Parasitic nematodes and hymenopteran parasitoids have been introduced and used extensively to control invasive Eurasian Sirex noctilio woodwasps in pine plantations in the Southern Hemisphere where no members of this community are native. Sirex noctilio has more recently invaded North America where Sirex-associated communities are native. Sirex noctilio and its parasitic nematode, Deladenus siricidicola, plus six native hymenopteran woodwasp parasitoids in New York and Pennsylvania, were sampled from 204 pines in 2011–2019. Sirex noctilio had become the most common woodwasp in this region and the native parasitoids associated with the native woodwasps had expanded their host ranges to use this invader. We investigated the distributions of these species among occupied trees and the interactions between S. noctilio and natural enemies as well as among the natural enemies. Sirex noctilio were strongly aggregated, with a few of the occupied trees hosting hundreds of woodwasps. Nematode parasitism was positively associated with S. noctilio density, and negatively associated with the density of rhyssine parasitoids. Parasitism by the parasitoid Ibalia leucospoides was positively associated with host (S. noctilio) density, while parasitism by the rhyssine parasitoids was negatively associated with density of S. noctilio. Thus, most S. noctilio come from a few attacked trees in a forest, and S. noctilio from those high-density trees experienced high parasitism by both the invasive nematode and the most abundant native parasitoid, I. l. ensiger. There is little evidence for direct competition between the nematodes and parasitoids. The negative association occurring between rhyssine parasitoids and I. l. ensiger suggests rhyssines may suffer from competition with I. l. ensiger which parasitize the host at an earlier life stage. In addition to direct competition with the native woodwasp Sirex nigricornis for suitable larval habitat within weakened trees, the large S. noctilio population increases the parasitoid and nematode populations, which may increase parasitism of S. nigricornis.

Aggregation, co-infection, competition, Deladenus, density dependence, forest pest, Ibalia, parasite community, pine, Rhyssa, Sirex noctilio, spillback

In any community there are multiple parasite species simultaneously associated with any host species (Shaw and Dobson 1995; Pedersen and Fenton 2007). Parasites share hosts by infecting different individuals in host populations and by co-infecting the same host individuals. Thus, interactions among parasites occur indirectly due to exploitative competition for shared hosts or directly through interaction outside the host (Ode et al. 2021), as well as within host individuals (van Nouhuys and Punju 2010; Harvey et al 2013). Because parasite species are each independently sensitive to host density, the strengths of their interactions with each other and with hosts can change with host density. The resulting interactions influence the population dynamics of both parasites and hosts, affecting the composition and structure of communities (Settle and Wilson 1990; Telfer et al. 2010; Holt and Bonsall 2017).

Invasive species may act as consumers, competitors, hosts to existing parasites, and as parasites of existing hosts. The introduced species may have a greater prevalence than in their native range (Callaway and Aschehoug 2000; Wolfe 2002), and their invasion can have a strong effect on the structure of native communities (Mitchell et al. 2006; Kelly et al. 2009; Kenis et al. 2009). The roles of multiple parasites, both native and introduced, in these new composite communities are understudied (Sarabeev et al. 2022) and complex (Dunn et al 2012; Llopis‐Belenguer et al. 2020). For instance, in the northeastern US, the invasive spongy moth, Lymantria dispar, interacts with two pathogens and numerous parasitoid species that have been accidentally and intentionally introduced (Fuester et al. 2014), resulting in many complex interspecific dynamics (Hajek and van Nouhuys 2016).

Larvae of the Eurasian woodwasp Sirex noctilio (Siricidae) develop in the xylem of pine trees in association with a symbiotic white rot fungus, Amylostereum areolatum, that assists in slowly killing attacked trees and acting as an external rumen for larvae (Hajek and Castrillo 2021; Gruner and Thompson 2021). In North America, this invasive woodwasp was first found to be established near Lake Ontario in New York state in 2004 and has been spreading since then (Liebhold and Hajek 2021). Sirex noctilio has previously invaded and caused severe damage in pine plantations in the Southern Hemisphere, where pines are not native (Hurley et al. 2007). In North America, where pines are native, S. noctilio often successfully develops within trees that are already suppressed or weakened (Dodds et al. 2010; Foelker and Haavik 2021). Native Sirex species in North America and their associated parasite communities were already present when S. noctilio arrived. In particular, in eastern North America, the native Sirex nigricornis and associated parasitoids and nematodes already infested weakened pine trees (Table 1). Since its introduction, S. noctilio has been found to develop within the same pines as S. nigricornis (e.g., Hajek et al. 2013, 2017) and is attacked by the same species of hymenopteran parasitoids (Foelker et al. 2016a, b) and parasitic nematodes (Morris et al. 2013; Haavik et al. 2016) (Table 1).

The community of parasitoids associated with S. noctilio and S. nigricornis in eastern North America is composed of up to six native hymenopteran species. Ibalia leucospoides ensiger (Ibaliidae) parasitizes eggs/early instars and is almost always the most common species. Four species of rhyssines (Ichneumonidae) parasitize later larval instars, and Pseudorhyssa nigricornis (also Ichneumonidae) is a kleptoparasitoid attacking the rhyssines (Foelker and Parry 2021) (Table 1). The biologies of these native parasitoid species are not well understood, although molecular analyses have shown that the rhyssines (analyzed as a group) and I. l. ensiger parasitize both S. nigricornis and S. noctilio (Foelker et al. 2016b).

The dimorphic nematode, Deladenus siricidicola, is also a parasite of S. noctilio. The impact of nematodes on host woodwasps is determined by both nematode and woodwasp species and genotypes (Bedding 1972). Different species and strains of Deladenus parasitizing differing species and strains of Sirex are known to either totally sterilize (i.e., kill all host eggs in adult females), partially sterilize, or not sterilize parasitized hosts (van Nouhuys et al. 2022). The INA (= Introduced to North America; formerly referred to as ‘non-sterilizing’) strain of D. siricidicola that was putatively accidentally introduced with S. noctilio to North America does not sterilize eggs of S. noctilio but causes a decrease in adult size and fecundity (Kroll et al. 2013; Hajek and Morris 2021). Deladenus siricidicola also occurs as a free-living mycophagous form that feeds on the Sirex-symbiotic fungus within infested pines. Several strains of S. noctilio have been introduced to North America from unknown sources and origins (Boissin et al. 2012; Bittner et al. 2017), and the INA strain of D. siricidicola is also from an unknown source, probably within the native distribution of S. noctilio (Morris et al. 2020). A native dimorphic nematode in pines, Deladenus proximus, described originally as a parasite of S. nigricornis (Bedding 1974), is never (Kroll et al. 2013; Haavik et al. 2016) or only rarely (Morris et al. 2013) recorded parasitizing S. noctilio.

Invasive S. noctilio in northeastern North America is therefore attacked by numerous species of native parasitoids as well as an introduced parasitic nematode (Table 1). This woodwasp species did not co-evolve with these parasitoids and it is questionable whether S. noctilio in any tree might have co-evolved with the nematode strain within that tree. Regardless, these natural enemies occupy the same community and rely on the same host as a resource, with the parasitoids killing woodwasps as they develop and nematodes living as internal parasites in S. noctilio larvae and adults.

Our overall goal in this study was to investigate relations among the parasitic nematode, the native parasitoid community, and the invasive host woodwasp in naturally occurring attacked trees. We evaluated emergence from 204 S. noctilio-infested trees from multiple forested areas in northeastern North America between 2011 and 2019. We tested the hypotheses that S. noctilio aggregate within a few of the infested trees in a stand, and that the natural enemies would each respond positively to host density. Further, since the natural enemies co-occur but cannot all successfully infect the same host individuals, there would be negative associations among them where their densities were high enough for competition to occur. Given the patterns of tree infestation and parasitism that we found, we discuss likely consequences to populations of the native woodwasp S. nigricornis from invasion of this community by S. noctilio. Results from this study will improve our understanding of population level interactions in mixed invasive and native communities, as well as the ecology of this host/parasite community, as the invasive host continues to expand its range in North America.

Between 2010 and 2018, 204 S. noctilio-infested pines were identified at sites where active infestations of S. noctilio were known or hypothesized to be present in pine forests at 21 sites in New York and Pennsylvania, USA (Table 2). Detection generally occurred from October to December. Nearly all of the pines were Pinus resinosa. Infested trees were patchily distributed within the forest stands. When searching for infested trees, trunks were inspected to 3 m for resin beads as these are indications of infestation (Haavik and Foelker 2021). In spring 2011–2019, before S. noctilio emergence, the trees that had been identified as infested were felled and areas of the trunks with resin beads were cut into 70 cm long bolts. Insects were reared from this wood and collected as they emerged, as described in Hajek et al. (2017). Throughout, the years presented for samples are the years in which emergence occurred.

Sirex and parasitoid adults that emerged were kept individually in vials at 4 °C. Sirex noctilio and S. nigricornis were identified using Schiff et al. (2006, 2012) and associated parasitoids were identified using Standley et al. (2012). In 2011, we were unable to distinguish between some rhyssines, and the rhyssines emerging in 2012 were not all identified, so these are listed as Rhyssa spp. (Suppl. material 1).

To evaluate nematode parasitism, subsamples of S. noctilio from each site and year were dissected as described in van Nouhuys et al. (2022) and presence of parasitic nematodes within bodies was recorded. A total of 3154 randomly chosen S. noctilio were dissected to detect whether they were parasitized by nematodes. Species of nematodes were not evaluated using molecular methods for this study but, based on past results (Kroll et al. 2013; Williams and Hajek 2017), we presume that the vast majority of parasitic nematodes were D. siricidicola INA. Another nematode species, Deladenus proximus, is primarily known from S. nigricornis hosts. Deladenus proximus was reported parasitizing a few S. noctilio collected in New York state in 2010 (Morris et al. 2013). However, D. proximus are usually found in the host eggs (Hartshorn 2021; van Nouhuys et al. 2022), and in our study nematodes were only seen within eggs from 4 S. noctilio from 2 sites in 2011. This was the first year of this study and there S. nigricornis were more abundant (Suppl. material 1).

Between 2011 and 2019, S. noctilio and their parasitoids were reared from 204 S. noctilio-infested trees from Pennsylvania and New York State (Table 2). There were 9852 adult woodwasps or parasitoids that emerged over the eight years of sampling. Of these, 3870 were parasitoids, 5804 were S. noctilio, and 178 were the native woodwasp S. nigricornis (Suppl. material 1). Most trees (130 of 204, 63.7%) were infested with 1–25 Sirex (adult Sirex + their parasitoids). However, 45.7% of the population occupied 12 of 204 trees. These trees each contained more than 198 S. noctilio, with 854 individuals emerging from the tree with the highest occupancy (Fig. 1). Thus, the distribution of Sirex equivalents (adult Sirex + their parasitoids) among trees was strongly aggregated, differing significantly from the expected Poisson distribution that would occur if they were distributed randomly among infested trees. This is true for the combined data (χ² = 1700.34; P < 0.0001) (Fig. 1), as well as in each of 12 sites analyzed separately (χ² ranged from ≈ 0.00 to χ² = 121.34, all are P < 0.0001). The distributions of individuals among trees in the six sites with the highest densities are presented in Fig. 2.

Figure 1.

The distribution of Sirex equivalents (Sirex adults + parasitoid adults) emerging from trees in the whole data set. The inset shows the distribution of Sirex equivalents among the same number of trees if the Sirex were distributed randomly among trees (a Poisson distribution). The * indicates the mean number of wasps per tree (48.3). The number of wasps along the x-axis is shown in a modified log scale in order to make both small and large numbers of wasps per tree visible.

Figure 2.

The distribution of Sirex equivalents (Sirex adults + parasitoid adults) emerging from the 6 sites with the most data (each had 398 or more Sirex equivalents and nine or more trees). The wasps were strongly aggregated in these and the other 6 sites tested (see text). * denotes the mean number of Sirex equivalents per tree. The horizontal gray bars represent the 95% CI range of wasps per tree if the Sirex were distributed randomly among the trees, which would be a Poisson distribution (as shown in the inset in Fig. 1). The number of wasps along the x-axis is shown in a modified log scale to make both small and large numbers of wasps visible. a) 2017 Hillsdale site, 2052 Sirex equivalents from 10 trees, b) 2013 Hypocrite Trail site, 400 Sirex equivalents from 12 trees, c) 2011 Government Rd. site, 1058 Sirex equivalents from 9 trees, D) 2012 Government Rd. site, 575 Sirex equivalents from 14 trees, e) 2014 Hypocrite Trail site, 2055 Sirex equivalents from 39 trees, f) 2012 Hypocrite Trail site, 398 Sirex equivalents from 18 trees.

Overall, 39% of the Sirex were parasitized by parasitoids. The most abundant parasitoid species was I. l. ensiger, which comprised 81% of the parasitoids reared and was found at every site. Next most abundant was Rhyssa lineolata, which made up 63% of the rhyssines (leaving out samples from 2012 when rhyssine species were not separated, Suppl. material 1) and was also found at every site, although densities were always less than I. l. ensiger. The other four rhyssines (two Rhyssa species: R. persuasoria and R. crevieri, Megarhyssa nortoni, and the kleptoparasitoid Pseudorhyssa nigricornis) occurred much less commonly. The 8 collected M. nortoni came from three sites in the years 2011 and 2013. Most of the 13 P. nigricornis came from Hypocrite Trail in Tioga State Forest (Pennsylvania), in two different years (Suppl. material 1).

Nematodes were found in 11% of the dissected S. noctilio. The proportion of nematode-parasitized hosts in a tree was positively associated with host density (χ² = 4.92; P = 0.0265; Table 3a; Fig. 3a). There was no association with parasitism by I. l. ensiger (χ² = 0.004; P = 0.9460; Table 3a), but a negative association occurred with parasitism by the rhyssine parasitoids (χ² = 31.57; P < 0.0001; Table 3a; Fig. 3b). This was especially true at low host densities (χ²_{rhyssines x host density} = 21.67; P = 0.0001; Table 3a). Nematode parasitism also differed between years (χ² = 166.04; P < 0.0001; Table 3a), as well as among sites within years (χ² =329.13; P < 0.0001; Table 3a). For instance, only 7% of hosts dissected from the Corsica and Hillsdale, PA sites sampled in 2017 were parasitized by nematodes, whereas 82% of hosts dissected at the Tanglewood, PA site in 2015 were parasitized by nematodes (Fig. 4).

Figure 3.

The association of parasitism by the nematode Deladenus siricidicola with a per tree Sirex density (Logistic regression model coefficient P < 0.026; Table 3a) and b fraction of Sirex in the tree parasitized by rhyssines (Logistic regression model coefficient P < 0.0001; Table 3c). Each point represents a dissected S. noctilio. The points are distributed vertically in order to show the amount of data at each host density. The points below the curves are parasitized D. siricidicola (Yes) and the points above the curve are not (No). See the text for explanation of the statistical models.

Figure 4.

The proportions of dissected S. noctilio parasitized by D. siricidicola at each site between 2011 and 2019. Sites with fewer than 10 dissected S. noctilio are excluded as well as all sites from 2012.

Overall parasitism by parasitoids was unrelated to tree-level host density (χ² = 0.003; P = 0.9522; Table 3b). However, parasitism by I. l. ensiger increased with host density (χ² = 31.89; P < 0.0001; Table 3c; Fig. 5a), while parasitism by the rhyssines decreased with host density (χ² = 12.60; P = 0.0004; Table 3d; Fig. 5b). Parasitism by I. l. ensiger had no overall association with rate of parasitism by the rhyssines (χ² = 3.04; P = 0.0813; Table 3c), except at low host densities where rhyssine parasitism was high (χ²_{rhyssines × ln(host)} = 62.39; P = < 0.0001; Table 3c). Similarly, parasitism by the rhyssines also had no overall association with rate of parasitism by I. l. ensiger (χ² = 3.00; P = 0.0832; Table 3d), except at low host density (χ² _{I. l. ensiger × ln(host)} = 82.44; P = < 0.0001; Table 3d). Parasitism by I. l. ensiger, rhyssines, and total parasitism each differed significantly among years and among sites (Table 3b, c, d).

Figure 5.

The association of parasitism by parasitoids with tree level host density. The curves show the logistic fit for a I. l. ensiger (P < 0.0001; Table 3c) and b rhyssines (P = 0.0004; Table 3d). The horizontal axis shows host density as the log of the number of Sirex equivalents (adult Sirex + all parasitoids) emerging from a tree. Each data point is one sample. The points are distributed vertically in order to show the amount of data at each host density. The points below the curve are parasitoids (Yes) a I. l. ensiger and b rhyssines, and the points above the curve are unparasitized Sirex (No). See text for explanation of the statistical models.

The invading S. noctilio exhibited very strong aggregation in specific trees within a stand, with up to 854 Sirex equivalents emerging from one tree and few S. noctilio emerging from the majority of infested trees (Figs 1, 2). In North America, Krivak-Tetley et al. (2022) and Haavik et al. (2018) reported a similar pattern of higher densities in few trees with lower densities in most. In Galicia, Spain, where S. noctilio is native, the pattern of aggregated emergence also occurred in pine plantations although the overall population was lower and the maximum emerging from one tree was 50 S. noctilio (Lombardero et al. 2016).

Generally, aggregated resource use can result from variation of resource quality (Woodcock et al. 2002), host susceptibility (Poulin 2013), or behavioral attraction to conspecifics (Cronin and Strong 1999). In this case, woodwasps are often attracted to weakened pines that are more susceptible to attack (Haavik and Foelker 2021); the occurrence of weakened trees can be a limiting and ephemeral resource for woodwasps. Additionally, S. noctilio females are attracted to the odor of the Sirex-associated fungal symbiont (Sarvary et al. 2015; Faal et al. 2021). Thus, S. noctilio may be strongly attracted to trees with many Sirex because those trees would already have been colonized by the Sirex-symbiotic fungus.

The nematode D. siricidicola was present in 11% of the S. noctilio which is less than reports from many smaller studies in North America (see Krivak-Tetley et al. 2022), though nematode parasitism levels were variable and were much higher in some sites (Fig. 4). The fraction of samples in a tree that were parasitized by D. siricidicola increased strongly with host density (Fig. 3a). This is in contrast to the negative association found by Kroll et al. (2013) when studying fewer trees (n = 25) over one year. Positive host density dependence could result from a higher chance of nematode-infected females ovipositing in the same trees as healthy females at high density. Additionally, spread of the nematode among woodwasps within trees would be facilitated by high host densities. One reason for this is that high woodwasp density would result in high densities of the fungus A. areolatum, which is eaten by these nematodes when they are mycophagous. More A. areolatum throughout trees would potentially facilitate greater dispersal by the mycophagous-phase nematodes within trees, which would result in more nematodes reaching and parasitizing Sirex larvae.

Nematodes may parasitize a woodwasp before or after it has been parasitized by a wasp. Once within a shared host, nematodes could, in principle, move from the woodwasp larva to a parasitoid larva. However, the nematode D. siricidicola is specific to Sirex as well as parasitizing a wood-dwelling genus of beetle associated with siricids (Serropalpus) (Bedding and Akhurst 1978; Bittner et al. 2016). Other species of parasitic nematodes have been reported from rhyssines in other locales (Bedding and Akhurst 1978). However, Morris et al. (2020) evaluated 388 adult rhyssines emerging from pines between 2012–2015 in the same research area that we studied and found no nematodes in them. Thus, we do not believe that D. siricidicola in this study were parasitizing the parasitoids.

Further, the nematode D. siricidicola cannot successfully develop in or be vectored by a Sirex parasitized by a parasitoid because the host is killed by the parasitoid. While the nematodes would not succeed in such a shared host, parasitoids likely can develop in Sirex larvae that are also parasitized by nematodes. There may be a cost to parasitoids due to reduced host size since nematode-parasitized Sirex are somewhat smaller than unparasitized individuals (Hajek and Morris 2021). Additionally, nematode infection has been shown in some systems to increase host immune response, including parasite encapsulation (Castillo et al. 2011), which could reduce the success of parasitoids. Alternatively, immune suppression by the nematode combined with immune suppression due to the parasitoid venom could be detrimental to the host, causing it to die or to have greater susceptibility to other parasites (Shaik et al. 2020). Nonetheless, we expect that when the nematode D. siricidicola and a parasitoid are using the same host individual, the parasitoid would be the superior competitor. We found that parasitism by nematodes was neither positively nor negatively associated with parasitism by the abundant parasitoid I. l. ensiger (Fig. 3b), but it was negatively associated with parasitism by the rhyssines (Fig. 3c).

The observed negative association of nematode parasitism with rhyssines may simply be because the nematodes are strongly positively host density dependent, so they are nearly absent from trees with low host density, which are the ones in which parasitism by rhyssines was highest. Alternatively, competition between nematodes and rhyssines at low host densities may decrease the number of nematode-infected Sirex, contributing to the pattern of host density dependence of the nematodes. While such competition may contribute to the pattern, it is not the full explanation because at low host density nematode parasitism is near 0 (Fig. 3a), and rhyssine parasitism reaches only 24% (Fig. 5b).

Both the nematode and the abundant parasitoid I. l. ensiger increase with host density within trees. Yet there is no evidence of direct competition, since even accounting for host density, high I. l. ensiger density was not associated with lower prevalence of nematodes. This suggests that there could be some spatial segregation of host use. For example, I. l. ensiger, which have short ovipositors, have been found to be most abundant high up in trees where bark is thinner (Foelker et al. 2016a; Long et al. 2009). Studies in South Africa found that the wood at the bases of pines was moister than that higher within the trunks, and D. siricidicola densities were also greater at the bases of trees; numerous studies have suggested that wood moisture content can influence activity of both D. siricidicola and A. areolatum (see Hurley et al. 2008). If the nematodes are less abundant higher in trees due to such an environmental variable, then direct competition with I. l. ensiger would be reduced. Alternatively, nematode mortality due to competition with the parasitoid could be masked by very strong host density dependence by the nematodes.

The native community of parasitoids that were present in the northeastern US before invasion by S. noctilio readily use S. noctilio as a host. On average, 37% of the Sirex from a tree was parasitized, ranging from 0 to 100%. This overall parasitism is similar to what was found by Krivak-Tetley et al. (2022) in a smaller study in this region, and by Haavik et al. (2015) using experimentally-exposed logs. We found the frequency of parasitism by I. l. ensiger increased with host density, and the frequency of parasitism by rhyssines decreased with increasing host density, leading to no association of overall parasitism with host density (Table 3). Different parasitoid species often respond differently to host density. Some species respond to high host density by aggregating, which leads to a positive dependent rate of parasitism (Walde and Murdoch 1988; Ives 1992). Indeed, I. l. ensiger has been shown to be attracted to the fungus associated with S. noctilio (Martinez et al. 2006; Faal et al. 2021) and to increase residence time on trees at increased S. noctilio density (Corley et al. 2010). However, in this complicated system, Foelker et al. (2016a) found no association of rate of parasitism by I. l. ensiger with host density while Krivak-Tetley et al. (2022) found a host density association for I. l. ensiger in Pinus resinosa, the pine predominantly sampled in this study, but not in Pinus sylvestris. Such differences between studies are not surprising where there is density dependence, especially when studies are conducted at different host densities. Our study was conducted over a wide range of host densities.

The rhyssine species showed higher parasitism at low host density. If parasitoids do not respond to host density the fraction parasitized can appear higher at very low host density than at high host density, simply reflecting the population size of foraging females (Walde and Murdoch 1988). This same pattern could also be due to competition between the rhyssines and I. l. ensiger. The latter species parasitizes Sirex eggs or young larvae (Foelker and Parry 2021) which makes it likely to be the superior competitor, since later-arriving endoparasitoids rarely are able to kill the parasitoid already developing in a host (van Nouhuys and Punju 2010; Cusumano et al. 2016). We only found a significant negative association between the parasitoid wasps at low host densities, where the fraction of hosts parasitized by rhyssines was highest. At higher host densities the frequency of parasitism by rhyssines was low, so the chance of both species parasitizing the same host individuals was unlikely, especially if there is some spatial segregation of parasitoid species within trees (Long et al. 2009; Foelker et al. 2016a).

As a caveat, the interactions between parasitoids and S. noctilio is complicated to evaluate due to variable voltinism of both rhyssines and S. noctilio (Foelker et al. 2016a). In particular, rhyssines have been thought to oviposit on late instar siricids in spring but then only emerge the following spring (Foelker and Parry 2021). However, it has also been suggested that two generations of rhyssines could possibly occur per year (Foelker and Parry 2021), which is supported by findings in this study, in which rhyssines emerged from the same wood in the same year as S. noctilio. In addition, there is the complication that S. noctilio has been found to oviposit in the same trees over multiple years (Haavik et al. 2018) and, once developing within a tree, some of the S. noctilio require more than one year for development (Ryan et al. 2012; Myers et al. 2014; Hajek et al. 2017).

In this study the native woodwasp, S. nigricornis occurred at much lower densities than S. noctilio, with the majority emerging early in the study, between 2011 and 2014 (Suppl. material 1); this supports the hypothesis that S. noctilio is the more successful competitor (Hajek et al. 2017). Of the 26 trees from which S. nigricornis emerged, S. noctilio also emerged from 88.5% of them. The local high densities of S. noctilio can impact the native woodwasp S. nigricornis in several ways. The large populations of S. noctilio could directly compete with S. nigricornis for trees in which to oviposit, especially because S. nigricornis is considered a less aggressive colonizer (i.e., needing weaker trees than S. noctilio) and most oviposit after S. noctilio (Hartshorn 2021). Secondly, when both woodwasps occupy the same tree, which creates higher woodwasp densities, parasitism by the nematode and I. l. ensiger, which are positively host density dependent, would be higher than if S. nigricornis occurred alone (van Nouhuys and Kraft 2012). Finally, high S. noctilio population densities can drive up the shared parasitoid and nematode populations in a forest, which could cause S. nigricornis to decline, an example of spillback by an invasive species (Kelly et al. 2009). Such apparent competition, driven by shared natural enemies, can have a strong impact on insect communities (van Nouhuys and Hanski 2000; Holt and Bonsall 2017).

In summary, this study of invasive S. noctilio and its introduced and native natural enemies emerging from 204 naturally infested trees over eight years, allowed for a robust analysis of the distribution of S. noctilio among trees, and for an exploration of the complex association of host and natural enemies over a large range of natural host densities. We found a strong pattern of aggregation by the woodwasp in a subset of the trees it occupied. Parasitism by the nematode D. siricidicola increased with host density, making its distribution also aggregated. Parasitism by the native woodwasp parasitoid I. l. ensiger was high and also increased with host density, suggesting that this wasp may contribute to the control of the woodwasp. Additionally, though both the nematode and I. l. ensiger were positively host density dependent, there was little evidence of direct competition between the parasitoid and the nematode. The suite of native rhyssine parasitoids of Sirex were less abundant and occurred mostly where host densities were low. While our study indicates that they suffer from competition with I. l. ensiger, their population level relationship with the host and the other parasitoids may be complicated by a variation of voltinism.

We thank Brad Regester, Bill Laubscher, and Sarah Johnson of the Pennsylvania Division of Forest Health for assistance with searching for and felling trees. We thank Cornell’s Arnot Forest and Christopher Foelker, Dylan Parry, and Melissa Fierke of SUNY, ESF for assistance with finding and sampling infested pines. Stefan Long, Stefanie Kroll, Jake Henry, and Chad Keyser as well as many additional people in the Hajek Lab assisted with emergence from wood, identification of parasitoids, and detection of nematodes in Sirex. SvN thanks The Israeli Institute for Advanced Studies for support during writing.