Fitness comparison of Plutella xylostella on original and marginal hosts using age‐stage, two‐sex life tables

Abstract The diamondback moth, Plutella xylostella, is an important agricultural pest that severely damages cruciferous vegetables. Although previously considered a threat only to Brassica species, P. xylostella has been observed to feed on noncruciferous vegetables. Here, we established a population of P. xylostella on the pea Pisum sativum (PxP population). We compared this PxP population's performance on the pea host plant to a population (PxR) reared on the original host plant radish (Raphanus sativus) for several generations using an age‐stage, two‐sex life table and analyzed the correlations between different fitness parameters. In the 1st generation of the PxP population, survival rate of immature stage was 17%, while the survival rate of PxR was 68%; the duration of the 4th larval instar (5.30 d) and mortality (25%) of this generation were significantly longer (2.8 d) and higher (1%) than that of PxR, respectively (both p < .001). Upon long‐term acclimation, the PxP fitness improved significantly, especially that the survival rate of immature stages increased to approximately 60% in the 15th, 30th, and 45th generations. However, PxP feeding on pea exhibited poorer fitness with longer larval developmental time, shorter total life span, lighter pupa, and lower fecundity in different generations compared with PxP feeding on radish. PxP feeding on pea also showed a significantly lower intrinsic rate of increase (r), net reproduction rate (R 0), finite increase rate (λ), and longer mean generation time (T) than PxP feeding on radish in all generations tested. Significant positive correlations were observed between pupal weight and female fecundity in pea‐fed populations, and between female longevity and female fecundity in pea‐fed and radish‐fed populations. Our findings suggest that P. xylostella adaptation to pea does not improve overall fitness compared with the original host radish, making pea a marginal host for P. xylostella.


| INTRODUC TI ON
The diamondback moth, Plutella xylostella (Lepidoptera: Plutellidae), is an economically significant pest, jeopardizing various vegetables in the Brassicaceae family and causing total yield losses and pest management costs of US$4-5 billion annually worldwide (Zalucki et al., 2012). The pest status of P. xylostella has increased on Brassicaceae crops in different parts of the world recently (Furlong et al., 2013), likely due to its short life cycle, resistance to various insecticide classes, and adaptability to different environments, and the increasing agricultural demand of Brassica vegetables and oilseed crops (Furlong et al., 2013;Li et al., 2016;Talekar & Shelton, 1993).
It is mainly considered an oligophagous pest, only feeding on Brassicaceae plants. However, its dietary behavior has been questioned. For example, Löhr and Gathu (2002) found a field of sugar snap pea, Pisum sativum (Fabaceae) being destroyed by a P. xylostella population, in Kenya, indicating polyphagous behavior. This population adapted quickly to peas under laboratory conditions, with larval survival rate increasing from 2.4% in the 1st generation to 49.7% in the 4th generation (Löhr & Gathu, 2002). The mean generation time of P. xylostella also showed a negative trend over successive generations (Henniges-Janssen et al., 2011), indicating adaptability to pea.
However, other population parameters of P. xylostella feeding on pea remain unknown.
An age-stage, two-sex life table is designed to study the demographics of an insect population in response to environmental variables, including host plant (Chi et al., 2020). Population parameters such as intrinsic rate of increase (r), finite rate of increase (λ), net reproductive rate (R 0 ), and mean generation time (T) describe characteristics of a population as values. These are effective estimators to predict the potential insect population sizes on different host plants. Guo et al. (2021) used an age-stage, two-sex life table to compare the population fitness of Spodoptera frugiperda populations feeding on maize, tobacco, and potato. Golizadeh et al. (2009a) compared the life tables of P. xylostella feeding on five different cultivated Brassica host plants. Fathipour et al. (2019) and Soufbaf et al. (2010) compared the performance of P. xylostella on five and ten canola cultivars, respectively. Nikooei et al. (2015) compared P. xylostella performance on different genetically manipulated Brassica plants (canola's progenitor, two cultivated canola cultivars, one hybrid, one gamma-ray mutant, and one transgenic genotype).
Extensive studies have reported the correlations between various fitness parameters among insect populations. Development time was correlated with body size in insects, but these correlations could be positive, zero, or negative (Nijhout et al., 2010;Teder et al., 2014). Fecundity and body weight or developmental rate were positively correlated in 60 insect species from eight orders, including Coleoptera and Lepidoptera (Calvo & Molina, 2005;Coyle et al., 1999). In addition, studies showed that diet can change the correlations between these parameters (Gebhardt & Stearns, 1988).
For example, the nutrient quantity and quality of insect foods could influence the correlation between development time and body size (Kause et al., 2001;Uhl et al., 2004). However, whether these correlations apply to P. xylostella feeding on different host plants remains unknown.
The objective of this study was to compare the population demographics of P. xylostella after long-term acclimation to its marginal host plant, pea, and to its original host plant, radish . Using an age-stage,   two-sex life table, we examined the life history parameters of survi-vorship, development time, and fecundity. We compared these parameters between different generations of P. xylostella feeding on radish and on pea, and studied the correlations between pupal weight and female fecundity, pupal weight and adult longevity, duration of larval stage and pupal weight, and female longevity and female fecundity.

| Insects and plants
Plutella xylostella pupae were initially collected from Brassica fields in July 2004 in the Fuzhou suburban area (26.08°N, 119.28°E) and reared in the laboratory on "Nanpanzhou" cultivar radish (Raphanus sativus) for more than 200 generations (referred to as PxR population). To establish a PxP population, PxR eggs were artificially placed on "Purple Flower-Soft Pod" cultivar pea (Pisum sativum) leaves. We named the 1st generation of P. xylostella after the host shift as PxP-1, and so on in a similar fashion (PxP-n). Upper generations of the PxP population feeding on radish were named PxP-n_R. Larvae were left to feed on radish or pea under controlled conditions of 23 ± 1°C, 65 ± 5% relative humidity (RH), and 16-hr light: 8-hr dark photoperiod. Adults were provided with a 10% w/v honey-water mixture.
Radish was cultivated in rectangular plastic trays (420 × 320 × 100 mm) with nutrition soil, and pea was cultivated in disposable nutrition bags (16 × 14 mm) with peat soil. Plants were kept in fine transparent cages with 0.1 mm mesh screens in a walk-in growth chamber at 23 ± 1°C and 65 ± 5% RH, under a 16-hr light: 8-hr dark photoperiod. Leaves of radish at 1 week and pea at 4 weeks were used to feed P. xylostella larvae.

| Investigation of fitness parameters
The development time, survival, and reproduction of P. xylostella feeding on radish and pea leaves were investigated and compared. The eggs of the PxR population and the 14th, 29th, and 44th generations of the PxP population were kept separately on radish and on pea, and life tables (PxR, PxP-1, PxP-15, PxP-15_R, PxP-30, PxP-30_R, PxP-45, and PxP-45_R) were studied ( Figure 1). One hundred newly laid (<2 hr) light yellow eggs were placed in plastic film dishes measuring 9 cm in diameter. The neonates were counted and individually transferred to film dishes measuring 60 mm in diameter, with fresh radish or pea leaves. Fresh leaves were provided daily, and uneaten leaves were removed. The survival and development times of each developmental stage were recorded daily. One pair of newly emerged female and male moths was confined in a plastic chamber (29 × 38 × 32 mm) capped by a fine mesh for ventilation. Chambers were lined with film papers with grooves serving as oviposition substrates. Eggs laid by each female moth were counted daily until it died. Parameters such as survivorship, fecundity, oviposition period, and total life span were recorded. Pupal weight was also recorded at the second day after pupation with 0.1 mg measurement accuracy (OHAUS CORPORATION ® AR224CN balance, China). Each newly hatched larva was considered as one replicate.

| Data analysis
Life history data of PxR and PxP populations were analyzed using TWOSEX-MSChar (V2018.05.04) (Chi, 2019;Chi & Liu, 1985;Chi et al., 2020). The age-stage survival rate (S xj ) was calculated based on the age-stage-structure matrix (Chi & Liu, 1985). The formula for the parameters was calculated as follows: where l x is age-specific survival rate, f xj is age-stage specific fecundity, m x is age-specific fecundity, R 0 is net reproductive rate, r is intrinsic rate of increase, λ is finite rate, and T is mean generation time.
The variances and standard errors of these life history parameters were calculated 100,000 times using the bootstrap technique.
Data for each parameter were analyzed separately using paired bootstrap tests, except for pupal weight, which was analyzed using one-way ANOVA followed by Tukey's test.
Pearson correlation coefficients were used to analyze the correlations between different fitness parameters (pupal weight and female fecundity, pupal weight and adult longevity, duration of larval stage and pupal weight, and female longevity and female fecundity) using packages ggplot2, ggpubr, and ggpmisc in the R software environment (V3.6.3; R Core Team, 2017). SigmaPlot (V12.0) was used for creating scientific graphs.

| Development, survivorship, and reproduction
The development time for each immature stage, adult longevity, total life span, pupal weight, and female fecundity of PxP feeding on radish and on pea in different generations are shown in Table 1. In the 1st generation of the PxP population, only 49 neonates out of 100 eggs successfully developed to the 2nd instar, compared with 82 in the PxR population. The duration of the 4th instar of PxP-1 (5.30 d) was longer than that of PxR (2.80 d) (p < .001). The mortality rate of PxP-1 was significantly higher (25%) than that of PxR (1%) (p < .001). The duration of the larval stage from the 1st to 4th instar was significantly shorter in radish-fed groups than in pea-fed groups (all p < .05). The duration of the pupa stage of PxP larvae feeding on pea in the 30th Note: Different capital letters within a separate column indicate significant differences between different host plants in the same generation and different lowercase letters within a row indicate significant differences between different generations in the same host plant using the paired bootstrap test (p < .05), while the same letters represent no significant difference. One-way ANOVA followed by Tukey's test was used to analyze pupal weight. PxR, P. xylostella reared on radish; PxP-1, PxP-15, PxP-30, and PxP-45: the 1st, 15th, 30th, and 45th generations of PxP population; PxP-15_R, PxP-30_R, and PxP-45_R: the 15th, 30th, and 45th generations of PxP population feeding on radish.
eggs in the 45th generation) (all p < .01). The mean female fecundities of PxP feeding on pea fluctuated with generations, with PxP-15 exhibiting the highest value (94.83 eggs). Meanwhile, the mean female fecundities of PxP feeding on radish showed an inverse correlation with generation.

| Age-stage survival rate and fecundity
The age-stage survival rate (S xj ), defined as the survivorship to age x and stage j, was plotted in Figure 2. A mortality rate of nearly 60% was observed in the 1st instar larvae of the 1st generation of the PxP generation, and 60.6% of immature stages in the 45th generation.
The observed age-specific survival rates (l x ), the age-stage specific fecundity (f xj ), and age-specific fecundity (m x ) are shown in The m x curve, describing the start times and duration of the reproductive phase, began at age 22nd d in PxP feeding on pea in the 1st generation, which was 3 days later than that of PxR. In addition, the maximal daily oviposition rate of PxP feeding on pea occurred at an average age of 30th d, with mean fecundity of 4.45 eggs per female, which was lower than that of PxR (20th d, 13.07 eggs per female). In the later generations, the duration, reproduction, and maximal daily oviposition showed no evident difference between the two host plants.

| Population parameters
The intrinsic rates of increase (r), finite rates of increase (λ), net reproductive rates (R 0 ), and mean generation times (T) for various groups are shown in Table 2. The r, λ, and R 0 values of PxP feeding on radish were significantly higher than those of PxP feeding on pea in all generations (all p <.05). In addition, the observed T values were significantly shorter in PxP feeding on radish than in PxP feeding on pea in all generations (all p < .05). The r, λ, and R 0 values of PxP feeding on pea or on radish in high generations were significantly higher than those in the 1st generation (all p < .05), but the R 0 values of PxP feeding on radish showed no difference across generations. In addition, the T values were significantly lower in high generations of the PxP population than in lower generations (all p < .05).

| Correlation between fitness parameters
Correlations between different fitness parameters are displayed in  and Taizhong 13) and leaf quality. Eventually, we found fresh leaves of "Purple Flower-Soft Pod" cultivar were generally suitable with 17% P. xylostella survival rate of immature stages in the 1st generation. One possible reason is the bottom-up effects of Brassica cultivars or genotypes on the performance of P. xylostella (Fathipour et al., 2019;Fathipour & Mirhosseini, 2017;Kianpour et al., 2014;Nikooei et al., 2015;Soufbaf et al., 2010). Löhr and Gathu (2002) found that the survival rate increased to 49.7% in the 4th generation. In the present study, the survival rate reached 60% by the 15th generation and thereafter remained stable, indicating quick adaptability of P. xylostella to peas. These differences may be caused by Sugar Pod" cultivar when consistently observed for more than 50 generations. In the present study, we also found that as acclimation time increased, the mean generation time shortened further indicating high adaptability of P. xylostella to peas. Moreover, we found that the duration of the 4th instar was prolonged and that the mortality was high in the 1st generation of P. xylostella feeding on pea, compared with that of PxR. A future study investigating the transcriptome as it relates to adaptability will further explain these observations.

| D ISCUSS I ON
Reproduction is a critical biological indicator, especially in insects that typically produce hundreds of offspring that survive without parental protection (Harano, 2011 (Golizadeh et al., 2009b), 27 eggs on NSA2 versus 5 eggs on Red-Rocky (Fathipour et al., 2019), and 61 eggs on Opera versus 8 eggs on PF (Nikooei et al., 2015). In the present study, the average fecundities of PxP feeding on pea were significantly lower than that of individuals feeding on radish in all generations. One possible explanation is that oviposition stimulators are low in pea. For example, glucosinolates can attract and stimulate P. xylostella to lay more F I G U R E 2 Age-stage survival rates (S xj ) of PxP population (Plutella xylostella reared on pea) feeding on radish and pea in different generations. L1, 1st instar; L2, 2nd instar; L3, 3st instar; L4, 4th instar. PxR, P. xylostella reared on radish; PxP-1, PxP-15, PxP-30, and PxP-45: the 1st, 15th, 30th, and 45th generations of PxP population; PxP-15_R, PxP-30_R, and PxP-45_R: the 15th, 30th, and 45th generations of PxP population feeding on radish F I G U R E 3 Age-specific survival rates (l x ), female age-stage specific fecundity (f xj ), and age-specific fecundity of total population (m x ) of PxP population (Plutella xylostella reared on pea) feeding on radish and pea in different generations. PxR, P. xylostella reared on radish; PxP-1,  eggs and shorten the prelaying period, but the pea plant has low glucosinolates content (Badenes-Perez et al., 2014, 2020. Another possible explanation is that pea lacks sufficient nutrients for the reproduction of P. xylostella. Darwin's fecundity advantage hypothesis suggests that larger females can reproduce more offspring (Afaq, 2013;Andersson, 1994;Darwin, 1874;Honěk, 1993). In the present study, PxP feeding on pea showed a significant positive correlation between pupal weight and fecundity, in accordance with insect species such as Streblote panda (Calvo & Molina, 2005), Peregrinus maidis (Wang et al., 2006), Sitobion avenae, Rhopalosiphum padi, and Schizaphis graminum (Hu et al., 2015). In contrast, a longer development period has been correlated with larger individuals (Nijhout et al., 2010;Teder et al., 2014). In addition, Chilo suppressalis (Huang et al., 2018) and Ostrinia furnacalis (Xia et al., 2019) showed a negative correlation between larval development time and pupal weight. However, no correlation was found between larval development time and pupal weight in our study. This variation in observations supports the conclusion of a previous study that development time and body mass could exhibit positive, negative, or zero correlation (Nijhout et al., 2010). Female fecundity and adult longevity are generally considered to have a "trade-off" relationship: the cost of reproduction shortens adult longevity (Bell, 1986;Williams, 1966). This has been demonstrated in Chorthippus brunneus (De Souza Santos and & Begon, 1987), Drosophila melanogaster (Sambucetti et al., 2015), and Helicoverpa armigera (Thyloor et al., 2016). However, we observed a significant positive correlation between female longevity and fecundity ( Figure 4).
In conclusion, we established a stable P. xylostella population on a marginal host plant upon long-term acclimation. However, the observed fitness in PxP population feeding on pea was lower than that feeding on its original host, radish. Correlations between different fitness parameters were influenced by host plants. Our results may facilitate the prediction of pest behavior when the preferred host is absent.

CO N FLI C T O F I NTE R E S T
None declared.