2.1 Study Sites

The study was conducted on three sites near Ajaccio namely, Loretto, Suartello and Vignola (Figure 1, Table 1). They constitute the ecological compensation zones for the Loregaz project of Engie and managed on its behalf by an association, the Conservatoire d'Espaces Naturels de Corse. The main vegetation on each site is the Mediterranean maquis and the sampling considered the environment differences in order to have a good representativeness of the ecosystems (Table 1).

The Loretto site, located a few hundred meters from the city centre of Ajaccio adjoining the industrial Loregaz site, is made up of a plant mosaic, alternating open areas and old olive groves (Figure 1). The Suartello site, located on the edge of a wooded area, is a plant mosaic environment made up of an open (e.g., grassland) and shrubland environments. The Vignola site facing the sea (ca. 200 m inland) was partly degraded by heavy rotary grinding in 2018, 4 years before the study. The sampling included the ‘recovering’ environment as well as the closed environment making up most of the site. The proximity of the sites to each other makes it possible to consider their average temperatures and precipitation as being similar (Figure 1). Thus, they have a warm temperate climate with an average annual temperature of 1°C and an average annual precipitation of 526 mm. However, some differences exist: Vignola is exposed to sea spray and Suartello is a bit shadier.

2.2 Sampling Method

On each of the three sites, every 2 weeks from mid-February to mid-November 2022, hoverflies were only collected when visiting flowers during the different time slots of the day (9 am–5 pm). Hoverflies were hence sampled dynamically using a butterfly net, rather than passive traps such as malaise tents, emergence traps or colour plates (O'Connor et al. 2019). This sampling method was motivated by our functional aim to document the plant species diversity visited by anthophilous hoverflies rather than producing a complete syrphid diversity list. In addition, the studied sites been preserved ecological areas we wanted to have a limited impact on their respective (pollinating) entomofauna.

Two complementary sampling methods were used to record insect floral visits during each field sampling on each site at different times. First, a dynamic census along 2 linear transects of 30 × 2 m at slow speed (30 min per transect), located in two characteristic areas of the different sites. All hoverflies spotted to visit floral structures were captured with a net and stored in a freezer until their identification. Afterwards, they were mounted and pinned in our collection or preserved in 70% ethanol. The flower/inflorescence visited by the insect was identified to species level. Second, static observations of visiting hoverflies were performed for 5 min on two successive plants of 6 different plant species, i.e., 12 observations for 1 h per site. These plants were chosen for each sampling period according to their abundance in the flowering community. Monitored plant species hence varied throughout the year according to their flowering pattern. Overall, the 9-month sampling at Vignola resulted in 60 sampled transects corresponding to ~30 h of dynamic observations and 25 h of static observations of flower visitors. The sites of Loretto and Suartello were sampled similarly, with a total of 56 transects sampled, corresponding to around 28 h of dynamic observations and 26 h of static observations.

2.3 Hoverfly Identification

The morphological identifications of the syrphid specimens were done under a stereomicroscope with the help of several articles and reference books on French hoverflies (Seguy 1961; Speight 2020; Speight and Langlois 2020; Speight and Lebard 2020; Speight and de Courcy Williams 2021), on Western Europe hoverflies (Van Veen 2004) and Mediterranean species (Grković et al. 2015; Van Steenis, Hauser, and van Zuijen 2017; Malidžan et al. 2022). Also, the list published by the French National Museum of Natural History (MNHM) from la ‘Planète revisitée’ Corsican missions was consulted (Touroult et al. 2020; Ichter et al. 2022). In addition, some identifications were confirmed by Syrphid specialists in the genus Eumerus (Cavailles. S., and Lebard. T., pers. comm.) and for Merodon chalybeus Wiedemann in Meigen, 1822 (Sarthou. V., pers. comm.). Seven specimens were not yet identified at the species level but are recognised, for the moment, as distinct morpho-species for the diversity analyses. Their identification is in progress.

2.4 Statistical Analysis

Statistical analyses and graphs were performed with Past statistical software version 4.16c (Hammer, Harper, and Ryan 2001). Beta-diversity differences amongst sites was tested on species abundances (Bray-Curtis distances) using the function betadisper from the package vegan in RStudio 12.1 (Posit team 2024). The network visualisation/analysis was performed using the function plotweb from the package bipartite (Dormann, Gruber, and Fruend 2008; Dormann et al. 2009).

3.1 Hoverfly Floral-Visitor Components

The Syrphidae represented 3.4% of the overall recorded floral visits considering the four main orders of pollinators (Hymenoptera, Coleoptera, Diptera and Lepidoptera), but 51% of the Diptera collected (Maestracci, Plume, and Gibernau 2024). Globally, the hoverfly family appears to be generalist visiting flowers from one-third of the plant community with a significant higher abundance on Foeniculum vulgare (see Section 3.5; Table S3). All the plant species visited by hoverflies were, in average, also visited by 10 other anthophilous insect families (range: 2 for the undetermined Poaceae to 25 for Foeniculum vulgare). Moreover, each hoverfly species appeared also to be generalist except probably for Eumerus basalis and Eristalinus taeniops (see Section 3.5). In terms of abundances, they represent < 20% of the total insect floral visitors for a given species (9% for Dittrichia vicosa, 12% for Foeniculum vulgare and 19% for Raphanus raphanistrum). Interestingly, they appear to be numerical important floral visitors in autumn (September to mid-November) almost equalling the Hymenoptera abundance (Maestracci, Plume, and Gibernau 2024; see Section 3.3).

3.2 Hoverfly Diversity and Richness

A total of 138 specimens were caught visiting flowers, representing 27 species from 16 genera (Table 2). The subfamily Syrphinae was the most diverse (9 genera, 13 species) compared to Milesinae (3 genera, 8 species) or Eristalinae (4 genera, 6 species). The seven most abundant species accounted for 67% of the abundance of the hoverfly community (Table 2), namely Eumerus barbarus (Coquebert, 1804) and Sphaerophoria scripta (Linnaeus, 1758) (19 specimens each), Chrysotoxum intermedium (15 specimens), Episyrphus balteatus De Geer, 1776 and Syritta pipiens (12 specimens each), Melanostoma mellinum (Linnaeus, 1758) (8 specimens) and Melanostoma scalare (Fabricius, 1794) (7 specimens). There was a high difference in abundance between the study sites, two-fold more hoverflies were collected at Loretto (65 individuals) than at Suartello (37 specimens) or Vignola (36 specimens). The compositional dominance was a bit lower in Loretto (D = 0.096) than in Suartello (D = 0.114) and Vignola (D = 0.127). In fact, the most abundant species per site represented 25% of the total floricolous hoverfly abundance at Loretto (Eumerus barbarus, N = 16) and at Vignola (Chrysotoxum intermedium, N = 10) and 30% at Suartello (Sphaerophoria scripta, N = 11).

It is interesting to note that the diversity of hoverflies visiting flowers at the three sites studied was different, with 13 species at Vignola, 15 species at Suartello and 19 species at Loretto (Table 2). Accordingly, the Shannon index at Vignola (2.35) was lower than at Suartello (2.53) or Loretto (2.68). About half the species diversity was in common for any combination of two sites (between 7 and 10 species), and Loretto appeared to have a significant higher number of ‘exclusive’ species (6) than Suartello or Vignola (3 for both, Figure 2). The beta diversity comparisons (Harrison 2 measure) showed lower differences between Loretto and Suartello or between Loretto and Vignola (0.263 each), than between Suartello and Vignola (0.40). Those differences are probably partially explained by differences in the geographical characteristics, the main vegetation and the anthropic pressure of these three sites (Table 1, study site section) but also by differences in species diversity. The estimate of total species richness of flower-visiting hoverflies was between 28.2 and 26.9 at Loretto using both Chao1's unbiased or abundance-based coverage estimators, respectively. These estimates were lower for Suartello (24.1 and 24.8) or Vignola (19.8 and 25.3), suggesting that the potential diversity at Loretto, with 19 reported species, was better covered (~68%) than the two other sites (52%–60%).

3.3 Seasonal Variations

Both the abundance and the species diversity of anthophilous hoverflies underwent the same monthly variations with a bimodal pattern presenting a first peak in March and the second in November (Figure 3). At the studied sites, October proved to be the month with the most observations of floral hoverflies, both in terms of species diversity (16) and abundance of specimens (54; Figure 3). Hoverflies were most abundant in autumn since 60% of our captures took place between September and mid-November (Table S1, Figure 3). Not only the abundance but also the diversity was higher during this period since 19 species were recorded visiting flowers out of the total number of 28 (Figure 3). Note that 9 species (33%) were singletons (Table S1). With the exception of March, hoverflies were less frequently observed on flowers at the end of winter and during spring (second-half of February to June), with 4–13 individuals collected monthly representing only 3–6 species (Figure 3). In March, the second highest diversity and abundance of hoverflies visiting flowers was recorded with 10 species collected represented by 25 individuals. Summer months (July and August) were the poorest with only one hoverfly observed on flowers during each of them (Figure 3).

Now considering the monthly species composition, only two-third of the data are relevant since 9 out 27 recorded species were singletons (Table S1). They were collected throughout the year except in summer and particularly in spring (March: 3 singletons, April: 1, and May: 1) and autumn (September: 1, October: 2, and November: 1). Seven species were active all ‘year-round’ visiting flowers from February–March until October–November, namely Chrysotoxum intermedium, Episyrphus balteatus, Eumerus sp1, Eupeodes corollae (Fabricius, 1794), Meliscaeva auricollis (Meigen, 1822), Sphaerophoria scripta and Syritta pipiens. Five species were recorded on flowers from May–June to October–November, namely Eumerus barbarus, Eumerus basalis Loew, 1848, Paragus bicolor, Paragus haemorrhous Meigen, 1822 and Paragus quadrifasciatus Meigen, 1822. Finally, the last six species, were recorded visiting flowers during only one season, in spring (Melanostoma mellinum (Linnaeus, 1758), Melanostoma scalare and Eristalis tenax (Linnaeus, 1758)) or in autumn (Eristalinus taeniops, Paragus sp1 and Syrphus ribesii).

3.4 Daily Activity

Hoverflies' abundance and richness varied similarly amongst the three sites on daily-basis observations (Figure 4). Hoverflies were mostly observed on the flowers in the morning, particularly on early hours, rather than in the afternoon (Figure 4). The two time slots with the highest abundances of hoverflies were 09:00–10:00 h (37 specimens) and 11:00–12:00 h (27 specimens). In summary, 58% of all hoverfly specimens were collected during the morning hours (09:00–12:00 h) after 12:00 h between 9 and 14 specimens were collected per hour (Figure 4, Table S2).

In terms of diversity, the greatest number of hoverfly species (70%) was also collected during the morning hours (09:00–12:00 h). The species collected only in the afternoon, and mostly (6 out of 10) between 14:00 and 15:00 h, were in fact singletons (Table S2). Six species were active on flowers all day long and four other species had an extended period of floral visits continuous as Eristalinus taeniops (11:00–17:00 h) or ‘discontinuous’ with one capture both in the morning and the afternoon (Eristalis tenax, Eumerus pulchellus, Myathropa florea). The remaining six species appeared to have a more specific daily period for visiting flowers, especially in the morning for Eupeodes corollae, Melanostoma mellinum, Paragus quadrifasciatus or Eumerus basalis, at midday for Syrphus ribesii, or in the afternoon Paragus sp1.

3.5 Diversity of Visited Flowers

We recorded a total of 21 angiosperm species visited by hoverflies during our study, distributed in 12 families (Table S3). The Asteraceae was by far the most diverse family visited by hoverflies with 10 species recorded, all the 11 other families were represented by one species each. The five most visited plants species, representing 73% of the total visit number, were Foeniculum vulgare (34%), Raphanus raphanistrum (16%), Dittrichia viscosa (11%), Reichardia picroides (7%) and Leontodon tuberosus (6%). Notable species were three daisies (Anthemis arvensis, Calendula arvensis, Glebionis segetum) and the rockrose Cistus monspeliensis (Figure 5A). In terms of species diversity, the same plant species listed above (minus Anthemis arvensis and Glebionis segetum) were also visited by a large number of hoverfly species (Figure 5B). In relation to the observed seasonal variation of both abundance and diversity of Syrphidae, flowering plants were visited by fewer hoverflies in spring and summer than in autumn (Foeniculum vulgare, Dittrichia viscosa, Reichardia picroides and Leontodon tuberosus) with the notable exception of Raphanus raphanistrum (May) (Table S3).

Interactions between hoverflies and flowering plants are highlighted on the bipartite network (Figure 6). Amongst the species present as singletons, four were captured on various Asteraceae (Anthemis arvensis, Dittrichia viscosa, Glebionis segetum and Reichardia picroides) and five on flowers of Foeniculum vulgare (2 spp.), Erica arborea (2 spp.) and Raphanus raphanistrum (1 sp) (Figure 6, Table S3). All these plant species were also visited by various other hoverfly species except the flowers of Erica arborea. Most of the remainder hoverfly species (18 spp.) behaved generalistically by visiting flowers from various plant species, particularly the most abundant ones (Figure 6). Sphaerophoria scripta visited mainly flowers of Raphanus raphanistrum (26% of the specimens) and Anthemis arvensis (16%) but also from nine other plant species belonging mainly to Asteraceae (7 spp.) (Figure 6). Episyrphus balteatus visited mainly flowers of Foeniculum vulgare (25%) and Leontodon tuberosus (25%) but also from six other species mainly Asteraceae (5 spp.). Chrysotoxum intermedium and Syritta pipiens, each visited flowers of six species with a higher number of visits on Dittrichia viscosa (20% and 33%, respectively) and Foeniculum vulgare (47% and 33%, respectively) (Figure 6, Table S3). Several species appeared to be oligolectic in our sampling. Melanostoma scalare and Melanostoma mellinum visited mainly flowers of Raphanus raphanistrum (88% and 71%, respectively) and only one or two others species (Figure 6). Melanostoma mellinum was the only hoverfly visiting Vicia villosa. Eumerus barbarus were mainly collected when visiting flowers of Foeniculum vulgare (89%), but this hoverfly species visited also in low frequency flowers from two other plant species (Figure 6). Finally, two species appeared to be specialist visitors in our sampling, Eristalinus taeniops on Dittrichia viscosa (4 specimens from Vignola) and Eumerus basalis on Foeniculum vulgare (5 specimens from Loretto and Suartello) but these results have to be confirmed considering the relative small sample sizes.

From the plant point of view, most of the plant species (76%) were visited by at least two hoverfly species. As mentioned before, four plants species (Foeniculum vulgare, Raphanus raphanistrum, Dittrichia viscosa and Reichardia picroides) were key floral resources since they were visited in total by 21 species representing 78% of the total observed hoverfly diversity (Figure 6). On the other hand, five plant species (Asphodelus ramosus, Sherardia arvensis, Urospermum dalechampii, Vicia villosa and an undetermined Poaceae) were each visited specifically by only one hoverfly species (Figure 6, Table S3).