp.nr inerme"; 4. Probably the mentioned species, but confirmatory work, such as comparison to type material, is required, "Aprostocetus cf.suevius"; 5. Identified to species beyond reasonable doubt.Only identifications at level 3 to 5 were included in statistics of new host-parasitoid relations and significant biogeographic records, unless the finding represented the first record for a gall midge host of the relevant parasitoid taxon.

Parasitoid name Gall midge host name Method Collection date Emg. month Coll. C. Int. nov. Biogeo. nov.
Hymenopteran parasitoids reared from European gall midges (Diptera, Cecidomyiidae) Hymenopteran parasitoids reared from European gall midges (Diptera, Cecidomyiidae) Several (i.e.85) of the records represents the first reported occurrence of the focal species in the relevant country, of which the majority (i.e.69) refer to Denmark (Table1).