Classical or introduction biological control is the intentional importation, release, and establishment of natural enemies for suppressing damaging populations of invasive non-native organisms to densities that no longer cause economic or ecological harm. This approach aims to reduce pest population densities by re-associating safe (i.e., host-specific) and efficacious natural enemies with the target pest (1). Host range and host specificity testing are important primary steps in identifying natural enemy species that may have deleterious impacts on pest populations while presenting minimal risk to non-target species (1). Host use evaluation studies are mandatory in the United States of America (U.S.) and provide safety data for review by Federal agencies (i.e., United States Department of Agriculture, Animal and Plant Health Inspection Service [USDA APHIS]), that regulate the importation and release of natural enemies for use in classical biological control programs (2). Host range and host specificity testing evaluations are time consuming, often taking years to complete (3). During this time, newly established invasive pest populations tend to increase in density and spread as management plans are slowly developed and implemented. Proactive biological control attempts to reduce or eliminate this window of opportunity for an invasive pest by evaluating candidate natural enemies for potential use in a classical biological control program in advance of the anticipated incursion and establishment of the target pest in the area of concern (3).

Spotted lanternfly, Lycorma delicatula (White) (Hemiptera: Fulgoridae), native to China (4), was detected for the first time in the U.S. in Berks County Pennsylvania, in September 2014 (5). By November 2022, L. delicatula infestations were confirmed in an additional 13 states in eastern and mid-western areas of the U.S. (Connecticut, Delaware, Indiana, Maryland, Massachusetts, Michigan, New Jersey, New York, North Carolina, Ohio, Rhode Island, Virginia, and West Virginia) (6). Lycorma delicatula is a phloem-feeding fulgorid that has a broad host range (7). Direct feeding damage can cause mortality to highly preferred hosts like Ailanthus altissima (Miller) (Sapindales: Simaroubaceae) and grapevines (Vitis vinifera L. [Vitales: Vitaceae]). Indirect damage results from the excretion of high quantities of honeydew that promote sooty mold growth (7, 8). Lycorma delicatula has been recorded infesting forest and ornamental shade tree species in natural and urban areas, respectively (7, 9), and presents a significant risk to economically important perennial agricultural crops (e.g., grapes and nuts) (10, 11). Long-distance dispersal by L. delicatula is almost entirely human-assisted. This occurs primarily through the accidental translocation of egg masses that are often laid indiscriminately on inert substrates (e.g., wooden pallets and railcars) that undergo subsequent transportation into uninfested areas (12, 13). This type of inadvertent relocation in the U.S. likely resulted in the establishment of invasion bridgeheads in Indiana (2021) and Michigan and North Carolina (2022) (6). Ecological niche models indicate that L. delicatula has a potential distribution that includes large areas of the west coast of the U.S., and other parts of the world (e.g., Europe) (14). For California, a western U.S. state, with an agricultural economy worth ~$50 billion per year (15), L. delicatula is viewed as a significant invasion threat that could cause major problems for producers of specialty crops, like grapes and nuts, which are multi-billion-dollar industries (15).

The egg parasitoid, Anastatus orientalis Yang and Choi (Hymenoptera: Eupelmidae), was found parasitizing L. delicatula eggs in northern China in 2011during foreign exploration for natural enemies for use in South Korea, where L. delicatula is also invasive (16, 17). Following the invasion and spread of L. delicatula on the east coast of the U.S. there was renewed interest in the potential use of A. orientalis as a classical biological control agent (18, 19). Molecular analyses identified six different haplotypes of A. orientalis collected from the native range. Importations of A. orientalis into the U.S. were initially comprised of Haplotype C, which was first evaluated as a classical biological control agent against L. delicatula (20). The majority of Anastatus spp. Motschulsky are primary endoparasitoids attacking eggs of Diptera, Dictyoptera, Coleoptera, Hemiptera, Lepidoptera, Orthoptera, and Mantodea (21–27). Numerous Anastatus species have been considered or used as classical biological control agents against various pests around the world. For example, A. japonicus Ashmead was released in the eastern U.S. for control of Lymantria dispar L. (Lepidoptera: Erebidae) (27, 28) and against Tessaratoma papillosa Drury (Hemiptera: Pentatomidae) in China (29, 30). Anastatus sp. was released in Nepal to control Rhynchocoris humeralis (Thunberg) (Hemiptera: Pentatomidae) (31). Anastatus bifasciatus (Geoffroy) was evaluated to study levels of biotic resistance of central European natural enemies against invasive Halyomorpha halys Stål (Hemiptera: Pentatomidae) populations (32, 33). In the eastern U.S., A. reduvii, a native species, has been detected as one of the most common parasitoids emerging from eggs of invasive H. halys (34). Some Anastatus spp. are commercially-available and used for augmentative biological control of Amblypelta nitida Stål and A. lutescens lutescens Distant (Hemiptera: Coreidae) in Australia (35, 36).

Given the obvious threat posed to California agriculture by the westward migration of L. delicatula in the U.S., a proactive biological control program was initiated. Proactive research efforts focused on the suitability of A. orientalis (Haplotype C) as a potential classical biological control agent of L. delicatula in advance of its expected establishment in California (3). Consequently, the objective of this study was to investigate the physiological host range of A. orientalis on native and non-native non-target species from the southwestern U.S. (i.e., California and Arizona) to determine whether or not this natural enemy would be a suitable candidate to release for classical biological control of L. delicatula should it eventually establish in California. The results of these studies are presented here.

This is the first study to assess the physiological host range and specificity of A. orientalis (Haplotype C) with respect to potential non-target species that occur in the southwestern U.S. In the quarantine laboratory, no-choice tests assessing the developmental suitability of non-target host species for A. orientalis demonstrated that 15 out of 34 non-target host species were suitable hosts for A. orientalis. This work indicated H. halys eggs could successfully support development of A. orientalis (offspring sex ratio was strongly male biased), a finding contrary to previous reports (43). Antheraea sp. (Saturniidae) eggs (i.e., A. pernyi) have been reported to be suitable reproductive hosts for A. orientalis (43). However, parasitization of Antheraea oculea Neumoegen eggs by A. orientalis was not recorded in this study. Five new hemipteran hosts in three families (Coreidae [A. thomasi {native}], Fulgoridae [P. fuliginosa {native}], and Pentatomidae [C. hilaris {native}, H. halys {invasive} N. viridula {invasive}]) and ten lepidopteran hosts in four families (Erebidae [P. edwardsii {native}], Lasiocampidae [G. arizonensis {native}], Saturniidae [A. luna {non-native}, A. anona {native}, A. cecrops pamina {native}, E. calleta {native}, H. euryalus {native}, R. cincta {native}, S. walterorum {native}] and Sphingidae [P. occidentalis {native}]) were identified as suitable hosts and are added to an increasing list of identified species that A. orientalis can successfully use as reproductive hosts in the laboratory ( Table 2 ).

In terms of percent parasitism, eggs of two non-target species, C. hilaris (Pentatomidae [native]) and A. anona (Saturniidae [native]), were similar to or better hosts, than the target, L. delicatula. In the first case, C. hilaris, percent parasitism (68.79% ± 12.41; sex ratio 0.22 ± 0.1) was similar to the target, L. delicatula (58.24 ± 10.09; sex ratio 0.77 ± 0.07). This result may need to be interpreted cautiously, because C. hilaris eggs masses had on average a smaller number of eggs (17.73 ± 2.51) when compared to the average number of eggs per egg mass (39.08 ± 3.84) from the target, L. delicatula. In the second case, A. anona, percent parasitism was higher (76.48 ± 14.10; sex ratio 0) than the target L. delicatula (58.24 ± 10.09), and both egg masses were approximately equal in size (i.e., 42.2 ± 7.43 eggs and 39.08 ± 3.84 eggs for A. anona and L. delicatula, respectively). Collectively, data reported here and findings from companion studies (i.e., Broadley et al. [USDA], Submitted) suggest that A. orientalis, is at a minimum oligophagous, but more likely to be a polyphagous species.

Similar results from host specificity tests from other Anastatus spp. have been reported further supporting findings that Anastatus spp. potentially have broad host ranges. Stahl et al. (32, 33) studied the physiological host range of A. bifasciatus (this species is native to Europe) for use as a biological control agent against H. halys, an invasive pest in Europe. In this study, A. bifasciatus successfully parasitized eggs of eight pentatomid species (including N. viridula [tested in this study]) and 14 lepidopteran species belonging to seven different families (Endromidae, Erebidae, Lasiocampidae, Notodontidae, Papilionidae, Saturniidae, and Sphingidae). Results reported here indicate that A. orientalis can also parasitize hosts eggs from Erebidae, Lasiocampidae, Saturniidae and Sphingidae. Host range tests that expose A. orientalis females to eggs of Notodontidae and Papilionidae, two lepidopteran families with species representation in the southwest U.S., may be warranted to further understanding of potential non-target host use (the family Endromidae is not present in North America). For Anastatus spp., results reported here and those of Stahl et al. (32, 33), indicate strongly that selecting non-target species which are restricted to close taxonomic relatedness to the target pest (e.g., families in the Fulgoroidea [Hemiptera: Auchenorrhyncha]) may be inadequate as Anastatus spp. may tend to be generalists capable of parasitizing species across different orders.

No additional significant levels of excess mortality (i.e., mortality from causes other than parasitism) to non-target species exposed to A. orientalis was observed ( Figure 3 ). Failure to detect excess mortality of non-target eggs exposed to A. orientalis when compared to levels of naturally-occurring mortality in control eggs not exposed to A. orientalis may have at least two explanations: i) parasitized non-target eggs develop successfully when parasitoid larvae and/or envenomation failed to kill the host egg (host eggs are incapable of encapsulating parasitoid eggs and larvae), or ii) A. orientalis females only host fed on eggs that were parasitized and it is possible that A. orientalis is a concurrent parasitoid (i.e., host feeds on parasitized eggs). Thus, egg mortality from host feeding alone was not observed and egg mortality was attributed to solely to parasitism. Additional studies confirming the lack of excess mortality due to unsuccessful parasitization or host feeding following exposure of non-target host species to A. orientalis are needed. Excess mortality of non-target host eggs following exposure to A. orientalis should be considered an important deleterious non-target impact if it occurs (44).

For three non-target host species tested, A. oslari (Saturniidae [native]), C. vanduzeei (Fulgoridae [native]) and C. vittiger (Coreidae [native]), all parasitoid females exposed to egg masses of these species failed to parasitize L. delicatula egg masses in sequential exposure tests. For C. vanduzeei and C. vittiger, too few non-target eggs were available for experiments and replication was low for each test species and therefore were not excluded from the results. Interestingly, for female A. orientalis exposed to A. oslari eggs excess egg mortality increased slightly but not significantly, and all females (n=16) died following exposure to A. oslari eggs prior to sequential exposure to L. delicatula egg masses. These results, a slight but non-significant increase in excess egg mortality and premature mortality of females, suggest that A. orientalis females may have host fed on A. oslari eggs and egg contents were possibly toxic to parasitoids. This could be explained by chemical protection of eggs by secondary plant compounds, like tannins, which are used as chemical defenses by host plants (i.e., Quercus spp.) of A. oslari (45). Sequestration of protective compounds, like tannins, in A. oslari eggs could reduce survivorship rates of third trophic level organisms like A. orientalis (46).

To evaluate a more realistic exposure time of A. orientalis to non-target species, 24 hour exposure tests (as opposed to a 7 day exposure period) were performed with five non-target species (C. hilaris, G. arizonensis, H. halys, N. viridula and S. walterorum). Anastatus orientalis was able to parasitize the five non-target species in this shorter exposure period following the seven-day pre-oviposition exposure period on L. delicatula eggs. These findings suggest that prior host exposure, especially to the target, L. delicatula, does not deter use of subsequent non-target host eggs. In addition, when non-target and target host eggs were exposed to A. orientalis females at the same time (i.e., choice experiments), that non-target species were attacked under both exposure time scenarios (i.e., 24 hours and 7 days). These results further suggest that A. orientalis is probably a generalist parasitoid capable of using any suitable non-target host species upon encounter. In many instances, female parasitoids engaged in parasitism within minutes of introduction into test arenas and contact with non-target eggs (pers. obs. F. Gomez Marco).

The age of non-target host eggs can affect the acceptance behavior of A. orientalis females and rates of successful parasitism. For example, percent parasitism by A. orientalis was higher on young eggs (<24 hours of age) vs. older eggs (>24 hours to 3 days) of H. halys. However, for two other non-target species, G. arizonensis and N. viridula, tested in egg age acceptance studies, no significant effect of egg age on parasitism was found. Therefore, age of non-target eggs and the non-target species may be an important covariables to consider when host tests are being designed and executed. Additionally, defensive chemical compounds (see above) may also affect the acceptance behavior of parasitoids (and survivorship rates) and this may also warrant consideration in design, analysis, and interpretation of host range tests (46).

The sex ratio of A. orientalis offspring was strongly dependent on host egg size. All non-target host eggs which were parasitized by A. orientalis with an average egg size visibly (not measured in this study) smaller than L. delicatula produced a male biased sex ratio (< 0.5) and smaller adult males and females. However, the sex ratio from hosts which produced smaller A. orientalis males decreased, indicating that there was a strong correlation between offspring size and offspring sex ratio. Only one non-target host with eggs larger than L. delicatula eggs (pers. obs. F. Gomez Marco) was used in our tests (A. thomasi) and this resulted in the largest female parasitoid (hind tibia size of 1.04 mm) obtained from host-range studies. These findings tentatively support conditional sex allocation theory, where, with decreasing host quality (i.e., host egg size in this study), parasitoid offspring sex ratio becomes more male biased and males are generally smaller, both of which correlate with decreased fitness (47). In support of results presented here, two previous studies, Hou et al. (48) and Stahl et al. (32), reported more male-biased sex ratios when Anastatus spp. were reared on host eggs that were smaller than the target host. Interestingly, differences in A. orientalis offspring sex ratio from two non-target host species, C. hilaris and H. halys, existed when egg masses were exposed for 24 hours (C. hilaris and H. halys: 0, no females) or seven days (C. hilaris: 0.22 ± 0.1, H. halys: 0.05 ± 0.02). This finding might indicate a preference of A. orientalis females to first oviposit (within at least the first 24 hours) non-fertilized eggs (i.e., produce male offspring) and then oviposit fertilized eggs (i.e., produce female offspring) when exposure times are longer and there is more time to repeatedly assess host quality. Sex ratio of offspring has important demographic implications as it affects rates of population grown. Male-biased progeny production on non-target hosts of marginal quality may limit or negate adverse population-level impacts on non-target species (32).

Anastatus orientalis offspring reared from non-target host species successfully parasitized L. delicatula egg masses. Percent parasitism and sex ratios resulting from mated couples that emerged from the same non-target host species were similar to values recorded for A. orientalis reared continuously on L. delicatula. Interestingly, males that emerged from G. arizonensis and P. fuliginosa when paired with unmated A. orientalis females that emerged from L. delicatula failed to produce any offspring. Due to the low number of repetitions for these experiments this finding should be viewed with caution. However, no obvious biological explanation (i.e., size of the males or inability to copulate with females) was observed to explain these results as unmated A. orientalis females should be able to oviposit unfertilized eggs that produce male offspring without mating with males that emerged from G. arizonensis and P. fuliginosa.

In laboratory studies in a quarantine facility, sequential no-choice and choice exposure studies that exposed female A. orientalis to non-target eggs of 34 native and non-native species in three orders and 12 families and target eggs (L. delicatula) for either 24 hours or 7 days, indicate that this egg parasitoid potentially has a wide host range as it successfully parasitized 15 species in 6 families in two orders (Hemiptera and Lepidoptera). Results presented here are for A. orientalis (Haplotype C). Other haplotypes of A. orientalis have been identified and are being assessed to determine if differences (i.e., greater specificity) in host preferences exist (20). A well-founded criticism of host-range tests is that parasitoids are constrained in small ventilated containers spaces with easily accessible hosts for long periods of time (i.e., 24 hours to 7 days) and are unable to engage in behaviors (e.g., rapid abandonment of patches with sub-optimal hosts) that could reduce or eliminate non-target use (49–52). Consequently, in the absence of comprehensive field data on host use and non-target species - target species - parasitoid phenology in the native range (i.e., China) from where A. orientalis was sourced, it is difficult to determine if high levels of non-target host use observed in host range tests reported here occurs in the field. Until detailed field data from the native range are available and based on results of host range tests reported here that suggest A. orientalis [and possibly most species of Anastatus (23–27, 32, 33)] has a very broad host range, use of this natural enemy in classical biological control of L. delicatula in the western U.S. should be assessed critically and with extreme caution.

The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.

FG-M conceived and designed the experiments, performed collections and experiments, analyzed the data, prepared figures and/or tables, authored or reviewed drafts of the paper, and approved the final draft. DY performed collections, reviewed drafts of the paper, and approved the final draft. MR performed collections, reviewed drafts of the paper, and approved the final draft. MH obtained the funding, conceived and designed the experiments, performed collections authored or reviewed drafts of the paper, and approved the final draft. All authors contributed to the article and approved the submitted version.