Differential Susceptibility of Coleomegilla maculata and Scymnus creperus Larvae to Aggression by Solenopsis invicta Workers

Simple Summary The biological control of aphids by aphid-eating lady beetles (coccinellids) is often hampered by aphid-tending ants or aggressive, invasive ants foraging for food; aggressive species, such as the imported fire ant, may attack and kill lady beetle larvae. This study tested the hypothesis that wax-secreting coccinellid larvae are less susceptible than non-wax-secreting coccinellid larvae to fire ant aggression. Laboratory experiments were set up using bird cherry-oat aphid nymphs and adults (as prey for coccinellids) on barley leaves with either coccinellid species with or without fire ant workers. Results revealed that the presence of fire ants reduced aphid predation by non-wax-secreting but not wax-secreting larvae. The frequency of fire ant attacks was greater on non-wax-secreting larvae; the mortality was significantly greater for non-wax-secreting larvae. The wax covering over coccinellid larvae reduced fire ant aggression. Unexpectedly, coccinellid larvae with the wax cover removed did not suffer greater fire ant attacks or increased mortality. In conclusion, the wax cover and, perhaps, the volatile or non-volatile compounds in the wax and on the integument of wax-secreting coccinellid larvae reduce fire ant aggression. Further research could identify the wax compounds and determine their role as potential repellents or irritants against fire ants. Abstract The biological control of aphids by aphidophagous coccinellids is often hampered by aphid-tending ants or aggressive, invasive ants foraging for food. Aggressive species such as the imported fire ant Solenopsis invicta Buren may attack and kill coccinellid larvae. This study tested the hypothesis that wax-secreting Scymnus creperus Mulsant larvae are less susceptible than non-wax-secreting Coleomegilla maculata (DeGeer) larvae to S. invicta aggression. Laboratory experiments were set up using bird cherry-oat aphid Rhopalosiphum padi (L.) nymphs and adults (as prey for coccinellids) on barley leaves in arenas with either coccinellid species and with or without S. invicta workers. The presence of S. invicta reduced aphid predation by C. maculata but not Sc. creperus. The frequency of S. invicta attack was greater for C. maculata than Sc. creperus; mortality was significantly greater for C. maculata than Sc. creperus. The wax covering on Sc. creperus reduced S. invicta aggression. Unexpectedly, Sc. creperus larvae with the wax cover removed did not suffer more S. invicta attacks or increased mortality. In conclusion, the wax cover and, perhaps, the volatile or non-volatile compounds in the wax and on the integument of Sc. creperus larvae reduce S. invicta aggression. Further research could identify the wax compounds and determine their functionality as semiochemicals against S. invicta.

Scymnus creperus Mulsant (Coleoptera: Coccinellidae: Scymnini) is native to North America and distributed primarily in the southern regions of the USA [19]. It is a predator of aphids [20,21]. The larvae produce a sticky, white-colored wax that exudes from their integument's pores. Coleomegilla maculata (DeGeer) (Coleoptera: Coccinellidae: Coccinellini) is native to North, Central, and South America [19,22]. It is a predator of aphid nymphs, adults, and lepidopteran and coleopteran eggs [23][24][25]. The larvae do not produce a wax covering but have bristle-like setae that cover their integument. The imported fire ant S. invicta is native to South America but is well established in the southern USA and northeastern Mexico and parts of the western USA [26,27].
Our preliminary research on aphid control by mass-reared C. maculata larvae has been challenged by the occupation of potted strawberry plants by ants, including S. invicta. In contrast, the larvae of an unidentified Scymnus species attacked and consumed sugarcane aphid Melanaphis sacchari (Zehntner) on potted sorghum plants in the presence of tending S. invicta in Stoneville, Mississippi (EWR, unpublished observations). This study tested the hypothesis that Sc. creperus larvae are less susceptible than C. maculata larvae to S. invicta aggression. The objectives of this study were to determine the predation potential of C. maculata and Sc. creperus larvae in the presence or absence of S. invicta workers and estimate the frequency of S. invicta attacks and incidence of mortality to C. maculata, Sc. creperus, or Sc. creperus with the wax covering removed. This research will help select the most effective coccinellid species for aphid biological control on crop plants in areas where S. invicta and other aggressive ants are problematic.

Plant and Insect Cultures
Barley seedlings infested with bird cherry-oat aphid Rhopalosiphum padi (L.) nymphs and adults (in banker plant system) were purchased from IPM Laboratories Inc., Locke, NY, USA (https://www.ipmlabs.com). Barley seeds were also purchased from IPM Laboratories to establish barley host plants to maintain a continuous culture of R. padi at the National Biological Control Laboratory (NBCL), ARS, USDA in Stoneville, MS, USA.
The Coleomegilla maculata 3rd instars used in this study were reared from eggs collected randomly from a stock culture maintained in an environmental rearing room (22-24 • C, 45-55% RH, and 16L: 8D L:D photoperiod) in the NBCL. This culture has been maintained for over 30 consecutive generations without any introduction of "feral" individuals. Larvae and adults were reared on factitious food based on a proprietary mixture of brine shrimp (Artemia franciscana Kellogg) eggs, microalgae (Chlorella vulgaris Beijerink), and fatty acids [28,29].
Scymnus creperus adults were purchased from IPM Laboratories Inc. Additional adults were obtained from an established laboratory colony at the University of Florida, in Gainesville, FL, USA. The adults and larvae were fed R. padi nymphs and adults, which had been reared on barley seedlings in the laboratory, for one generation before experimentation. Just before the experiments, the R. padi nymphs and adults, 3rd instar C. maculata, and 3rd instar Sc. creperus were maintained in a plant growth chamber (22 • C, 60% RH, and 16:8 L:D photoperiod).
Imported fire ants S. invicta were collected near roadsides and agricultural fields in Washington Co., Stoneville, MS, USA, in the spring and summer of 2018. Fire ants are mound-building ant species. Ants with mound soil were collected by shoveling mound soil into 5-gallon buckets. The water-drip method was then used to separate the ants from the soil [30]. Queen-right colonies were established in rearing trays and maintained in an environmental rearing room (22-24 • C, 45-55% RH, and 16: 8 L:D photoperiod). Distilled water in glass vials, stoppered with cotton, was always provided. Frozen, then thawed house crickets Acheta domesticus (L.) were used as a food source. Before experimentation, rearing trays were held on the laboratory benchtop and S. invicta workers were gently removed and placed in smaller containers coated with fluon. Workers actively moving around in the containers were selected at random for experiments.
2.2. Expt. 1: Estimating Predation of R. padi by C. maculata or Sc. creperus in Arenas with or without S. invicta Using a completely randomized design, this laboratory experiment was set up to estimate predation, i.e., the killing capacity, of 3rd instar C. maculata or Sc. creperus on R. padi nymphs and adults on barley leaves in arenas with or without S. invicta workers. Rhopalosiphum padi mid-to-late instar nymphs and adults were randomly selected from colony cages and placed on a barley leaf in the center of 10 arenas (plastic Petri dishes; 2.5 cm high, 9.0 cm diam., 159 cm 3 , with screened lids). Because 3rd instar C. maculata are larger than 3rd instar Sc. creperus (and can purportedly consume more prey), 60 and 50 R. padi were randomly placed in arenas in the C. maculata and Sc. creperus trials, respectively. In preliminary observations, these prey densities exceeded the 7-h killing rate of the coccinellids. Next, five S. invicta workers, presumably of approximately the same age but variable in size, were randomly selected from an in-house colony (collected in Washington Co., Stoneville, MS, USA on 23 May 2018) and carefully added into 5 of the 10 arenas in the C. maculata tests. Note that 10 S. invicta workers were used in the Sc. creperus tests because preliminary observations indicated that S. invicta were typically not aggressive to Sc. creperus larvae. In addition, S. invicta workers had no prior exposure to C. maculata or Sc. creperus larvae at any time in the laboratory. Within 1 h, a single 3rd instar C. maculata or 3rd instar Sc. creperus was gently added onto the barley leaf in each arena. At approximately 30-min intervals, within a 7-h time frame, the number of R. padi killed (including partially or completely consumed) by C. maculata or Sc. creperus was determined from each arena. This experiment was replicated four times, i.e., four trials, and the total sample size was 40 observations (or arenas) for C. maculata and Sc. creperus. Trials with C. maculata were conducted on 7, 11, 25 June, and 6 July 2018; those with Sc. creperus were on 16, 27, 31 July, and 9 August 2018. All arenas were maintained in a plant growth chamber (24 • C, 60% RH, and 16 h photophase), but removed on the day of experimentation. Experiments were performed on a laboratory bench at room temperature (23-24 • C, 40% RH).

Expt. 2: Estimating Frequency of S. invicta Attack and Mortality of C. maculata or Sc. creperus
Using a completely randomized design, this laboratory experiment was carried out to estimate the S. invicta attack rate and mortality of 3rd instar C. maculata versus 3rd instar Sc. creperus. Approximately 20-30 R. padi nymphs and adults (randomly selected from the colony) were placed on a barley leaf, positioned inside a medium-sized Petri dish arena (2.5 cm high, 9.0 cm diam., 159 cm 3 , with screened lids), and replicated 10 times; five arenas had one C. maculata and the other five arenas had one Sc. creperus. Next, 10 S. invicta workers (randomly selected from an in-house colony established from workers and queens collected in Washington Co., on 23 May 2018) were gently added to each arena. Note that the S. invicta workers had no prior exposure to C. maculata or Sc. creperus larvae at any time in the laboratory. Within 1 h, a single 3rd instar C. maculata or 3rd instar Sc. creperus was gently added onto the barley leaf in each arena. At approximately 30-min intervals, the number (and proportion) of C. maculata versus Sc. creperus attacked and subsequently killed by S. invicta was recorded over a 7-h time frame. This experiment was replicated three times, i.e., three trials, and the total sample size was 30 observations (or arenas). Trial dates were on 1, 2, and 3 August 2018. All arenas were maintained in a plant growth chamber (24 • C, 60% RH, and 16 h photophase). On the day of experimentation, the arenas were removed and the experiments were performed on a laboratory bench at room temperature (23-24 • C, 40% RH).

Expt. 3: Frequency of Attack by S. invicta and Mortality of C. maculata, Sc. creperus, or Sc. creperus with Wax Removed
This experiment was identical to Expt. 2, except an additional treatment was added to determine if removing the waxy material from the integument of Sc. creperus affected the S. invicta attack rate and mortality of Sc. creperus. The wax filaments on the dorsal and lateral integument of sample 3rd instar Sc. creperus were carefully removed with a fine camel's hair paintbrush. The three coccinellid treatments in this experiment were 3rd instar C. maculata, 3rd instar Sc. creperus, or 3rd instar Sc. creperus without wax. The prey density (20-30 R. padi) and arenas (medium-sized Petri dishes) were the same as before. The S. invicta density was 10 workers per arena for all three treatments. The procedures of adding R. padi onto a barley leaf, then S. invicta into each arena were the same. At approximately 30-min intervals, the numbers (and proportion) of C. maculata versus Sc. creperus versus Sc. creperus without wax attacked and subsequently killed by S. invicta were recorded over a 7-h time frame. A total of 5 replicate arenas were used for each treatment, for 15 arenas per trial. Trial dates were 16, 17, and 21 August 2018. The three trials resulted in a sample size of 45 observations (i.e., 45 arenas) for the entire experiment. All arenas were maintained in a plant growth chamber (24 • C, 60% RH, and 16 h photophase). On the day of experimentation, the arenas were removed, and the experiments were conducted on a laboratory bench at room temperature (23-24 • C, 40% RH).

Statistical Analysis
Before analysis, all data were checked for normality (using the Shapiro-Wilk test) and equal variances. For almost all datasets, the assumptions of normality and equal variances were met. Therefore, the square root transformation of absolute data and arcsine transformation of proportional data were not performed. In separate experiments, a Student's t-test was used to test the significance of S. invicta presence in arenas on predation of R. padi by C. maculata or Sc. creperus. A Pearson Product Moment Correlation Analysis was used to detect any correlation between the approximate time (within a 7-h time frame) that S. invicta killed C. maculata and the number of R. padi killed by C. maculata. A Student's ttest was used to test the significance of S. invicta attack frequency and subsequent mortality of C. maculata and Sc. creperus in the same experiment. A one-way analysis of variance (one-way ANOVA) was used to test the significance of the S. invicta attack frequency and subsequent mortality of C. maculata, Sc. creperus, or Sc. creperus without wax covering in the same experiment. The Holm-Sidak Method was used after the one-way ANOVA to separate the mean values. Means were considered significantly different at p < 0.05.   attack frequency and subsequent mortality of C. maculata, Sc. creperus, or Sc. creperus without wax covering in the same experiment. The Holm-Sidak Method was used after the one-way ANOVA to separate the mean values. Means were considered significantly different at p < 0.05. Computer software programs SigmaPlot 12.0 interfaced with SigmaStat (SYSTAT Software, Inc., San Jose, CA, USA, 2010) and JMP ® 14.2.0 (SAS Institute, Inc., Cary, NC, USA, 2018) were used for data analysis. Figure 1 illustrates R. padi nymphs and adults on a barley leaf. Companion images illustrate R. padi being consumed by 3rd instar C. maculata, S. invicta stinging and biting 3rd instar C. maculata, and S. invicta not attacking 3rd instar Sc. creperus. The mean ± SE number of R. padi killed by 3rd instar C. maculata was significantly lower in the presence of S. invicta (t = 3.32, df = 38, p = 0.002; Figure 2A). Solenopsis invicta did not kill or tend R. padi nymphs or adults. The mean (± SE) proportion of R. padi killed during the 7-h test period was 0.46 (± 0.05) and 0.66 (± 0.02) in the C. maculata and S. invicta treatment versus C. maculata alone treatment, respectively. There were 60 R. padi nymphs and adults in each arena at the onset of the experiment with C. maculata. Solenopsis invicta Insects 2023, 14, 318 6 of 10 worker aggression reduced the predation capacity of C. maculata. Moreover, the number of R. padi killed by C. maculata decreased as the time required for S. invicta to detect and kill C. maculata decreased (R c = 0.898; p < 0.001; n = 13 observations; Figure 2B).

Expt. 1: Predation of R. padi by C. maculata or Sc. creperus in the Presence/Absence of S. invicta
Images of 3rd instar Sc. creperus not being attacked by S. invicta workers (D,E).
The mean ± SE number of R. padi killed by 3rd instar C. maculata was significantly lower in the presence of S. invicta (t = 3.32, df = 38, p = 0.002; Figure 2A). Solenopsis invicta did not kill or tend R. padi nymphs or adults. The mean (± SE) proportion of R. padi killed during the 7-h test period was 0.46 (± 0.05) and 0.66 (± 0.02) in the C. maculata and S. invicta treatment versus C. maculata alone treatment, respectively. There were 60 R. padi nymphs and adults in each arena at the onset of the experiment with C. maculata. Solenopsis invicta worker aggression reduced the predation capacity of C. maculata. Moreover, the number of R. padi killed by C. maculata decreased as the time required for S. invicta to detect and kill C. maculata decreased (Rc = 0.898; p < 0.001; n = 13 observations; Figure 2B). The mean ± SE number of R. padi killed by 3rd instar Sc. creperus was not affected by the presence of S. invicta (t = 0.45, df = 38, p = 0.651; Figure 2C). Moreover, the mean (± SE) proportion of R. padi killed during the 7-h test period was 0.52 (± 0.009) and 0.53 (± 0.01) in the Sc. creperus and S. invicta treatment or Sc. creperus alone, respectively. There were 50 R. padi nymphs and adults in each arena at the onset of the experiment with Sc. creperus. Generally, S. invicta workers were not aggressive to Sc. creperus; no Sc. creperus were killed in any of the arenas. The mean ± SE number of R. padi killed by 3rd instar Sc. creperus was not affected by the presence of S. invicta (t = 0.45, df = 38, p = 0.651; Figure 2C). Moreover, the mean (±SE) proportion of R. padi killed during the 7-h test period was 0.52 (±0.009) and 0.53 (±0.01) in the Sc. creperus and S. invicta treatment or Sc. creperus alone, respectively. There were 50 R. padi nymphs and adults in each arena at the onset of the experiment with Sc. creperus. Generally, S. invicta workers were not aggressive to Sc. creperus; no Sc. creperus were killed in any of the arenas.

Expt. 2: Frequency of Attack by S. invicta and Mortality of C. maculata or Sc. creperus
As illustrated in Figure 3A,B, S. invicta attacked and killed 3rd instar C. maculata rather than Sc. creperus in experimental arenas. The attack frequency was significantly greater on C. maculata than Sc. creperus (t = 3.58, df = 4; p = 0.023). Mortality resulting from S. invicta attacks was significantly different between C. maculata and Sc. creperus (t = 13, df = 4; p < 0.001). The attack frequency and mortality of C. maculata were approximately the same, indicating that S. invicta killed approximately all the C. maculata they attacked. The mortality of 3rd instar C. maculata was from the biting and stinging behavior of S. invicta. Moreover, in any of the three trials, S. invicta never consumed C. maculata during the 7-h test period. Although S. invicta attacked a small proportion of Sc. creperus during the 7-h test period ( Figure 3A), none of them were killed ( Figure 3B). Solenopsis invicta workers were seen biting the wax filaments covering the integument of 3rd instar Sc. creperus. Several seconds later, S. invicta released its grip on Sc. creperus and immediately began to rub its mandibles against the substrate, attempting to remove the wax particles. mortality of 3rd instar C. maculata was from the biting and stinging behavior of S. in Moreover, in any of the three trials, S. invicta never consumed C. maculata during th test period. Although S. invicta attacked a small proportion of Sc. creperus during th test period (Figure 3A), none of them were killed ( Figure 3B). Solenopsis invicta wo were seen biting the wax filaments covering the integument of 3rd instar Sc. cre Several seconds later, S. invicta released its grip on Sc. creperus and immediately beg rub its mandibles against the substrate, attempting to remove the wax particles. Figure 3. The mean ± SE proportional frequency of attack by S. invicta and mortality of C. ma or Sc. creperus (A,B). The sample size was six observations for proportional attack frequenc proportional mortality estimates for C. maculata or Sc. creperus with or without S. invicta. The ± SE proportional frequency of attack by S. invicta and mortality of C. maculata, Sc. creperus, o wax covered Sc. creperus (C,D). The sample size was nine observations for proportional frequency, and mortality estimates for C. maculata, Sc. creperus, or non-wax-covered Sc. creperu or without S. invicta.

Expt. 3: Frequency of Attack by S. invicta and Mortality of C. maculata, Sc. creperus, o creperus with Wax Removed
The artificial removal of the wax covering from 3rd instar Sc. creperus before exp to S. invicta workers did not increase the attack frequency. It was greatest on C. ma than on wax-covered or non-wax-covered Sc. creperus (F = 13.95, df = 2, 6; p = 0.006; F 3C). Moreover, significant differences were not detected in attack frequency between covered and non-wax-covered Sc. creperus. Removal of the wax filaments from integument of Sc. creperus did not increase the mortality of Sc. creperus. Attacks fro invicta resulted in the mortality of C. maculata but never of Sc. creperus (F = 169.63, df p < 0.001; Figure 3D). The artificial removal of the wax covering from 3rd instar Sc. creperus before exposure to S. invicta workers did not increase the attack frequency. It was greatest on C. maculata than on wax-covered or non-wax-covered Sc. creperus (F = 13.95, df = 2, 6; p = 0.006; Figure 3C). Moreover, significant differences were not detected in attack frequency between wax-covered and non-wax-covered Sc. creperus. Removal of the wax filaments from the integument of Sc. creperus did not increase the mortality of Sc. creperus. Attacks from S. invicta resulted in the mortality of C. maculata but never of Sc. creperus (F = 169.63, df = 2, 6; p < 0.001; Figure 3D).

Discussion
This laboratory study clearly observed the differential susceptibility of C. maculata versus Sc. creperus to S. invicta aggression. The predation capacity of C. maculata larvae was reduced by S. invicta attacks and subsequent mortality. In contrast, the predation capacity of Sc. creperus was not affected by S. invicta. In previous studies, the wax covering the integument of Scymnus larvae reduced the aggressive behavior of ants tending the aphids for their honeydew or foraging for food [15,17]. Coccinellids that lack the wax covering, as in the case of C. maculata and relatives in the tribe Coccinellini, are susceptible to ant aggression. For example, the ant Monomorium minimum (Buckley) tended to the soybean aphid Aphis glycines Matsumura in the laboratory and consequently hindered predation by the coccinellid Harmonia axyridis (Pallas) [2]. Larvae of Ha. axyridis are devoid of a wax covering and are members of the tribe Coccinellini. In another study, S. invicta reduced the predation efficiency of H. convergens (tribe Coccinellini) 3rd instars more than Sc. louisianae (tribe Scymnini) 3rd instars in laboratory experiments with the cotton aphid A. gossypii [5]. The susceptibility to ant aggression suggests that biological control success will not be as effective with coccinellid larvae that do not possess a wax cover.
Although C. maculata larvae possess morphological defenses (spines, or setae) on the dorsal surface of their integument and chemical defenses (alkaloids) in their hemolymph, these defenses provide little protection from S. invicta aggression. Alkaloids discovered in C. maculata adults might function as semiochemicals [31,32]. Precoccinellin was thought to defend against some predators [31], whereas 2,4,6-trimethylpyridine may function as a pheromone [32]. To the best of our knowledge, the presence or concentration of either compound in C. maculata larvae has not been determined.
Wax structures on the integument of Sc. louisianae larvae provided a morphological and perhaps chemical defense against attacks from the ant L. neoniger; removal of wax from the integument of Sc. louisianae larvae did not attenuate L. neoniger aggression [15]. Dissolving the wax structures in hexane, then applying the solution to the integument of Sc. louisianae larvae reduced ant aggression, suggesting that chemical cues in the wax were partly responsible for lowering L. neoniger aggression [15]. In this study, removing the wax covering did not increase S. invicta attacks or cause mortality of Sc. creperus, suggesting that chemical compounds on the surface of their integument, in addition to the wax, might play a defensive role, perhaps as a repellent or an irritant. Moreover, S. invicta workers rubbed their mandibles against the substrate immediately after biting, then released wax-covered Sc. creperus larvae, apparently cleaning their mandibles. This S. invicta cleaning behavior further suggests a repellent or irritant functionality of the wax. Future research could identify the chemical compounds in the wax covering and on the surface of the integument of Sc. creperus larvae. In addition, determining the function of the isolated compounds in bioassays against S. invicta workers would be a worthwhile investigation potentially leading to the discovery of novel repellents.
Complimentary research to develop effective baits and bio-based insecticides against S. invicta is needed. Perhaps, combining tactics to control aphids by releasing the most effective coccinellids and using baits and repellents to keep S. invicta workers away from crop plants in cultured systems, e.g., greenhouses, would be the best approach.
In conclusion, this study provides evidence for the differential susceptibility of C. maculata and Sc. creperus larvae to aggressive behavior from S. invicta workers in laboratory arenas. Wax covering the dorsal and lateral surfaces of the Sc. creperus integument provides protection and prevents mortality from S. invicta attacks. The selection of the most effective aphidophagous coccinellids would ensure the success of the biological control of aphids on crop plants. The selection of wax-covered larvae (Scymnini) rather than non-wax-covered larvae (Coccinellini) would be the best choice in areas, such as greenhouses or glasshouses, where aphid-tending ants or aggressive ants foraging for food on or near crop plants are problematic [33].  Data Availability Statement: At the authors' discretion, data supporting the article's results can be made available on ResearchGate.