The mechanisms and ecological circumstances of adult diapause in Ichneumonidae are poorly studied. An overview is presented of what observations and research have been carried out on ichneumonid diapause to date, and new ecological and distributional data are presented. The new data primarily concerns species that hibernate in association with trees, based on observations made in Belgium and the Netherlands. A preliminary checklist of the 50 species that are now known to hibernate is provided for both these countries. Auberteterus alternecoloratus (Cushman, 1929), Dicaelotus montanus (de Stefani, 1885), Dicaelotus pictus (Schmiedeknecht, 1903) and Orthocentrus sannio Holmgren, 1858 are reported as adult hibernators for the first time. Four species are newly recorded for the Belgian or Dutch faunas.

Darwin wasps, diapause, ecology, hibernation

Diapause is a broad concept used to designate a dynamic process in several successive phases of low metabolic activity, at least partially genetically determined, with the neuro-hormonal system as mediator (Denlinger 2002; Kostál 2006). It serves to help face and survive seasonal changes (mostly in temperature and humidity) at a low energetic cost (Hahn and Denlinger 2007; Tougeron 2019). Diapause may occur at any stage of life: egg, larval, pupal or adult; the latter is also called reproductive diapause (Denlinger 2002; Quicke 2015). During adult diapause, respiration rate declines and the reproductive functions are put on hold (Fielding 2008; Chen et al. 2012; Hodek 2012).

Depending on the seasonal characteristics, specific terminology is used such as summer diapause or aestivation and winter diapause or hibernation. The latter has generally received most attention (Tougeron 2019). It is important to note these terms do not correspond well to the timing of the seasons in the field: winter diapause may begin as early as midsummer for example (Hodek 2012). In obligate diapause, species arrest development regardless of environmental conditions, whereas in species with facultative diapause, external conditions do have an impact and determine whether the species undergoes diapause or not (Hahn and Denlinger 2007). Recent studies suggest both tendencies are highly flexible, experiencing phenotypic plasticity, and may thus be mixed at species or even population level, inherently resisting any kind of generalization (Hodek 2012; Tougeron et al. 2018a, b).

It is agreed that photoperiodism, the physiological reaction of organisms to day length, is the key factor in regulating adult diapause (Hodek 2012; Tougeron 2019). Consequently, a different geographic latitude possibly results in differing diapause strategies, caused by the specific photoperiodic regulation at population level (Hodek 2012). A second crucial aspect seems to be temperature. Several studies suggest how, next to photoperiodism, temperature plays an important role in regulating adult diapause (Sgolastra et al. 2010; Chen et al. 2012), even in areas with a similar latitude (Tougeron et al. 2018a). On a biochemical level, it is known how glycerol levels, which contribute significantly to maintain metabolism and the utilization of metabolic reserves (Storey and Storey 1990; Denlinger 2002), will increase through a specific phosphorylation mechanism, starting from 0–5 °C (Hahn and Denlinger 2007). Furthermore, respiration rates decrease at lower temperatures. In general, energetic costs are reduced in that case (Hahn and Denlinger 2007). Other possible factors governing diapause are food availability and quality, humidity and population density (Hodek 2012).

While adult diapause is very common in some orders, especially Coleoptera and Hemiptera (Heteroptera), the incidence of this phenomenon is far lower among the Lepidoptera and Hymenoptera, involving approximately 5% of the temperate species in each (Hodek 2012). In the Ichneumonidae, although there has been some research on diapause (mainly) in early life phases (Claret 1973, 1978; Aeschlimann 1974; Griffiths 1975; Renfer 1975; Arthur and Mason 1985; Shaw and Wahl 1989; Coop and Croft 1990; Zijp and Blommers 2002a; Humble 2006), the majority of studies have focused on adult hibernation. Some studies on hibernating ichneumonids have contributed significantly to our more general understanding of diapause. For example, Duffield and Nordin (1970) demonstrated the relationship between presence or absence of glycerol synthesis and hibernation among North American species. Hibernating ichneumonids mainly belong to the subfamily Ichneumoninae, although a few examples from Cryptinae, Orthocentrinae, Metopiinae, Phygadeuontinae and Pimplinae are also known (Quicke 2015 and see Discussion). It is important to note this review principally addresses ‘free living’ adult wasps. More exceptionally species pass much of the year as cocooned adults (e.g. several Campopleginae) or as cocooned pharate adults (e.g. Agriotypinae and some Lissonota spp.) (see Quicke 2015 and Shaw et al. 2016 for some examples).

The hibernating ‘free living’ adult wasps are always fertilized females, having mated before entering diapause. During hibernation the female’s abdomen sometimes appears to be swollen, due to the ripening of the ovaria (Bischoff 1927; Leruth 1939; Selfa and Escola 1991). When diapause reaches its end the females are ready to start searching for hosts. Tereshkin (1996) suggests how adult hibernators benefit from using specific mechanisms for coping with harsh(er) climatic conditions although these have been little studied. Further, little attention has been given to phenological synchronicity between adult (hibernating) parasitoid wasps and their hosts, an interesting and unique relationship, which could be destabilized by climate change (Godfray 1994; Tougeron et al. 2021).

To date, most studies have been on hibernating species themselves and their different diapause sites or hibernacula, which are known to affect diapause energetics and costs by, for example, decreasing metabolic rates (Hahn and Denlinger 2007). These sites are useful to structure observations because species tend to be site-specific, meaning they specialize in a specific ecological niche to hibernate. Examples are (old) buildings, caves, under stones or moss, among grass tussocks or even near the top soil layer between dense conifer foliage (e.g. pine needles) or leaf litter. Trees, both living and dead, are also used as hibernation sites.

Even within this range of hibernacula types, it is important to consider all conditions, as there can be small but important differences that may influence abundance and usage. For example, hibernating ichneumonids are possibly more common in caves with higher temperatures (Novak et al. 2010). Publications on hibernating ichneumonid wasps unfortunately do not always specify the hibernacula, and often implicitly focus on tree hibernation. Important works for the Palaearctic are summarised by country in Table 1. In addition, there are a number of unpublished observations we are aware of, and many other publications in which hibernation is mentioned only as a side note (see Verheyde 2022, in prep.). Studies related to particular types of hibernacula are discussed below and listed in Table 2. The total number of ichneumonid species that have been reported to hibernate is unknown but is at least 393 species worldwide (Yu et al. 2012).

Additional to existing literature, two datasets were used. The first dataset consists of the data exclusively gathered by the first author, specifically by searching for ichneumonid wasps on dead wood. To have an overview of his own field excursions, a fixed template in Microsoft Excel was used. So far, 53 excursions were entered, dating from 3 November 2018 to 20 February 2021 with a total searching time of 72.5 hours. Each excursion lasted for at least 30 minutes, and at least 10 trees had to be available to check. Subsequent variables were noted: location, date, temperature, daytime, searching time, amount of people searching, trees checked (approximation margin of 5), dominant tree species (> 5 present), other tree species, numbers of individual wasps found, amount of species found. Approximately 1515 tree trunks were checked and the average temperature was 9 °C. On average, it took three trees and 8 minutes and 30 secs to find a hibernating wasp. In one area, on average, 10 individuals of 3 different species were found. In order to limit ecological impact, the same area was never investigated twice in one season.

The second dataset consists of validated data entered in the citizen science portals waarnemingen.be and waarneming.nl. For the first author’s own observations, which were also entered, extra information was noted down at the level of individuals: species found together, tree species, tree diameter and any other interesting information (dead specimens, the position of the tree, etc.). Coordinates and exact timestamps were automatically linked when adding the findings via mobile phone. The second dataset also contains (live) data from other observers on waarnemingen.be and waarneming.nl. All data has been validated by the author and is freely available on the portals and partially on GBIF (https://doi.org/10.15468/r0fx1v). The set was harvested and analysed manually for all data entered before 1/10/2021. Both datasets will be used and merged more profoundly in a future study, and published with open access, which is why this study is preliminary. Specific observations from the portals are referred to with a unique identifier, called the ObsID (= unique identifier on GBIF); which can be placed in the browser to search for; for example https://waarnemingen.be/waarneming/view/169516420 or https://waarneming.nl/waarneming/view/226649126

Combining both datasets, in total 1120 records involving a total of 50 different species, were included in this study. A record is defined as a finding at one specific locality at one moment in time, unrelated to individual numbers at that locality (see also Table 3). Moreover, a record was deemed valid when a) the species could be safely identified to species (or species-complex) and b) it could be seen as hibernating. Hibernation was mainly defined by the presence of the wasp in a hibernaculum and/or exceptionally other contextual aspects such as temperature and phenology (see also Discussion). In cases of doubt we did not include the record. At the level of individual specimens, 524 wasps of 16 species were found by the first author.

Related to our results, especially those concerning tree hibernation, several ecological questions will be investigated and hopefully answered in the future. Rasnitsyn (1959) and Valemberg and Vago (1974a) investigated the presence of ichneumonid wasps in respect to the orientation of dead wood, stating the northern (and eastern) side of the tree is more often used because it has more stable temperature and humidity, especially in regions where temperatures are more extreme. Similarly, it is likely that species specializing in dead wood prefer trunks deeper in the forest, where the sun seldom raises substrate temperatures (Dasch 1971). Modern technology will allow easy measurement of these ecological parameters.

Some authors, for example Sebald et al. (2001) and Tereshkin (2002) have commented on habitat types and that the suitability of wood may depend on whether it is standing or lying flat but both questions have not yet been explored fully. The hibernaculum used may not be fixed for the whole winter. When disturbed, wasps may move. Some have been observed to move further down the tree and shelter under undamaged bark, while others have fallen to the ground and dug themselves into the litter where they would not hibernate normally. It is possible that some more accessible types of hibernacula, for example the clay and mud between the roots of a fallen tree (DTCL; Table 3), provide space for ‘temporary refugees’ or species that need an escape route separate from their ideal or permanent hibernaculum and might explain some rare examples of species crossing over ‘hibernation’ niches.

Another question relates to body size. It is obvious that ichneumonids hibernating on wood (often Ichneumonini) are much larger on average (12–18 mm) in comparison to ichneumonid wasps hibernating in, for example, grass tussocks (often Phaeogenini; 6–10 mm; see also Pénigot 2020). As Pénigot states, the most plausible factor at play here seems to be cost efficiency and avoidance of energy loss. On average, smaller species have less fat reserves and are thus more sensitive to energy loss (Lease and Wolf 2011). Getting access to certain hibernacula such as dead wood can be more demanding than access to various other hibernacula such as decaying vegetation. Other physiological barriers could also play a role, for example, limits of glycerol production or more generally the lack of antifreeze capability. On this subject more detailed research is crucial, for example to see if there are any differences among certain species or subfamilies (e.g. Shaw and Quicke 2000 on the braconid wasp Acampsis alternipes (Nees)). It is unclear how several surprisingly small ichneumonid wasps, such as Orthocentrus spp. (Orthocentrinae), often measuring around 5 mm, are able to hibernate in dead wood. Group clustering could be one part of the explanation (Rasnitsyn 1959 mentions up to 190 individuals in one case), but clearly does not suffice as many larger species use the same strategy (see above). The scarcity of findings suggests these species hibernate deeper inside the dead wood, several centimetres beneath the bark, rather than among the bark humus. Hypothetically then, they profit from thermal constancy deeper in the wood. It is unclear whether there is any difference in total overwintering time in comparison to larger ichneumonid wasps (presumably not). Further, it is not known why some larger ichneumonids, such as Ichneumon sarcitorius Linnaeus, 1758 and Diphyus amatorius (Müller, 1776), are only found in or under decaying vegetation and not on tree trunks. Perhaps biochemical analyses might provide insight.

A third aspect is the distribution and facultative nature of the diapause within some ichneumonid species. Some species are generally rare but can be locally abundant, and different local populations may show different diapause strategies. For example, Ichneumon xanthorius Forster, 1771 is extremely commonly observed in Belgium and the Netherlands, with close to a thousand registered records every year, and is known to hibernate under stones (S; Table 1). However, this wasp has only been found hibernating in those countries a few times. The same can be said for many other species which are known to hibernate in other parts of the Western Palearctic (Verheyde 2022, in prep.), but are only caught during the flying season in Belgium and the Netherlands. This seems to suggest there are indeed intriguing differences at the population level, which were emphasized by, for example, Hodek (2012) and Tougeron et al. (2018b).

There is still much to learn about what life style features, such as nutritional ecology and oogenesis, may be shared by the taxa involved. Host-parasitoid relationships could be of particular importance because of the above-mentioned differences among populations or facultative nature of the diapause. For example, many Ichneumon spp. oviposit into host (pre-)pupae, whereas the other ichneumonine wasps mentioned in this paper attack late instar host larvae. The latter would allow slight koinobiosis to an extent, enabling some degree of nutrient absorption from host haemolymph by eggs that hibernating adults might have had limited capacity to yolk (Hinz and Horstmann 2007; M.R. Shaw, pers. comm.).

Finally, a critical analysis needs to be made of the past literature and checklists mentioned in the tables. This is in particular related to certain subfamilies, such as Cryptinae, Phygadeuontinae and to a lesser extent Pimplinae. For practical reasons, most researchers (including ourselves) implicitly accept a species as ‘hibernating’ when it is found in a hibernaculum (for example, Rasnitsyn 1959: ‘In addition, several species of Ichneumoninae […] have been found in hibernation sites’) and/or shows visual signs of diapause. However, the discovery of an adult ichneumonid wasp during the winter does not automatically mean that it is hibernating. For example, it could be active during the winter (possibly polyvoltine), it could be a species surviving in artificial conditions; optionally being connected to a host in urban structures (e.g. in a heated house), or it could be imported from another climate. For the Low Countries, there are several cases that could be classified as such. Even more troublesome is the presence of active wasps near natural hibernacula. In the Netherlands, the cryptine Agrothereutes abbreviatus (Fabricius, 1794), was found under bark on 16^th December 2017 (ObsID: 146787701). Normally, this would be interpreted as hibernation, but it is possible the wasp was just actively searching for hosts there, instigated by changing climatic conditions. The same applies to (often apterous) females of several other cryptine and phygadeuontine genera and species, which have a long tradition of being accepted as adult hibernators (see for example Bischoff 1927). Careful examination is necessary case by case: some species, for example, do not hibernate as adults, some species hibernate as adults in the host cocoon and some hibernate as adults in litter (e.g. Gelis proximus (Förster 1850); Schwarz and Shaw 1999; Schwarz 2002). In the end, only molecular or biochemical analysis may provide a solution here.

We sincerely want to thank everyone who made any observations of hibernating ichneumonid wasps and posted those on citizen science portals, and especially those who made rare findings and made an extra effort to send specimens to the first author: Koenraad Bracke (Dicaelotus pictus, Hoplismenus bidentatus, Ichneumon stramentarius), Dirk Duytschaever (Orthocentrus sannio), Tom Kruize (Zanthojoppa lutea) and Karel Schoonvaere (Ichneumon confusor). It is great to see the passion of this incredible world spreading. Special thanks in this respect to Willem Vergoossen (Diphyus castanopyga/ palliatorius/ quadripunctorius; Exephanes ischioxanthus/ riesei) for his amazing efforts of monitoring the cave complexes in the Netherlands at least annually and thus establishing long-term data sets of Ichneumonidae, which is very rare in general.

FV is grateful to anyone who has accompanied him so far at least once on field trips: Rudy Claeys, David De Grave, Augustijn De Ketelaere, Sofie Declercq, Eli Devos, Kurt Geeraerts, Lowie Lams, Hannelore Theite, Thibaud Vandaudenard and Winfried Vertommen. Special thanks to Patrick Debeuf and Wim Duran who made an extra effort in this respect. Jean-Luc Vago is thanked for sending the papers of Joël Valemberg; Filippo Di Giovanni is thanked for sending additional papers related to Ichneumonidae and diapause.

We also thank those who helped with identifications: Geir Ørsnes and William Pénigot for Ichneumonini, Erich Diller for Phaeogenini and Andrei Humala for Orthocentrinae; thanks also to Augustijn De Ketelaere, William Pénigot, Karel Schoonvaere and Malcolm Storey for sharing data and exchanging ideas. Lennart Bendixen, Mark Shaw and Alexandra Viertler are thanked for their very useful feedback as reviewers of the paper.

DLJQ was supported by a senior postdoctoral fellowship from the Rachadapisek-Somphot Fund, The Graduate School, Chulalongkorn University.