Scrapter is a genus of colletid bees with a primary distribution centered in Southern Africa. The genus currently comprises 68 recognized species, which are divided into nine species groups, ranging from one to 29 included species. The Scrapter heterodoxus group is presently considered to be the only monotypic group, because of synonymization of Scrapter heterodoxus with Scrapter peringueyi in a previous revision of the genus. A comparative examination of these two species using both morphological assessment and molecular sequence data from the COI barcode region supported the recognition of S. peringueyi as a valid species, which we accordingly resurrect as the second species of the Scrapter heterodoxus species group. We provide high resolution images of the type specimens for both species and updated diagnoses to enable their separation from all other species of Scrapter.

Bees, COI, DNA barcoding, Afrotropical, morphology

Scrapter Lepeletier & Serville, 1828 is a genus of colletid bees endemic to Africa (Eardley 1996; Davies et al. 2005; Michener 2007; Eardley et al. 2010). A species-rich genus, Scrapter currently comprises 68 recognized species (Kuhlmann and Friehs 2020), and new species are being discovered at a remarkable pace. In the past 25 years since Eardley’s (1996) revision of the genus, in which he recognized 31 valid species, more than 30 new species have been described (Davies et al. 2005; Davies and Brothers 2006; Kuhlmann 2014; Kuhlmann and Friehs 2020). As many species of Scrapter seem to be ephemeral in their biology, including tight hostplant associations and narrow flight periods in highly seasonal environments (Kuhlmann 2009; Kuhlmann and Eardley 2012; Kuhlmann et al. 2012; Kuhlmann and Friehs 2020), this increase of newly discovered species can be expected to continue into the upcoming decades.

In the past, the phylogenetic relationships of Scrapter to other colletid lineages proved difficult to establish based on morphology alone (McGinley 1981; Alexander and Michener 1995; Plant and Paulus 2016). Previously considered to be part of Paracolletini (e.g., Michener 1944), albeit without strong morphological evidence (McGinley 1981), analyses of nucleotide sequence data strongly indicated a sister group relationship of Scraptrinae and Euryglossinae, an Australian-endemic lineage of Colletidae (Almeida and Danforth 2009; Almeida et al. 2012; Kayaalp et al. 2017; Cardinal et al. 2018). Interestingly, a close relationship to Euryglossinae has been discussed as early as 1933 (Cockerell and Ireland 1933; specifically discussing Euryglossidia Cockerell) and a sister-group relationship is supported by certain morphological characters of the mature larvae (McGinley 1981). Fossil-based divergence-time estimates agree that the lineages forming the present-day Scraptrinae and Euryglossinae split in the early Eocene, around ~55 million years ago (Almeida et al. 2012; Kayaalp et al. 2017). Scrapter is the only colletid lineage endemic to Africa, and with the exception of one species reported from Kenya (Davies et al. 2005), it is geographically restricted to southern Africa. With Euryglossinae being endemic to Australia, the biogeographical puzzle leading to this exceptional distribution has been difficult to explain (Almeida et al. 2012; Kayaalp et al. 2017).

Scrapter is a morphologically heterogeneous genus (Davies and Brothers 2006) with great differences in body size (e.g., 3.5 mm, Scrapter minutissimus Kuhlmann, 2014 vs. 14 mm, Scrapter heterodoxus (Cockerell, 1921)). Eardley (1996) presented the first attempt to systematically revise the genus, including redescriptions, synonymizations, and type designations. He established eight species groups based on similar characteristics, and Kuhlmann (2014) added the ‘euryglossiform’ species as a ninth group. One group, the Scrapter heterodoxus group is currently understood as being monotypic, because Eardley (1996) synonymized Scrapter peringueyi (Cockerell, 1921) with Scrapter heterodoxus. Both these species were described by Cockerell (1921) in the same article, but based on different sexes. In the present article, we re-examine the type specimens of the two species and additional material of the Scrapter heterodoxus group. Studying both sexes of both species, we found strong morphological evidence that S. peringueyi is not a synonym of S. heterodoxus, but a valid species. Nucleotide sequence data from COI barcodes of both morphotypes show a divergence of > 6%, underlining the significant discrepancy between the two lineages. Based on this combined evidence, we resurrect S. peringueyi as a valid species.

We located and examined the type specimens of S. heterodoxus and S. peringueyi in the collections of the Iziko South African Museum, Cape Town (SAMC), and in the South African National Collection of Insects, Pretoria (SANC). Additional specimens of both species were collected during field work by Bryan Danforth in South Africa in September 2001 and were deposited in the Cornell University Insect Collection (CUIC). In total, we examined 133 specimens, which includes all specimens deposited in the collections of the SAMC and the CUIC. We mapped the distributions of the two recognized species using SimpleMappr (Shorthouse 2010). The terminology of surface sculpturing follows Harris (1979).

To compare the DNA sequences of Scrapter heterodoxus and S. peringueyi, we obtained sequence data for both species. One COI sequence for S. heterodoxus was retrieved from GenBank (identifier MH578427) and we extracted another COI barcode from a UCE assembly of the same species (from Branstetter et al. 2017). This DNA sequence data is associated with a male voucher specimen which is deposited in the CUIC. This voucher specimen was examined in the present study and is part of a series of specimens consisting of male and female S. heterodoxus: males and females were collected jointly at the same time and place. Since both sexes share distinct morphological features that distinguish them from what was described as S. peringueyi, we deemed them conspecific. For S. peringueyi, we generated two new barcode sequences. Two male specimens collected on 28 September 2001 and listed below under additional material were used for DNA extraction. These specimens are vouchered in the CUIC as well and are labelled with a green colored extraction code label. DNA was extracted from ground-up thorax tissue using a CTAB phenol-chloroform protocol. We used the DNA extractions to sequence ultraconserved elements (UCEs) as detailed in Bossert et al. (2021), as the S. peringueyi samples were processed jointly with the samples of that study. After assembling the raw read data with SPAdes (ver. 3.13.1; Bankevich et al. 2012), we extracted the COI barcode region using the script phyluce_assembly_match_contigs_to_barcodes of the Phyluce package (ver. 1.7.1; Faircloth 2016) and a reference COI barcode of Apis mellifera L., 1758. The four sequences consisting of two representative barcodes for each of S. heterodoxus and S. peringueyi were aligned with MUSCLE (ver. 3.8.425; Edgar 2004).

Examination of the sequence alignment by eye revealed that for each species the two barcode sequences differed in length but were otherwise identical across the shared positions. Since we were interested in examining the interspecific distance between S. heterodoxus and S. peringueyi, we only retained the longest DNA sequence for each species. This led to a sequence alignment of 658 nucleotides (658 present positions for S. heterodoxus and 657 for S. peringueyi). We estimated the evolutionary distance between these two sequences by quantifying the proportion of sites at which nucleotides differed (p-distance). We uploaded the reference COI sequence of S. peringueyi to NCBI GenBank where it can be retrieved under identifier MZ682106. The sequence alignment of all four sequences can be found as Suppl. material 1.

Images were acquired at SAMC with a Leica LAS 4.9 imaging system, comprising a Leica Z16 microscope (using either a 2× or 5× objective) with a Leica DFC450 Camera and 0.63× video objective attached. The imaging process, using an automated Z-stepper, was managed using the Leica Application Suite V 4.9 software installed on a desktop computer. Diffused lighting was achieved using a Leica LED5000 HDI dome. All images presented in this paper, as well as supplementary images, are available on WaspWeb at www.waspweb.org.

Scrapter heterodoxus and S. peringueyi were described as species by Cockerell in the same publication (Cockerell 1921). Subsequently, in his major revisionary work on Scrapter, Eardley (1996) synonymized the two species and regarded only Scrapter heterodoxus as valid. Thus, the Scrapter heterodoxus species ‘group’ was regarded as monotypic over the past decades. In the present study, we reassess the status of both species using morphological and molecular methods, and find strong support for the re-recognition of S. peringueyi as a valid second species in the Scrapter heterodoxus group. While both species are morphologically very similar, clearly closely related, and not particularly like any other species of Scrapter, they can be readily separated using morphological characters and molecular data. Differentiation of the species is particularly clear in the male sex based on the species-specific shape of the hind tibia (Fig. 5), which shows no intraspecific variation among the examined specimens or in illustrations in the literature (Brauns 1929; Eardley 1996). Other structures that are often diagnostic for species-recognition of Scrapter, such as the genital capsule and terminal sterna, seem identical between the two species and cannot be used to separate S. heterodoxus and S. peringueyi. However, both sexes can also be separated by the different surface sculpturing of the basal zone of the propodeum and the polished or shagreened interspaces on the mesoscutum. These morphological differences that distinguish both males and females of S. heterodoxus from S. peringueyi allowed us to associate the female sex for both species, since we only generated COI sequence data from male individuals. In line with these morphological differences is the significant genetic distance between the examined specimens, which is 6.1% for the 657 base-pair long COI barcode region. Species delimitation based on pairwise genetic distances of this partial gene region is common practice in modern insect systematics and has been routinely applied for many insect groups such as Lepidoptera (e.g., Hausmann et al. 2011; Nneji et al. 2020), Coleoptera (e.g., Oba et al. 2015; Huang et al. 2020), Hymenoptera (e.g., Sheffield et al. 2009; Stahlhut et al. 2013), and specifically for certain African bees (Bossert et al. 2020). While the threshold for delimiting species boundaries is not universal, varies among studies, and is not ultimate proof, a distance of 2–3% is common practice to recognize a barcoding gap (Ratnasingham and Hebert 2013; Hebert et al. 2003; and references above). The calculated distance between S. heterodoxus and S. peringueyi exceeds such thresholds considerably, underlining the need to recognize them as distinct species.

Prior and subsequent to Eardley’s (1996) revision of Scrapter and the synonymization of S. peringueyi with S. heterodoxus, samples of ‘Scrapter heterodoxus’ have been included in a number of morphological-phylogenetic studies (Alexander and Michener 1995; Davies and Brothers 2006; Packer 2008; Mthethwa 2016; Plant and Paulus 2016; Porto and Almeida 2019), or assessments of pollinator communities (Tribe 2007; Goldblatt et al. 2009). As the distinguishing characters between S. heterodoxus and S. peringueyi are not specifically mentioned in these works, it is not immediately clear which of the two species were actually included in the respective studies. It is therefore possible that some of the examined specimens may in fact have corresponded to S. peringueyi. For example, the material examined in Mthethwa (2016) almost certainly consists of a mixed sample of both S. heterodoxus and S. peringueyi, since the specimens for morphological study were collected in Citrusdal and Hermanus. According to the distributional patterns discussed below, these collection localities make it very likely that both species were included. However, given the overall very similar, or seemingly identical shape of most examined morphological structures and the close evolutionary relationship of the two species, we do not expect that this combined interpretation could significantly impact results and conclusions of any of these phylogenetic studies. More care would need to be taken in assessing pollination networks, given that there may be disparity in host plant fidelity between the two species. Interestingly, the two species were confused early on: in one of the very first treatments of S. heterodoxus after Cockerell’s description (Cockerell 1921), Brauns (1929) redescribed the species and illustrated the hindleg tibia based on a male specimen. The shape of the tibia, however, clearly corresponds to that of S. peringueyi (cf., Fig. 5A) and not of S. heterodoxus.

Mapping the distributions of S. peringueyi and S. heterodoxus based on the 133 examined specimens reveals slightly different distribution patterns for the two species (Fig. 6). Scrapter peringueyi is a southern Cape coastal species, without any records north of the Cape Town area. All localities are in close proximity to the shoreline, without any records from inland regions. Scrapter heterodoxus in turn extends from Cape Town northwards up the south-western coast of South Africa, with most occurrences recorded from inland of the western coastline. The two species are sympatric in the Cape Town vicinity. Additionally, we recovered two isolated records for S. heterodoxus, one from the interior of the Eastern Cape (Katberg), and another one from Kakamas in the interior Northern Cape region. These records are particularly interesting as they significantly expand the distributional range of S. heterodoxus, but they also warrant further study: the surface sculpturing of these two specimens is slightly less rugulose than in the females from the Cape Town region, which is where the type locality is located. Additional study of specimens from the interior Northern and Eastern Cape regions is required to determine the degree of variation of this propodeal character and could possibly reveal additional, yet to be described species of the S. heterodoxus species group.

Figure 6. 

Distribution map of Scrapter peringueyi (Cockerell, 1921) and Scrapter heterodoxus (Cockerell, 1921) based on 133 examined specimens. If several specimens were collected at the same site, they are shown as a single occurrence.

We thank Bryan N. Danforth (Cornell University) for providing us access to the specimens deposited in the Cornell University Insect Collection and Werner Strumpher (South African National Collection of Insects) for help in locating the type specimens. We thank the reviewers and the editorial team of African Invertebrates for their constructive feedback on the manuscript. This work was supported by U.S. National Science Foundation grants DEB-2127744 and DEB-0814544.