1. Introduction
Legumes comprise almost 20,000 species in 800 genera [
1], of which soybeans, peanuts, broad beans, peas, mung beans, cowpeas, and lentils are the main legume crops in agriculture and play a central role in maintaining ecological sustainability and crop diversity [
2,
3,
4]. Legumes are essential food for humans and animals, and their role in the global food supply cannot be underestimated [
5]. Specifically, legumes, such as cowpeas, soybeans, and jack beans, are considered equivalent to meat by the poor and are an important source of protein, vitamins, and minerals for them [
6,
7,
8].
Aphis craccivora Koch,
Myzus persicae Sulzer, and
Megoura japonica Matsumura (Homoptera: Aphididae) are major pests in agricultural production and are especially harmful to legume plants [
9,
10,
11].
Aphis craccivora is a pest of nineteen plant families with a preference for legumes, such as broad beans, peas, mung beans, and soybeans [
12]. The aphid’s excretion of “honeydew” can cause sooty blotch, which can seriously affect crop yield and quality [
13].
Myzus persicae, known as the tobacco aphid, can harm more than 400 kinds of plants in addition to tobacco, such as peas, soybeans, peaches, cruciferous vegetables, etc.; it is an important economic pest and the first insect in the world that was found to spread viruses [
14,
15,
16].
Megoura japonica is also one of the main pests of legumes [
17] and is widely distributed around the world [
18]. It sucks the sap of its host plant with piercing–sucking mouthparts, causing the plant stems and leaves to shrink, which inhibits crop growth [
11]. Thus, these three species of aphids pose a serious threat to agricultural production and safety.
At present, the treatment of aphids mainly relies on chemical pesticides [
19,
20], which not only lead to pesticide resistance in aphids but also cause pesticide residues, environmental pollution, and the killing of non-target insects [
21,
22], which does not meet the developmental requirements of modern green agriculture. In order to better carry out green pest control and maintain the sustainable development of ecosystems, it is urgent to seek biological control measures for aphids [
11], and the use of predatory natural enemies is one of the important methods of achieving this [
23,
24].
Hoverflies are seen as candidates for controlling aphid populations because of their voracity [
25].
Eupeodes corollae Fabricius (Diptera: Syrphidae), one of the most common hoverfly species, has a wide range of ecological niches and is a predatory natural enemy insect worldwide [
26]. It is distributed throughout China, India, Myanmar, America, Australia, Japan, and so on, and is the dominant species of hoverfly in many places [
27].
E. corollae adults visit flowers and are important pollinators in the natural ecosystem. In one study, the seed set of sweet peppers and the yield of fruit pollinated by
E. corollae increased by 395% and 390%, respectively, compared to the control group (not pollinated by hoverflies) [
28].
E. corollae larvae are polyphagous, can feed on a variety of aphids and small lepidopteran larvae, and have the potential to perform biological control functions in agroecosystems [
29,
30,
31].
In recent years, research on aphid predation and biological control has mostly focused on ladybirds [
32,
33,
34]. As an important natural enemy insect, it is imperative to scientifically evaluate the predation and control effect of
E. corollae on aphids. This study mainly evaluated the spawning and predation preference, predation function responses, searching efficiency, and prevention and control capabilities of
E. corollae on
A. craccivora,
M. persicae, and
M. japonica. The results have good theoretical and application value for the biological control of aphids by
E. corollae.
2. Materials and Methods
2.1. Test Aphids
Aphis craccivora Koch, M. persicae, and M. japonica aphids were collected from an experimental field at the Langfang Experimental Station, Chinese Academy of Agricultural Sciences (CAAS; 39°30′29″ N, 116°36′8″ E), in Hebei Province in 2019. The three species of aphids were all reared on broad bean plantlets for more than 20 generations in a greenhouse at 25 ± 1 °C, 50 ± 5% RH, and 16:8 (L:D) h. All plantlets were grown in nutrient soil and vermiculite in plastic boxes (50 × 40 × 20 cm).
2.2. Test Hoverfly
A total of 30 adults of E. corollae (♀:♂ ≈ 1:1) were collected from an experimental field at the Langfang Experimental Station, Chinese Academy of Agricultural Sciences (CAAS; 39°30′29″ N, 116°36′8″ E), in Hebei Province in 2019. The adults were fed with 10% v/v honey water and a mixture of pollen (rape:corn = 3:1) in nylon gauze cages (50 × 40 × 30 cm). Broad bean plantlets infested with mixed aphids (A. craccivora:M. persicae:M. japonica ≈ 1:1:1) were placed in the cages for laying eggs. The larvae were fed on mixed aphids in plastic containers (50 × 40 × 15 cm) for 20 consecutive generations in a greenhouse at 25 ± 1 °C, 50 ± 5% RH, and 16:8 (L:D) h.
2.3. Spawning Selection
The adults of E. corollae (♀:♂ = 5:5) were reared in 1 × 1 × 1 m cages (200-mesh nylon) with broad bean plantlets infested with A. craccivora, M. persicae, or M. japonica individually (200 aphids/plantlet, 5 plantlets/pot), and were also provided daily with 10% v/v honey water and a mixture of pollen (rape:corn = 3:1). The broad bean plantlets were grown in vermiculite and nutrient soil in plastic pots (12 cm upper diameter × 9 cm lower diameter × 10 cm height). The number of E. corollae eggs on the broad bean plantlets was recorded daily. The spawning selection rate was (no. of eggs on a type of aphid plantlet)/(total no. of eggs) × 100. The experiment was repeated five times.
2.4. Predatory Selection
Eupeodes corollae larvae were starved for 24 h and then transferred onto an excised Cucurbita pepo leaf disc (same area as Petri dish) individually. The leaf disc was placed on a layer of 2% agarose (to slow leaf water loss) within a Petri dish (9 cm diameter × 1.5 cm height). Three species of aphids (
A. craccivora:
M. persicae:
M. japonica = 1:1:1) were attached to the Petri dish; for 2nd-instar larvae, 10 aphids of each species were placed, and for 3rd-instar larvae, 30 aphids of each species were placed. The amount of predation by the larvae on each species of aphid was recorded separately after 24 h, and the experiment was repeated three times. When a variety of prey coexist, the predation preference (
Ci) is expressed by the equation proposed by Zhou et al. [
35]:
where
Fi is the proportion of the
ith prey in the environment, and
Qi is the predation ratio of the predator to the ith prey. When
Ci = 0, it means that the predator has no preference for the ith prey; 0 <
Ci < 1 indicates that the predator has a positive preference for the
ith prey; and −1 <
Ci < 0 indicates that the predator has a negative preference for the
ith prey.
2.5. Predatory Response
The method was the same as described in
Section 2.4. Adult aphids of a similar size (
A. craccivora,
M. persicae, or
M. japonica) were inserted into Petri dishes as prey, and the number of larvae preying on aphids was recorded 24 h later. Each treatment was repeated 5 times. The prey density settings are shown in
Table 1.
2.6. The Control Effect of E. corollae on Aphids in Enclosed Spaces
One five-day-old female adult of
E. corollae that was starting to lay eggs was reared in a cage (50 cm × 35 cm × 45 cm, 200-mesh nylon) with broad bean plantlets infested with aphids (
A. craccivora,
M. persicae, or
M. japonica). The initial density of the aphids was set to 5 gradients to form different hoverfly–aphid release ratios (
Table 2). The experiment was repeated 3 times per treatment; 3 pots of broad bean seedlings were used per replicate, and there were 20 plants per pot. The numbers of aphids, eggs, larvae, and pupae were investigated on days 3, 6, 9, and 12, and the rate of aphid population decline was calculated. Population decline rate (%) = [(No. of pre-treatment insect population − No. of post-treatment insect population)/No. of pre-treatment insect population] × 100.
2.7. Data Analysis
There are two main steps in the research analysis of predatory functional responses. First, based on prey density and actual predation quantity, the type of functional response is determined according to the following binomial logistic regression equation:
where
Na is the number of prey individuals consumed;
N0 is the initial prey number, and
P0 (intercept),
P1 (linear),
P2 (quadratic), and
P3 (cubic) are the estimated coefficients. Type I responses are characterized by a constant positive slope and an intercept. Type II responses are characterized by a proportion of prey consumption that decreases monotonically with the initial prey number (
P1 < 0). In type III responses,
P1 > 0, and
P2 < 0, and this type of response is described by a proportion that first increases and then decreases with an increasing initial prey number [
36]. Second, the parameter estimation was performed according to the type of predatory functional response (type II or type III).
where
Na is the number of prey individuals consumed;
N is the prey density;
a is the instantaneous attack rate;
Th is the prey handling time;
T is the available time for the predator to feed (1 d);
S is the searching efficiency, and b, c, and d are the fitted constants [
37].
Differences in the spawning selection rate, the amount of predation, the predation preference, and the population decline rate were determined using a one-way analysis of variance (ANOVA), followed by Tukey’s honestly significant difference (HSD) test, with proportional data first arcsine square-root-transformed to meet the assumptions of normality and heteroscedasticity. All tests were performed in the program SPSS version 25 (IBM, Armonk, NY, USA), except for the polynomial function, which was fitted using the glm function in R version 2.0.1 (R Development Core Team 2008). The charts were generated using OriginPro 2021 (OriginLab Corporation, Northampton, MA, USA).
4. Discussion
Eupeodes corollae, a well-documented beneficial insect group in agricultural ecosystems [
30], provides biological control and pollination services that help maintain biodiversity and improve crop yield [
38]. The results of this experiment showed that the spawning and predation preferences of
E. corollae were consistent and that the second- and third-instar larvae all had a predation ability and showed a Holling’s type II functional response when feeding on
A. craccivora,
M. persicae, and
M. japonica. At the same time,
E. corollae had a certain inhibitory effect on the growth of these three aphid populations. Therefore,
E. corollae can be considered a natural enemy insect with a high biocontrol potential and field application value.
Predators often have predatory preferences under the condition of many kinds of prey or different morphological types of the same prey [
39].
Neoseiulus californicus prefers to prey on the
Eotetranychus lewisi rather than
Tetranychus urticae [
40].
Neoseiulus bicaudus prefers to prey on the
Tetranychus turkestani rather than
Bemisia tabaci [
41]. The predatory preference of predatory insects is affected by multiple factors, such as prey size, nutrient level, exoskeleton “hardness”, and genetic characteristics [
42,
43,
44,
45], and predators usually choose prey with the best energy return [
46]. The predation rate for the summer form of
Cacopsylla chinensis (higher levels of protein) by
Harmonia axyridis was higher than that of the winter forms [
39]. The spiders that had been maintained on high-quality flies exhibited significant positive selection for the high-quality flies (higher levels of nitrogen and protein) compared to the low-quality flies [
44]. The highest
Ci was found for
A. craccivora in the study, probably due to its higher nutritional return among these three species of aphids. At the same time, the spawning selection rate of
E. corollae adults in
A. craccivora populations was the highest, and the predation and spawning preferences were consistent, which also existed in a study on
Spodoptera frugiperda (both favoring maize) by Zhang et al. [
47]. The
Ci of the third-instar larvae for
A. craccivora was higher than that of the second-instar larvae (0.15 > 0.08), which may be due to the differences of the insects themselves at different growth stages, as was also found in a study on the predation selection of
C. undecimpunctata by Cao et al. [
48].
The predatory functional response is often used to quantitatively assess the ability of predatory natural enemies to control pests [
49,
50]. In general, predatory insects tend to have a Holling’s type II functional response [
51,
52]. As a natural enemy of aphids, the functional response of
E. corollae can better evaluate its ability to control aphids. The second- and third-instar larvae of
E. corollae all showed a Holling’s type II functional response when feeding on
A. craccivora,
M. persicae, or
M. japonica, which was consistent with the predation trend of
E. corollae on
Brevicoryne brassicae,
Lipaphis erysimi, and
M. persicae found by Li et al. The 1/
Th values of the second- and third-instar larvae of
E. corollae on
M. persicae were 142.86 and 200.00 in our study, respectively, while the numbers reported by Li et al. were 55.71 and 166.67, respectively. This difference is likely related to the different geographical populations [
53]. The predators’ voracity on prey increased with their ontogenetic development, as shown in previous studies [
54,
55]. For these three aphid species, the 1/
Th of the third-instar larvae was higher than that of the second-instar larvae. Thus, it is recommended to select third-instar larvae for the control of aphids in field applications. The data on the functional response cannot be fully used to reflect the foraging behavior of predators in the wild due to the complexity of wild spaces [
56,
57], but they can be used under conditions of high-density prey [
58]. Therefore, our results provide a valuable theoretical basis for the biological control of aphids using
E. corollae. The
S of the second- and third-instar larvae on these three species of aphids gradually decreased with an increase in prey density, and when the prey density was the same, the
S of the third-instar larvae was greater than that of the second-instar larvae, as also found for
H. axyridis and
C. undecimpunctata [
59,
60].
Natural enemies that exhibit a Holling’s type II response can be applied for the biological control of pests [
61]. The benefit–harm ratio of natural enemy placement is an important parameter for this application. This study showed that the control effect and release ratios of
E. corollae on three species of aphids were positively correlated, which is consistent with a study on the effects of
E. corollae on the cotton aphid
Aphis gossypii by Li et al. [
62]. The more natural enemies are released, the better the control effect under certain conditions. Considering the application cost, the recommended release ratios of
E. corollae for the control of
A. craccivora,
M. persicae, and
M. japonica are 1:2000, 1:6000, and 1:1000, respectively. Sun et al. found that the benefit–harm ratio of
E. corollae larvae to aphids needed to be as high as 1:30 to achieve a 91.05% control effect [
63]. In comparison, it can be considered that the adults of
E. corollae have great biocontrol potential. Adult hoverflies lay eggs; the eggs hatch into larvae, and the larvae feed on the aphids to control the growth of their populations. The density of the aphid population is too large in the late stages of the treatment, and the feeding of a large number of aphids leads to a wilted and dry state of broad bean seedlings. This also causes the
E. corollae adults to lay very few eggs or even stop laying eggs, and the population development of aphids and
E. corollae affects each other, as also found between aphids and
Hippodamia variegata [
64].
Aphids are extremely harmful to agricultural production, and biological control is an important part of comprehensive aphid control and plays a vital role [
65]. Among the natural enemies of aphids, hoverflies do not have a summer diapause as large ladybugs, such as
H. axyridis and
Coccinella septempunctata [
66,
67], so they are more suitable predators in the summer, especially when large ladybugs are inactive. This study on the predation and control effect of
E. corollae on three species of aphids enriched the biological control theory of “insect control” in order to provide a reference for the application and promotion of the use of
E. corollae to control aphids.
This experiment was performed in a greenhouse that represented a semi-closed condition with a constant temperature and humidity, which is not fully representative of the behavior of hoverflies under complex conditions in the wild [
68]. The predation behavior of hoverflies on aphids is not only related to the insects’ own characteristics but is also affected by such environmental factors as temperature, humidity, and light under natural conditions. At the same time, the intraspecific competition and interference should also be of concern because the occurrence of hoverflies is highly overlapping with other natural enemies of aphids (ladybirds and grass flies) [
69]. Under natural conditions, unfavorable environmental conditions are easy to have a negative impact on the mating and egg-laying behaviors of adult hoverflies [
70], and the lack of food (aphids, pollen, or nectar) and host plant species are also susceptible to influence [
71]. The study by Li et al. found that hoverfly adults did not forage or lay eggs on tomato plants (flowerless), and the
E. corollae populations may not be able to sustain themselves in greenhouse systems [
62]. Using chemical pesticides is currently the main means of control of aphids, which not only pollutes biological habitats but also directly causes certain lethal effects on hoverflies [
72,
73]. Every problem in the practical application of
E. corollae as natural enemies requires our attention, and further research is necessary to identify best practice applications.