Is the Multicolored Asian Ladybeetle, Harmonia axyridis, the Most Abundant Natural Enemy to Aphids in Agroecosystems?

The multicolored Asian ladybeetle, Harmonia axyridis Pallas (Coleoptera: Coccinellidae), was introduced into Western Europe in the late 1990s. Since the late 2000s, this species has been commonly considered one of the most abundant aphid predators in most Western European countries. In spite of the large amount of research on H. axyridis, information concerning its relative abundance in agroecosystems is lacking. This study aims to evaluate the abundance of H. axyridis within the aphidophage community in four crops situated in southern Belgium: wheat, Triticum aestivum L. (Poales: Poaceae), corn, Zea mays, potato, Solanum tuberosum (Solanales: Solanaceae), and broad bean Vicia faba (Fabales: Fabaceae). In order to assess the species diversity, the collected data were analyzed by considering (1) the species richness and (2) the evenness according to the Shannon diversity index. Eleven aphidophages were observed in every inventoried agroecosystem, including five abundant species: three coccinellids, the seven-spotted ladybug, Coccinella septempunctata L. (Coleoptera: Coccinellidae), the 14-spotted Ladybird, Propylea quatuordecimpunctata, and H. axyridis; one hoverfly, the marmalade hoverfly, Episyrphus balteatus De Geer (Diptera: Syrphidae); and one lacewing, the common green lacewing, Chrysoperla carnea Stephens sensu lato (= s.l.) (Neuroptera: Chrysopidae). Harmonia axyridis has been observed to thrive, breed, and reproduce on the four studied crops. Harmonia axyridis is the most abundant predator of aphids in corn followed by C. septempunctata, which is the main aphid predator observed in the three other inventoried crops. In wheat and potato fields, H. axyridis occurs in low numbers compared to other aphidophage. These observations suggest that H. axyridis could be considered an invasive species of agrosystems, and that potato and wheat may intermittently act as refuges for other aphidophages vulnerable to intraguild predation by this invader. Harmonia axyridis is not the most abundant aphid predator in the main Belgian crops.


Introduction
The multicolored Asian ladybeetle, Harmonia axyridis Pallas (Coleoptera: Coccinellidae), was introduced from Asia into Western Europe and other parts of the world to control aphid and coccid populations Brown et al. 2008). In Belgium, H. axyridis was used as a biological control agent beginning in 1997 and was first observed in the wild in 2001. Since then H. axyridis populations have increased and gradually expanded into Belgium .
This species presents all the characteristics shared by an efficient aphid predator: large body size, high voracity, high predation efficiency (Labrie et al. 2006), high colonization aptitude , rapid development, high fecundity, and low susceptibility to pathogens or natural enemies (Marco et al. 2002). Harmonia axyridis has become ubiquitous in many parts of the world, including America, Europe, and Africa (Lombaert et al. 2010;Brown et al. 2011b), and has been reported in many different habitats, such as agroecosystems, gardens, and arboreal habitats (Majerus et al. 2006).
Due to its large body and efficient physical and chemical defenses, H. axyridis has become an intraguild predator (Sato and Dixon 2004;Ware and Majerus 2008). Intraguild predation has been observed among other ladybeetle species Ware and Majerus 2008); other aphid natural enemies, including syrphids, chrysopids, and parasitoids (Phoofolo and Obrycki 1998;Wells et al. 2010;Ingels and De Clercq 2011); and aphid pathogenic fungus ). This intraguild predation behavior is thought to have led to a decrease in native species (Brown and Miller 1998;Harmon et al. 2007;Ware et al. 2009;Brown et al. 2011a;Roy et al. 2012). In Belgian urban areas, Adriaens et al. (2010) found a significant decline in native species, including the two-spot ladybird, Adalia bipunctata L. (Coleoptera: Coccinellidae), and the 10 spotted ladybird, Adalia decempunctata L., on pine, lime, and maple trees following the arrival of H. axyridis. The decline of native species can possibly be explained by the decline in number of their principal prey, resulting in reduced survivorship in local habitats and altered dynamics of habitat use and dispersal (Evans 2004).
Our study was conducted eight years after the first observation of H. axyridis in the wild in Belgium (Adriaens et al. 2003). Following aphidophagous decline highlighted by several studies, the current study was conducted in order to assess the relative abundance of H. axyridis through the quantification of the abundance of this exotic species and other aphidophages in four important Belgian crops (wheat, Triticum aestivum L. (Poales: Poaceae), corn, Zea mays, potato, Solanum tuberosum (Solanales: Solanaceae), and broad bean Vicia faba (Fabales: Fabaceae)) using a three-year inventory.

Study site
Aphidophagous insect populations were sampled from 2009 to 2011 in Hesbaye, an intensive agricultural production area in Wallonia, the southern region of Belgium (individual sites given in Table 1). Four crops were chosen for their agronomic importance: wheat, corn, potato, and broad bean Vicia faba. The sampling period ran from mid-May to late September. Every week, nine fields of each crop were sampled.

Sampling methods
The sampling methods used to assess the numbers of aphidophagous predators and aphids consisted of whole-plant visual inspections, using 1 m² quadrats distributed randomly throughout the whole field. In order to avoid the influence of surrounding crops, a 20 m buffer zone around the edge of each field was not sampled. Visual sampling was conducted, as it provides an easy and accurate method for the estimation of larval and adult densities of coccinellids in agroecosystems (Michels and Behle 1992). Forty-eight quadrats were examined in each crop every week. Quadrats were located along transect lines across each field and spaced 20 m apart. All leaves and stems within the quadrat were examined, and all aphidophagous stages were recorded. Aphid species were also quantified on all leaves and stems. Larvae and pupae were brought to the laboratory to develop under laboratory conditions (24 ± 1° C, 75 ± 5% RH) for identification at the species level. All aphid predators were identified, with the exception of members of the common green lacewing, Chrysoperla carnea Stephens (Neuroptera: Chrysopidae) species complex, which were pooled together. This group comprises three cryptic species, C. kolthoffi Navas, C. lucasina Lacroix, and C. carnea Stephens, which can only be differentiated using molecular techniques (Bozsik et al. 2003;Lourenco et al. 2006).

Statistical analysis
In order to assess the species diversity, the collected data were analyzed by considering (1) species richness and (2) evenness according to the Shannon diversity index (H) (Magurran 1988 Because mean densities per m² were low, these values are presented per 100 m². The mean abundance per species was analyzed with an analysis of variance (ANOVA: general linear models) with crops (q = 4) and years (n = 3) used as factors (α = 0.05). Within crops, densities of species were compared using the least square difference (LSD; α = 0.05). To account for the variations of predator abundances, the abundance per species was analyzed using an analysis of covariance (ANCOVA: general linear models) with crops (q = 4) and years (n = 3) used as factors (α = 0.05) and aphid densities used as the covariable. Prior to analyses, a log 10 (x + 1) was used to transform the data distribution (counting) due to its asymmetry (Dagnelie 2011). Although statistical analyses were performed on transformed data, untransformed data are presented in Tables 4  and 5. Statistical analyses were performed using Minitab ® release 1.5 (www.minitab.com).
From 2009 to 2011, species richness increased in broad bean, corn, and wheat, reaching 6, 8, and 7 species respectively (Table 3). Species richness did not evolve in potato crops, remaining at five species during the entire period. Overall, five aphidophages were continually observed during the three-year period in each crop and represented 95% of all the aphidophage observed in 2009 and 99% in both 2010 and 2011: E. balteatus, C. carnea s.l., C. septempunctata, P. quatuordecimpunctata, and H. axyridis.

Relative abundance of aphidophage in four crops
The relative abundance of both adult and larval stages of the five aphidophage within each crop showed significant differences (Tables 4  and 5

Effect of aphid densities and sampling year on relative abundance of aphidophage
Abundances of H. axyridis in wheat and potatoes were not analyzed, due to very low numbers of individuals observed during the three-year inventory. Over the three-year sampling, adult populations of H. axyridis in corn significantly increased (F 2,1709 = 14.5, p < 0.001) (Table 6) from 7.8 ± 2.0 in 2009 to 19.9 ± 1.8 individuals per 100 m² in 2011 (Table 4). Larval populations in the same crop also increased statistically (F 2,1709 = 39.9, p < 0.001) ( Table 6), rising from 11.9 ± 4.2 to 70.8 ± 6.6 larvae per 100 m² (Table 5). In broad bean, relative abundance of H. axyridis was not significantly different among the three years, neither at the adult (Table 6) nor larval (Table 7) stages.
The abundances of three other aphidophage showed variable changes (Tables 6 and 7).
The ANCOVA analyses showed a linear relationship between aphid and predator populations in 55% of adult populations ( Table 6) and 35% of larvae populations (Table 7) (p aphids < 0.05, ANCOVA). In these cases, aphid densities influenced the predator abundance. Results (p years ) comparison between ANOVA and ANCOVA showed that the influence of aphid populations on predator abundance variations between years was not statistically significant.; p years of the two statistic analyses showed the same results.

Discussion
Since the invasive coccinellid H. axyridis spread over Europe (Brown et al. 2008), imposing negative impacts on native aphidophage and affecting composition of several guilds (Soares et al. 2008;Roy et al. 2012), studies have evaluated the population spread of this coccinellid. The current study's sampling of aphid predators in Belgian agroecosystems from 2009 to 2011 showed that H. axyridis lives and reproduces more efficiently in corn and broad bean than in wheat and potato. In corn, the evenness during the three years decreased when H. axyridis population increased strongly and was higher than the population of other species.
During the three-year sampling, 11 aphidophage were observed in these agroecosystems, but five of them predominated: E. balteatus, C. septempunctata, P. quatuordecimpunctata, H. axyridis, and C. carnea s.l. Five dominant species in agroecosystems is a common observation (Hodek and Honěk 1996). Observations on predator densities highlight that H. axyridis was not the numerically dominant species in every crop: in wheat and potato, C. septempunctata was more abundant than H. axyridis. In many European agricultural crops, both C. septempunctata and P. quatuordecimpunctata were dominant prior to the arrival of H. axyridis (Honěk 1979;Bode 1980;Chambers et al. 1982), and it appears that these two species have maintained their dominance in spite of being prone to intraguild predation by H. axyridis in the field ).
In our study, larvae of E. balteatus were the most abundant observed predators in wheat, which has already been reported by Tenhumberg and Poehling (1995) prior to the arrival of H. axyridis. Episyrphus balteatus has also been previously reported as one of the most abundant aphidophage in vegetable crops, such as broad beans (Colignon et al. 2001;Colignon et al. 2002). This could be explained by abiotic conditions (high density cereal crop, with high humidity and low temperature) that are more favorable to the larvae of E. balteatus (Honěk 1983).
The fact that H. axyridis is not the most abundant aphidophage in agrosystems is probably due to its generalist behavior and arboreal habitat selection (Hodek 1973;Chapin and Brou 1991;LaMana and Miller 1996;Brown and Miller 1998;Labrie 2007). However, it has been reported that H. axyridis can also thrive in agrosystems such as wheat, corn, and potato (LaMana and Miller 1996;Buntin and Bouton 1997;Colunga-Garcia and Gage 1998;Michaud 2002;Brown 2003;Nault and Kennedy 2003;Snyder et al. 2004;Jansen and Hautier 2008), as well as in herbaceous habitats (LaMana and Miller 1996;Koch et al. 2006;Alhmedi et al. 2007).
There were evident changes in the abundance of aphidophages in crops through the years, but this study does not propose to identify a global evolution (increase or decline) in any of the species that were observed in this study. The causes of such fluctuations are diverse and may include factors such as the diversity and abundance of aphid species (Wright and Laing 1980;Honěk 1982;Thalji 2006). The results of the ANCOVA showed that there was a linear relationship between prey and predator populations, but additional biotic and abiotic factors contribute to the annual variability of predator abundance. Climate could be one such factor, due to its influence on natural enemies, overwintering mortality, and aphid populations (Hodek and Honěk 1996;Szentkirályi 2001;Rotheray and Gilbert 2011). Several other factors could also explain the variation between crops: insolation, humidity (Honěk 1985), quantity and quality of host plants (Alhmedi et al. 2009), and adjacent habitats (Colignon et al. 2001;Alhmedi et al. 2009).
A particularly interesting finding is that although H. axyridis breeding occurred in all four inventoried crops to some extent, adults of this species are not ubiquitous; few immature individuals were recovered from potato and wheat. Assuming that declines in native species are caused by H. axyridis (Roy et al. 2012), this suggests that certain crops, such as wheat and potato, could act as refuges from H. axyridis at certain times, while native species, such as E. balteatus and C. septempunctata, are able to breed with a lower risk of intra-guild predation or other forms of competition from the invaders. Such habitats could become even more important as native species adapt to the invader by evolving to avoid habitats where H. axyridis occurs in high numbers, as has been seen in co-occurring aphidophages in their native habitats (Sloggett 2012).
In conclusion, our study indicates that H. axyridis was not the most frequently observed aphidophage in the four most important Belgian agronomical crops. In future studies, longer samplings would be preferable in order to eventually identify quantitative changes in the native fauna suggested from other studies. Agroecosystems may even constitute an ecological reservoir for certain native aphidophage.  +: presence of aphidophages; bold is used to call attention to specific species Table 3. Species richness and diversity index (H= Shannon-Weiner diversity index, where absolute diversity = 1.00; J = evenness or relative diversity (H/Hmax), where absolute evenness = 1.00).
Southern Ontario. Canadian Entomologist 112(10): 977-988. Table 4. Abundance (means and SE) of aphidophage (adults) per 100 m 2 in four agroecosystems from 2009 to 2011. Means within a crop followed by the same letter were not significantly different for the same field (p > 0.05; LSD test).
Bold is used to call attention to specific species Table 5. Abundance (means and SE) of aphidophage (larvae) per 100 m 2 in four agroecosystems from 2009 to 2011. Means within a crop followed by the same letter were not significantly different for the same field (p > 0.05; LSD test).
Bold is used to call attention to specific species Table 6. ANOVA and ANCOVA summary of the effects of aphid abundance and year sampling (2009,2010,2011) on the abundane of five predators at the adult stage in four crops (corn, wheat, potato, and broad bean).
p values come from GLM, *** p < 0.001, ** p < 0.01, * p < 0.05, ns = not significant, p > 0.05; bold is used to call attention to specific species Table 7. ANOVA and ANCOVA summary of the effects of aphid abundance and year sampling (2009,2010,2011) on the abundance of five predators at the larval stage in four crops (corn, wheat, potato, and broad bean).
p values come from GLM, *** p < 0.001, ** p < 0.01, * p < 0.05, ns = not significant; bold is used to call attention to specific species Table 8. Mean numbers and SE of aphids/100 m² observed in four fields (wheat, broad bean, corn, and potato) from 2009 to 2011.