Introduction
The bee genus Hylaeus Fabricius, 1793 (Hymenoptera: Colletidae) consists of generally small, black-and-yellow, superficially wasp-like bees, of which three species have established non-native populations in North America (Sheffield et al. Reference Sheffield, Dumesh and Cheryomina2011). Hylaeus (Hylaeus) leptocephalus (Morawitz, 1870) was the first species to be recorded in 1912 in North Dakota, United States of America (Snelling Reference Snelling1970, as H. stevensi) followed by H. (Spatulariella) punctatus (Brullé, 1832) in 1981 in Los Angeles County, California, United States of America (Snelling Reference Snelling1983) and finally H. (Spatulariella) hyalinatus Smith, 1842, collected in 2001 in central New York State, United States of America (Ascher Reference Ascher2001). These species have since been inventoried in cities across the continent with records from California (Snelling Reference Snelling1970; Snelling Reference Snelling1983; Ascher Reference Ascher2001) and eastern North America including New York City, New York (Matteson et al. Reference Matteson, Ascher and Langellotto2008); Ontario, Canada (Buck et al. Reference Buck, Paiero and Marshall2005; Sheffield et al. Reference Sheffield, Dumesh and Cheryomina2011); and Washington, District of Columbia, United States of America (Ascher et al. Reference Ascher, Gambino and Droege2006). The two more recently detected species belonging to the subgenus Spatulariella Popov, 1939 are not yet generally distributed across North America but H. punctatus has been detected in interior cities in Colorado, United States of America (Scott et al. Reference Scott, Ascher, Griswold and Nufio2011) and both species are known from Chicago, Illinois, United States of America (Tonietto and Ascher Reference Tonietto and Ascher2009) and southern Ontario (Richards et al. Reference Richards, Rutgers-Kelly, Gibbs, Vickruck, Rehan and Sheffield2011; Onuferko et al. Reference Onuferko, Kutby and Richards2015). Exotic bees are a concern because of their competition with indigenous bees for floral and nesting resources, their unpredictable impact on pollination networks, and their potential to transmit pathogens to indigenous populations (Cane Reference Cane2003; Aizen et al. Reference Aizen, Morales and Morales2008; Vergara et al. Reference Vergara, James and Pitts-Singer2008; Stout and Morales Reference Stout and Morales2009). They have the potential to spread widely even if introduced as a singly mated female (Zayed et al. Reference Zayed, Constantin and Packer2007). The ongoing increase in accidental introduction and establishment of exotic bees (Bartomeus et al. Reference Bartomeus, Ascher, Gibbs, Danforth, Wagner, Hedtke and Winfree2013) compounded by deliberate introduction of additional generalised solitary bee species (Batra Reference Batra2003) demands concerted biomonitoring in invaded areas. We report, for the first time, the presence of a new exotic bee species, H. (Hylaeus) communis Nylander, 1852, for North America, assess its threat to native pollinator networks and evaluate potential pathways for its continental dispersal.
The success of an exotic bee species in a newly colonised area will be mediated by the extent to which it can adapt to novel floristic and nesting opportunities (Barthell et al. Reference Barthell, Frankie and Thorp1998). For instance, specialist bees that restrict pollen collecting to a single floral host taxon (e.g., genus) and with immobile nesting requirements (e.g., obligate association with a particular natural substrate) are unlikely to establish themselves easily because they would be restricted to their host plants and the topography of their local environment (Giannini et al. Reference Giannini, Chapman, Saraiva, Alves dos Santos and Biesmeijer2013; Silva et al. Reference Silva, Gonzalez, Melo, Lucia and Alvarez2014). Conversely, polylectic diets allow bees to find food in newly colonised areas (Gibbs and Sheffield Reference Gibbs and Sheffield2009) and most exotic bee species in North America occupy a variety of nesting substrates that can be transported easily through international trade (e.g., pithy stems, cracks in window frames) (Cane Reference Cane2003). Notably, however, two Palearctic species of Chelostoma Latreille, 1809 (Hymenoptera: Megachilidae) oligolectic on Campanula Linnaeus (Campanulaceae), Hoplitis (Hoplitis) anthocopoides (Schenck, 1853) (Hymenoptera: Megachilidae) oligolectic on Echium Linnaeus (Boraginaceae), and Lithurgus chrysurus Fonscolombe, 1834 (Hymenoptera: Megachilidae) oligolectic on Centaurea Linnaeus (Asteraceae) are well established locally in a limited area of eastern North America where their host plants occur (Eickwort Reference Eickwort1970, Reference Eickwort1980; Sheffield et al. Reference Sheffield, Dumesh and Cheryomina2011). Whether H. communis is likely to have large impacts on the North American bee fauna and on pollination networks, as may be the case for certain exotic bees such as Megachile (Callomegachile) sculpturalis Smith, 1853 (Hymenoptera: Megachilidae) (Laport and Minckley Reference Laport and Minckley2012), may be assessed through consideration of the abiotic and biotic factors relevant to delimiting its range and expansion from its currently limited sites of adventive occurrence.
We began by profiling H. communis in terms of the flexibility of its climatic, floristic, habitat, and nesting affinities. We then compared its floral preferences in both European and North American pollination networks as well as its habitat preferences in both locations to assess how and whether H. communis might spread across the continent. To facilitate future biomonitoring efforts, we provided diagnostic material to distinguish H. communis from other Hylaeus found in North America.
Identification
Hylaeus communis belongs to subgenus Hylaeus, males of which are recognised in North America by the combination of omaulus rounded (as opposed to carinate in subgenus Spatulariella); margins of interantennal elevation nearly parallel between antennal sockets, terminating on frons well above level of upper margins of antennal sockets (as opposed to margins of interantennal elevations sharply convergent between antennal sockets and ending little, if any, above upper margins of antennal sockets in subgenera Metziella Michener, 1942, Paraprosopis Popov, 1939, and Prosopis Fabricius, 1804); apical lobes of S7 pectinate with four apical lobes, apical process of eighth sternum bilobate (as opposed to spatulate in Spatulariella or with apex rounded or truncate in Prosopis); and scape often broadened apically (but unlike in subgenus Cephalylaeus Michener, 1942, in which it is not much broader than long) and concave beneath (Michener Reference Michener2007). Females of subgenus Hylaeus found in North America can be recognised by omaulus not carinate (carinate in Spatulariella); gena conspicuously narrower than the eye; T1 polished with punctures fine and scattered and without apicolateral patch of highly plumose, appressed pale pubescence (punctures more conspicuous and well defined in subgenera Paraprosopis and T1 with a conspicuous apicolateral setae patch in Prosopis); and first tergum facial foveae ending nearer to inner eye margin than to lateral ocellus (ending nearer to lateral ocellus in subgenus Paraprosopis). With reference to Macek et al. (Reference Macek, Straka, Bogusch, Dvořák, Bezděčka and Tyrner2010), the Dathe (Reference Dathe1980, Reference Dathe2000) keys to Palearctic Hylaeus sensu stricto and Snelling’s (Reference Snelling1970) key to North American species, males and females of H. communis can be shown to be distinctive among European and North American members of subgenus Hylaeus in possessing the following unique combination of characters (Fig. 1).
Images of Hylaeus communis (male and female). Photographs of side-view and face of Hylaeus communis female (A, B) and male (C, D) specimens.
Males
Yellow markings on clypeus reduced to a variable extent, usually not extending to lateral margins, so that the yellow of the clypeus forms an isolated, longitudinal, I-shaped bar with irregular borders, but with yellow covering at least the centre third; supraclypeal mark often restricted to a yellow patch, so the extension of the interantennal elevation above the level of the antennal sockets is less conspicuous than in other species of the subgenus that have larger supraclypeal markings; lateral face marks not fully separated from eye margin but with pointed median extension ending over antennal sockets; scape sometimes spotted yellow and moderately broad, with width greater than 0.5 times its length; second flagellomere about as long as broad; flagellum long and black; metatibiae yellow on less than basal half; and sternum 8 not widened apically and glabrous.
Females
Face relatively elongate and with eyes strongly converging below; clypeus with punctures distinct and rather uniformly distributed throughout including basal half; lateral face marks angled slightly inwards towards the antennae above and transversally truncate beneath; fovea short; malar space shorter than broad; pronotum mostly black, with yellow spot on distal portion of pronotal lobes; tegula with yellow spot; mesepisternum sharply angulate but not carinate anteriorly; propodeum with enclosure divided by strong transverse carina, with dorsal surface of enclosure with strongly elevated longitudinal ridges separating smooth interspaces; tergite 1 polished with fine punctures; subsequent tergites have finer, denser punctures; abdomen (T1) without lateral fringes; and range in head width and length between 1.5 and 1.6 mm.
Challenges to identification
Males of H. communis found to date in eastern North America have been readily identified due to their distinctive yellow facial markings, as the reduction of markings on the clypeus laterally and on the supraclypeal area result in three separated yellow patches on the face. This is unlike other species with which it might be confused where the yellow markings on the lower face are contiguous (H. saniculae (Robertson, 1896), with more reduced facial markings, has broad depressions on the upper face and is otherwise unlike the much larger H. communis). However, males of H. communis in Europe often have more fully developed yellow facial markings. Identification of such individuals, if found in North America, would be more challenging. They would most likely be confused with the similarly sized H. annulatus, another yellow-faced species of subgenus Hylaeus (unlike consubgeneric H. leptocephalus, which differs conspicuously in having white or pale cream markings). Hylaeus annulatus differs from H. communis in having a broader scape extensively marked with yellow laterally (scape black or with small sublateral yellow spot in H. communis) and a broader face with inner eye margins less strongly convergent below.
Females of H. communis are more difficult to identify, as they are similar to H. annulatus. Noteworthy features of H. communis include the longer face with inner eye margins more strongly converging below, presence of a yellow spot on the tegula (usually absent in H. annulatus and in H. modestus modestus, a superficially similar species of subgenus Prosopis). The clypeus in H. communis is rather uniformly punctate throughout whereas in H. annulatus only the apical half has distinct, regular punctures. The propodeal enclosure of H. communis has the dorsal enclosure regularly striate whereas that of H. annulatus is more irregularly rugose. Both of these Hylaeus (Hylaeus) species differ conspicuously in propodeal sculpture from yellow-marked Hylaeus (Prosopis) species such as H. modestus in that only the latter lacks the strong transverse carina that delimits the dorsal and posterior face of the propodeal enclosure in the former. Hylaeus leptocephalus females are readily distinguished by the white rather than yellow integumental markings and also differ in other characters such as the duller sculpturing of the mesepisternum.
Natural history
Hylaeus communis is adapted to a wide range of spatiotemporal environmental conditions in its native range, being found throughout Europe, North Africa, and Asia Minor (Koster Reference Koster1986; Ornosa and Ortiz-Sánchez Reference Ornosa and Ortiz-Sánchez2004). It is recorded from the United Kingdom and Ireland north to Sweden, Finland, and Russia, south to Morocco, Tunisia, and Lebanon, and east to Russia and Afghanistan. Dathe (Reference Dathe2000) clarified the distribution of the Mediterranean species H. (H.) deceptorius (Benoist, 1959), which had historically been confounded with H. communis group.
Hylaeus communis has a broad niche encompassing many different habitats, nesting substrates, and floral resources. It is bivoltine and has an extended activity period, with adults flying from May to early October (Koster Reference Koster1986; Westrich Reference Westrich1989; Amiet et al. Reference Amiet, Müller and Neumeyer1999). The species occurs in Eurasian montane conditions, forest glades and clearings, hedgerows, thickets, sand and clay pits, and railway embankments, and is found regularly in urban gardens, parks, and waste areas (Koster Reference Koster1986; Westrich Reference Westrich1989; Ornosa and Ortiz-Sánchez Reference Ornosa and Ortiz-Sánchez2004; Banaszak et al. Reference Banaszak, Kowalczyk and Nauk2007). In these areas, it occupies a variety of nest sites, including: existing cavities, e.g., insect burrows, cracks in old window frames; pithy stems, e.g., of Rubus Linnaeus (Rosaceae), Sambucus Linnaeus (Adoxaceae), Rosa Linnaeus (Rosaceae), Prunus Linnaeus (Rosaceae), Arundo donax Linnaeus (Poaceae), and Rhamnus Linnaeus (Rhamnaceae); tree bark, e.g., of Pinus Linnaeus (Pinaceae) and abandoned oak (Quercus Linnaeus (Fagaceae); galls of Andricus Hartig, 1840 (Hymenoptera: Cynipidae) and reed galls of Lipara lucens Meigen, 1830 (Diptera: Chloropidae); and nest boxes (Westrich Reference Westrich1989; Janvier Reference Janvier2012). Preferred substrates include empty stalks of Leonurus Linnaeus (Lamiaceae) and Rubus idaeus Linnaeus (Rosaceae), selecting cavities with a canal diameter of 2–6 mm (mainly 3–4 mm) and a canal length of 1–20 cm (most often around 15 cm) (Gosek et al. Reference Gosek, Ruszkowski and Kaczmarska1996). The bee is rather broadly polylectic, having been observed on 40 different plant taxa (Table 1) (Westrich Reference Westrich1989; Ornosa and Ortiz-Sánchez Reference Ornosa and Ortiz-Sánchez2004; Carvalheiro et al. Reference Carvalheiro, Barbosa and Memmott2008), 37 of which have naturalised populations in the United States of America and Canada (United States Department of Agriculture, Natural Resources Conservation Service 2016).
Floral hosts for Hylaeus communis in its native European range as listed in Westrich (Reference Westrich1989), Ornosa and Ortiz-Sánchez (Reference Ornosa and Ortiz-Sánchez2004), and Carvalheiro et al. (Reference Carvalheiro, Barbosa and Memmott2008).
Note: Species and genera with established populations in North America are in bold.
Methods
Generalism of Hylaeus communis in Europe
To assess whether H. communis exhibits high adaptability within its native range, we compared its habitat and floristic affinities with those of other European species. We obtained data from Koster (Reference Koster1986) on the Hylaeus of the Netherlands. Over 1000 specimens consisting of 13 species were studied, having been captured between 1950 and 1980. We examined this historical data set using two bipartite networks (Dormann et al. Reference Dormann, Gruber and Fruend2008; Dormann Reference Dormann2011), one linking the Hylaeus species with 20 different floral hosts and the second associating the bees with the habitats in which they were found. Specifically, the data set consisted of the percentage of the total number of interactions for each bee species and floral host or for each bee species and habitat pair. Percentage values were rescaled from one to 10 in Koster (Reference Koster1986). Analyses were restricted to nine species that had at least 20 observations each to avoid drawing false inferences on species generalism from inadequate sample sizes (Dormann Reference Dormann2011). Among the species examined was H. hyalinatus, providing a useful reference point for invasiveness given that it is currently established in North America (Ascher et al. Reference Ascher, Gambino and Droege2006). We calculated the paired difference index (PDI) (Poisot et al. Reference Poisot, Canard, Mouquet and Hochberg2012) for these Hylaeus species to estimate the degree of generalism that they exhibited in each of the two networks. Poisot et al. (Reference Poisot, Canard, Mouquet and Hochberg2012) demonstrated that this index is robust to under sampling, has a high signal-to-noise ratio across simulated trials, and can be compared over a range of bipartite networks with absolute values below 0.5 indicating species generalism. The significance of PDI in the two networks could not be tested in this analysis because the floral and habitat visitation data in Koster (Reference Koster1986) were rescaled, precluding the generation of valid null models that are usually based on frequency values (Dormann et al. Reference Dormann, Fruend, Bluethgen and Gruber2009).
Hylaeus communis in North America
In 2012, specimens of H. communis were discovered in Montréal and Mont-Saint-Hilaire, Québec, Canada by E.N. and K.T.M., respectively (Figs. 1–2). We observed Hylaeus during sunny and calm weather conditions between 12:00 PM and 06:00 PM, with temperatures ranging from 24 °C to 33 °C. In Montréal, bee samples were collected every two weeks from May to September with pan traps (400 mL yellow, blue, and white Krylon® (Krylon Products Group, Cleveland, Ohio, United States of America) painted bowls filled with soapy water with three bowls per 1000 m2 sampling area) and sweep netting. In Mont-St-Hilaire, bee samples were taken from July to August using sweep netting. Floral hosts were only recorded for specimens obtained from Mont-Saint-Hilaire. Samples were identified by J.S.A. as H. communis in March 2013 with reference to taxonomic revisions (Dathe Reference Dathe1980, Reference Dathe2000) and reference specimens in the American Museum of Natural History (New York, New York, United States of America). All voucher specimens have been deposited at McGill University’s Lyman Entomological Museum in Sainte-Anne-de-Bellevue, Québec.
Map of the study sites where Hylaeus communis was found. Sites are coded as garden, park, or semi-natural habitats. The island of Montréal and city of Mont-Saint-Hilaire are highlighted in yellow. The inset situates the two cities in the context of Québec, Canada.
In Montréal, H. communis was collected by E.N. in community gardens, cemeteries, and parks. In Mont-Saint-Hilaire, the species was collected by K.T.M. in semi-natural areas (e.g., meadows, roadsides) and in residential gardens (Supplementary Table S1). Semi-natural areas consisted of abandoned agricultural lots and drainage ditches dominated by a mixture of grasses and forbs, including Ageratina altissima (Linnaeus) King and Robinson (Asteraceae), Daucus carota Linnaeus (Apiaceae), Lythrum salicaria Linnaeus (Lythraceae), Solidago Linnaeus (Asteraceae), and Trifolium Linnaeus (Fabaceae). Residential and community gardens were characterised by beds of horticultural flowers, fruits, and vegetables, commonly having Astilbe Buchanan-Hamilton ex. Don (Saxifragaceae), Cucumis sativa Linnaeus (Solanaceae), Echinacea purpurea (Linnaeus) Moench (Asteraceae), Hydrangea Linnaeus (Hydrangeaceae), and Solanum lycopersicum Linnaeus (Solanaceae). Parks and cemeteries comprised open grassy expanses dotted with small woodlots and patches of either weeds or horticultural flowers.
Given that we were the first to record H. communis in North America, we sought to understand how it was adapting its foraging behaviour to its novel environment and which habitats it preferred. We chose not to use niche modelling to determine its potential range in North America due to possible data quality issues arising from the European occurrences upon which the analysis would have been based. Hylaeus are notoriously difficult to identify and there are numerous examples of unresolved species complexes in this group (Dathe Reference Dathe1980, Reference Dathe2000; Magnacca and Brown Reference Magnacca and Brown2010). Instead, we focussed on data gathered across 29 sites in Mont-Saint-Hilaire, where native and exotic Hylaeus were found; examining the bipartite network between the Hylaeus species that we surveyed and their respective floral hosts (Dormann et al. Reference Dormann, Gruber and Fruend2008). The same index, PDI, was calculated as with the European networks. Paired difference index calculations, in this case, were based on frequency values so we could test their significance by contrasting observed visitations against expectations from a null model (Dormann et al. Reference Dormann, Fruend, Bluethgen and Gruber2009). The null models scattered the observed number of interactions over all flowering plant species while keeping the number of links constant and the marginal totals identical to those observed at each site. To help interpret results, flower species were classed as being native, naturalised, or horticultural using the PLANTS Database (United States Department of Agriculture, Natural Resources Conservation Service 2016).
We examined the habitat affinities of the Hylaeus community in Montréal and Mont-Saint-Hilaire using a partial redundancy analysis (RDA) (Legendre and Legendre Reference Legendre and Legendre2012). We began by subjecting the frequency matrix (52 sites×8 species) of bee observations to a Hellinger transformation (square root of the relative abundance of each species at each site) to alleviate the masking effect of dominant species on the covariance structure of the ordination (Legendre and Gallagher Reference Legendre and Gallagher2001). To control for differences in sampling methods between inventories in Montréal and Mont-Saint-Hilaire, we treated a factor coding for city as a conditioning variable in the RDA. This removed any city-specific sampling biases from the community matrix; the residuals were then analysed using principle components analysis (PCA). We coded sites as belonging to one of three habitat categories, namely gardens (n=23; community and residential gardens), parks (n=14; cemeteries and municipal parks), or semi-natural areas (n=15; drainage ditches, fallow fields). Habitats were regrouped in this way so that sample sizes per category would be comparable and to reflect the predominating floral cover gradient: gardens have mostly horticultural plants, semi-natural areas native and naturalised forbs, whereas parks generally comprise a mix of flower types. To see if community composition was distinct in these three habitat types, we projected 95% confidence ellipses for garden, parks, and semi-natural sites onto the final PCA biplot.
All statistical analyses were conducted using R version 3.2.0 (R Core Team 2015) with the packages bipartite version 2.05 (Dormann et al. Reference Dormann, Gruber and Fruend2008) and vegan (Oksanen et al. Reference Oksanen, Blanchet, Kindt, Legendre, Minchin and O’Hara2015). Data used in the analysis are given in Supplementary Tables S2–S5.
Results
European network analysis
Results from the habitat and floral bipartite network analyses (Table 2) show that H. communis is, overall, the greatest generalist of all the Hylaeus species studied by Koster (Reference Koster1986) (Fig. 3). The PDI values for H. communis were the only ones approaching the cutoff value of 0.5 for species generalism in both networks (habitat network: 0.58; floral network: 0.54). Moreover, H. communis may be a greater generalist than the other European species with previously established North America populations, H. hyalinatus, which has an equal PDI value to that of H. communis in the floral network and a higher PDI value in the habitat network. Despite being highly adaptable, H. communis was most frequently observed foraging from the flowers of Rubus fruticosus Linnaeus (Rosaceae) and was most often found in gardens and parks.
Habitat and floral bipartite networks for the European Hylaeus community. Bipartite graphs for bee communities studied in Koster (Reference Koster1986). Panel (A) depicts the network between European flora (left-hand boxes) and Hylaeus species (right-hand boxes); panel (B) shows the network between European habitats (right hand boxes) and Hylaeus species (left-hand boxes). The two networks were based on different sets of data. The size of each box is scaled to the prevalence of a given bee, flower, or habitat in their respective networks. The percentage of the total number of bee-flower and bee-habitat interactions per species, rescaled from 1 to 10, is indicated by the thickness of connecting lines. Flowers and habitats visited by H. communis are highlighted in yellow.
Paired difference index (PDI) values for Hylaeus species in three separate networks: (1) between bees and European flora, (2) between bees and European habitats, (3) between bees and Québec flora.
Note: The significance of PDI values was tested for the Québec network only, with ** indicating a P-value<0.01.
North American floral network analysis
A total of 172 specimens of H. communis were discovered in Montréal (123 individuals) and Mont-Saint-Hilaire (49 individuals), representing 23% and 29% of all Hylaeus captured in each location, respectively. Hylaeus communis was the most abundant Hylaeus in Montréal and was second most abundant in Mont-Saint-Hilaire following H. (Prosopis) modestus modestus Say, 1837 (55 individuals). The bee was widespread, found in 15 sites in Montréal and 16 sites at Mont-Saint-Hilaire (Fig. 2). A total of seven Hylaeus species were inventoried in 2012 in addition to H. communis, of which four native species were found in both cities: H. (Prosopis) affinis (Smith, 1853), H. (P.) modestus modestus, H. (Hylaeus) annulatus (Linnaeus, 1758), and H. (H.) mesillae cressoni (Cockerell, 1907). Three exotic species were found only in Montréal: H. hyalinatus Smith, 1842, H. leptocephalus (Morawitz, 1871), and H. punctatus (Brullé, 1832).
Results from the bipartite analysis (Fig. 4; Table 2) of the Mont-Saint-Hilaire community indicated that H. communis had significantly (P<0.01) the lowest PDI (0.81) of all the Hylaeus in the network but that it was still higher than the 0.5 threshold for species generalism. Hylaeus communis was found on the greatest proportion of the 28 floral species (63%) in this network but it tended to prefer both naturalised and horticultural plants. A total of 69% of all H. communis captures were on non-native plants, half of which were on flowers unique to residential gardens. Furthermore, the bee visited seven exotic species not visited by the other Hylaeus species at these study sites (but Melilotus albus Medikus (Fabaceae) is known to be a favoured host of H. leptocephalus elsewhere (Snelling Reference Snelling1970)). Of all the flowers visited, H. communis had the strongest affinity for D. carota, then Hydrangea paniculata Siebold, Melilotus albus, and Rudbeckia hirta Linnaeus (Asteraceae), of which only R. hirta is native to North America.
Floral bipartite network for the Québec Hylaeus community. Bipartite graph for the Québec Hylaeus community sampled in 2012, following the same layout as Figure 3. Boxes on the left represent flower species and boxes on the right species of Hylaeus. Dashed lines distinguish between horticultural plants, exotic/naturalised plants, and native plants.
North American habitat preferences
A total of 63% of the total residual variance in the bee community matrix was explained by the first two principal components of the PCA (Fig. 5). The first principle component (PC1; 40% variance explained) polarises garden sites against parks and semi-natural sites. The goodness of fit of the factor coding for habitat type with the ordination is significant (R 2=0.26, P<0.001). The 95% confidence intervals surrounding group centroids show that the community composition of garden sites is distinct from that of semi-natural areas and parks but that parks and semi-natural areas are not distinct from each other. Consistent with the preference of H. communis for horticultural flowers, the bee is strongly associated with residential and community gardens: the species vector for H. communis clearly points towards the centroid for garden sites. All exotic Hylaeus tend to be found in gardens and all native Hylaeus in parks and semi-natural areas. Noteworthy is the negative correlation between H. communis and the native H. affinis along PC1, with the latter bee favouring semi-natural habitats.
Transform-based principal component analysis of the Québec Hylaeus community. Biplot from the principal component analysis of the Hylaeus community matrix (52 sites×8 species). The solid circles represent sampling sites and have been colour-coded according to habitat type (garden, park, or semi-natural); 95% confidence ellipses surround the centroids of each habitat type. Vectors depict the top three bee species with the greatest correlation with the axes of the ordinal space. Crosses are used for remaining Hylaeus. Abbreviations for species labels take the first letter of the genus (capitalised) and the first four letters of the specific epithet (lowercase). Exotic Hylaeus are indicated with asterisks.
Discussion
Cane (Reference Cane2003) predicted that in addition to the 15 alien bee species already naturalised in North America, new arrivals would likely predominately be temperate-adapted stem nesters from Europe with broad larval pollen diets. Hylaeus communis certainly conforms to this description, having a wide range across the temperate Palaearctic and both a generalist foraging habit and flexible nesting preference. It is predisposed to successful anthropogenic dispersal in that it likely survives long-distance transport by hibernating in nests as prepupae (Gosek et al. Reference Gosek, Ruszkowski and Kaczmarska1996) and by using multiple nesting materials easily available in cargo holding areas (e.g., pithy stems, old wood). The bee would thus not only meet its resource needs in novel North American environments but also take advantage of human-mediated transport to disperse.
The potential of H. communis to spread throughout North America is high given that it is just as adaptable in its use of floral hosts as an exotic bee species already well established in North America, H. hyalinatus. Moreover, we found that H. communis exploits more habitats than any other European Hylaeus species, adapting to the widest range of environmental conditions. The introduction of the bee will be facilitated by the existence of large naturalised populations in North America of many of the preferred floral hosts in the native range.
Close examination of the habitat proclivities of H. communis in its newly established Québec populations does, however, indicate that the potential distribution of the bee is not without its constraints. Consistent with observations on European populations, it tends to prefer urban settings. Hylaeus communis also appears to be associated with non-native plants, as are most other exotic bees (MacIvor et al. Reference MacIvor, Ruttan and Salehi2015). Whether it may be excluded from residential parks and semi-natural areas due to niche segregation from native species such as H. affinis is a subject warranting further investigation.
Future biomonitoring of the spread of H. communis across the continent will be essential to understanding its effect on native fauna and flora. In planning these efforts, our results show that the dispersal pattern of H. communis may emanate out radially from sites of initial detection in cities of southern Québec through urban and suburban areas. The bee may, alternatively, follow the more discontinuous North American distributions of H. hyalinatus and H. punctatus as per detection at disconnected sites that may represent multiple human-mediated dispersal events (Ascher et al. Reference Ascher, Gambino and Droege2006). Molecular diagnostic studies are needed to determine whether the occurrence of these Hylaeus represents independent colonisation from Europe or spread within North American following a single introduction.
As honey bees (Apis mellifera Linnaeus; Hymenoptera: Apidae) decline in North American, the arrival of a widely adaptable bee species can be considered from some perspectives (see Batra Reference Batra2003) a positive, albeit accidental, intervention. Hylaeus may not be the most archetypal crop pollinator, lacking the pollen-carrying scopal setae of honey bees and bumble bees, and Hylaeus are not managed as pollinators, but they do, nonetheless, provide pollination services. Hylaeus communis is an important carrot pollinator in its native range and can increase seed yield by at least 50% (Ruszkowski and Gosek Reference Ruszkowski and Gosek1999). Its potential use in agricultural management is limited as it does not domesticate easily (Gosek et al. Reference Gosek, Ruszkowski and Kaczmarska1996), but given its broad nectar and pollen palette the species may react positively and be attracted to complementary floral strips used to boost wild bee pollinators of commercial crops (Scheper et al. Reference Scheper, Bommarco, Holzschuh, Potts, Riedinger and Roberts2015). It then becomes a matter of comparing the advantage gained in having one additional unmanaged crop pollinator with the possible indirect and negative effects H. communis may have on pollination services by undermining native Hylaeus or indigenous bee fauna as a whole. In doing so, investigations should be directed at its competition with native species for floral and nesting resources, risk of pathogen transmission, reproductive disruption via interspecific mating with congeners, and finally, its effects on plant-pollinator networks (Stout and Morales Reference Stout and Morales2009).
Acknowledgements
The authors would like to thank Julien Massé-Jodoin and Sarah Saldanha, who assisted with sampling Hylaeus species around Mont-Saint-Hilaire. Anna Luz was kind to help make the map of the study region. K.T.M. was funded by the Natural Science and Engineering Research Council of Canada and the Fonds de Recherche, Nature et Technologies du Québec, when the bee was first discovered in 2012. Holger Dathe helped to identify voucher specimens referred to in this study and generously shared unpublished data and expertise consulted during this project.
Supplementary material
To view supplementary material for this article, please visit https://doi.org/10.4039/tce.2016.62
