Imidacloprid soil drenches indirectly weaken the selection and predatory ability of the coccinellid predator Propylea japonica (Coleoptera: Coccinellidae)

Abstract

Systemic insecticides are widely used to control sap-feeding pests, and their risks for natural enemies have become well-assessed in the past decades. However, most of these risk assessments focused on the direct effects when natural enemies come in contact with the sprayed leaves or forge contaminated prey. Consequently, the indirect impacts of systemic insecticides on regulating plant volatile organic compounds (VOCs) and how do these affect natural enemies have been overlooked. In the current study, we examined the indirect impact of imidacloprid soil drenches on prey selection behavior and functional response of the ladybird Propylea japonica. Our findings demonstrated that systemically applied insecticides could negatively affect natural enemy behaviors by disturbing VOCs emitted from plants. Compared with those exposed to fresh air and VOCs emitted from water-treated plants, P. japonica adults exposed to VOCs emitted from imidacloprid-treated plants were less chosen, and their attack rate (a) and handling times (h) markedly decreased, which potentially weaken biological control service. Meanwhile, we also found Frankliniella intonsa was less frequently chosen for imidacloprid-treated plants. Therefore, utilizing systemic imidacloprid to control pests needs to be cautiously considered when integrated pest management strategies are set up.

Introduction

Integrated pest management (IPM) combines biological and chemical control to minimize the non-target effects of chemical management, such as insecticide resistance and disruption of ecosystem services^1–3 and provides an alternative approach to conventional, chemical insecticides on management. Yet, chemical control is still the most efficient measure in production for controlling thrips^4,5most commonly with the application of synthetic insecticides such as imidacloprid. However, since this application of insecticides often results in subsequent pest outbreaks due to the elimination of natural enemies, finding alternative measures to chemical pest control has received increasing research attention over the past decade. Nevertheless, a thorough risk assessment from a variety of perspectives should be the first step towards this direction.

Neonicotinoid insecticides are the most commonly used insecticides in the world and are effectively control a broad range of pests^6,7. Meanwhile, long-term application of neonicotinoid systemic insecticides has resulted in a series of harmful side effects on natural enemies, especially the predatory enemies that may be adversely affected through direct contact^6,8feeding on contaminated prey^9,10pollen^11,12 and honeydew^13,14. However, to our best knowledge, evaluations of the side effects have mainly focused on the direct effects of insecticides on natural enemies, particularly when they contact the sprayed leaves or forage on contaminated prey, and the non-lethal effects of different insecticide treatments on survival rate, reproduction, and functional response to prey density of predators were also well-investigated^15–17. For instance, a study found that thiamethoxam significantly reduced adult emergence, fecundity, and fertility of the Coccinella septempunctata Linnaeus, 1758 (Coleoptera: Coccinellidae) at sublethal doses¹⁸. Insecticide applications (diazinon, fenitrothion and chlorpyrifos) caused a decrease in the attack rate and an increase in the handling time of exposed bugs Andrallus spinidens (Fabricius, 1787) (Hemiptera: Pentatomidae) at sublethal doses¹⁹. Cypermethrin reduced prey consumption and decreased the fecundity and fertility drastically of Alpaida veniliae (Keyserling, 1865) (Araneae, Araneidae) in sublethal concentrations²⁰. Harmonia axyridis (Pallas, 1773) (Coleoptera: Coccinellidae) exposed to thiamethoxam at LC₅₀ showed a decreased attack rate and an increased handling time⁸. Studies have also confirmed negative impacts of neonicotinoids on other beneficial arthropods^21–23. Therefore, the application of neonicotinoid insecticide in IPM schemes requires to be cautiously considered because even sublethal concentrations of these insecticides are likely to be harmful to enemy population growth or predation ability.

Systemic insecticides, such as imidacloprid soil drenches, absorbed by plants via their roots and then are transported throughout the tissues of the plant, which tends to reduce exposure to non-target organism. Specifically, it would decrease the risk of direct contact to insecticides and residues on leaves for carnivorous predators. However, even with systemic insecticide applications there are concerns of risks to omnivorous predators^15,24 and pollinators^6,25. To avoid the negative side effects of conventional foliar spraying, systemic neonicotinoids are commonly used to control important sap-feeding pests, such as thrips, whiteflies, and aphids^26–30. Previous studies have demonstrated that systemic insecticides were highly effective in controlling Asian citrus psyllid Diaphorina citri Kuwayama, 1908 (Hemiptera: Psylloidea) in Curry Leaf³¹thrips Megalurothrips usitatus (Bagnall, 1913) (Thysanoptera: Thripidae) in cowpea³⁰ and whitefly Bemisia tabaci (Gennadius, 1889) (Hemiptera: Aleyrodidae) in tomato plants²⁹. Though they tend to reduce the exposure of non-target organisms and weaken the direct effects on natural enemies, the overuse of systemic insecticides has also raised concerns over a series of side effects. Previous studies showed that imidacloprid soil drenches can negatively impact zoophytophagous natural enemies, including decreased longevity, fecundity and weight^15,24. On the other hand, systemic insecticides can also interfere with the metabolites of the host plants to which they were applied, which leads to changing the volatile organic compounds (VOCs) emitted from plants both in terms of quantity and quality^7,28,32,33. This can be attributed to the oxidative cleavage of neonicotinoids (e.g.imidacloprid) among the carboxylic acid metabolites in plants, which activate SA-associated plant defense responses^34,35 or suppress the critical synthesis-related genes³³. Since VOCs emitted from plants are commonly used by herbivores/enemies when searching for suitable host plants, the predatory behavior of enemies will also likely be disturbed^36,37. Though numerous studies have demonstrated that herbivore-induced plant volatiles attract ladybird beetles^38,39few have simultaneously assessed the impact on the predators’ functional response. Study showed that herbivore-induced rice VOCs would attract wolf spiders Pirata subpiraticus (Bösenberg & Strand, 1906) (Araneae: Lycosidae) and Pardosa pseudoannulata (Bösenberg & Strand, 1906) (Araneae: Lycosidae) and increase their daily predation capacity or shorten their predation latency⁴⁰. Thus, the predatory behavior of natural enemies can be potentially disturbed by VOCs emitted from plants on which systemic neonicotinoid insecticides were applied, which would compromise the biological control performance of these enemies²⁴. Therefore, evaluating odor-induced changes to coccinellid functional responses is of significant practical importance in biological control, especially during the early colonization phase of prey or the pre-flowering stage in the field. However, this indirect effect of systemic insecticides on the predatory ability, specifically for functional response, of natural enemies through modifying VOC profiles is still largely unknown.

Pomelo (Citrus grandis (L.)) is a popular citrus fruit that is widely cultivated around the world. In China, more than 50% of the pomelo is produced in Fujian Province and exported to more than 50 countries or regions⁴¹. However, cultivated pomelo is attacked by multiple insect pests, especially thrips Frankliniella intonsa that cause wounds on flower buds, flower organs and young fruits^42–44which would decrease the economic value. Coccinellid predators play a key role in controlling several pests in citrus orchards in China⁴⁵particularly, the locally common ladybird, Propylea japonica (Thunberg, 1781) (Coleoptera: Coccinellidae) contributes greatly to suppressing sap-sucking pests in both natural and agricultural landscapes. Meanwhile, the application of imidacloprid soil drenches for thrips control on honey pomelo has increased significantly, particularly within IPM orchards. Thus, by investigating the indirect effects of systemic insecticides on P. japonica, our work aims to provide a basis for future guidelines for optimizing IPM programs involving both natural enemies and systemic insecticides in controlling sap-sucking pests. Here, we assess (1) how imidacloprid soil drenches at recommended concentration affect the prey selection behavior of P. japonica, and (2) how imidacloprid soil drenches affect the functional response of P. japonica on pest Frankliniella intonsa (Trybom, 1895) (Thysanoptera: Thripidae). We hypothesize that:

1. imidacloprid soil drenches to plants decrease the selection behavior by P. japonica;

2. imidacloprid soil drenches negatively affect the functional response of P. japonica.

Results

Prey selection behavior of ladybird Propylea japonica in different treatments

Volatile organic compounds (VOCs) emitted from imidacloprid-treated plants were less chosen by ladybirds, which induced about 20.0% of the recruitments, that was significantly less than those for VOCs emitted from water-treated plants (χ² = 7.5385, df = 1, P = 0.006, Fig. 1A). Similarly, about 26.7% of the recruited ladybirds were attracted by VOCs emitted from imidacloprid-treated plants, which was significantly less than those attracted by VOCs emitted from fresh air (χ² = 4.1667, df = 1, P = 0.0412, Fig. 1A). However, there was no significant preference for ladybird between VOCs emitted from water-treated plants and fresh air ( χ² = 0.0370, df = 1, P = 0.8474, Fig. 1A).

Fig. 1.

Olfactory analysis results for Propylea japonica (A) and Frankliniella intonsa (B) to response choice to different treatments. Comparisons are made among treatments by one-way chi-squared test. “*” indicates a significant difference at P < 0.05 level, “**”at P < 0.01 level, and “***” at P < 0.001 level. “NS” represents no significant difference at P > 0.05 level.

Likewise, thrips were less chosen by VOCs emitted from imidacloprid-treated plants. Only 18.7% and 22.0% of the overall number included in the experiment were attracted to VOCs emitted from imidacloprid-treated plants, which was significantly less than that to VOCs emitted from water-treated plants (χ² = 16.802, df = 1, P = 0.00004 < 0.001, Fig. 1B) and fresh air (χ² = 13.815, df = 1, P = 0.0002 < 0.001, Fig. 1B). However, compared with ladybirds, thrips markedly preferred VOCs emitted from water-treated plants to fresh air (χ² = 16.697, df = 1, P = 0.00004 < 0.001, Fig. 1B).

Functional response of ladybird P. japonica at different treatments

Propylea japonica adults exhibited similar functional responses to all treatments. The average number of thrips consumed per predator increased as thrips density increased. Consumption was the highest at initial densities of 100 individuals (Table 1). In addition, imidacloprid exposure conditions did not affect the consumption of P. japonica at high densities of F. intonsa (80, 100, 120 individuals) but markedly decreased it at low densities of F. intonsa (40, 60 individuals, Table 1).

Table 1.

Consumption (means ± SE) of Propylea Japonica at different densities of Frankliniella Intonsa pupals.

Treatment	Thrips density							χ2	df	p
	20	40	60	80	100	120	140
Fresh air	13.14 ± 1.28	25.29 ± 1.44	35.20 ± 1.83	38.25 ± 2.58	49.60 ± 4.34	40.80 ± 4.14		28.88	5	< 0.001
Water	7.60 ± 0.68	22.00 ± 2.00	36.00 ± 2.77	33.67 ± 3.49	46.50 ± 5.88	41.80 ± 3.33	38.20 ± 3.28	23.39	6	< 0.001
Imidacloprid	11.75 ± 1.01	17.00 ± 0.84	21.67 ± 1.50	27.17 ± 2.82	50.80 ± 2.78	44.80 ± 1.16	48.20 ± 5.42	34.82	6	< 0.001
χ2	9.16	8.88	10.09	5.62	0.64	0.31
df	2	2	2	2	2	2
p	0.01	0.01	< 0.01	0.06	0.73	0.86

According to the logistic regression analysis, P. japonica adults exhibited a type II functional response that best fitted a quadratic model with a negative first-order term at all treatments (Table 2; Fig. 2). The estimated parameters indicated that the attacking rate changed from 1.528 h^− 1 in fresh air to 0.687 h^− 1 in the treatment with VOCs emitted from imidacloprid-treated plants. The handling times were calculated to be 0.014 h, 0.013 h, and 0.007 h, at fresh air, VOCs emitted from water-treated plants, and VOCs emitted from imidacloprid-treated plants, respectively (Table 2).

Table 2.

Effects of different treatments on estimates and respective standard error of the linear coefficient of logistic regression analysis and functional response parameters of Propylea Japonica towards Frankliniella Intonsa model.

Treatment	Linear Coefficient		Type	Parameter	Estimate	SE	z-Value	Pr(z)
	P	Pr(z)
Fresh air	− 0.013	< 0.0001	Type II	a	1.528	0.150	10.193	< 0.0001
				h	0.014	0.001	11.507	< 0.0001
Water	− 0.010	< 0.0001	Type II	a	1.099	0.107	10.276	< 0.0001
				h	0.013	0.001	10.427	< 0.0001
Imidacloprid	− 0.004	< 0.001	Type II	a	0.687	0.063	10.883	< 0.0001
				h	0.007	0.002	3.944	0.00008

a, attack rate; h, handling time.

Fig. 2.

Functional response of Propylea japonica to different densities of Frankliniella intonsa at three exposure conditions. Dots represent the number of preys consumed at each initial prey density. Black lines were predicted using Roger’s random predator equation; the shaded areas represent the limits of the bootstrapped 95% CIs.

The comparison of type II functional response parameters showed that the attack rate of P. japonica exposed to VOCs emitted from imidacloprid-treated plants was significantly reduced, and was lower than of those exposed to either VOCs emitted from water-treated plants or fresh air (Table 3). Similarly, the handling time of P. japonica exposed to VOCs emitted from imidacloprid-treated plants was markedly shorter than of those exposed to either VOCs emitted from water-treated plants or fresh air. However, no significant differences in the handling time of P. japonica were found between VOCs emitted from water-treated plants or fresh air (Table 3).

Table 3.

Estimates of the differences in attack rate (Da) and handling time (Dh) between different treatments based on the difference method.

Treatment compared	Parameter	Estimate	SE	z-value	Pr(z)
Fresh air vs. imidacloprid	Da	0.841	0.163	5.168	< 0.0001 ***
	Dh	-0.007	0.002	3.384	0.00072 ***
Fresh air vs. water	Da	0.429	0.184	2.332	0.020 *
	Dh	0.002	0.002	0.416	0.678
Water vs. imidacloprid	Da	0.412	0.124	3.315	0.00092 ***
	Dh	0.007	0.002	2.992	0.00277 **

Note: “* " indicates a significant difference at P < 0.05 level, “* *"at P < 0.01 level, and “* ** “at P < 0.001 level.

Discussion

Previous studies have confirmed the negative effects of systemically applied neonicotinoid insecticides on coccinellid predators directly contacting the sprayed leaves or foraging on contaminated prey^16,46,47. Here, to gain a comprehensive understanding of the impacts of systemic imidacloprid on coccinellid predators, we tested the indirect effects of systemically applied imidacloprid on the predatory behavior of the generalist predator P. japonica.

Our findings confirmed our first hypothesis, that P. japonica does not prefer VOCs emitted from imidacloprid-treated plants. Unlike zoophytophagous predators which would simultaneously feed on plant sap and insect prey, the carnivorous P. japonica would not be affected by directly feeding on imidacloprid-treated plants, which further supports that, indeed, systemic insecticide-induced VOCs repel carnivore natural enemies and not any other factors. Since herbivore-induced VOCs attracting natural enemies to pest-feeding locations is an important component of biological control^40,48this effect likely hampers biological control. However, this repellent effect also offers ecological benefits by reducing coccinellid exposure risk—both through direct toxic contact and consumption of contaminated prey or pollen. This is particularly important for zoophytophagous predators, such as Orius insidiosus (Say, 1832) (Hemiptera: Anthocoridae) and Podisus maculiventris (Say, 1832) (Hemiptera: Pentatomidae)^15,24as repellency may additionally lower their risk of feeding on contaminated plant tissues. Meanwhile, we also found a repelling effect of VOCs emitted from imidacloprid-treated plants on the pest F. intonsa. This is in line with a previous study that showed that imidacloprid-treated Ficus microcarpa L.fil., (Rosales: Moraceae) plants attracted significantly less Gynaikothrips uzeli (Zimmermann, 1900) (Thysanoptera: Phlaeothripidae) than those from untreated plants²⁸. The repelling effect of VOCs emitted from imidacloprid-treated plants could due to imidacloprid treatment suppress the emission of herbivore-induced volatiles that are crucial for attracting predators, leading to apparent repellency. Previous study suggested that imidacloprid treatment can have a negative effect on the emission of VOCs due to suppressing the critical jasmonic acid synthesis-related gene, consequently affecting plant indirect defense^33,44. Simultaneously, the altered VOCs can sometimes directly repel or deter certain pests. Such as methyl salicylate, a volatile form of salicylic acid, can have repellent effects on insect herbivores^49,50. Although insecticides generally reduce pest abundance by direct mortality or by reducing the fecundity of the target species, our results highlighted an indirect way of pest control by the repelling effects of VOCs. Thus, since systemically applied imidacloprid simultaneously repels pests and natural predators by mediating VOCs emitted from plants, it makes it difficult to predict the net impact of its use on pest control output.

Our findings showed that exposing to VOCs emitted from imidacloprid soil-drenched plants markedly decreased the consumption of P. japonica at low densities of F. intonsa (40, 60 individuals). It is highly probable that the exposure to the VOCs emitted from imidacloprid soil-drenched plants caused disorientation in P. japonica, which made them unable to develop accurate attack, capture or feeding abilities at lower density. While when experiments were carried out with the high prey density of F. intonsa (80, 100, 120 individuals), there were practically no statistical differences among treatments. This result suggests that the high availability and abundance of prey can “mask” the various effect of VOCs⁵¹. Ladybird females exhibited a typical type II functional response in all treatments, which is the most common response in coccinellids^46,52 and other enemies^53–55. The regression curves suggested increased predation with increased density of F. intonsa until reaching a plateau (100 individuals), and then, followed by a gradual decline⁵⁶. The decline in predation at high prey densities could be due to predator saturation effects⁵⁷digestive constraints⁵⁸or optimal foraging behavioral shifts⁵⁹which may amplify the variability. At the same time, both the attack rate and handling time of P. japonica significantly decreased after exposing the adult beetles to VOCs emitted from plants treated with imidacloprid soil-drenching, which can reduce the predator’s ability to control agricultural pests. This is consistent with a previous study reporting that the attack rate and handling time of H. axyridis were markedly reduced in systemic treatments¹⁶. The reduction in attack rate and handling time might be the result of an antifeedant effect caused by some specific VOCs that were emitted from imidacloprid-treated plants^60,61.Furthermore, the previously observed repulsive effects may also disturb the navigation/orientation of the enemies which as well can result in decreased attack rate and handling time. However, an opposing result, that systemic treatment increased the handling time of ladybird S. japonicum on whitefly B. tabacci was reported¹⁷ for which the reason may lay in that neonicotinoids in contaminated preys caused neurotoxic symptoms, such as trembling, paralysis and loss of coordination, which increased the resting and preening time of enemies. In conclusion, the functional responses of P. japonica were negatively affected by imidacloprid soil drenches.

However, because of the lack of field experiments, our results could have over- or underestimated the field-realistic negative effects on the functional responses of P. japonica and thus, further, field-based, studies are needed to verify the indirect impacts of systemic imidacloprid on P. japonica in honey pomelo gardens. Since field conditions are often suboptimal for natural enemies, we also recommend estimating the insecticide-dependent predator biocontrol activity in combination with other stressors (e.g., temperature, starvation, varying photoperiods)^52,62. Moreover, further studies are necessary to identify the specific active compounds emitted from imidacloprid-treated plants that influence thrips and ladybird behavior. On the other hand, although our results demonstrate that imidacloprid soil drenches can indirectly reduce coccinellid predation capacity, field-based predator-prey dynamics between coccinellids and thrips are unlikely to be mediated solely by imidacloprid-induced changes in plant volatiles. Rather, these dynamics may be more strongly influenced by direct acute^6,8 or subacute toxic effects from consuming contaminated prey or pollen^10,12. Future research should quantify the relative contributions of these direct and indirect impacts.

In summary, our study confirmed that systemic imidacloprid indirectly affects the predatory behavior of the coccinellid predator P. japonica. We provide evidence that VOCs emitted from systemic imidacloprid treated plants were less chosen by P. japonica and impair its predatory ability which can weaken its biological control service. Therefore, utilizing systemic insecticides to control pests needs to be cautiously considered when integrated pest management strategies are set up.

Methods

Insect colonies

Laboratory colonies of thrips F. intonsa and ladybird P. japonica were collected from a honey pomelo plantation and the surrounding grassland of Ageratum conyzoides L. (Asterales: Asteraceae), respectively, in Pinghe County, Fujian Province. The thrips population has been reared on kidney beans since 2021. The kidney beans were soaked in water and then air-dried. Insects were kept in artificial climate incubators at 27 ± 1 °C and 60 ± 5% RH, under a photoperiod of 16 L: 8D. P. japonica was reared on aphid-infested beans planted in a fine mesh net plastic cage (50 × 50 × 50 cm). Young P. japonica adults, isolated for 24 h in a Petri dish (5 cm diameter), were used for the experiments.

Soil drenching treatment

Pomelo seedlings were planted one per pot. Plants at the flower-bud stage were selected for this experiment. Imidacloprid (active ingredient = 10%, Zhejiang Haizheng Chemical Co., Ltd) was used at the field recommended rate against thrips in vegetable and fruit crops (200 mg of active ingredient per liter of water, according to the label indications). One hundred ml of diluted insecticide solution was applied to the substrate of each pomelo plant seedling. The same amount of pure water was used in the control treatment. Plants were used for our experiment after 24 h of drenching period.

Olfactory bioassays

To investigate the response of F. intonsa and P. japonica to the odor sources, double-choice tests in closed Y-tube olfactometer systems were used, following the method described by Lin et al. (2018). Three combined treatments were used to test the variable response of F. intonsa and P. japonica to VOCs emitted from different treatments. (1) imidacloprid-treated plant versus fresh air; (2) imidacloprid-treated plant versus water-treated plants; and (3) water-treated plants versus fresh air. Only adult insects were used in our olfactory tests. Treated plants were placed in dome-shaped VOC collection chambers (35 × 25 cm) as described by Lin et al. (2018) and purified air was blown into the chambers from the inlet at a rate of 200 ± 10 ml/min. The behavioral responses of adult insects were monitored over 300s after individually releasing them at the entry of the Y-tube. The olfactory choice was recorded when the adult thrips or ladybirds reached more than 1/2 of the length of the arm, and no choice response was recorded if the insects did not reach 1/2 of the length of an arm. For each of the three combined treatments, 50 adult female thrips and 10 adult female ladybirds were used. As their predation is often higher than that of males, only female ladybirds were used in our study. The olfactory bioassay was repeated three times.

Functional response bioassays

Adult female ladybirds were individually placed into 15 ml plastic vials and starved for 24 h. Each beetle was offered F. intonsa pupae in seven different prey densities in a Petri dish (80 mm Ø, 15 mm high) with a drilled hole (40 mm Ø) on the top half of the dish. Prey densities of either 20, 40, 60, 80, 100, 120 or 140 pupae were available for the beetles for a 24 h period. All dishes were put into dome-shaped chambers modified from Lin et al. (Fig. 3) for 24 h, then unconsumed preys were counted under the microscope to quantify predation. Between 5 and 8 replicates were tested at each prey density in all three treatments. The experiments were conducted under ambient conditions (25 °C temperature, 60% relative humidity) regulated using an air conditioner and humidifier.

Fig. 3.

Schematic diagram of the predation experiment. (A) Purifying fresh air cups; (B) Collecting VOCs from treated plants; (C) Observing predation activity.

Data analysis

Olfactory bioassays

The number of ladybirds responding to the odor sources was recorded to test how the prey selection behaviour changes with VOCs emitted from different treatments. A one-way chi-squared test with the null hypothesis of equal proportion (50: 50) was performed to analyze the preferences of adult insects in the dual-choice behavioral assay.

Functional response bioassays

Testing functional responses is a useful tool to evaluate the efficacy of natural enemies in controlling pests in response to ecological change^63,64. It indicates how the individual predation rate changes with increasing prey availability and are commonly used to evaluate the predatory capacity of predators⁶⁵. The attack rate and handling time are used to explain the magnitude of functional responses of natural enemies to their prey⁶⁶. These parameters determine the foraging behavior of a predator and explain how efficient a predator is against a given density of prey.

Firstly, prey consumption were compared using Kruskal-Wallis non-parametric tests, then the functional response analysis was performed using the R package “frair”⁵⁶. Logistic regression of the proportion of consumed thrips as a function of the initial density was used to determine the type of functional response. The type II functional response is characterized by a significant negative first-order term in the logistic regression in which the prey consumption declines monotonically with the increasing prey density. In contrast, a type III functional response is described by a significant positive first-order term and followed by a negative second-order term in the logistic regression in which the proportionate prey consumption is positively density-dependent.

Type II functional response model was described by Rogers⁶⁷:

where is the number of consumed prey, is the original prey density, is the attack rate, is handling time and T is experimental period.

Type III functional response model was described by Real⁶⁸:

where is the number of consumed prey, is the original prey density, is the attack rate, is handling time, T is experimental period, is scaling component (q = 0 in Type II and q > 0 in Type III), is search coefficient.

For these models, the assumptions of homogeneity of variance were tested using Levene’s test and when necessary, bootstrapping was used to reflect the likely range of values by calculating 95% confidence intervals (CIs) using frair_boot() function. To select the best model between type II and type III, frair_test() function was used to distinguish the overall shape of functional response curves. This method fits a polynomial logistic function of the proportion of prey consumed (/) that is suitable to detect slight differences in curve shapes between type II and type III⁶⁹.

where is the number of preys consumed, is the initial prey density, and , , and are the constant, linear, and quadratic coefficients.

Since our data fit a Type II functional response, the Rogers Type II equation was used in our study. The component parameters (attack rate and handling time) were estimated by using frair_fit() function, and compared their difference using the “difference method” by frair_compare() function.

Bootstrapping was used to construct 95% CIs to visualize variability around the fitted curves by frair_boot() function. Functional response curves were plotted with their respective 95% CIs by drawpoly() function.

All statistical analyses were conducted in R 4.3.1 (R Core Team 2023).

Author contributions

J. Z.: Conceptualization; methodology; software; formal analysis and visualization; writing—original draft preparation; writing—review and editing; funding acquisition. P. H.: Methodology. Y. P. and Y. Z.: Investigation; resources. J. Y.: writing—review and editing. D. Y.: Conceptualization; writing—review and editing; supervision and funding acquisition.

Data availability

The datasets generated during the current study are available from the corresponding author on reasonable request.

Declarations

Competing interests

The authors declare no competing interests.