Occurrence and ecological data on an exotic solitary bee accidentally introduced in Brazil

Currently, there is a global concern regarding exotic species due to, among other factors, their great ability to reproduce and spread rapidly through the novel environment. As such, these species often compete for nesting places and food resources or convey pathogens. Anthidium manicatum (Linnaeus) (Hymenoptera: Megachilidae) is a non-native solitary bee occurring in Brazil. This study aimed to collect data about the occurrence sites of this species to investigate the historical sequence of its spread throughout the country. Based on this, we estimated population data such as the number of males and females, phenology and bioclimatic niche overlap with native species. The occurrence records were retrieved from speciesLink and Global Biodiversity Information Facility. All analyses were performed in R. The collected data demonstrate that, except for the 1960s, the records of the occurrence of A. manicatum in Brazil are few and constant, being notified since the mid-1930s in at least nine Brazilian states. In total, 778 individuals were sampled, with males being recorded about 1.7 times more than females. This species seems to be bivoltine, with generations in May and November. Anthidium manicatum showed a low and moderate bioclimatic niche overlap with two native species, Anthidium sertanicola Moure & Urban and Anthidium latum Schrottky, respectively. These data provide relevant information on the biology and status of A. manicatum in Brazil. However, since most Brazilian scientific collections have not digitalized their data in the platforms consulted here, some ecological features described here may be underestimated.

The introduction (intended action) or invasion (non-intended action) of bee species to places where they historically do not occur may bring serious ecological effects to the local fauna (russO 2016;VOllEt-nEtO et al. 2018). Overall, such organisms are r-strategists, i.e., they show a high fecundity rate and an elevated dispersal capacity (saKai et al. 2001). As a result, their effects on the environment may be disastrous since they (i) compete for nest substrates, (ii) compete for food resources, (iii) transmit pathogens, (iv) alter plantinsect interaction, among others (russO 2016;VOllEt-nEtO et al. 2018). In an attempt to avoid such issues, most countries are regulating the management and transport of bees within their territories to diminish the chances of introductions or accidental invasions of novel (exotic) species (IPBES 2016).
Brazil is a large nation with an area of 8,516,000 km 2 . Thus, monitoring exotic bee species becomes an arduous task. Almost 65 years ago (1956), the state of São Paulo (municipality of Rio Claro) was an epicenter of the most famous bee introduction (Apidae: Apini: Apis mellifera scutellata Lepeletier), the African honey bee, in the world (michEnEr 1973). In a few decades, this exotic social bee was well-established, spread and hybridized (Africanized honey bees) in Brazil and other Latin American countries (michEnEr 1973). Due to this well-known case, there is currently a new concern with the commercial importation of bumblebees (Apidae: Bombini: Bombus terrestris Linnaeus) from Europe to South America. This exotic bumblebee is negatively affecting native species of the same genus in Chile and Argentina (mOralEs et al. 2013;aizEn et al. 2019). A recent study has suggested that B. terrestris has a high potential to invade Brazil in the oncoming years (acOsta et al. 2016 Males and females of A. manicatum are conspicuous due to both their large size (♂: 12.3-17.7 mm; ♀: 9.2-12.2 mm) and their yellow-banded black abdomens (stranGE et al. 2011). Anthidium manicatum males are as active as females (wirtz et al. 1992). Yet, A. manicatum females have a prolonged sexual receptivity and nest on pre-existing cavities of the aforementioned substrates, in which they store food (pollen and nectar) and, immediately after, lay their eggs (wirtz et al. 1992;stranGE et al. 2011).
Anthidium manicatum males defend plants that are attractive to females against conspecific males and mate with females as soon as they enter their territories (wirtz et al. 1992;muEllEr & wOlf-muEllEr 1993). Males defend their territories using thorny projections of the distal region of their abdomens against the wings of competitors (stranGE et al. 2011;GOnzalEz & GriswOld 2013). This behavior facilitates the detection of males in environments they inhabit (stranGE et al. 2011;GOnzalEz & GriswOld 2013).
Although interesting, A. manicatum is an exotic species in Brazil. Thus, by considering the ecological threats that a nonnative bee species may represent to local populations, the main goal of this study was to conduct a chronological survey of the occurrence of A. manicatum to investigate its current status in Brazil. Therefore, we evaluated (i) the total number of occurrences in data available in digital platforms; (ii) the male-female ratio; (iii) the Brazilian states where A. manicatum has been recorded according to the consulted database; (iv) the number of sampled individuals over the decades; (v) the species phenology (i.e., annual activity pattern) using a circular analysis; and (vi) potential bioclimatic niche overlap between A. manicatum and two native species of Anthidium with enough data to carry out such an analysis. We believe that the results herein provided will help future research and drive field works particularly devoted to monitoring this exotic solitary bee species in Brazil.

MATERIAL AND METHODS
Occurrence data. We surveyed the occurrence of A. manicatum in Brazil in two digital databases: speciesLink (http://splink.cria.org.br/) and Global Biodiversity Information Facility (GBIF -https://www.gbif.org/). Data of the latter were retrieved from GIBF using the function 'occ_search' (country='BR') of the package rgibf (chambErlain et al. 2020) for R (ihaKa & GEntlEman 1996; R cOrE tEam 2018). In both databases, we searched for the following data: geographic coordinates, names of collections and museums where specimens were deposited, recorded date, number of sampled individuals and whether they were male or female.

Data analysis. Circular analysis:
The activity period of A. manicatum in Brazil was evaluated with a circular analysis using the function 'circular' from the package circular (aGOstinElli & lund 2017). Our temporal data were the sampling months, whereas the angles were our reference measure and the hours were our measure units. The directionality was tested with Rao's Spacing using the function 'rao.spacing.test' of circular.
Bioclimatic niche overlap: Since A. manicatum is an exotic bee, it most likely inhabits a fraction of or the entire region already inhabited by native Anthidium bees. Thus, we evaluated a potential bioclimatic niche overlap (see below) between exotic (A. manicatum) and native (A. isabelae, A. larocai, A. latum, A. sanguinicaudum, and A. sertanicola) Anthidium bees. The occurrence data of native species were obtained using the functions 'name_backbone' and 'occ_ search' [country="BR"] of the package rgibf (chambErlain et al. 2020). Since georeferenced data were not available for A. isabelae and A. larocai, and since only two occurrences were retrieved for A. sanguinicaudum, niche overlap was assessed between A. manicatum (n = 57) and both A. latum (n = 63) and A. sertanicola (n = 22). Numbers inside brackets indicate unique (non-repeated) localities, totaling 142 geographic coordinates.
Bioclimatic variables (Bio1-Bio19; https://worldclim.org/ data/bioclim.html) of each occurrence point were obtained using the function 'getData' of the package raster (hijmans 2017). These variables were then incorporated into the data frame of all three Anthidium species using the functions 'SpatialPoints' of the package sp (biVand et al. 2013), and 'extract' of the package raster. However, before proceeding with the analysis, we evaluated the collinearity between all bioclimatic variables using the function 'vifstep' of the package usdm (naimi et al. 2014). Only non-collinear variables were used to analyze niche overlap (see results).
Finally, the bioclimatic niche overlap between the three Anthidium species was performed using the functions 'overlap' and 'overlap.plot' of the package nicheROVER (lysy et al. 2014) after 2,000 iterations. A high overlap suggests a similar use of their bioclimatic region. This metric is referred to as Niche region (N R ), which is defined as a probability of 95% of an individual of group A to be found within the N R of group B (swansOn et al. 2015). Since this measure is asymmetric, the probability of A. manicatum individuals to be A B

e-ISSN 1983-0572
Volume 13, 2020 -www.entomobrasilis.org EntomoBrasilis 13: e891 found in the bioclimatic N R of A. latum and A. sertanicola may be different from the probability of the latter to be found in the bioclimatic N R of A. manicatum. It will depend on how these bee species use their niche areas (swansOn et al. 2015).

RESULTS
We obtained 1,235 occurrence points of A. manicatum in Brazil. However, after we removed incongruent and overlapped data, the final number was 778, of which 489 were of males and 289 of females. This proportion indicates that, in digital platforms (speciesLink, GBIF), there are 1.7 more males than females. These specimens are deposited at the following institutions: (a) "Coleção Entomológica Paulo Nogueira-Neto", Instituto According to the available data, A. manicatum in Brazil has been observed in, at least, nine states: Minas Gerais (59.7%), São Paulo (15.6%), Santa Catarina (12%), Paraná (6.8%), Rio Grande do Sul (2.3%), Bahia (1.6%), Rio de Janeiro (1.3%), Tocantins (0.5%) and Ceará (0.2%). Most available records seem to be distributed along the Brazilian coast ( Figure 2). The first recorded specimen of A. manicatum is from 1934 with a single individual in Rio de Janeiro, and the last one is from 2020, in Paraná, again with a single individual ( Figure 3); remembering that such records are based on data available on digital platforms. The year with the largest number of sampled individuals is 1963, thirty years after the first observation, with 418, most from Minas Gerais (Figure 3).  The annual activity pattern of A. manicatum may take place throughout the year with a clear directional tendency to be bivoltine (two generations per year) in their phenology (Rao´s Spacing = 351.74, p < 0.001) since the peaks of activity were in May (late autumn) and November (late spring) (Figure 4).
Of the 19 bioclimatic variables, nine were non-collinear (Table 1) and adequate for the next analysis. Our data show that A. manicatum may overlap, on average, 4.7% (2 -8: 95% confidence interval) of the bioclimatic N R of A. sertanicola. Thus, there is a low probability of a randomly sampled exotic individual to be found in the bioclimatic N R of this native Anthidium species. On the other hand, the probability of A. manicatum to be found in the bioclimatic N R of A. latum is on average almost three times higher (15.2%; 10 -21: 95% confidence interval), Figure 5.

DISCUSSION
Our data show that A. manicatum remains a little sampled species in Brazil since its first observation almost nine decades ago , except for 1963 when a peak of individuals was recorded. The few samplings of this species may be due  BIO2 -Mean diurnal range; BIO3 -Isothermality (oscillation of day-to-night temperatures relative to the summer-to-winter (annual) oscillations); BIO8 -Mean temperature in the wettest trimester; BIO9 -Mean temperature in the driest trimester; BIO13 -Precipitation in the wettest month; BIO14 -Precipitation in the driest month; BIO15 -Precipitation seasonality (coefficient of variation); BIO18 -Precipitation in the warmest trimester; BIO19 -Precipitation in the coldest trimester. Table 1. Nine non-collinear bioclimatic variables and their means across the range of three Anthidium species (Hymenoptera: Megachilidae), two native (*) and one exotic (**). Note: Mean ± standard deviation; color indicates lower (yellow), medium (light green) and higher (dark green) values of bioclimatic variables. to both their solitary habits and because field collections were not particularly focused on this species. Yet, the peak of records in the 1960s may be related to an increased collection effort on that period, most likely due to funding of research for this purpose, which was not repeated in the following decades.
The amount of occurrence data of A. manicatum individuals (n = 778) represents ca. 9.8% of global data, which reaches 7,925 occurrence points according to data extracted from GBIF in our analysis. However, these data may be underestimated for Brazil because this bee species is solitary and its long-term continuous sampling is impaired. Furthermore, Brazilian scientific collections and museums may not have yet made all occurrence data of A. manicatum available in the digital platforms consulted by us.
We also cannot neglect the possibility that researchers are Therefore, to better understand the current occurrence and distribution range of A. manicatum in Brazil, new field works are welcome if focused on sampling this exotic bee species. Even those localities where only one individual was collected may represent a research bias. We also suggest that Brazilian scientific institutions should receive stimulus and more financial support to make the data inside their collections digitally available to a broad audience. This may promote large-scale monitoring of both native and exotic Anthidium bees.
According to our phenology analysis, A. manicatum seems to be active throughout the year, although it shows a clear tendency of presenting two generations (bivoltine), one in May and one in November. It is known that A. manicatum may modify its voltinism depending on the region its population inhabits. For example, it may be univoltine in a non-native place such as Utah, USA, in which generations are active from July to October (stranGE et al. 2011), as well as in native regions, such as Germany, where it is active in August (wirtz et al. 1992). On the other hand, it may be bivoltine in Italy with one generation in May, as in Brazil, and one in October (Mueller & Wolf-Mueller 1993). Knowing that A. manicatum most likely has two clear generations in Brazil (May, November) may help researchers increase their chances of observation and collection in future field works.
The wide occurrence of A. manicatum in several places of the world has been attributed to its nesting behavior and its large bioclimatic amplitude (stranGE et al. 2011). It appears to negatively affect the acquisition of floral resources and may cause the transmission of parasites and pathogens to other bee species (russO 2016). With occurrence data and their corresponding bioclimatic values, it was possible to infer how and to which extent A. manicatum occupies the bioclimatic niche region of native Anthidium bees such as A. sertanicola and, more widely, A. latum. This indicates that A. latum most likely suffers a stronger ecological pressure than A. sertanicola by A. manicatum. Therefore, we suggest that field works could corroborate this finding and its implications to Anthidium bees.
The successful settlement and subsequent dispersal of exotic species may be attributed to features such as: (1) extended longevity of adults; (2) broad diet, i.e., few restrictions to specific host plants; (3) eusociality, which does not involve the studied bee species; (4) passive dispersal by human activities that accelerate the displacement; (5)  Our study demonstrates the importance of scientific collections since their digitalized data allowed us to historically track the occurrence of an exotic solitary bee species in Brazil. Despite the small number of sampled individuals over the decades, A. manicatum has been recorded from their first notification almost 90 years ago to the present. As we observed, most specimens were males (1.7x). Nevertheless, this may be an underreporting because most removed data were of specimens without information on sex. Similarly, the larger sampling of A. manicatum in Brazil's south and southeast regions seems to indicate that these places harbor a larger number and/or older scientific collections, causing a bias in our data.
In summary, our data suggest that the best periods to sample A. manicatum in Brazil are May (late autumn) and November (late spring) when these generations seem to be more active.
Our study also supports that A. manicatum may occupy a fraction of bioclimatic niche region of, at least, two of five native Anthidium bees. Thus, while with A. sertanicola, whose distribution range appears to be further south in Brazil, the overlap is low (< 5%), with A. latum, whose distribution range is larger, the probability of overlap reached about 15%. Therefore, we strongly suggest that a field monitoring of A. manicatum in Brazil should occur together with the survey of native Anthidium species, especially where co-occurrence is common. Finally, we recommend the maintenance of both the training of young people and the financial support to scientific institutions since the efficiency and quality of digitalized data allow a remote assessment of biodiversity that needs to be continuously monitored.

ACKNOWLEDGMENTS
The authors are grateful to Favízia Freitas de Oliveira for consulting databases of possible Anthidium manicatum specimens not yet inserted in platforms such as speciesLink e Global Biodiversity Information Facility (GBIF). CFS would like to thank the Coordination for the Improvement of Higher Education Personnel (CAPES) of the Ministry of Education (MEC) for funding the postdoctoral fellowship of the National Postdoctoral Program (PNPD, Finance Code 88882.314829/2019-01). CAB is grateful to the Programa