The fall armyworm (FAW) Spodoptera frugiperda (J. E. Smith), is a noctuid moth pest made up of two morphologically indistinguishable strains (C- and R-strains) that is native to tropical and subtropical regions of the Americas. It is highly polyphagous and feeds on host species from at least 76 plant families, principally Poaceae (106 spp.), Asteraceae (31), and Fabaceae (31) (Montezano et al. 2018). In the Americas, it causes significant economic damage to maize, rice, sorghum, millet, soya bean, wheat, alfalfa, cotton, turf, and fodder crops (Nagoshi et al. 2019a). Recognized globally as one of the top 20 arthropod pests (Willis 2017), S. frugiperda was first officially confirmed outside of its native range in Western Africa in early 2016 (Goergen et al. 2016) and then detected across virtually all of Sub-Saharan Africa by February 2018 (Nagoshi et al. 2018, Otim et al. 2018, Botha et al. 2019). In July 2018, it was confirmed in Yemen and India, and by December 2018 in Bangladesh, Sri Lanka, Myanmar, and Thailand, followed closely by China in January 2019. Nepal confirmed its presence in May 2019, and in July 2019 it was also reported in Egypt, Indonesia, Laos, Cambodia, Malaysia, Vietnam, Taiwan, The Republic of Korea, and Japan (FAO 2021). The moth's strong flight ability (Rose et al. 1975, Sparks 1979, Jones et al. 2019, Xiao et al. 2020) its potential to contaminate certain commodities, and the ability to act as a ‘hitchhiker’ in trade (Early et al. 2018) contribute to its rapid spread. Whole genome analyses of invasive S. frugiperda populations from various African nations, India, China, Southeast Asia, and Australia also suggested that a high proportion of S. frugiperda individuals were hybrids of the C- and R-strains (Gui et al. 2020; Zhang et al. 2020; Schlum et al. 2021; Tay et al. 2021, 2022a; Rane et al. 2022).

Much of South and Southeast Asia, and much of Northern Australia, are highly suitable climatically for S. frugiperda year-round (du Plessis et al. 2018, Early et al. 2018). It was confirmed in Queensland (Qld; Bamaga Northern Peninsula Area), Australia following clearance of lure traps on 31 January 2020, followed by reports also in north-Western Australia (WA; Kununurra) and in the Northern Territory (NT) in March 2020. Within two weeks of the Bamaga detection, there were reports of S. frugiperda damaging maize fields in Strathmore (Qld; ca. 1,000 km from the initial Bamaga detection site), followed by successive southward detections of S. frugiperda in Qld, marking the pest's south-eastward antipodean expansion into New South Wales (NSW) in September 2020. Through CLIMEX simulation analysis based on irrigation patterns and rainfall records, S. frugiperda is expected to undertake seasonal migrations and may reach as far south as Victoria and Tasmania (du Plessis et al. 2018). Indeed, S. frugiperda was reported in Victoria in December 2020 and in Tasmania in April, 2021. Concurrently, and from the Western Australia Kununurra detection, S. frugiperda followed the CLIMEX simulated southward expansion patterns along WA's western coast, with larval populations detected in the Gingin area (<80 km north of the city of Perth) by February 2021.

The incursion of S. frugiperda into Australia is believed to have occurred through natural dispersal across the Torres Strait/Timor Sea into northern Australia via a single-entry point (Jing et al. 2021, Qi et al. 2021). However, incursion pathways to Australia could also involve multiple entry points as reported for the biting midges Culicoides bravitarsis (Tay et al. 2016), other anthropogenic routes such as infested commodities in trade, as well as through seasonal or on-going new arrivals of individuals from populations that have successfully established in Asia and Southeast Asia (SEA). The eastward spread of S. frugiperda across sub-Saharan Africa, the Middle East, the Indian subcontinent, SEA, China, and the Far East (South Korea, Japan) mostly followed chronological detection and confirmation which led to a widely accepted belief that an invasive population established in west Africa (Goergen et al. 2016) through a single founder, similar to the ‘invasive bridgehead effect’ (Guillemaud et al. 2011), was the starting point for its subsequent global expansion. This hypothesis is supported by an overwhelmingly homogeneous genetic signature based on single partial gene markers (e.g., Cock et al. 2017, Nagoshi et al. 2018).

Whole genome analysis and genome-wide single nucleotide polymorphic marker analysis of invasive populations from Africa (Benin, Uganda, Malawi, Tanzania) and Asia (India, China) instead suggest multiple founding events (Zhang et al. 2020; Schlum et al. 2021; Tay et al. 2021, 2022a); a similar conclusion reached also from analysis of multiple S. frugiperda populations from across China based on microsatellite DNA markers (Jiang et al. 2022). Recently, population genomic analysis of S. frugiperda populations from SEA (i.e., Philippines, Vietnam, Laos, Malaysia, Myanmar), from East Asia (i.e., South Korea), and Pacific/Oceania (i.e., Papua New Guinea; Australia) identified significant genetic diversity in these invasive S. frugiperda populations, and distinct population signatures between populations in close proximity [e.g., Kedah and Penang populations from Malaysia, and between Qld, WA and NT populations in Australia (Rane et al. 2022); between populations within e.g., Yunnan province (Tay et al. 2022a), and populations from Anhui and Jiangsu (Jiang et al. 2022)]. Such distinct population genomic structure contradicted the expected gene flow signatures of a single introduction and a west-to-east spread of this pest. Instead, they suggested relatively limited and localized spread of populations while also identified the likely multiple independent introduction pathways of S. frugiperda into the region including in SEA, Asia and Australia (Rane et al. 2022).

Distinguishing among single versus multiple introductions within Australia (and elsewhere), and understanding the resistance profiles of existing and any new incursions, is therefore critical for informing the future management of this pest. For example, this information will assist with forecasting likely resistance profiles of invasive populations selected outside of Australia, and could prioritize efforts to prevent the introgression (potentially of multiple separate introductions) of known insecticide resistance genes and alleles from endemic populations (e.g., Carvalho et al. 2013, Banerjee et al. 2017, Boaventura et al. 2020a, Guan et al. 2021) into invasive populations, or through reciprocal introgression of newly selected/developed resistance traits from invasive populations to native populations.

In this study, we report on the first bioassay experiments and resistance gene characterization that aimed to understand how the initial invasive populations of S. frugiperda in Australia responded to selected approved insecticide compounds and Bacillus thuringiensis (Bt) toxins in commercial transgenic plant varieties and available as sprays for the horticultural, grains and cotton industries. Since populations were presumed collected before selection occurred in Australia, these responses served as the first indication of selection against insecticides and Bt toxins used elsewhere [i.e., in the species' native range and in its recent invasive ranges from Southeast Asia, Asia, and/or Africa (Schlum et al. 2021, Jiang et al. 2022, Rane et al. 2022, Tay et al. 2022a)] before arrival. While sampling from multiple Australian regions would be ideal (but impossible at the time as our laboratory had only been provided with these two lines during the pandemic travel restriction), comparing the response differences between these two S. frugiperda populations from different northern Australia regions nevertheless could demonstrate the use of bioassay data to distinguish signatures of possible multiple points of entry compared with a single entry and spread event as generally assumed (e.g., Jing et al. 2021, Qi et al. 2021). To provide a first insight into the efficacy of key compounds and toxins we compared the responses of S. frugiperda with two related major endemic crop pests in Australia, Helicoverpa armigera (subspecies conferta; Hardwick 1965; Anderson et al. 2016; Pearce et al. 2017a,b; Zhang et al. 2022) and S. litura, which are currently managed through adaptive resistance management plans. Finally, we provided a critical review of resistance profiles in selected invasive populations of S. frugiperda globally to better understand its invasion biology, and consider our work alongside a subsequent more comprehensive examination in Australia of insecticide resistance profiles (Bird et al. 2022). Our study is the only available information for Australian populations on the vulnerability of new arrived S. frugiperda to relevant Bt toxins.

Material and Methods

Spodoptera frugiperda Specimens for Genomic Analyses

Populations of S. frugiperda from South Korea (SK), Papua New Guinea (PNG), and Peru were included for whole genome sequencing (Table 2). The methods for SK and PNG specimen preservation involved collection from fields and placing larvae directly into high concentrations of ethanol (95–99.9%) to transport to the laboratory where they were then stored at –20°C and replaced with fresh ethanol after 24–48 hr. Peru specimens were samples intercepted from Australian pre-border inspections of imported agricultural/horticultural commodities between 2016 and 2019 (see Tay et al. 2021, 2022a). Australian samples (Sf20-1 from Qld, Sf20-4 from WA) were F₀ individuals and represented individuals obtained directly from fields. Samples from SK, PNG, and Peru were sent to CSIRO and stored at –20°C until DNA extraction.

Table 1.

List of insecticides and Bt toxins used in the bioassay experiments

Table 2.

Spodoptera frugiperda samples from Australia, Papua New Guinea, South Korea, and Peru used in whole genome sequencing for resistance gene characterization

Results

Resistance Alleles by Whole Genome Sequencing

The loci examined for potential resistance alleles were present in the data set with the appropriate level of coverage to accurately call the genotype. Resistance alleles to carbamate/organophosphates (ACE-1) were the only resistance alleles identified in the Kununnura, Strathmore, South Korea, Papua New Guinea, and Peru populations. No resistance alleles associated with target site mutations were detected for pyrethroid (VGSC) or for diamide insecticides (RyR), although there are likely other genes (e.g., detoxification genes) that confer resistance to synthetic pyrethroid in these laboratory lines of S. frugiperda (Bird et al. 2022). Resistance allele profile differences between the two Australian populations were evident between the A201S and the F290V amino acid substitutions. Similar ACE-1 resistance allele profiles were detected between Strathmore and South Korean populations, while the Papua New Guinea population shared ACE-1 allele profiles with Strathmore population (A201S) and with Kununnura (F290V). The Peruvian S. frugiperda population was the only population that had heterozygous individuals with resistance allele for G227A amino acid substitution (Table 6).

Table 3.

Summary bioassay data on Spodoptera frugiperda populations from Queensland and Western Australia, S. litura, and Helicoverpa armigera involving surface treatment of the diet with Bt toxins and products

Table 4.

Summary bioassay data on Spodoptera frugiperda populations and Helicoverpa armigera (GR laboratory line) involving topical application to the insect with contact insecticides

Resistance allele characterization by sequencing approaches from this study and published studies for the ACE-1 gene is summarized in (Fig. 1, Supp  Table 2 [online only] (toac151_suppl_supplementary_materials.zip) and references therein). The most common resistance allele detected in the invasive and native population (456 individuals examined in total) was the F290V mutation with 66 homozygous and 222 heterozygous resistant genotypes detected. This mutation is a T to G single nucleotide polymorphism (SNP) that changes the codon encoding the amino acid from TTT to GTT leading to a phenylalanine (F) to valine (V) change in the protein sequence encoded by the ACE-1 gene. This mutation was present at all locations and evenly distributed between the invasive and native populations. In Australia, of the sequenced individuals (N = 146), 24.7% (N = 36) were heterozygous, and 8% (N = 12) were homozygous for the resistance allele. Heterozygous and homozygous individuals were found in populations from all four states (i.e., WA, NT, Queensland, and NSW) as well as from a single individual from Erub Island, suggesting it is common across Australia.

The second most common allele detected was A201S (456 individuals examined) with 90 heterozygous individuals detected. This mutation is a C to A SNP which changes the codon from GCG to TCG leading to an alanine (A) to serine (S) amino acid change. While less common, this mutation also appears to be in both the invasive and native range with no obvious pattern. In Australia, of the sequenced individuals (N = 146), 15.8% (N = 23) were heterozygous, and 0 were homozygous for the resistance allele. Heterozygous individuals were found in WA and Queensland, also suggesting it is common across Australia.

The G227A mutation was the least common (from 456 individuals) with 20 heterozygous individuals and 2 homozygous individuals detected. This mutation is a G to a C SNP which alters the codon from GGA to CGA encoding a glycine (G) to alanine (A) amino acid change. Interestingly G227A was only present in individuals from the native range (Brazil, USA, Puerto Rico, Peru) but absent in individuals from across the invasive range that was surveyed in this study and from the related studies of Tay et al. (2021, 2022a). When compiling allele frequencies for both the VGSC and ACE-1 resistant genes from this study and from published whole genome sequencing and targeted PCR/Sanger sequencing data (Fig. 1; see also Supp  Table 1 [online only] (toac151_suppl_supplementary_materials.zip)), the rare G227A resistance allele was present only in the Indonesian (Boaventura et al. 2020a) and Hubei populations (Guo et al. 2020) but absent in African, Australian, and other Asian (e.g., Indian, South Korea) populations, including populations from six other Chinese provinces.

No target site mutation alleles predicted to cause resistance to pyrethroids or the group 28 pesticides were detected in this work. However, while the previously identified resistance alleles (Bolzan et al. 2019, Boaventura et al. 2020b) were not detected, considerable variation was present in the RyR gene at the potential resistance loci. This should be further investigated in conjunction with bioassays to establish whether any of the variants could contribute to resistance.

ABCC2 Resistance Alleles in Australia, Papua New Guinea, South Korea, and Peru Populations

None of the known and validated Cry1 resistance ABCC2 mutations were identified in the individuals sequenced for this work however, some variation from the reference was observed (Supp  Fig. 13 [online only] (toac151_suppl_supplementary_materials.zip)). Numerous synonymous mutations and a total of 31 nonsynonymous mutations were observed in the coding sequence of the ABCC2 gene in all sequenced individuals. Several of these nonsynonymous mutations were associated with a common deletion and insertion motif in the first exon where an 11 bp deletion and a 2 bp insertion maintain the reading frame but replace and change several amino acids. Most of the other mutations are the result of one (i.e., single nucleotide variant; SNV) or two (i.e., multiple nucleotide variant; MNV) nucleotide changes. Several of these mutations are found in other assemblies of S. frugiperda and likely reflect natural variation. An alternative explanation for the variation is that they are associated with the c-strain which is thought to make up at least some of the genome of the invasive populations. Only one mutation, a 2 bp deletion in one individual collected from the Burdekin in Queensland was identified that might cause a frame shift mutation (see  Supp Figs. 14 and 15 [online only] (toac151_suppl_supplementary_materials.zip)) as has been identified in Bt resistant individuals in other studies. It was present as a heterozygote in the individual and has not been experimentally validated.

Table 5.

Summary bioassay data on Spodoptera frugiperda and Helicoverpa armigera involving diet incorporation of insecticides

Discussion

In this study, we showed insecticide and Bt response differences in two of the first reported S. frugiperda populations in Queensland (i.e., SF20-1) and Western Australia (Sf20-4) in Australia, with the Queensland (Sf20-1) population being less tolerant to various insecticides such as methomyl, chlorantraniliprole, and indoxacarb compared to the Western Australian (Sf20-4) population. On the other hand, the Queensland population exhibited at least a two-fold higher tolerance to Cry1Ac, Cry2Ab, Cry1F, and Vip3A Bt toxins than the WA population. Characterization of the ABCC2 resistance gene identified one S. frugiperda individual from Queensland as potentially being heterozygous with a 2 bp deletion that could underpin Cry1F resistance, although confirmation of the detected mutation and of the resistance phenotype is required. This could be accomplished via PCR and Sanger sequencing (e.g., Guan et al. 2021) as well as via the CRISPR/Cas9 gene editing approach (e.g., see Wang et al. 2017). The response differences to chemical insecticides and Bt toxins of the two studied invasive S. frugiperda populations in Australia and their ACE-1 resistance allele profile differences (Table 6) may reflect the diverse genetic composition across the pest's recent expanding range (Schlum et al. 2021, Rane et al. 2022, Tay et al. 2022a), and suggest that separate pathways were involved in the establishment of these Queensland and Western Australian populations. This is contrary to the current postulation of a single introduction pathway for the arrival of S. frugiperda to Australia based on an assumption (Jing et al. 2021) or reverse trajectory simulation (Qi et al. 2021) and highlights the importance of harmonizing simulation studies with genomic and phenotypic evidence.

With significant economic impacts on agriculture from S. frugiperda reported in over 80 countries (excluding the New World native range) from Africa, Middle East, Asia, Southeast Asia, and Oceania, its response to different insecticidal and Bt control agents is increasingly investigated at recently impacted localities (e.g., Worku and Ebabuye 2019, Deshmukh et al. 2020, Zhang et al. 2020, Kulye et al. 2021, Lv et al. 2021). However, the diverse bioassay methods (approach, larval stage, scoring criteria) used in these studies complicate the meaningfully comparison of findings. Taking emamectin benzoate and indoxacarb as examples, Zhang et al. (2022) used a topical application bioassay on 3rd instar larvae, Deshmukh et al. (2020) used leaf-dip bioassays on 2nd instar larvae, while Hardke et al. (2011) used a diet-incorporation approach on 3rd instar larvae. Responses have been expressed as LD₅₀ (e.g., Deshmukh et al. 2020) or as EC₅₀ (e.g., Kulye et al. 2021), further making it difficult to rigorously compare outcomes.

Table 6.

ACE-1 locus of Spodoptera frugiperda individuals from Australia, South Korea, Papua New Guinea, and Peru characterized via whole genome

It is important to stress the challenge and difficulty to compare bioassay results between studies due to the different genetic background of the test samples (e.g., due to different number of individuals used to establish test populations), methods and approaches between research groups, and the general different rearing conditions of laboratory cultures that could contribute to varied response outcomes. A further challenge for invasive S. frugiperda management is the lack of base-line values representing susceptible responses in newly populated areas which makes it difficult to monitor changes through time in insecticide efficacy due to resistance. Knowledge of whether the introduction occurred once (e.g., single ‘invasive bridgehead effect’; Guillemaud et al. 2011) versus multiple times (e.g., multiple ‘mass dispersal’; Wilson et al. 2009) and patterns of gene flow also impact the long-term monitoring of insecticide resistance evolution. Using native S. frugiperda colonies established in 2005 from cotton fields in Louisiana, USA, Wilson et al. (2009) undertook diet-incorporation bioassays (on 3rd instar larvae), using insecticides including chlorantraniliprole, indoxacarb, and spinetoram (see Table 7). Yu (1991) undertook topical application bioassays against methomyl (on 4th instar larvae) which included comparison with a susceptible population free from insecticide exposure since 1975, and a resistant population collected from a maize field in Gainesville, Florida (see Table 8). In comparing insecticide responses in Indian invasive S. frugiperda populations with diet incorporation assays (on 3rd instar larvae), Yu (1991) also included a native susceptible S, frugiperda population from Brazil to assist with interpreting changes to insecticide responses (including chlorantraniliprole and spinetoram) at spatial and temporal scales.

In the absence of native S. frugiperda populations in Australia, and in addition to the information collected herein for H. armigera and S. litura, we used the information from these global studies on native and newly invaded populations to assist interpretation of the early data collected in this study on invasive S. frugiperda populations in Australia. These examples used broadly similar approaches, instars, and scoring criteria to our study for bioassays with specific insecticides. Relative to the indoxacarb results from Hardke et al. (2011) on native S. frugiperda populations from the USA, our bioassay findings for the invasive Western Australian and Queensland populations suggest a 28- and 3-folds difference, respectively (Table 7). For spinetoram, resistance ratios of the two Australia populations (which were like each other) is around 10 and 1.5 times higher than for the native populations from Brazil and the USA respectively. In contrast the ratios for the invasive populations from India were 1.3 and 0.3 relative to the Brazil and USA native populations. The differences are even more pronounced for chlorantraniliprole where the Sf20-4 WA population exhibited ratios that were 32 and 2.4 times higher than both the native Brazilian and American populations, which are at least around 3-fold higher than for the invasive populations in South Africa and India (Table 7) and also for the invasive Indian population studied by Deshmukh et al. (2020) using a leaf dip assay (and hence not reported in Table 7).

Methomyl resistance alleles have been reported in invasive populations from China, Indonesia, Africa, and in this study. Comparisons between susceptible and resistant strains of native S. frugiperda populations from Florida USA (Yu 1991) and with the South African invasive population of S. frugiperda (Eriksson 2019) showed that the Queensland population as similar in its response as the susceptible Florida S. frugiperda, while the Western Australia population was around 5 times more tolerant. This contrasted with the South African population which was 110 times more tolerant than the Florida susceptible S. frugiperda strain, and around 8 times more tolerant than the resistant S. frugiperda strain from Florida (Table 8). Taken as a whole, similar bioassay results between native (i.e., Brazil, USA) and various invasive populations suggested potential significant genetic diversity in introduced populations. This concurs with population genomic and genetic analyses (Zhang et al. 2020, Schlum et al. 2021, Jiang et al. 2022, Rane et al. 2022, Tay et al. 2022a) that suggested multiple origins for the invasive African, Asian (Indian, Chinese), and Southeast Asian (e.g., Malaysia) S. frugiperda populations.

It is possible that the differential response between the two Australian S. frugiperda populations to insecticides and Bt toxins represents natural variation in the national population which has emerged from a single founding incursion. However, it is unlikely that the differential responses within Australian populations represents local selection pressures because the period between populations establishing and being collected for this study was presumed to be too short to enable this opportunity. It could be that separate incursions involving different source populations occurred in WA compared to the eastern states of Australia, as supported by genome-wide SNP marker population genomic studies (Rane et al. 2022). Different selection pressures on the global population which recently originated from multiple Asian, Southeast Asian, and African incursions (Zhang et al. 2020, Schlum et al. 2021, Rane et al. 2022, Tay et al. 2022a) may have driven different phenotypes which entered the expanding ranges tested in our study and others reported herein. However, Kulye et al. (2021) demonstrated in S. frugiperda populations collected in India during 2018, 2019, and 2020 that large response changes such as those observed in South African (e.g., methomyl; Eriksson 2019) and Western Australian (chlorantraniliprole, indoxacarb; this study) populations were unlikely considering the short time frame since the very recent arrival of S. frugiperda especially assuming a west-to-east spread (Goergen et al. 2016; Cock et al. 2017; Nagoshi et al. 2018, 2019b). Another explanation for the differential responses of the two Australian populations that we studied is that they have different fitness because of variation in genetic diversity which reflects the number of individuals founding the populations. However, if this was the case one might expect sensitivity levels to be consistently higher or lower to all of the agents that we tested but this was not the case.

Fig. 1.

Summary of Acytylcholinesterase (ACE-1) susceptible and resistance allele frequencies in invasive range Spodoptera frugiperda populations from: (a) six African countries (i.e., Benin (#1), Uganda (#2), Kenya (#3), Tanzania (#4), Zambia (#5), Malawi (#6)), (b) Asia (i.e., India, Indonesia, Papua New Guinea (PNG), South Korea), (c) China, and (d) Australia. A total of 1,177 individuals representing 75 populations in  Supp Table 2 (online only) (toac151_suppl_supplementary_materials.zip) were used to compile the data below. Population identity is as provided in Supp  Table 2 (online only) (toac151_suppl_supplementary_materials.zip) that combined data from this study (Australia SF20-1 and SF20-4 (generation 0 representing field-collected individuals), PNG, South Korea) and from published studies (Boaventura et al. 2020a, Zhang et al. 2020, Zhao et al. 2020, Guan et al. 2021, Nguyen et al. 2021, Tay et al. 2021, Yainna et al. 2021, Rane et al. 2022). Susceptible and resistant alleles from the three previously reported loci (i.e., F290V, G227A, A201S) from the ACE-1 gene provided evidence to support multiple independent introductions across the invasive S. frugiperda populations, such as in (b) Indonesia (#2; G227A) and South Korea (#4; F290V), and in (c) China (Hubei province (#10-13); G227A) as indicated by the arrows. In (d) Australia, newly established S. frugiperda populations between Queensland (e.g., #27, #28 (Walkamin), #29 (Strathmore), #32 (Burdekin), and Western Australia (i.e., #2 (Kununurra)/Northern Territory (e.g., #21 (Bluey's Farm)) suggested this likely involved multiple introductions from diverse populations from neighboring countries and likely arrived via separate pathways and entry points (see also Rane et al. 2022).

Fig. 1. Continued

Alien invasive agricultural pests are increasingly being shown to carry novel insecticide resistance genes (e.g., Anderson et al. 2018, Walsh et al. 2018, Tay and Gordon 2019). This includes S. frugiperda in which invasive populations have been confirmed via whole genome sequence analyses (e.g., Zhang et al. 2020, Guan et al. 2021, Yainna et al. 2021) and molecular characterization to harbor selected resistance genes (Boaventura et al. 2020a,b; Zhao et al. 2020). Our review of reported ACE-1 and VGSC resistance allele frequency differences (Fig. 1; Supp  Table 2 [online only] (toac151_suppl_supplementary_materials.zip) and references therein) suggests that the invasive S. frugiperda populations within China (Guo et al. 2020), Indonesia (Boaventura et al. 2020a), Queensland, Australia (Tay et al. 2021; Rane et al. 2022), and South Korea (this study) were genetically diverse and likely originated from different native populations (Fig. 1). This further supports the perceived rapid spread of S. frugiperda across Africa, Asia, and Oceania as likely to also involve multiple independent introduction events.

Given the insecticide resistance allele frequency differences in S. frugiperda (e.g., Boaventura et al. 2020a, Guan et al. 2021, Lv et al. 2021, Yainna et al. 2021) and variation between the two populations sampled herein in bioassay responses to some of the approved insecticides for broadacre cropping, effort in Australia and indeed, for other invasive regions where possible, should now be directed to establishing baseline susceptibilities against key chemistries for multiple populations across geographies. For instance, pyrethroid resistance is common in the field in H. armigera (Walsh et al. 2018), and our comparisons with H. armigera and S. litura suggest that pyrethroids like cypermethrin are unlikely to provide good control in Australia against S. frugiperda. It also indicates some heterogeneity in the response to pyrethroids in the first invasive S. frugiperda populations in Australia. The large discrepancies between WA and Qld populations for methomyl and indoxacarb also require study of further populations before drawing firm conclusions. This information will be critical for on-going monitoring of resistance allele frequencies to key chemistries and determining their field application efficiencies, among geographies which are required to inform resistance management plans. It will also be important to understand gene flow patterns between different populations. Early detection of potential future introductions into Australia of novel resistance genes/alleles should also be a priority; for example, the VGSC L1014 resistance allele and the ACE G227A resistance allele from Southeast Asia (Boaventura et al. 2020a) and China (Guo et al. 2020), ryanodine receptor (RyR) resistance alleles from Brazil and China (Bolzan et al. 2019; Boaventura et al. 2020b; Lv et al. 2021) and ABCC2 resistance alleles from the Americas (Banerjee et al. 2017, Flagel et al. 2018, Guan et al. 2021, Yainna et al. 2021).

Being alert to new S. frugiperda incursions carrying novel resistance genes is relevant also for the Bt toxins. Our bioassay findings suggest Australia's S. frugiperda populations likely do not carry Cry1F resistance alleles in the homozygous state known to exist in native range S. frugiperda populations (e.g., Banerjee et al. 2017, Flagel et al. 2018, Guan et al. 2021, Yainna et al. 2021), however, the 2 bp deletion identified via whole genome sequencing in a single heterozygous individual from Burdekin (Queensland) will require further confirmation. Field-selected lines of VIP3Aa20 resistant S. frugiperda have been reported in Brazil and the USA where they occur as a recessive trait (Yang et al. 2013, 2018, 2019; Bernardi et al. 2015, 2016). The candidate resistance gene(s) and associated mutation(s) underpinning this resistance are yet to be identified, therefore phenotypic bioassays are required to detect resistance to VIP3A. Using what was essentially an F₀ screen of material we did not detect resistance to VIP3A in our field-derived laboratory-maintained S. frugiperda colonies. However, bioassays involving F₂-crosses (Andow and Alstad 1998) would be needed to further confirm the status of VIP3A resistance allele in both populations. The F₂-crosses approach should be a valuable tool in protecting the Australian cotton industry against S. frugiperda, given the >90% up-take of Bollgard III cotton containing Cry1Ac, Cry2Ab, and Vip3A proteins. To protect agriculture in Australia (and elsewhere) from an accidental introduction of VIP3A resistant populations from the Americas (e.g., from Brazil, Bernardi et al. 2016) into global invasive populations, national (e.g., pre-border) and industry biosecurity preparedness strategies must be coordinated to prevent and to increase the chances of early detection of such novel introductions of new traits.

Table 7.

Comparisons of bioassay results for selected insecticides fed to laboratory populations of Spodoptera frugiperda via diet incorporation approaches

Novel introductions leading to unique population structure has been reported for western African S. frugiperda (Nagoshi et al. 2022), in populations from China (Jiang et al. 2022), in Africa (e.g., Benin vs. Malawi; Tay et al. 2022a), and in, e.g., Australia, Malaysia, and Myanmar versus China populations (Rane et al. 2022), suggesting that the widely anticipated long distance migration of S. frugiperda especially in the invasive range could likely be less and may be impacted by localized ecological and climatic determinants (Tay et al. 2022b). Factors impacting the founding for the two Australian populations (e.g., founding number i.e., involving one/few individuals vs. many); gene flow dynamics at spatial and temporal scales, and frequencies and impact from potential novel and on-going introductions) used in our bioassay studies is at present unknown. Bioassays testing of a larger population pooled from several sites or several individual populations from multiple sites would be especially relevant in a pest capable of high dispersal across the landscape, however, the spread of S. frugiperda in the invasive range is increasingly being recognized as may not be as rapid and widespread as originally believed (e.g., Jiang et al. 2022, Nagoshi et al. 2022, Rane et al. 2022, Tay et al. 2022a). As such, we advocate cautionary approaches to avoid premature assumptions that extensive population admixture via gene flow in Australia landscape would lead to homogenized populations and therefore the approach to combine multiple distantly sampled populations in bioassay studies (e.g., Bird et al. 2022).

Finally, to enable meaningful comparisons of bioassay findings from across the S. frugiperda invasive range, there is a need to globally standardize approaches for testing of insecticides and Bt toxins. Localized novel resistance traits in populations of S. frugiperda would likely serve as a new resistance management challenge to other neighboring regions (Kalyebi 2020), and movements of S. frugiperda in the new invasive ranges could lead to as yet unknown and complex gene flow patterns that could significantly hinder the development of suitable pest and resistance management strategies for this global pest complex.

Table 8.

Comparisons of topically applied methomyl bioassay on S. frugiperda between resistant or susceptible native populations from USA (FL) and invasive populations from South and Australia