Syrphidae of Southern Illinois: Diversity, floral associations, and preliminary assessment of their efficacy as pollinators

Abstract Syrphid flies (Diptera: Syrphidae) are a cosmopolitan group of flower-visiting insects, though their diversity and importance as pollinators is understudied and often unappreciated. Data on 1,477 Syrphid occurrences and floral associations from three years of pollinator collection (2017-2019) in the Southern Illinois region of Illinois, United States, are here compiled and analyzed. We collected 69 species in 36 genera off of the flowers of 157 plant species. While a richness of 69 species is greater than most other families of flower-visiting insects in our region, a species accumulation curve and regional species pool estimators suggest that at least 33 species are yet uncollected. In order to further the understanding of Syrphidae as pollinators in the Southern Illinois region, we produced a NMDS ordination of floral associations for the most common syrphid species. The NMDS did not sort syrphid species into discrete ecological guilds, and syrphid floral associations generally fit those predicted by traditional pollination syndromes. We also conducted a preliminary analysis of the pollen-carrying capacity of different syrphid taxa, which found several Eristalis species to carry pollen loads comparable to the European Honey Bee, Apismellifera, and showed significant differences in the pollen-carrying capacity of various syrphid species. Notably, the extremely common genus Toxomerus and other small Syrphinae species carried very little pollen, while large and pilose Eristalinae species carried large pollen loads.


Introduction
Syrphidae, known as syrphids, hover flies, or flower flies, is one of the largest families of flies (Diptera), represented by almost 6,000 described species worldwide and 812 in the United States and Canada (Miranda et al. 2013). Syrphids are common flower visitors, though their role in pollinator communities is understudied and often unappreciated (Inouye et al. 2015, Klecka et al. 2018. Syrphids are pollinators of many wild plants (Ssymank et al. 2008), in some cases as important as bees (Forup et al. 2007, Levesque and Burger 1982, Ornduff 1975, Scribailo and Posluszny 1984. Compared to bees, syrphids as a group tend to exhibit higher degrees of generalization (Klecka et al. 2018, Lucas et al. 2018, a propensity to visit flowers considered anemophilous (Ssymank and Gilbert 1993, Inouye et al. 2015, Holloway 1976, and a capacity for temporary floral constancy similar to that of bees (Inouye et al. 2015, Raguso 2020. Syrphid adults feed almost exclusively on pollen, nectar, or honeydew (Rotheray and Gilbert 2011), and at least one species is known to possess branched palynophilic hairs and pollen combing behavior similar to that of bees (Holloway 1976). Because syrphids do not provision their offspring in a nest as do bees, syrphids are able to range over more of the landscape and may carry pollen longer distances than bees (Lysenkov 2009. However, the wide diversity in ecology and morphology makes generalizations about the efficacy of Syrphidae as pollinators difficult (Raguso 2020).
Evaluating the efficacy of flower-visiting taxa as pollinators can be difficult and laborintensive. Often, visitation data alone is used as an indicator of pollinator efficacy, but this assumption may lead to overestimation of the importance of abundant species that do not actually transport, or transport little, pollen between conspecific flowers (King et al. 2013). As such, some measure of pollinator 'quality' is necessary in order to evaluate the importance of different pollinators to plant reproduction (Herrera 1987). Factors influencing pollinator quality include visitation rate, floral constancy, and the amount of pollen carried on a visitor's body (Herrera 1987, Herrera 1989, Tepedino et al. 2011, Erhardt 2008. A species' mean pollen load size has been found to correlate positively with pollen deposition on stigma . In general, flies carry less pollen than bees (Orford et al. 2015, Inouye et al. 2015. While some authors have suggested that the abundance of flies as flower visitors may make up for their inefficiency as pollinators, this may not be true, as visitors that consume floral resources with little or no pollination services may negatively impact plant pollination. Because pollinator efficacy is variable within Syrphidae (Raguso 2020), measures of pollinator quality should be conducted on a per-species basis.
Comprising six level IV US EPA Ecoregions (Woods et al. 2006), Southern Illinois is an area of high biodiversity within the Midwestern US (Stein et al. 2000). Syrphids, however, have historically been little sampled in the Southern Illinois region, with just 8 museum records compiled on GBIF for the 16 southernmost counties of Illinois as of May 2020 (GBIF.org 2020a). The objectives of this paper are 1) to report on Syrphidae diversity in Southern Illinois using data of floral visitors from a region-wide pollinator inventory, and 2) to develop a baseline of understanding of the efficacy of Syrphidae as pollinators of wild plants in Southern Illinois by establishing measures of both the abundance and potential pollinator quality of syrphid species in the region. To our knowledge, this is the first study to provide a regional inventory of syrphid species in Southern Illinois.

Collection methods
Collection methodology was consistent for each of the three studies contributing data, and followed standardized procedures for bee sampling (Droege et al. 2016), with slight modifications made to accommodate the objectives of the studies. Collection events all consisted of targeted hand-netting of floral visitors plus pan trapping. These events were supplemented with additional opportunistic hand-netting. The use of both pan trap and hand-net methods has been shown to be complementary and offset the taxonomic biases of each method alone (Baum and Wallen 2011). Hand-netting was conducted for 80 person-minutes per collection event. Flower-visiting hymenopterans, dipterans, coleopterans, and lepidopterans were collected with aerial insect nets and euthanized in cyanide kill jars. Insects were kept separate by floral associations. Netting was conducted primarily during clear, sunny days. All netting was carried out between 7:00 and 17:00. Plants from which floral visitors were collected were identified to species using published keys of Mohlenbrock (2014).
Pan traps were 7 cm diameter polypropylene bowls (DART manufacturer, stock number 325PC) painted fluorescent blue, yellow, or white and filled with a dilute detergent solution (Dawn Blue dish soap). Traps were set out in sets of three along a transect at a spacing of 10 m. Each set consisted of a blue, a yellow, and a white bowl placed along a line perpendicular to the transect and spaced 5 m apart. Pan trap sets were set along two 50 m transects. Pan traps were left from 4-6 hours during daylight hours, all between 7:15 and 18:00.

Study sites & dates
All specimens reported here were collected during surveys of all flower-visiting taxa. We sampled throughout the southernmost eleven counties of Illinois as well as Randolph county. Sampling focused on federally managed lands (Shawnee National Forest and Crab Orchard Wildlife Refuge) but also included state and private lands. Sites were stratified with respect to land use and major habitat types, and included upland and bottomland hardwood forests, open areas, roadsides, agricultural fields, reclaimed strip mines, limestone glades, and wetlands. The majority of the specimens were collected during three studies: a 2017-2019 regional inventory of flower-visiting insects of Southern Illinois focused on the USFS Shawnee National Forest and USFWS Crab Orchard Wildlife Refuge (Figs 1, 2), a 2017 study of pollinators on agricultural weeds and clover cover crops in agrosystems within Crab Orchard Wildlife Refuge in Williamson County (Fig. 4), and a 2019 inventory of the Illinois National Guard Sparta Training Area in Randolph County, IL (Fig. 3). Taxa other than Syrphidae collected during these studies will be reported elsewhere.
Floral visitors were collected from April-September of 2017, February-September of 2018, and March-July of 2019. Over the three years of collection, 292 sites were visited and 756 collection events conducted, with 55% of these events conducted for the regional inventory, 40% for the agricultural study, and 5% for the Sparta Training Area inventory. Syrphids were collected at 222 of the sites and 445 collection events (Figs 5, 6; Suppl. material 1). The 70 sites and 311 collection events that did not yield any syrphids are not included in these analyses.

Species identification
Specimens were identified to species by JLC and GFGM using published keys of Skevington et al. (2019) and Miranda et al. (2013). Two percent of specimens were identified only to genus level (female Sphaerophoria and Paragus), and 6% were unidentified due to damage. Species-level circumscriptions follow Skevington et al. (2019). All collected specimens are deposited at the Southern Illinois University Carbondale.

Species accumulation and species pool estimators
To determine if syrphids were sufficiently surveyed to capture the species richness of Southern Illinois, a species accumulation curve was generated based on individuals sampled using the rarefaction method (rationale in Gotelli and Colwell (2001). Specimens unidentified to species were not included in this analysis. The regional species pool was estimated by first-order jackknife and bootstrap estimators. Jackknife estimators have been shown to perform better than other estimators where a small proportion of the total species richness has been sampled (Fattorini 2013), as is suggested by the species accumulation curve for this study. These analyses were conducted in the package "vegan" version 2.5-4 in R version 3.5.1 (Oksanen et al. 2019).

Pollen load estimation
To survey the potential efficacy of syrphids as pollinators, we assessed pollen carried on specimens using a modification of the methods of Tepedino et al. (2011). Syrphid pollen coverage was estimated for eight regions of the body: dorsal head, anterior head, ventral head, dorsal thorax, legs, ventral thorax, dorsal abdomen, and ventral abdomen. Syrphids were examined under a dissecting microscope and the pollen coverage for each region was scored either 0 (no pollen grains present on region), 1 (1-several pollen grains on region), 2 (pollen grains separated by >1mm), 3 (pollen grains separated by <1mm), 4 (near complete pollen coverage of region) or 5 (multiple layers of pollen covering region). For the two most abundant species, Toxomerus marginatus and T. geminatus, a subsample of 39 and 25 undamaged specimens were selected for examination, respectively. For all other species, all specimens collected off flowers were examined (386 specimens total). A selection of 30 specimens of Apis mellifera (European Honey Bee) were also examined for comparison to the syrphids. Scopal pollen was ignored in A. mellifera pollen scoring.   To test for significant differences in pollen load size between species, a weighted mean score for each specimen was calculated by downweighting the scores for the three head regions by 1/3, and then averaging all 8 scores. Downweighting the head regions corrected the bias of having three head regions versus two abdominal and thoracic regions. These weighted mean scores were used to run a Kruskal-Wallis non-parametric test and post-hoc Bonferroni corrected pairwise Wilcoxon Rank Sum Tests in R (alpha=.05). Only those species with >6 specimens examined were included in the analysis (18 species and 338 specimens). Apis mellifera was not included in statistical analyses.

Floral association ordination
A non-metric multidimensional scaling (NMDS) ordination of syrphid species by floral association genera was produced to identify guilds of floral visitors (Bray-Curtis distance). The 18 most abundant floral visiting syrphid species were included, excepting Toxomerus jussiaeae, a specialist that was collected only off of Ludwigia peploides.

Data resources
A table of coordinates for collection sites is given in Suppl. material 1. A spreadsheet of syrphid occurrence data reported here is in Suppl. material 2. Raw values for the syrphid pollen analysis are reported in Suppl. material 3.

Faunal composition
The floral visitor surveys used in this study yielded a total of 33,563 insect specimens, of which 1,477 were Syrphidae (4.40% of the total collection). The rest of the collection was comprised of 60.21% bees, 5.66% non-bee Hymenopterans, 12.92% non-syrphid Diptera, 5.07% Coleoptera, and 11.74% Lepidoptera; these will be reported on elsewhere.
The 1477 syrphids represent 69 species belonging to 36 genera (Table 1). Sixty-four of these species were identified as valid described species, 1 identified as a currentlyundescribed species ('Palpada undescribed species 1' according to Skevington et al. 2019), and 4 taxa identified to species groups or affinities. Collections of note include Microdon aurulentus Fig. 7 , which is known only from <10 records from US States Ohio-Georgia-Pennsylvania (GBIF.org 2020c, Skevington et al. 2019). Palpada agrorum Fig. 8, represented by 6 records in this collection, is a common species along the Gulf Coast and into Oklahoma but had previously not been collected as far north along the Mississippi river as Illinois. Introduced species comprised 1.9% of the total collection and include Eristalis arbustorum (n=1), Eristalis tenax (n=3), Syritta flaviventris (n=1), and Syritta pipiens (n=23).

Species pool estimate
The species accumulation curve (Fig. 9) does not approach asymptotic, suggesting that surveying was not adequate to capture the full syrphid species richness of Southern Illinois. First-order jackknife estimates the total species pool to be 101.93 (standard error 6.07). Bootstrap estimate predicted a lower number, 82.47 species (standard error 3.12). These values may underestimate the real regiona richness, as discussed below.

Floral associations
Of the 1477 syrphid specimens collected, 1047 (70.89%) were collected by hand-netting off of flowers and 107 (7.24%) were collected by hand-netting while flying. Syrphids were collected from the flowers of 157 plant species representing 47 plant families in Table 2     List of all floral taxa from which syrphids were collected. Plant species codes (as reported in Table  1) are comprised of the first three letters of the genus and specific epithet, and family codes are comprised of the first four letters of the family name. Floral associations were occasionally identified only to genus level, and these are reported in Table 1 as the first three letters of the genus name (except Tradescantia and Tragopogon, for which the first four letters are used).

Floral association NMDS
The NMDS of floral associations is given in Fig. 10.

Pollen load comparison
Pollen scores for each of the 18 species analyzed are summarized in Fig. 11. Species ranged from mean scores of 2.33 (Eristalis stipator) to 0.15 (Toxomerus politus).

Discussion
The results of this inventory have provided a baseline of Syrphidae species richness and relative abundance in the Southern Illinois region. The genus Toxomerus represents the majority of the flower-visiting syrphids, comprising 69% of all syrphid individuals collected. Though sampling for this inventory was thorough (756 collection events), the species accumulation curve (Fig. 9) suggests that sampling failed to capture much of the regional syrphid diversity; the curve rises at a nearly constant slope after the 400th individual, rather than leveling off as expected if the full regional richness was captured. The first-order jackknife estimate (102 species) suggests that as little as 68% of the regional species pool may be known. This is, however, very likely an underestimation, as the jackknife estimates the species pool of the sites, rather than the region as a whole. Additionally, some habitats may have been undersampled, wetlands, which generally contain high syrphid diversity. The first-order jackknife estimation of 102 species may be used as a lower limit for the regional species pool, though more sampling will be required to fully document the syrphid richness of the Southern Illinois region. While collections for this study encompassed a broad range of floral visitors, collection targeting syrphids would be more productive; malaise traps should be employed, which have been shown to be efficient in capturing syrphid diversity (Burgio and Sommaggio 2007). One group likely to be undersampled by our methodology is the genus Microdon, a group of ant nest predators which do not regularly associate with flowers (Duffield 1981). The single record of Microdon globosus visiting Medicago lupulina is of note, as there are very few observations of Microdontinae visiting flowers (M. Reemer, personal communication). Just four Microdon specimens were collected (3 in pan traps), constituting four different species. As Microdon are not typically flower visitors, however, they are unlikely to be pollinators of any import in our region. At 69 species, Syrphidae is one of the most diverse groups of floral visitors collected in our 2017-2019 surveys. Bee families yielded from 19 (Colletidae) to 67 (Apidae) species, and butterflies including skippers (Lepidoptera: Rhopalocera) yielded 72 species. One reason for the high syrphid richness documented may be that Southern Illinois is predominantly rural; syrphid abundance and richness have both been shown to decline with increasing urbanization (Udy et al. 2020). Seventeen species (25% of total) reported in this study have not been recorded in Illinois before (Table 1), according to records in Skevington et al. (2019) and 10 datasets in GBIF.org (2020d). This demonstrates the large gap in our knowledge of syrphid distribution in the Eastern US, stressing the need for further studies of this diverse group of pollinators.
Syrphids were collected from a wide range of flowers (157 species). Floral associations generally followed the predicted pattern for non-carrion fly pollination syndromes: white, yellow, green, or brown flowers in color, radial symmetry, exposed pollen and nectar (Faegri and Pijl 1979). Over half of flower visits observed were to Asteraceae (40 floral species, 28 syrphid species collected from). The 25 most common floral associations (all those with more than 7 syrphids collected) have either white perianths, yellow perianths, or both (as in the bicolored capitula of Erigeron and Leucanthemum) except for Trifolium pratense (pink flowers) and Plantago lanceolata (anemophilous without showy flowers, though anthers are large and white). However, the pink flowers of T. pratense may not differ visually from white flowers to syrphids, as flies exhibit low sensitivity to red light (Lunau 2014). The frequency of Fabaceous flowers as floral associations (7.39% of the Mean weighted pollen scores for each species analyzed, with standard error bars. Blue bars represent species in subfamily Eristalinae, orange bars Syrphinae, and grey bar Apis mellifera. Shared letters above bars denote no significant pairwise difference (pairwise Wilcoxon Rank Sum Test, alpha=0.05 with Bonferroni correction). Apis mellifera is included for comparison to syrphids but was not included in pairwise tests.
total; second most commonly visited plant family) is of note as the Fabaceous flowers collected off of (mostly Trifolium) possess tubular rather than open corollae, contrasting with the classical fly pollination syndrome. However, pollination syndromes have been shown to be poor predictors of floral visitation (Ollerton et al. 2009), and syrphids have been documented to forage on Trifolium species (Larson et al. 2014) and are likely pollinators.
The NMDS of floral associations failed to sort syrphid species into discrete guilds (Fig. 10), though some clustering is apparent. Syrphinae species are all (except for Ocyptamus fuscipennis) within ±0.5 of axis 1, whereas Eristalinae is far more evenly distributed in the ordination space. Eristalinae does cluster into two long groups on either side of axis 1, though this grouping is loose and does not reflect strong similarity in floral association within the Eristalinae. Five small predatory Syrphines (Paragus haemorrhous, Sphaerophoria contigua, Toxomerus boscii, T. geminatus, T. marginatus) and two small Eristalines (Orthonevra nittida, Syritta pipiens) are grouped around the Erigeron vector, the strongest vector in the ordination (p<0.001). Each of these species are common and exhibit low (below 1 mean pollen load score) pollen-carrying ability (Fig. 11). This group may act as abundant but low-quality pollinators of Erigeron and other weedy plant species.  (Fig. 11), which does not support grouping the whole genus as an ecologically similar guild. The grouping of Toxomerus (except for T. politus) in ordination space is in contrast to Eristalis, the three species of which are widely separated in the NMDS. Milesia virginiensis and Ocyptamus fuscipennis are at the most positive values of axis 1; both species inhabit forests (Skevington et al. 2019), which is reflected by the vectors in their quadrant of the ordination (forest plants such as Tradescantia and Actinomeris).
Examination of pollen loads showed significant differences in pollen carrying capacity of syrphid species (Fig. 11). Of note, the pollen analysis scored pollen coverage rather than number of pollen grains. Pollen coverage may be more important than pollen count in successful pollination, though we are aware of no studies assessing this. Many pairwise comparisons were not significant likely due to low sample sizes. Even so, some trends are clear. The six syrphids with the highest pollen scores were all in the tribe Eristalini of the Eristalinae, large bodied, and pilose: Eristalis stipator, E. transversa, E. dimidiata, Mallota bautias, Helophilus fasciatus, and Palpada vinetorum. This is expected, as pilosity and size are positively correlated with pollen load in flies (Inouye et al. 2015). This generalization is not a rule, however; Milesia virginiensis is large and pilose yet scored in the bottom five of the 18 species analyzed. The three Eristalis species analyzed all scored within ±0.25 of Apis mellifera, with Eristalis dimidiata even scoring slightly above. The high pollen load scores of these Eristalines has definite implications for the quality of the species as pollinators. Still, other factors such as floral constancy and pollen deposition on stigmas need be considered in order to further quantify their efficacy (Herrera 1987).
Several syrphids were collected in February, extremely early for floral visitors in the region: Eupeodes cf. americanus, Eristalis dimidiata, and Syrphus torvus. These specimens were collected off of a cultivated Hamamelis virginiana (American witch-hazel) on the SIUC campus. Considering the high pollen scores of Eristalis dimidiata and Eupeodes, these species may be important pollinators in the very early spring, before bees and most other floral visitors are flying.
The high pollen scores of the tribe Eristalini contrast greatly with many of the Syrphinae and less pilose Eristalinae. Orthonevra nitida and Toxomerus politus carried almost no pollen, and are thus unlikely to pollinate with any consistency. Toxomerus marginatus and T. boscii each scored ~0.5 on average, frequently carrying no pollen at all. T. geminatus scored slightly higher, though pairwise tests between the Toxomerus species were not significant. This is of note because Toxomerus was the most abundant genus of syrphids by far (69% of total). The similarity in pollen load size and floral association (Fig. 10) suggests that Toxomerus marginatus, T. geminatus, and T. boscii may be treated as a guild of similar pollinators. While their abundance may compensate for their low quality, consumption of floral resources by Toxomerus without pollen deposition on stigmas may harm plant reproduction. In contrast, Eristalis spp and other large pilose Eristalini syrphids are likely to be important pollinators where they occur, though their relatively low abundance means that these species are not ubiquitous across the Southern Illinois landscape and their importance as pollinators will be localized to where they are abundant.