Oviposition preference not necessarily predicts offspring performance in the fall armyworm, Spodoptera frugiperda (Lepidoptera: Noctuidae) on vegetable crops

Given the new spread and potential damage of the fall armyworm (FAW), Spodoptera frugiperda (J. E. Smith) (Lepidoptera: Noctuidae) in Asia, it has become imperative to understand the development biology of this invasive species on selected vegetable crops in newer geographical regions. In this study, we investigated the ovipositional preference of FAW females on different host plants, under choice- and non-choice tests. In addition, using the age-stage, two-sex life table theory, we assessed the performance of immature FAW individuals fed and reared on selected vegetable crops to get information related to development time, survival, reproduction and longevity. Fall armyworm females had an oviposition preference on maize compared to other vegetable crops, including cabbage and soybean, and reluctance for tomato, which was confirmed during the choice and non-choice tests. In contrast to the oviposition preference, our results also suggest that despite low preference for cabbage, soybean, and tomato, these crops seemed to provide a high benefit for an appropriate offspring performance, exceeding in some cases the benefits from a maize-based diet. Information from this study was discussed in terms of FAW ecology and how female’s decision affects their reproductive fitness, and the survival and performance of its offspring.


Fall armyworm oviposition behavior under a no-choice scenario. Under a no-choice scenario,
FAW females were reluctant to lay their eggs in tomato, but in contrast significantly more eggs were recorded from the net wall of the cage used for the experiment (Fig. 1a). No differences were observed for the number of eggs laid in maize plants compared to those laid in the cage wall (Fig. 1b). When the "masking" effect was estimated, FAW females still preferred to lay their eggs in the cage wall compared to the tomato sprayed with maize extract (Fig. 1c). In contrast, maize plants sprayed with tomato extract did not deter the preference of females for oviposition (Fig. 1d).
Life table analysis. The results of development time, survival, reproduction and longevity of FAW fed on cabbage, maize (ear and leaves), soybean, tomato, and artificial diet are represented in Table 2. No statistical analysis was conducted on egg stage, since the study started with new larval hatching at the same time, and 3 days after oviposition. L1 stage was significantly shorter in cabbage, soybean and maize ear treatments, compared to maize leaves. L2 stage was significantly shorter in soybean compared to artificial diet. L3, L4, L5, L6 + L7 stages were significantly shorter in maize (ear and leaves) compared to artificial diet. In addition, the pre-pupal stage was shorter in soybean-reared individuals compared to maize leaves-reared individuals, whereas the pupal stage was significantly shorter in maize leaves compared to artificial diet and maize ear treatments ( Table 2, Fig. 2). Table 1. Average number of egg masses and eggs (mean ± SD, min.-max. range) laid by females of the fall armyworm, Spodoptera frugiperda and recorded under a choice test on cabbage, maize, soybean and tomato. Means followed by the same letter(s) in a column are not significantly different (P < 0.05) by Tukey's HSD.
Not all females were reproductively active, and this was observed among all treatments. Therefore, the preoviposition period (APOP), total pre-oviposition period (TPOP), oviposition period, and fecundity parameters were calculated based on the information gathered from the reproductive females and not for the total number of females recorded (Table 3). Moreover, for cabbage, only 11 (44%) out of the 25 initially recorded females produced eggs, whereas in the case of maize ear, maize leaves, artificial diet, soybean, and tomato, the reproductive females were 25, 68, 29, 55, and 60% out of the total amount of females recorded for each treatment, respectively (Table 3).
Feeding on different host plants also affected other parameters, including adult APOP, TPOP, oviposition period, and fecundity of the Fall armyworm (Table 3, Fig. 4). Females had the shortest APOP (i.e., number of days before a female start laying eggs and counted from adult emergence) in cabbage, artificial diet, maize ear, and tomato (2.5-2.6 days), whereas females reared in maize leaves started laying their eggs 5 days after emergence. In line with this, the TPOP (i.e., number of days that includes the pre-adult stage + APOP) was 4 days shorter in females fed on cabbage, maize ear, and soybean, compared to females reared on maize leaves (Table 3, Fig. 4). FAW females reared on soybean and cabbage had the longest oviposition period (i.e., the number of days that female produced offspring), almost double compared to maize treatments. In addition, females reared on cabbage, soybean, tomato, and maize ear laid 2.4-4.5 times more viable eggs compared to females reared on maize leaves and diet treatments ( Table 3, Fig. 4).
The age-specific survival rate (l x ), fecundity (m x ), and maternity (l x m x ) values of the FAW reared on six different treatments are plotted in Fig. 4. Highest peaks on fecundity were observed for females fed and ovipositing on cabbage, soybean, tomato, followed by artificial diet and maize with the lowest fecundity values. Maternity curves (l x m x ) had a higher peak on soybean, tomato, and cabbage, followed by lower curves in maize and artificial diet treatments. In addition, information on cumulative fecundity including all reproductive females in each treatment showed a higher total contribution of viable eggs (i.e. larvae) in soybean, cabbage and tomato treatments, compared to maize treatments ( Table 3, Fig. 5).
Net reproductive rate (R 0 ) was 5.7 and 11.4 times bigger in soybean compared to maize ear and artificial diet, respectively ( Table 4). The intrinsic rate of increase (r) and finite rate of increase (λ) of FAW were higher on soybean, tomato, and cabbage compared to maize (intermediate), and artificial diet (lowest), which suggest FAW populations will grow faster in the vegetable crops tested compared to the artificial diet. In contrast, a shorter mean generation time was observed in females from the maize treatment compared to the other vegetable crops tested.
In this study, we also did a general assessment of the nutritional quality of the treatments tested (Table 5), and we found that the amount of protein and nitrogen contained in the artificial diet treatment was 5-and 10-fold higher compared to soybean, tomato, cabbage, and maize, whereas the amount of fat was 2-5 higher for the same treatments compared to diet. In the case of chlorogenic acid, the amount contained in tomato was 3-and www.nature.com/scientificreports/ 10-fold higher compared to soybean and cabbage, respectively, and this compound was not detected in maize and artificial diet. Caffeic acid, and quercetin were not detected in soybean, maize and artificial diet, whereas the amount in tomato was 2-and 10-fold higher compared to cabbage, respectively.     www.nature.com/scientificreports/ soybean, tomato, and an artificial diet, in order to get information related to development time, survival, reproduction and longevity of FAW. As we expected, FAW females had an oviposition preference for maize compared to other vegetable crops, including cabbage and soybean. However, we did not expect FAW reluctance for tomato. Therefore, the no-choice test was performed in order to confirm the preference or avoidance of the most and least preferred host plants, maize and tomato, respectively. FAW laid similar amount of egg masses on the inner walls of the cage as well as in the maize plants, but, in contrast, FAW highly preferred the cage wall to lay their eggs instead of the tomato plants. Oviposition on non-plant material (i.e., cage walls) despite the presence of host plants has been previously recorded for FAW 29,30 . More specifically, Rojas et al. 29 found strong tactile stimuli for grooved and pitted surfaces on which FAW oviposit regardless the presence of chemical cues. Similar observation has been recorded for the generalist moths, Spodoptera exigua 31 , and Epiphyas postvittana (Tortricidae) 32 , where tactile stimuli seem to be as important as the chemical cues in terms of ovipositional behavior and preference.
When the masking effect was estimated, FAW females still preferred to lay their eggs in the cage wall compared to the tomato sprayed with maize extract, suggesting tomato plants were still avoided, regardless of the additional olfactory/gustatorial cues provided by the maize extract. It is also possible that the most important oviposition stimulant(s) from maize may not have been present in the extract, which was sprayed on to tomato plants. It should be noted that FAW female moths laid more eggs on control areas than on substrates treated with leaf extracts of maize. Therefore, tomato plants were not favored by maize extract. On the other hand, when maize plants were sprayed with tomato extract, FAW were not deterred and the oviposition did not decrease. An earlier study also confirmed that substrates treated with tomato leaf extracts were less preferred for egg laying by FAW female moths 29 . This result suggests that all antixenotic factors from tomato (i.e., including trichomes and chemical exudates from them) were not present in the extract to deter FAW from choosing maize plants as ovipositional substrate. It was also found that ovipositional deterrents were present in tomato plants 29 .
Based on the preference-performance hypothesis 22,23 , this part of our study suggests that female moths choose to oviposit in hosts with the highest nutritional quality for offspring, leading to shorter developmental time, increased biomass, and higher reproductive potential in FAW offspring reared on maize. Moreover, it is expected that the ovipositional preference of FAW female moths would align with the performance of their offspring. Thus, we assumed that FAW female moths reared on maize plants (ears and leaves) would have decreased developmental time, prolonged ovipositional period, and increased reproductive potential, which were demonstrated not only in terms of fecundity numbers, but also in terms of positive impact on reproductive parameters, including higher net reproductive rate (R 0 ), intrinsic rate of increase (r) and finite rate of increase (λ). However, our results did not completely align with the preference-performance hypothesis. First, in terms of gained biomass, pupae reared on diet treatment were heavier than those reared on other vegetable crops. Second, in terms of reproductive potential and parameters, we got mixed results, with interesting outcome in terms of total reproductive females (i.e., egg producing females) observed in maize (i.e., ear and leaves), accounting between 25 and 68% of the total females reared on those treatments. Moreover, FAW females reared on soybean and cabbage almost doubled their oviposition period compared to maize plants. In addition, females reared on cabbage, soybean, tomato, and maize ear laid 4-6 times more viable eggs compared to females reared on maize leaves and artificial diet treatments. The higher number of reproductive females reared on the vegetable crop treatments would suggest Table 4. Mean (± SE) of the net reproductive rate (R 0 ), intrinsic rate of increase (r), finite rate (λ), and mean generation time (T) of the fall armyworm, Spodoptera frugiperda fed on cabbage, maize ear, maize leaves, artificial diet, soybean, and tomato. Standard errors were estimated using 100,000 bootstrap resampling. Data followed by the same lower-case letter within a column were not significantly different based on a paired bootstrap test at 5% significance level. www.nature.com/scientificreports/ in theory, a significant population growth in these crops under field conditions. However, and although the life table studies were carried out simultaneously and under controlled conditions, in the future it is also necessary to confirm these results under field and free choice conditions. Second, in terms of reproductive parameters, the net reproductive rate (R 0 ) was higher in soybean compared to maize, whereas the intrinsic rate of increase (r) and finite rate of increase (λ) of FAW were higher on vegetable crops compared to maize (intermediate), and diet (lowest). The evaluated ecological parameters suggest that FAW populations grow faster in the vegetable crops compared to maize. Furthermore, results from this study suggest that cabbage, soybean, and tomato seemed to provide a high benefit for an appropriate offspring performance, exceeding in some cases the benefits from its usual maize-based diet (Fig. 6). In a previous study, the host preference by female moths and offspring performance were documented for the FAW by comparing the discrimination among three maize cultivars with different degrees of resistance 33 . In a choice test, it was found that moths laid more eggs on the highly resistant cultivar and least on the susceptible cultivar, whereas larval performance was poor in the former but good in the susceptible. In addition, the same study found that older FAW larvae did not show any discrimination or preference among the tested maize cultivars, nor even for real versus model maize plants, suggesting larvae may have a random foraging behavior independent of volatile cues, with low discrimination among hosts and based on visual cues. An earlier study also confirmed that plant volatiles did not influence host finding by FAW, but tactile cues may be of important during the selection of oviposition sites 29 .
In contrast, a recent paper using the age-stage two-sex life table and adult female oviposition preference for the FAW 30 found consistent results with the preference performance hypothesis on FAW, when comparing the larval performance and oviposition preference of the FAW among maize, potato and tobacco. More specifically, the FAW reared on maize had the strongest development performance (i.e. shortest development time, prolonged longevity and higher reproductive rate) compared to potato (intermediate), and tobacco (least suitable host). More recently, another paper comparing the fitness of the FAW on three solanaceous crops, found that FAW was able to complete the life cycle on pepper and tomato, with exception of eggplant. Moreover, the authors also found that FAW larvae reared on tomato had a better fitness in terms of development of duration, survival rate, fecundity, and R 0 based on the developmental duration, compared to maize 34 .
In all, a random foraging behavior that seems to be independent of volatile cues but linked to a strong tactile stimulus seems to be driving the ovipositional behavior and preference for specific host plants. However, in terms of larval development and performance, the fitness of the resulting FAW progeny is not necessarily align with the mother's choice, and therefore, results cannot be explained under a preference performance hypothesis. www.nature.com/scientificreports/ In terms of nutritional aspects and differences among treatments, chlorogenic acid, which was present in tomato and cabbage, is an antioxidant polyphenol, a conjugate of caffeic acid and quinic acid, and is recognized as an efficient defense molecule against several insect herbivores, including Helicoverpa armigera, Spodoptera litura, Spodoptera exigua, thrips, and leaf beetles 35,36 , since this compound reduce the availability of amino acids and in consequence decreases the digestibility of dietary proteins, leading to a lower nutritional value, and thus resulting in growth reduction [36][37][38] . In addition, the presence of chlorogenic acid in tomato and cabbage would partially explain the low weight of female pupae on these treatments 39 .
In contrast, previous information suggested that despite the defensive compounds such as glycoalkaloids, phenolics and terpenes, as well as other anti-nutritive proteins 40 present in tomato, FAW was reported to feed and successfully complete its life cycle on tomato leaves 41 . However, a more recent study looking at the alterations to flight muscle protein structure and flight-related fitness traits of two armyworms caused by host plant defenses 42 suggested that FAW seems to show some degree of specialization for monocot crops (grasses), since its average survival was lower when feeding on tomato compared to a related species, the southern armyworm (SAW), Spodoptera eridania, a species that feeds on leafy dicots. The study also reported that FAW was less adapted than SAW to feed on tomato and cope with its defensive secondary metabolites.
Regarding other metabolites found in this study, the effect of flavonoids on insects has been extensively discussed. The mortality of the tobacco armyworm (S. litura) was previously linked to quercetin present on peanuts 43 . In addition, kaempferol and quercetin among other flavonols have been studied for their deterrent activity against the nematodes Radopholus similis and Meloidogyne incognita and for their anti-pathogenic activities on different cereal and vegetable crops [ 44 and references therein]. Furthermore, previous studies had suggested that a reduction of various flavonoids, including kaempferol and quercetin metabolites may facilitate the puncturing of the symplast and favoring a subsequent aggregation of the whitefly, Bemisia tabaci 45 . However, given that our scope in this study is not directly related to the nutritional content and differences among plants, a more detailed study would be needed to properly provide evidence in how the presence of these nutritional/ anti-nutritional/defensive compounds directly affect the performance of the feeding larvae and the impact on their reproductive fitness.
From an insect pest management perspective, our results are relevant because understanding the discrimination of FAW female moths for suitable hosts for oviposition vs. larval performance would give us a better idea of the risk for vegetable crops located in the vicinity of maize crops in newly invaded geographical regions. High and random dispersal of larvae would potentially move to alternate hosts to complete development. In addition, the above-mentioned vegetables may serve as appropriate host to improve the reproductive fitness of the next generations, since reproductive parameters would be positively impacted compared to the benefits offered from a maize-based diet. Therefore, the use of appropriate crop rotations in the vicinity of maize crops must consider the potential damage and the need to use alternative strategies, like the use of resistant cultivars or appropriate trap crops. A similar concern has been also raised in terms of intercropping systems with maize, potato, tobacco, pepper, and tomato 30,34 since late instars of the FAW would be able to move to less preferred host plants after consuming maize, posing a serious threat to less preferred crops, but on which FAW would be capable of completing its development. At the moment, more information needs to be gathered in Taiwan to assess the risk and potential economic damage of FAW on vegetable crops.
Lastly, it is also important to discuss our results in terms of the potential threats of the invasive FAW under island conditions. As pointed out by Kairo et al. 46 and Mochizuki 47 , the greatest threat to island biodiversity and habitat loss is the growing problem of invasive species. Hence, the arrival of the FAW to Taiwan would potentially affect the niche spaces and the competition with other insect pests on the area, but also, the lack of natural enemies will provide a significant advantage for its establishment, and in addition, adaptive changes to particular climatic conditions plus isolation may provide further opportunities for the emergence of additional FAW populations, which could also expand its host range including dicot crops such as vegetables. Furthermore, economically important insect pests including the FAW are already resistant to chemical pesticides in regions of original distribution, so conventional chemical control of this insect pest may become difficult in newly invaded areas. Therefore, there is a high priority to work with integrated pest management strategies that allow the use of alternative safe control strategies to keep FAW numbers low.

Conclusions
Overall, our results suggest that despite low preference of FAW females for vegetable crops, these crops seemed to provide a high benefit for an appropriate offspring performance, exceeding in some cases the benefits from its usual maize-based diet. In that sense, the reproductive fitness of the next generations may be more positively impacted compared to the benefits offered from a maize-based diet, suggesting the need to consider appropriate crop rotations and other sustainable management strategies in the vicinity of maize crops in order to reduce the potential damage. Host plants. Commercially  Oviposition preference: no-choice test for fall armyworm females. A follow up experiment was conducted to confirm the oviposition preference of female adults with the least and most preferred host plants (i.e., tomato and maize). In this experiment, we used additional treatments to assess a potential odor masking of both plants, and how it may affect pest insects' ability to find hosts. In that sense, tomato plants coated with maize extract and maize plants coated with tomato aqueous extract were tested. For this, 100 g of tomato or maize leaves from 1.5 to 2-month-old seedlings were collected and mixed with 400 ml of water and then liquefied. The resulting aqueous extract was applied to the treated plants (i.e., maize and tomato, respectively) by coating all surfaces by dipping or brushing. The plants were allowed to dry for 4 h before the start of the experiment. A total of three seedlings per host plant were offered to four virgin pairs released at the center of each cage, for a total of 12 replications. Female adults were allowed to oviposit for 5 days, and pure honey was provided as adlibitum as food resource for FAW adults. Egg masses and egg numbers were recorded on each crop and averaged within each cage after 6 days.

Biological traits and demographic parameters of fall armyworm reared on different host plants.
Due to the invasive status of this insect pest in Taiwan, the egg batches used in this study were derived from first generation of laboratory stocks from field caught FAW females from the initially eggs provided by the Taiwan Agricultural Research Institute (TARI), Council of Agriculture (COA) as previously mentioned. Furthermore, different biological traits including the development, survival, reproduction, and ecological parameters of FAW fed on cabbage, maize ear, maize leaves, soybean, tomato, and artificial diet were compared. Egg batches received from TARI were placed under controlled conditions and newly hatched larvae were individually placed and kept in plastic containers (sauce cup; 4 cm × 3 cm; 40 ml) and provided with fresh leaves of the selected crops (approx. 1-2 g or 30 cm 2 of food) on a daily basis and until they reached pre-pupal stage. Cabbage, soybean, and tomato were given as fresh leaves, and in the case of maize, larvae were offered fresh leaves and maize ear. Each day, the development and survival of larvae were observed for each treatment. Larval stages were L1, L2, L3, L4, L5, and L6-7. Pupal weight was collected from the surviving insects, and sex determination was conducted at the pupal stage. Pupae were individually kept and after adult emergence, one female and one male were placed in plastic containers covered with plastic cups (drinking cup; 5.5 cm × 10 cm; 350 ml). Oviposition substrate used to collect the resulting eggs corresponded to the same treatment food they were offered as immatures. In the case of the artificial diets, resulting adults were offered with an unscented corrugated paper and adults from all treatments were fed with a 10% honey solution. Observations of insect biological parameters and development including duration of instars, sex, and longevity of individuals were recorded on a daily basis.
In line with this, fecundity and survival data were recorded daily until the death of each individual and for each environmental condition tested.
Nutritional quality. The nutritional quality of uninfested dry leaves of the tested vegetables was estimated through various biochemical analysis (modified), including dry matter 49 , crude fat 50 , total nitrogen 51 , flavonoids 52 , and total phenolics 53 .
Statistical analysis. One way analysis of variance (ANOVA) and Tukey's HSD post-hoc test were used to assess the oviposition preference and pupal weight of FAW among cabbage, maize, soybean, and tomato. For the second experiment (i.e. non-choice test), FAW egg masses and number of eggs recorded in the plants and the cage walls were compared by paired t-test. Differences were considered highly significant at P < 0.001; and www.nature.com/scientificreports/