Implications for management of wireworms: insect pest suppressiveness of soils with plant functional diversity

Plant species diversity in agricultural systems has been shown to provide multiple ecological and agronomic benefits. How it affects the soil microbial communities, especially the guild of entomopathogens capable of controlling insect pests, has not been studied. We used the set-up of the “Experimentation in Ecosystem Research”, at Jena, Germany, to study the role of plant species diversity in facilitating insect pest suppressiveness in soils via entomopathogens such as insect-killing fungi (EPF) and nematodes (EPN). The aim was to gain insights for improving the natural control of wireworms and other pests by supporting the levels of pest antagonistic organisms in the soil. Soils were sampled from plots with a fixed number (from zero to 16) of plant species from different functional groups. Samples were baited with mealworm Tenebrio molitor larvae and the mortality (k-factor) was determined. A supplementary study was conducted to relate the findings to a soil pest, wireworm ( Agriotes sp.), using potting soil and a high and low dose of entomopathogenic fungi and nematodes. Field-collected wireworms were subjected additionally to a maximum challenge test, using a Metarhizium strain naturally occurring in the soil from which the wireworm larvae were collected. Our results showed that the increase in the number of plant species at the growing site can significantly increase the killing power (k-factor) of the soil. Plots with legumes had significantly higher k-values than plots without. EPF occurred significantly more often in plots with legumes than in plots without. Wireworms survived entomopathogen treatments where all mealworms died, but developed mycosis in the maximum challenge test. These findings need to be incorporated into cropping system design as a component in maximizing ecosystem services via ecostacking.


Introduction
Wireworms, the soil-dwelling larvae of click beetles (Coleoptera: Elateridae), are a cosmopolitan family consisting of 13 subfamilies and over 10,000 described species (Johnson 2002;Han et al. 2016). Only a few species are harmful to agricultural crops (Poggi et al. 2021), but these have been notorious as major pests worldwide for a long time. Pestiferous wireworms are extremely polyphagous and can cause severe economic damage on major arable crops such as potato, maize, and cereals. In North America about 965 valid species of wireworms exist, including about a dozen introduced species, six of which are known pests (Johnson 2002;Douglas 2011).
The increasing pressure towards a reduction in pesticide use has resulted in strong demand for alternative methods to control wireworm populations (reviewed by Poggi et al. 2021). Banning of several conventional pesticides has aggravated the wireworm problem, as effective natural control appears to be rare (Poggi et al. 2021). In our current study we focus on possible improvement of wireworm control by addressing one of the key components of natural biological regulation of pest insects: entomopathogenic fungi and nematodes present in the soil. Together with predators and parasitoids, entomopathogens can play a crucial role in the natural regulation of pest numbers.
Pest insects entering the soil at some stage in their lifecycle (often for pupation), or living in the soil as larvae (e.g., wireworms, root flies), often suffer high levels of mortality while in the soil due to entomopathogens or other factors (e.g., Razinger et al. 2014, Alyokhin et al. 2020, Erasmus et al. 2021, Menzler-Hokkanen et al. 2022. The extent of this mortality often remains unknown, and the soil component is frequently described as a "black box" in terms of population dynamics (e.g., Andrén and Balandreau 1999;Briones 2014;Custer et al. 2020). As this component of insect population regulation is practically absent in current agricultural fields (e.g., Zec-Vojinovic et al. 2006, Hokkanen andMenzler-Hokkanen 2017), there is an opportunity to utilize it for improved pest control by taking advantage of insect pest suppressive soils in the overall ecostacking approach (Hokkanen 2017;Hokkanen and Menzler-Hokkanen 2020). The improved biocontrol in the soil complements the mortality caused by predators and parasitoids, as well as other factors limiting the population size including competitors, plant secondary metabolites and induced resistance.
The role of various forms of plant diversity in providing ecosystem services, promoting community stability, and supporting sustainable management of crop pests and diseases in agricultural ecosystems has received increasing attention (e.g., Roscher et al. 2004, Letourneau et al. 2011, Ratnadass et al. 2012, Staudacher et al. 2013, Schipanski et al. 2014, Vukicevich et al. 2016, Farooq et al. 2022). It has been firmly established that selective, functional plant diversity in agricultural environments has multiple benefits, and for example can increase soil microbial diversity and mitigate decline of agricultural productivity (Vukicevich et al. 2016), provide community stability and overall species richness (Farooq et al. 2022), and affect the behavior of generalist root herbivores, reducing their damage and thereby enhancing crop yield (Staudacher et al. 2013).
However, to our knowledge, none of these studies have examined the relationship between plant diversity and soilborne entomopathogens providing biological control of crop insect pests. In natural ecosystems biological control services are provided at multiple life stages of pests, and in all environments including the soil, in which most insect species spend at least one life stage (e.g., pupation) (see Hokkanen and Menzler-Hokkanen 2017). The soil component in this respect is incompletely known, its role is often overlooked, and for example in many life-table studies only rough approximations are used (see Menzler-Hokkanen et al. 2022, and references therein).
Extensive surveys of agricultural field soils all over the world have shown that conventionally managed fields growing annual crops are practically void of these antagonists (Zec-Vojinovic et al. 2006). In contrast, soils in natural ecosystems typically harbor entomopathogens in quantities at several orders of magnitude higher than those in agricultural fields (Vänninen et al. 1989, Hokkanen andMenzler-Hokkanen 2017). The low levels of entomopathogens in annual cropping systems therefore imply that these environments do not sufficiently support the occurrence and survival of these important natural antagonists of pest insects. Entomopathogens, however, easily survive natural abiotic conditions in the field. Therefore, it appears that their absence in field soils is caused by some current agricultural practices, which are not compatible with insect pathogens. Soil-dwelling entomopathogens are likely to suffer from multiple stresses, such as frequent exposure to strong UV light when fields are without vegetation cover, dilution in soil via inversion tillage, lack of suitable host insects over long periods of time during the rotation, and the use of suppressive agrochemicals (e.g., fungicides, herbicides, mineral fertilizers) (Hokkanen and Menzler-Hokkanen 2017).
Our overall aim is to determine how crop, vegetation, and soil management in agricultural systems affect the natural occurrence in the soil of the most important soil-borne natural antagonists of crop pests -entomopathogenic fungi and nematodes. These components of functional biodiversity will determine the "suppressiveness" of soils towards pest insects. In our earlier studies (e.g., Vänninen et al. 2000;Zec-Vojinovic et al. 2006,) we have shown that different crop and soil management strategies can support the existence of insect antagonistic biota in the soil. We also know based on our earlier work that natural, undisturbed habitats such as forests and orchards harbor high levels of activity by these microbial antagonists of crop pests (Vänninen et al. 1989, Hokkanen andMenzler-Hokkanen 2017).
The primary objective of the current study was to determine the role of plant functional diversity in supporting the natural occurrence of entomopathogenic fungi and nematodes in the soil, using Tenebrio molitor as an indicator of the killing power of the soil. Keller and Schweizer (2001) already proposed that wireworm problems in agriculture may be due to lack of effective entomopathogens in the soil. Therefore, a further objective of our study was to relate our findings to the management of a group of serious crop pests-the wireworms-with entomopathogenic fungi and nematodes; see Shah et al. (2023) for a similar approach. Furthermore, our aim was to gain insights into whether the history of crop and soil management ('legacy') influences the diversity and strength of the antagonistic capacity (functional biodiversity) in soils.
We expected to discover what type and level of botanical biodiversity is needed for the natural microbial antagonists of insect pests to survive at levels great enough to be of importance in crop pest management.

Study site and description of study plots
The study was conducted by making use of the experimental set-up of the Jena Experiment in Germany (Roscher et al. 2004). This is a large long-term biodiversity experiment situated in the floodplain of the river Saale near the city of Jena (Thuringia, Germany; 130 m above sea level; 50° 55′ N, 11° 35′ E). The region around Jena has a mean annual temperature of 9.9 °C and mean annual precipitation of 610 mm (1980( -2010( , Hoffmann et al. 2014. The soil of the experimental site is a Eutric Fluvisol, almost free of stones. The soil texture ranges from sandy loam close to the river to silty clay with increasing distance from the river. The target community to create a species pool for the biodiversity experiment was Central European mesophilic grassland of the Arrhenatherion type (Ellenberg 1988). To establish the experiment, sixty species typical for these grasslands were selected, and assigned to four functional groups: grasses (16 species), small herbs (dicots) (12 species), tall herbs (20 species) and legumes (12 species). The Jena Experiment controls plant species richness (1, 2, 4, 8 and 16 species) and functional group number (1, 2, 3 and 4 functional groups) in a near orthogonal design with the restriction that it is not possible to have more functional groups than species in a community. Each species richness level has 16 replicates with the exception of the 16-species mixtures, which had 14 replicates because enough species were not available in the species pool to have 16-species mixtures with only legumes or small herbs. The experimental communities were created by random draws from the respective functional group pools. Plots were arranged in four blocks parallel to the river, with replicates running perpendicular to the river (see: http:// the-jena-exper iment. de/ index. php/ locat ion/) to account for the gradient in soil characteristics. The experimental communities were established by an initial sowing in May 2002 with a total of 1000 viable seed per m 2 , distributed equally among species in the mixtures. Some poorly established species were resown a single time in autumn 2002 (for details, see Roscher et al. 2004). The plot size at the time of our sampling was 6 × 6 m (= 36 m 2 ). The biodiversity experiment has been managed as it is typical for extensive hay meadows, that is, the plots are mown two times a year, in early June and in early September and the mown plant material is removed. The plots were not fertilized. To maintain the original species composition, the plots were weeded two or three times a year (April, July and October).
For this study, soils were sampled from nine different plant species mixes, ranging from 0 to 16 plant species.
Description of the botanical composition of each of our study plots is given in Table 1. In addition, an agricultural field next to the Jena experiment, was sampled. All sampled plots were situated in similar soil, equidistant from the river Saale.
Estimating the recycling capacity for entomopathogens based on plant functional biodiversity A list of herbivorous insects known to utilize the plant species growing in our study plots, and to spend at least some part of their life-cycle in the soil, was collated from two main sources: https:// www. bioin fo. org. uk/ and https:// www. comma nster. eu/ Comma nster/ Plants/. The aim was to estimate the extent to which these plants are capable of maintaining populations of insects that could support the recycling of EPN and EPF in the soil. This assessment is complicated by the variation in susceptibility and abundance of the listed herbivorous insects. A high number of regularly occurring, susceptible insect species can maintain populations of entomopathogens in the soil much better than irregularly occurring rare species (own, unpublished data). The number of insect species known to live on the study plot plants, and which could sustain populations of entomopathogenic fungi and nematodes in the soil, are given in Table 1.

Soil sampling and processing
The first soil sampling at Jena was carried out on 15-16 March 2015, and the second on 6-7 March 2016. In both years, the same plots were sampled. The sampling was carried out with a soil borer (40 mm diameter) to a depth of 5 cm. From each plot, 5 samples at each of 5 different spots on the plot were taken (25 soil cores per plot in total). Samples were placed in a plastic container, and homogenized to one homogenous sample per plot. All samples were transported to our laboratory in Helsinki, and processed within two weeks, maintaining them at + 7 °C.
The bioassays followed the principles described by Bedding and Akhurst (1975) and by Zimmermann (1986). For the bioassay, the moisture content of the soils was adjusted to 40% (vol/vol), corresponding to the typical natural moisture level of the soil type. Each sample was divided into four replicates of about 2 dl volume. These were placed into a plastic container with a perforated lid. Three larvae of the mealworm T. molitor-routinely reared and maintained in our laboratory-were placed as a bait into each container. The containers were stored at room temperature of 22 ± 2 °C for three weeks, and inspected every second day for mortality of the bait larvae. Dead larvae were surface sterilized with 70% ethanol, and incubated at 100% RH in moist Petri

Comparative bioassay with larvae of wireworms Agriotes sp. and mealworm T. molitor
To relate the findings obtained from the Jena field study site, where we used a highly susceptible bait insect species (T. molitor), to a serious soil dwelling crop pest, we conducted comparative bioassays using larvae of wireworms Agriotes sp. Wireworm larvae were field-collected on 6-7 October 2022 from a grassy permanent field margin in southern Bavaria, Germany (425 m above sea level; 48° 9' N, 12° 19' E) and immediately taken into bioassays. Mealworm larvae were obtained from a pet-food shop. Untreated commercial gardening soil was used for the experiments, conducted in 1 dl plastic containers with a perforated lid. These bioassays were conducted as described above. Seven treatments with three replicates were included, one set with the mealworm bait larvae, and another set with wireworm larvae. The selected EPF have previously been shown to be infective to wireworms (Reddy et al. 2014;Sharma et al. 2020). The range of entomopathogen doses used in this part of our study relate to the range of their densities reported in the literature for agricultural and natural soils, including our own earlier studies (e.g., Vänninen et al. 1989Vänninen et al. , 2000Hokkanen and Menzler-Hokkanen 2017). The treatments were as follows: 1. Control with untreated potting soil only 2. Soil inoculated with a high dose of the entomopathogenic fungus Beauveria bassiana strain GHA, obtained from Dr. Stefan Jaronski (USDA/ARS, Sidney, MT, USA), 5 × 10 5 cfu/gram of soil (cfu = colony forming units) 3. Soil inoculated with a low dose of B. bassiana, 5 × 10 3 cfu/gram of soil 4. Soil inoculated with a high dose of the entomopathogenic fungus Metarhizium robertsii strain DWR356, obtained from Dr. Stefan Jaronski, 5 × 10 5 cfu/gram of soil 5. Soil inoculated with a low dose of M. robertsii, 5 × 10 3 cfu/gram of soil 6. Soil inoculated with a high dose of the entomopathogenic nematode Steinernema feltiae commercial product Nemaplus (e-nema, Schwentinental, Germany); 10,000 infective juveniles (IJ) per pot 7. Soil inoculated with a low dose of S. feltiae; 1000 IJ/pot Inoculation of the potting soil with the entomopathogens was carried out by diluting the formulated products (granular for the EPF, powder for the EPN) in a small amount of water and mixing thoroughly with the potting soil prior to the potting. After preparing the experimental pots, three test larvae were placed in the pots, and the pots were incubated at room temperature of 22 ± 2 °C for three weeks.
In addition, soil from the field edge from where the wireworm larvae had been collected, was tested with the mealworm larvae for natural presence of entomopathogens. Three replicates with 10 T. molitor larvae in each replicate were used for this assay (conducted in 3 dl containers). A further set of three soil samples was collected from the agricultural field itself, at 2 m distance to the field edge, and baited with 10 mealworm larvae each.
As the field-margin soil, from which the wireworm larvae were collected, contained effective Metarhizium sp. rapidly killing all 30 mealworm bait larvae, we subjected the *Data from Marquard et al. (2016) except for the plot B0 A1 **Number of known insect herbivore species for each plant taxon, having a life stage in the soil. Data collated from https:// www. bioin fo. org. uk/ and https:// www. comma nster. eu/ Comma nster/ Plants/ surviving wireworm larvae to an additional "maximum challenge test" with the Metarhizium from their native soil. All mealworm cadavers sporulating clearly with green Metarhizium spores were combined in one plastic container with the original infective soil, following the basic idea presented by Shapiro-Ilan et al. (2003) to augment entomopathogens in the soil. The wireworm larvae were added to this container to investigate whether a high (but unquantified) dose of the native Metarhizium would infect the wireworms.

Collation of bait mortality results
The 'killing power' of the tested soils, the k-value, was calculated based on Varley and Gradwell (1970). The number of bait insect individuals which died was compared with the initial number, and expressed logarithmically: where a x is the number of individuals at the start of the experiment, time x, and a x+1 is the number at the end of the experiment, time x + 1. While the primary aim was to determine the killing power (k-value) of the examined soils with respect to the tested bait larvae, the cause of mortality was recorded when this was obvious. Thus, the presence of entomopathogenic nematodes was recorded as "EPN", and the presence of entomopathogenic fungi as "EPF". No effort was made to determine the species of the EPN, but EPF were classified as Metarhizium sp., Beauveria sp., or Isaria sp. based on spore morphology and color, when obvious. k = log 10 a x − log 10 a x+1

Statistical methods
Linear regression and logistic regression were performed using GraphPad Prism version 9.3.1 for Mac, GraphPad Software, San Diego, California USA.
Logistic regression was used for binary classification of presence versus absence of leguminous plant species, whereas linear regression was used to model how the number of plant species affect k-values, and the detection of entomopathogenic fungi (EPF) and nematodes (EPN). For linear regression we report the linear equation, R 2 as a metric for goodness of fit, and the F-test-derived p value testing the null hypothesis that the overall slope is zero. For logistic regression we report the curve equation, Tjur's R 2 as a metric for goodness of fit, and a likelihood ratio test (LRT) with derived p value to test the null hypothesis of non-superiority of a model without the given predictor. Statistical significance was considered as p < 0.05.

Results
Baiting the soil samples with T. molitor resulted in larval mortality between 17 and 83%, and k-values between 0.079 and 1.041, depending on the sample plot and the year ( Table 2). Most larvae did not produce identifiable indications regarding the cause of mortality. Of the 22 larvae showing signs of EPF, 12 were classified as Beauveria, 7 as Metarhizium, and 3 as Isaria. EPN were found from 13 dead mealworm larvae in 9 different samples (Table 2).

Impact of plant presence, diversity and species composition on k-values
Plant presence increased the k-values of the soil, compared to bare soil (Fig. 1). Increasing the plant diversity on the plot significantly increased the k-values. The increase was described by linear regression as Y = 0.02141*X + 0.1980. The goodness of fit R 2 is 0.3135 (n = 18) and the slope was significantly non-zero (p = 0.0157), indicating a positive correlation between the number of plant species and k-value (Fig. 1). The presence of legumes had a highly significant effect on k-values compared to plots without legumes. A logistic regression for the presence vs absence of legumes (represented in Fig. 2 as 1 and 0, respectively) as a function of the k-values was described as: log odds = -3.066 + 9.203*X. The Tjur's R 2 was 0.3971. The LRT rejected the null hypothesis of non-superiority of a model without the legumes as a predictor, with a p-value of 0.0037 (n = 18) (Fig. 2).

Impact on detection of entomopathogenic fungi (EPF) and nematodes (EPN)
The detection of the occurrence of entomopathogenic fungi was not significantly related to the number of plant species in the plot, although a positive trend was apparent.
However, the R 2 of the linear regression (0.1549) was not statistically different from a zero slope (p = 0.1061). Therefore, we could not show that the number of plants correlated with the number of bait insects with EPF (Fig. 3).
The presence of leguminous plants in the plot, in contrast, was a significant predictor for the detection of EPF in the plot (Fig. 4). A logistic regression for the  presence vs absence of legumes as a function of the number of bait insects with EPF was described as: log odds = − 1.442 + 1.185*X. The Tjur's R 2 was 0.3570. The LRT rejected the null hypothesis of non-superiority of a model without the legumes as a predictor, with a p-value of 0.0083 (Fig. 4).
Similar to the results with EPF, the detection of the occurrence of entomopathogenic nematodes was not significantly related to the number of plant species in the plot (R 2 = 0.1819; p = 0.0776), although a positive trend was apparent in this case as well. Therefore, we could not show that the number of plants correlate with the number of bait insects with EPN (Fig. 5).
For the detection of EPN, the presence of leguminous plants in the plot was not a significant predictor (Fig. 6). A logistic regression for the presence vs absence of legumes as a function of the number of bait insects with EPN was performed. The Tjur's R 2 was 0.02878. The LRT was performed, with the result that we could not reject the null hypothesis of non-superiority of a model without the legumes as a predictor (p = 0.4700; (Fig. 6).

Comparison between mealworms and wireworms as bait
Baiting untreated and entomopathogen-treated commercial potting soil with T. molitor and Agriotes sp. larvae showed that all mealworm larvae died within the first 16 days of the test in most treatments, while all but one wireworm larvae survived until the end of the assay 36 days after treatment (Table 3). Two mealworm larvae (out of 54 larvae) survived the entomopathogen treatments, both in treatments with M. robertsii. In contrast, only one wireworm larva died (in a M. robertsii treatment). Unfortunately, the commercial potting soil used in our experiment contained naturally occurring entomopathogenic fungi, which killed all but two mealworm larvae during the experiment (several showed signs of Beauveria infection).
The test with the soil from which the Agriotes larvae initially were collected for this study (field margin), showed that it was highly infective to T. molitor larvae: all 30 test mealworm larvae had died within 8 days after baiting the soil. Most of them showed clear signs of Metarhizium infection. In contrast, soil samples just 2 m away into the agricultural field itself showed completely different results: all 30 mealworm larvae were still alive at 8 days after treatment. At day 16, two bait larvae had died (k-value 0.030), one being infected with Metarhizium.
All Metarhizium-infected cadavers of T. molitor were used for a further 'maximum challenge' test with the wireworms. Out of 36 larvae in this test, 4 (11.1%) had died within 7 days, and further 10 within 16 days after treatment. At the final check on day 24, a total of 17 (47.2%) wireworm larvae had died (k-value 0.278). Almost all showed signs of Metarhizium infection.

The role of plant functional diversity for recycling of entomopathogens
Our study showed that the soil antagonistic ability ('suppressiveness' sensu Hokkanen and Menzler-Hokkanen 2018) towards pest insects is connected to the functional plant biodiversity of the site. The presence of leguminous plants appeared particularly important in promoting and maintaining this biocontrol capacity. We believe that the critical functional characteristic of the vegetation in this respect is the ability of the plant species to serve as host to naturally occurring insect species, which have a life stage living in the soil. Leguminous plants in Central Europe harbor several dozen species of such insects (see Table 1), many of which are known to be good hosts for entomopathogenic fungi and nematodes. For example, Trifolium species are hosts for at least 36 species of weevils, and several other species such as beetles, gall midges, and sawflies (data from sources cited in Table 1). In contrast, grassy species have much fewer herbivorous insects with a life stage in soil (Table 1). In addition, the majority of these species are various small flies or midges, which are poor hosts at least for entomopathogenic nematodes and cannot effectively recycle them (e.g., Sulistyanto et al. 1994;Nielsen and Philipsen 2004a, b). Various species of herbs included in the study plots are intermediate between legumes and grasses, with respect to being able to host insect species suitable for supporting recycling of entomopathogens in the soil (Table 1). Ratnadass et al. (2012) reviewed the possible role of plant species diversity for the sustainable management of crop pests and diseases in agroecosystems. They concluded that introducing a selected plant species diversity might be a better option for building up beneficial microbial populations than directly inoculating the soil with beneficial microorganisms. In their study, however, they did not discuss the enhancement of entomopathogens via alternative insect hosts. Recently, the interaction among soil micro-organisms, the plant, and the associated insects have been shown to have a large impact on insect herbivores, signifying the importance of the soil component for the plant and associated insects (e.g., Eichholtzer et al. 2021;Nenadić et al. 2022;Chitty and Gange 2022).
Only a few studies have addressed in any detail the key issue of ensuring recycling of entomopathogens in the soil. Nielsen and Philipsen (2004a) examined how various soil-dwelling life stages of common pest insects can support EPN recycling. They found that the cabbage root fly Delia radicum larvae could produce new generation of infective S. feltiae juveniles ranging from 359 per larva (from the smallest larvae) to 3549 (from the biggest larvae). The cabbage seedpod weevil Ceutorhynchus assimilis yielded an average of 1423 IJ/larva when infected with S. feltiae, and the pollen beetle Meligethes spp. yielded 1263 IJ/larva when infected with S. feltiae and 1260 IJ when infected with Heterorhabditis bacteriophora.
Cropping system management using techniques such as cover crops, crop rotations, and habitat management, has recently been investigated intensively to improve ecosystem services to the crop, such as biological control of pests (e.g., Letourneau et al. 2011;Ratnadass et al. 2012;Farooq et al. 2022). Unfortunately, most studies do not consider the soil ecosystem, and even fewer address the possible role of entomopathogens as antagonists of pests (but see Staudacher et al 2013). Ahmad et al. (2020), however, present an interesting study suggesting that winter cover crops may help to conserve Metarhizium spp. in annual cropping systems. Vukicevich et al. (2016) reviewed the possibility to manipulate soil microbial diversity through selection and management of cover crop mixtures. They concluded that increasing plant diversity in the field increases soil microbial diversity and increases populations of beneficial microbes. This could be achieved by increasing plant functional group richness via incorporating legumes, C4 grasses, C3 grasses, and non-leguminous forbs in the system. They also concluded that frequent tillage, herbicide use, and fungicides often harm populations of beneficial microbes. When cover crops were introduced into a 3-yr soybean-wheat-corn  (Schipanski et al. 2014). Specifically, Nielsen and Philipsen (2004b) showed experimentally that the inoculated EPN S. feltiae established well when pea was grown in the same or in the following year, whereas the nematodes were less able or unable to establish in other crops (cabbage, carrots, barley). The successful establishment of EPN, and their abundance in general, correlated well with the presence of plant specific insects, with Sitona lineatus being the key species (associated with leguminous plants).

Targeting soil pests with suppressive soils
We used a typical soil-borne pest, the wireworm Agriotes sp., as a comparison to our standard bait insect, the mealworm T. molitor, for assessing comparative killing-power of soils for insect larvae. Wireworms are quite resistant to soil inhabiting entomopathogens, and under normal field conditions suffer only minimally from natural infections: in a large survey Kleespies et al. (2013) found a natural infection rate of only 0.66%. Similarly, Keller and Schweizer (2001) found that arable fields in Switzerland had only low densities of entomopathogenic fungi while adjacent meadows had significantly higher densities of Metarhizium. They suggested that wireworm problems in arable fields may be due to a lack of effective biocontrol by EPF.
Our study confirmed that Agriotes sp. larvae are difficult to be killed by augmented entomopathogens. Many authors have demonstrated that specific strains of Metarhizium significantly increase wireworm mortality (e.g., Reddy et al. 2014, Razinger et al. 2020, Bourdon et al. 2021. Kabaluk and Eriksson (2007), and  consider that factors such as temperature, time exposed to Metarhizium, conidia soil concentration, and food availability affect mortality rates of wireworms, and are likely to influence field performance of Metarhizium as a biological control agent. In a study published in 2017, Kabaluk et al. found that of asymptomatic larvae of Agriotes collected from the field, 13-100% were latently infected with Metarhizium brunneum. Furthermore, they discovered that wireworm larvae hosted symbiotic bacteria, which appear to suppress the infection by Metarhizium. This might explain the difficulties of controlling wireworms with entomopathogens in the field.
Our result from baiting the soil from which the wireworms were collected indicated that the soil was highly suppressive to the mealworm larvae because of high natural presence of Metarhizium sp. Despite that, some wireworms were living in the grassy field margins. Exposing the wireworms to the 'maximum challenge' conditions, however, resulted in high mortality of the wireworms-possibly their defences (provided by the mutualistic bacteria, as proposed by Kabaluk et al. 2017) were overwhelmed, and infection could take place.
Many other soil pests, including the larvae of cabbage root flies Delia spp. and Tipula spp., have proven to be difficult targets for biocontrol by entomopathogens (e.g., Sulistyanto et al. 1994;Vänninen et al. 1999). It is conceivable that their ability to resist infection is related to the mechanism proposed for wireworms, as was shown by Zhou et al. (2021) for B. bassiana on the onion fly Delia antigua.

Integrated crop management practices and ecostacking
One of the overall aims of our work is to find ways to enhance the presence of entomopathogens in agricultural soils, as one critical component to the ecostacking approach (see Hokkanen 2017; Hokkanen and Menzler-Hokkanen 2020). Our results show again that agricultural fields under conventional management and with mainly cereals (grassy species) as the cultivated crop, cannot sustain antagonist populations that would be of importance in pest control. In order to build up and maintain soil suppressiveness to insect pests, appropriate management practices including the growing of leguminous species for example as undercrop, or as a cover crop, have to be developed. In addition, inversion tillage should be minimized (Hokkanen and Menzler-Hokkanen 2020).
Only few studies have been conducted to assess whether the prevalence and killing power of entomopathogenic organisms can be enhanced in agricultural systems by specific management practices. Zec-Vojinovic et al. (2006) reported results from the EU project MASTER (Management Strategies for European Oilseed Rape Pests), which included a comparison between conventional (STN) and integrated crop management (ICM) practices in oilseed rape cropping system over the project period of three field seasons in four different countries. The ICM treatment differed from the STN treatment by a reduced number of pesticide sprays, and by substituting inversion tillage with minimum tillage. In the first year (baseline, 2003), when the experimental fields were established, the STN system plots had a higher killing power by insect antagonists than the ICM system plots. The killing power indices for EPN and EPF were not significantly different between the systems in any country. In their field study, practicing the relatively simple ICM system resulted in a significant increase in the killing power by insect antagonists overall, and of nematodes in the following 2 years, and in a nonsignificant increase in the killing power of EPF. A significant increase in the mortality proportions of the bait insect larvae occurred from 2003 to 2004, followed by a slight (nonsignificant) increase from 2004 to 2005. This indicated that the ICM system effect could be more stable than the STN system, which showed significant variation over the years for each of the examined agents. The ICM system affected the examined agents similarly in each country. This suggested that different climatic conditions and also different base lines did not have an impact on the system effect (Zec-Vojinovic et al. 2006).
Another detailed study on the impact of different management systems on both the prevalence and killing power of naturally occurring EPN was conducted by Campos-Herrera et al. (2008), with similar results as those from the MASTER project. They observed that the abundance and recovery frequency of naturally occurring EPN significantly decreased with an increase in agricultural disturbance (or intervention), and that EPN could not be recovered from conventional annual crops. Besides the strong influence of agricultural management practices on EPN, Campos-Herrera et al. (2008) suggested that certain types of soil environments (such as heavy clay soils having compact structure) are unsuitable for these organisms. They also speculated that agronomic management could affect the natural activity of EPN by reducing their virulence and reproductive potential.
Our current study provides one plausible explanation for the lack of effective insect antagonist populations in agricultural soils: entomopathogens cannot sustain their populations at effective levels if their recycling is not ensured (c.f. Vänninen et al. 1989). This might be improved by introducing appropriate plant functional diversity into the cropping systems, with emphasis of including legumes when-ever possible as undercrops, cover crops, and/or rotation crops. Besides apparent effect of improving insect pest biocontrol, legumes provide a multitude of other ecosystem services, all critical for maximally synergizing the impact of various ecosystem services for agricultural productivity (ecostacking).
Acknowledgements The main part of this study was supported by the EC Grant agreement from the "Experimentation in Ecosystem Research" (ExpeER), https:// www. expee ronli ne. eu/ , that financed the user group's access to the research infrastructure. The grant number is 262060, and project title "Legacy effects of crop management on pest suppressiveness of soils" (acronym: SUPERSOILS). We thank two anonymous reviewers and the APIS Special Issue Editors for their constructive comments, which helped to strengthen our manuscript.
Funding Open access funding provided by University of Eastern Finland (UEF) including Kuopio University Hospital.

Data availability
The data reported in this article is available upon request from the corresponding author.

Conflict of interest
The authors have no competing interests to declare that are relevant to the content of this article.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http:// creat iveco mmons. org/ licen ses/ by/4. 0/.