Introduction

Fall armyworm, Spodoptera frugiperda (J. E. Smith) (Lepidoptera: Noctuidae), is a highly polyphagous insect native to tropical and sub-tropical areas of the Americas (Kenis et al. 2023; Luginbill 1928; Pitre 1988; Sparks 1979). Spodoptera frugiperda does not undergo diapause and therefore requires suitable temperatures and host plant availability year-round to maintain populations (Du Plessis et al. 2020; Prasanna et al. 2018). Prior to its invasion into Australia in early 2020, it was predicted that S. frugiperda would likely persist in northern Australia throughout the year, while migration southward was expected as temperatures in the south warmed (Abrahams et al. 2017). In northern Western Australia where S. frugiperda persists year-round (Department of Primary Industries and Regional Development WA 2020b), there is a tropical climate with a wet (hot humid with monsoonal rains) and dry season (warm and little rain). Temperatures are suitable year-round and the irrigated grain crops maize (Zea mays L. (Poaceae), sorghum (Sorghum bicolor L. (Poaceae) and Rhodes grass (Chloris gayana Kunth (Poaceae) grown as fodder are used as host plants by S. frugiperda primarily during the dry season (May to October). In the wet season, S. frugiperda larvae may use sorghum and Rhodes grass which may be grown for animal fodder but otherwise, there is a need for alternative hosts for development. Information on alternative hosts that can support S. frugiperda development during the wet season in northern Western Australia remains unknown. In this study, we sought to identify some of the possible host plants that S. frugiperda may use for development, with a focus on grasses and legumes used as fodder and available during the wet season (November to April).

Spodoptera frugiperda has been found to attack a wide host range of more than 353 recorded plants from 76 families, primarily from Poaceae (Gramineae), Asteraceae, and Fabaceae (de Freitas Bueno et al. 2011; Hardke et al. 2015; He et al. 2020; Kenis et al. 2023; Li et al. 2020; Montezano et al. 2018; Nagoshi and Meagher 2004; Pitre and Hogg 1983). Overall, it has strong preference for the grasses of Poaceae including maize, rice (Oryza sativa L. (Poaceae), sorghum, pasture grasses, turf grasses, and sugarcane (Saccharum officinarum L. (Poaceae) (Montezano et al. 2018; Dumas et al. 2015). To date, in northern Western Australian cropping systems, S. frugiperda has been found and reported to complete development primarily in maize, sorghum, and Rhodes grass. It has also been found in millet (Panicum miliaceum L. (Poaceae) and reported on other crops such as cotton, Gossypium sp, L. (Malvaceae), and cantaloupe, Cucumis melo L. (Cucurbitaceae), but there is no evidence, at this point, that S. frugiperda is completing development on these other hosts in Western Australia.

A wide range of grasses and legumes are available for rain-fed and irrigated production systems in the Kimberley region of Western Australia (Department of Primary Industries and Regional Development WA 2020a, 2021). Grasses are widely utilised as pastoral fodders because of their reliable establishment, very high biomass production under irrigation, ease of management and its resilience (Department of Primary Industries and Regional Development WA, 2020c; Hacker and McDonald 2021; Waters et al. 1997). In addition, legume species are well known for their high-quality feeding value and ability to improve soil fertility through nitrogen fixation (Finn et al. 2013; Peoples et al. 2012). Rangelands in Western Australia occupies an extensive area (approximately 2.2 million acres) and forage pastures contributing to sustainable animal production as harvested and grazed fodder, are valued over $3 billion annually (Department of Primary Industries and Regional Development WA 2020c). It is anticipated that the variety of pasture grasses and legumes used in the Kimberley region as fodder may provide alternative hosts for S. frugiperda during the wet season.

Given the capability of S. frugiperda to attack and establish in wider range of grasses as well as in legumes in their native range (Montezano et al. 2018), it remains unknown if S. frugiperda can utilize pastoral grasses and legumes available in northern rangeland to complete development. Through a no-choice bioassay we aimed to assess larval survival and development to the adult stage on several pastoral grasses and legumes to evaluate their potential as suitable host plants for S. frugiperda. We then determined if the emergent female moths reared on the different plant species could produce viable eggs.

Materials and methods

Insect colony

The study insects, S. frugiperda, used in this experiment were sourced as egg masses from a colony established in 2020 and maintained in a laboratory at the Western Australia Department of Primary Industries and Regional Development (DPIRD) research station at Kununurra. The colony was established with wild larvae collected from sweet corn, maize, and sorghum grown on local farms and the research station in 2020. New larvae and egg masses collected from the field were regularly introduced to the colony. The colony was maintained in controlled environment at a temperature of 24–26 °C and a photoperiod of 12:12 (L: D). At the time the study began, the colony had completed 11–13 generations. Egg masses were obtained from a panmictic mating of approximately 25–40 S. frugiperda adult moths from the 13th generation housed in a 47.5 × 47.5 × 47.5 cm mesh cage (Bugdorm 44,545, Megaview, Taiwan). Adults in the cage were provided ad libitum access to 10% honey solution fed through a cotton wick in a 30 ml container with lid. The cages were checked daily for oviposition and egg masses were collected from the cage and placed in 30 ml plastic containers with a lid (Solo plastic Ltd, Highland Park, Illinois, USA).

Larvae were randomly selected from each egg mass to be maintained in the colony and reared on artificial diet (modified from soybean flour and wheat bran diet; Greene et al. 1976) (1 L diet comprises navy bean flour-120 g, wheat germ- 60 g, Brewer’s yeast-36 g, ascorbic acid-3.6 g, sorbic acid-2 g, methyl-4 hydroxybenzoate-2.2 g, Vanderzant vitamin mixture-5 g (Sigma Aldrich, North Ryde, NSW, Australia), liquid formaldehyde-3 ml, agar-10 g, distilled water-550 ml). Each larva was provided 2 g of artificial diet and the diet was replaced every four days until the larvae pupated. Upon pupation they were placed individually in a 30 ml plastic cup, until adult emergence, and then placed in the cage for mating.

Test plants

Nine plant species i.e., maize, sorghum, Rhodes grass, sabi grass (Urochloa mosambicensis P.Beauv (Poaceae)), barnyard grass (Echinochloa crus-galli P.Beauv (Poaceae)), creeping blue grass (Bothriochloa insculpta Kuntze (Poaceae)), Cavalcade (Centrosema pascuorum var Cavalcade Benth (Fabaceae)), butterfly pea (Clitoria ternatea Linnaeus (Fabaceae)), and siratro (Macroptilium atropurpureum Urban (Fabaceae)) were used in the study as host plants. Maize, sorghum, Rhodes grass, creeping blue grass, Cavalcade, butterfly pea, and siratro were sourced as seeds of locally cultivated strains, whereas sabi and barnyard grasses were collected as rootstock from roadside areas of the DPIRD Kununurra research station. Plants were grown outdoors in a raised bed (200 cm*100 cm*60 cm) with multi-purpose premium potting mix (Debco Pty Ltd, Tyabb, Victoria, Australia). The raised bed was divided into ten small blocks (approx. 50 cm*40 cm) where seeds or rootstock of each plant species were sown or transplanted. The plants were watered daily throughout the experiments using an automated sprinkler system. No agro-chemicals were applied to the study plants. Foliage of two- to four-week-old plants were used to feed S. frugiperda larvae in the bioassay.

Bioassay

A freshly hatched neonate larva (0–12 h old) was selected from an egg mass and individually placed in a 30 ml plastic cup with lid. Each larva was provided with fresh foliage of one of the nine test plant species. When an egg mass was produced in the colony, a single neonate from each egg mass was placed in a cup with any plant species. This was done to ensure that any plant species did not receive multiple neonates from an individual egg mass. Thus, the cohort used in the bioassay was gathered from egg masses laid over a 7-day period. A total of thirty larvae were tested for each of the host plants except blue grass. The establishment of blue grass was inconsistent; therefore, only 24 larvae for blue grass were tested. Initially 1st and 2nd instar larvae were provided approximately 2 g of young fresh leaves based on the work by Fei et al. (2017), and from the third instar onward this amount was increased to 10 g of leaves for each larva. Foliage was replaced every day and thus, larvae received fresh plant material until the onset of pupation. Leaves were randomly collected from the plants and then cut into small pieces before being placed in the cups with the larvae. To minimise handling in the younger instars, the plant material and frass were removed daily once the larvae reached the third instar.

Larval death was checked daily to assess the proportion of larvae completing development. Although all attempts to minimise handling damage were made, there were deaths due to handling. This was determined if any specimen displayed physical damage in its body e.g., body was crushed, impaled, or decapitated. The date of death was recorded for any larvae dying during the experiment. We also recorded days to pupation for those larvae that successfully pupated, days to adult emergence (days from pupation to adult eclosion), and time for moth emergence. After adult emergence, sex of the moth was recorded to assess adult sex ratio. At 14 and 21 days after hatching, the larvae were individually weighed to the nearest 0.0001 g using an analytical balance (OHAUS; model- Pioneer IC-PX84/E) (Sharma et al. 2005). Following pupation, pupae were individually weighed using same analytical balance at eight days after pupation (He et al. 2021). Larvae in some treatments pupated before day 21, these larvae were not weighed at day 21 and therefore those data points are missing.

To test the reproductive capability of the S. frugiperda moths on the different diets, upon emergence, a single pair of adults was released into a mating cage. The cage was constructed from round plastic buckets (1.3 L capacity that were removed from a standard unitrap (Bugs for Bugs Ltd., Toowoomba, Queensland) and covered with black seamless fine mesh cloth (sourced from Wildtrak Kununurra) that was secured to the bucket with rubber bands. Adults were provided with 10% honey solution placed in a 30 ml plastic container with lid; a cotton wick was inserted with one end in the solution and the other end providing access to the solution above the lid. The containers were checked daily to record any egg laying. However, no mating observations were carried out. Once any egg masses were found, they were collected by carefully removing the mesh cloth and placed in a 30 ml plastic container with lid to be held until hatching. Egg hatching was recorded daily. The oviposition assays continued until the last pair of adults survived. In addition, if any moth laid eggs multiple times, this was also recorded. A total of six pairs moths from each plant species were tested for reproductive capability except Siratro. Only three pairs of moths emerged from the cohort of larvae provided siratro. There were insufficient moths emerging from the cohorts provided butterfly pea and Cavalcade to test for reproductive potential.

Statistical analysis

A survival analysis was conducted using the days to death for those individuals that died during the experiment and the date of adult emergence as the censoring date for those that completed development. Kaplan-Meier estimates and a test for equality of the survival curves for each cohort on the different diets was calculated using Genstat (v20.1.23942, VSN International Ltd, UK).

The proportion of larval survival on the different host plants was calculated as the total number of larvae completed development divided by the total number of larvae placed in the assay for the respective host plants and expressed as percentage. Individuals that died owing to handling damage were not included to determine the proportion of larval survival. The influence of host plant on the larval development period, larval mass at 14 days and 21 days, pupal period, and pupal mass were analysed performing a general linear model. Post-hoc pairwise comparisons were performed using Tukey’s HSD tests. For all models, data normality was assessed through graphical analyses of the residuals. All analyses were conducted using JMP Statistical Software Version 16.1. There was no survival when larvae were provided Cavalcade therefore, these host plants were not included in subsequent analyses. There were very few larvae (n = 5) that pupated when placed on butterfly pea diet. These were not included in analyses of pupal development time or pupal mass, but the means are presented. The larvae that were provided maize and sorghum pupated before they reached 21 days, therefore maize and sorghum were not included for the analysis of larval mass at 21 days. Finally, proportion of adult emergence, proportion of females ovipositing, and egg hatching were also calculated as percentage value. Sex ratio of the emerged adults was also estimated.

Results

Of the 264 larvae selected for this experiment, 47.3% completed development to the adult stage. Survivorship was influenced by larval diet (Log-rank = 204.447; df = 8; p-value < 0.001) with S. frugiperda fed maize (88%) having a much higher probability of survival than larvae provided with any other diet. Nonetheless, 81% fed sorghum, 79% fed blue grass, and 71% on barnyard grass completed development. Just over 50% of S. frugiperda completed development on Rhodes grass (55%) and sabi grass (53%). Fewer than 50% of S. frugiperda on all the other host plants completed development, with none surviving to adulthood on cavalcade, and only two surviving on butterfly pea.

The average period to complete development from larva to adult emergence was 31.17 ± 0.06 days and varied with larval diet (F7,117 = 38.6; p < 0.001). Overall development time was shorter when larvae were fed maize and sorghum than in any other plants. The full development period took about 4 days longer on average when larvae were fed blue grass and 5 days longer on barnyard grass, 6 days on Rhodes grass, and 7 or more days on sabi grass, butterfly pea and siratro. The average period of egg laying to hatching in this study was 4.29 ± 0.12. Most of the deaths occurred in the larval stage (36%); 11.4% died in the pupal stage. A small percentage (5.3%) died due to handling damage.

Effect of host plants on larval development

All the larvae provided with maize completed development to the pupal stage by day 20 (Table 1). The proportion of larval survival was greater when were reared on maize, sorghum, and blue grass, than larvae fed on siratro, butterfly pea, and cavalcade (Table 1). However, the proportion of larval survival provided with Rhodes, sabi, and barnyard grass was greater than those reared on legumes (Table 1). No larvae completed development on Cavalcade (Table 1). The average lifespan of larvae fed Cavalcade was 10.5 days (Table 1) with the longest lifespan being 19 days.

The average period of larval development in this study was 19.7 ± 0.3 days. However, larval development period varied with the host plant species (F7,159= 43.91; p < 0.0001; Table 1). Larva fed on maize and sorghum completed development in shorter time than those on the other host plants. The larvae fed siratro and butterfly pea took the longest to complete larval development compared to the larvae fed the other host plants (Table 1). Larval developmental time on the other grasses (blue, Rhodes, sabi, and barnyard grass) did not differ amongst those species (Table 1).

The average larval mass at 14 days in this study was 0.219 ± 0.011 g. Feeding on different host plants significantly influenced the larval weight (F7,209=112.52; p < 0.0001; Table 1). Larval mass was greater when larvae were fed sorghum followed by those provided maize but statistically, the 14-day mass of larvae in these two groups were not different from each other (Table 1). However, they were heavier than larvae that were reared on barnyard grass, blue grass, Rhodes grass, sabi grass, and siratro and butterfly pea (Table 1). Larvae reared on butterfly pea and siratro weighed less than larvae on all other plants (Table 1). Overall, the larval mass at 21 days (0.292 ± 0.014 g) did not vary among the different host plants in this analysis (F5,89=1.19; p = 0.32; Table 1). However, this analysis only included a limited number of groups, as larvae fed Cavalcade had died, and those fed maize or sorghum had already pupated.

Effect of host plants on pupal development and adult emergence

Pupal period was on average 11.91 ± 0.13 days in this study and significantly varied with the larval host plants (F6,116 = 5.59; p < 0.001; Table 2). Larvae that were reared on sorghum had a shorter pupation period compared to those reared on other larval hosts (Table 2). Similarly, pupal weight varied with the larval host plants (F6,135 = 9.34; p < 0.001; Table 2). Pupae that were obtained from larvae fed on maize and sorghum had the greatest mass, whereas larvae reared on siratro had the least mass (Table 2).

The proportion of moth emergence from the pupa varied according to larval host plants. Moth emergence was lower only for the pupae obtained from Rhodes grass (69.56%) and siratro (53.84%). Overall, adult emergence from other host plants were comparatively higher (> 80%) with 95% of adults emerging from barnyard grass-fed larvae (Table 2). Only two larve fed on butterfly pea were able to complete development to the adult stage.

Effect of host plants on reproductive capability

Overall, sex-ratio of the emerged moths (male: female) was 1.18:1 There was some variation around this ratio (Table 3). More moths fed as larvae on sorghum were female (1:2.67), while Rhodes grass yielded male-biased moths (3:1).

A majority of the adult females laid eggs irrespective of their larval host (Table 3). The proportion of females laying eggs was higher if they were fed blue grass and sabi grass as larvae (> 90%) than those reared on other hosts (> 75%) (Table 3). In addition, hatching rate was relatively higher for the eggs laid by the females provided maize at their larval stage (89%). Overall egg hatching proportion ranged from 71 to 89% (Table 3).

Discussion

The reported wide host range of S. frugiperda has generated considerable concern in the Australian grains and horticulture industries. Given the possibility that S. frugiperda can complete their development on other important crops grown in northern Australia such as cotton, melons, capsicum, and other vegetables (Hardke et al. 2015; Wu et al. 2021), the present study focused on the potential host use of pasture grasses and fodder due to the significant area coverage and value of the pastoral industry in northern Western Australia.

In this study, we found significant differences in the survival and developmental time of S. frugiperda larvae and pupae among the host plants tested. Maize and sorghum were found as premium hosts for S. frugiperda development (Day et al. 2017). It is already known that the larval development and reproductive potential of S. frugiperda females are significantly greater when larvae feed on maize and sorghum (Sotelo-Cardona et al. 2021). Generally, larval fitness as well as nutritional access mediate the attributes of the pupal stage (e.g., weight, size) (Eduardo et al. 2018). Individuals fed on these host plants generally weighed more and took less time to develop. Most adult females were able to lay viable eggs. We did not observe mating. Therefore, we cannot confirm if the egg masses that did not hatch were unable to do so because they were not fertilized or because embryos were unable to resource nutrients from the egg. A lack of nutrients in the egg may occur as a result of a poor larval diet. There was an unexpected female-bias in the sex ratio in the sorghum fed cohort and a male-bias in the Rhodes grass cohort.

Following maize and sorghum, despite their slower development on Rhodes grass, sabi grass, barnyard grass, and creeping blue grass, more than half of the larval population fed on those grasses, eventually developed into adults. The longer development time did not seem to impact on reproductive capability. The oviposition performance of the females was comparable if not higher than the females reared on maize and sorghum, with more females laying eggs proportionally (Table 3). In the field, the longer development time may incur other penalties as it increases vulnerability to parasitism and predation (Uesugi 2015), particularly in habitats with these grasses where the larvae are forced to hide in leaf litter and soil rather than entrench in a whorl. Both S. frugiperda larvae and adults experience a complex fluctuating natural environment with extensive spatial and temporal variations in plant resources (Carrasco et al. 2015). In this larval bioassay, we reared individuals in a no-choice scenario in a laboratory. It may be that in a more complex environment they would switch to preferred or alternate hosts and thus be able to complete development. Further studies are suggested to elucidate the behaviour of these insects through choice scenarios and alternating temperatures.

Legume pastures (Cavalcade and butterfly pea and siratro) were found to be the least suitable hosts in this bioassay. Of these 23% of S. frugiperda larvae could complete development on siratro but they took much longer and did not produce egg laying adults. More than 30 legumes from the Fabaceae family have been recorded as hosts of S. frugiperda worldwide (Montezano et al. 2018; Hailu et al. 2018) but the plants tested in this study do not appear to be capable of supporting development. This study was performed in a laboratory environment and no comprehensive field surveys have been conducted to assess the presence of S. frugiperda on any of the host plants tested here, or other potential hosts in pastoral systems. Therefore, surveys should be conducted to assess the presence of S. frugiperda in these host plants particularly over the wet season. These surveys should also assess the presence of potential natural enemies of S. frugiperda, as these may reduce the population over the wet season.

The development of S. frugiperda larvae on these grasses used in this bioassay is not surprising as S. frugiperda have been found to establish in bermuda, barnyard, Rhodes grass, Johnson grass and other closely related grasses in its established habitat (Montezano et al. 2018; Wiseman and Davis 1979). Given that S. frugiperda can complete development and produce viable eggs after feeding on pastoral grasses, it may become a risk to rangeland pastoralism. However, In the present study, we only concentrated on the larval host plant as single factor to evaluate its potential to support the development of S. frugiperda. Future studies on the oviposition preferences of the S. frugiperda females are needed prior drawing conclusions on the host plant range of FA S. frugiperda in Australia. A bioeconomic model and cost-benefit assessment of S. frugiperda impact on pastoral hosts could also be conducted to determine the risk to pastoral systems and livestock health (Similar to Cook et al. 2021). If, for example, S. frugiperda was to outbreak and lead to declines of pasture condition, there may be impacts for rangeland livestock. Determination of those conditions and assessment of the impacts through further study can gain further insight into S. frugiperda host choice.

Overall, the present study reveals the potential of rangelands pastoral grasses as wet season host plants for S. frugiperda when their favoured and ideals hosts maize and sorghum are unavailable.

Table 1 The proportion of fall armyworm larvae surviving to pupation, the development time and the larval mass at 14 and 21 days when placed on one of nine different host plants in a no-choice bioassay
Full size table
Table 2 Fall armyworm pupal development time, the pupal mass at 8 days and proportion of pupae emerging as adults when placed on one of nine different host plants in a no-choice bioassay
Full size table
Table 3 Fall armyworm adult sex ratio, proportion of adults laying eggs and proportion of egg hatching when placed on different host plants in a no-choice bioassay
Full size table