Received: 15 July 2019 | Revised: 8 November 2019 | Accepted: 15 November 2019 DOI: 10.1111/zsc.12402 ORIGINAL ARTICLE Too strict or too loose? Integrative taxonomic assessment of Bombus lapidarius complex (Hymenoptera: Apidae) Thomas Lecocq1,2 | Paolo Biella3 | Baptiste Martinet1 | Pierre Rasmont1 1Laboratory of Zoology, Research Institute of Biosciences, University of Mons, Mons, Abstract Belgium The latest progress of the taxonomy is the use of integrative approach for species 2Inra, URAFPA, Université de Lorraine, delimitation based on a multisource dataset. However, the taxonomic decision that Nancy, France should be made when convergence between the different lines of evidence is not ob- 3Department of Biotechnology and served remains debated. Here, we investigate the consequences of the application of Biosciences, University of Milano-Bicocca, Milano, Italy an ‘integration by cumulation’ approach on the taxonomic statuses within the Bombus lapidarius complex when using an integrative taxonomic framework (i.e. nuclear and Correspondence Thomas Lecocq, Unit Research Animal mitochondrial markers along with reproductive traits) compared with a strict ‘in- and Functionality of Animal Products, tegration by congruence’ method. Our results show similar taxonomic conclusions Université de Lorraine, Boulevard des whatever the decision-making approach used except for one taxon. According to the Aiguillettes, B.P. 70239, 54506 Vandœuvre- lès-Nancy, France. differentiation observed in our integrative taxonomic framework, recent divergence Email: [email protected] time and other field observations for this taxon, we assume that a too strict decision- making method could fail to detect recently diverged species. This is exemplified by Funding information F.R.S.-FNRS the new species Bombus bisiculus sp. n. occurring in South Italy and Sicily. KEYWORDS bumblebee, integration by cumulation approach, integrative taxonomy 1 | INTRODUCTION evolving lineages of population or meta-populations but diverge in criteria for delimiting species (De Queiroz, Since the beginning of the Linnaean nomenclature, taxono- 2007; Padial et al., 2010). The concept maintains the nu- mists have described and named an ever-increasing number merous delimiting species criteria as operational criteria of species thanks to the perpetual integration in taxonomy (De Queiroz, 2007). Separation of metapopulation lineages of new analytic tools or methods (e.g. population genetics, could thus be inferred from evidence for reproductive isola- phylogeography, phylogeny or chemical ecology) (Padial & tion, phylogenetic divergence, or ecological differentiation. De La Riva, 2010; Padial, Miralles, Riva, & Vences, 2010). Integrative taxonomy considers, these operational criteria The latest culmination of the taxonomy progress is the spe- to be separate lines of evidence when assigning species cies delimitation based on a multisource approach to gather status (Schlick-Steiner et al., 2010). This approach aims at different lines of evidence of speciation (Dayrat, 2005; proposing strongly supported taxonomic statuses, but it has Schlick-Steiner et al., 2010). This integrative taxonomy is not led to standardized species delimitation. Indeed, the de- a framework that unifies new conceptual and methodolog- gree of congruence between operational criteria required ical developments in qualitative/quantitative assessment of to regard a taxon as a distinct species remains debated (De species status (Schlick-Steiner et al., 2010). The approach's Queiroz, 2007; Padial et al., 2010), although species diag- paradigm is strongly related to the unified species con- nose is more likely in multiple evidence detection. There cept (USC) (De Queiroz, 2007). The USC states authors are two conflicting decision frameworks: the ‘integration nearly universally acknowledge that species are separately by congruence’ and the ‘integration by cumulation’ (Padial Zoologica Scripta. 2019;00:1–10. wileyonlinelibrary.com/journal/zsc © 2019 Royal Swedish Academy of Sciences | 1 2 | LECOCQ ET AL. et al., 2010). The integration by congruence defines species 1859 and B. lapidarius (L.)); (b) three poorly differenti- through the convergence of divergence in many or all oper- ated subspecies (Bombus lapidarius atlanticus Benoist, ational criteria. This a restrictive approach that recognizes 1928 from Maroccan Atlas, Bombus lapidarius decipiens only the strongly supported species (Padial et al., 2010). Pérez, 1890 from Iberian peninsula and Bombus lapidar- However, this could underestimate the number of species ius lapidarius (L.) from the Mainland of Europe and W. because the speciation process is not always accompanied Turkey); and (c) one highly differentiated taxon currently by change of all diagnostic characters (Adams, Berns, included in B. lapidarius (S. Italian and Sicilian decipi- Kozak, & Wiens, 2009). Moreover, another risk could be ens-like). Bombus lapidarius includes another subspecies that species that diverged in the distant past could have a (Bombus lapidarius eriophorus) in the Caucasus but it has increasing probability of divergence in all diagnostic traits not been revised with a multisource approach. Bombus resulting in a bias towards uncovering older species (Padial caucasicus is a species from the North East Anatolia and et al., 2010). In contrast, the integration by cumulation con- Caucasus while B. lapidarius occurs in most of the West- siders that divergence in any or few operational criteria can Palearctic regions. Molecular clock and phylogeographic suggest speciation (Padial et al., 2010). This allows uncov- analyses showed that the divergence between each B. lap- ering recently diverged and poorly differentiated species idarius subspecies and S. Italian-Sicilian decipiens-like but the uncritical use of a single line of evidence can lead occurred during Quaternary climatic oscillations (Lecocq to overestimation of species number (Padial et al., 2010). et al., 2013). In contrast, the divergence of B. caucasicus Recently, the confusing taxonomy of bumblebees happened likely earlier but cannot be currently related to a (Hymenoptera: Apidae: Bombus) has reaped the benefits of particular past biogeographic event (Lecocq et al., 2013). integration of several independent lines of evidence in the Here, we (a) assess the taxonomic status of taxa included species delimitation framework (Bertsch, 1997; Bertsch, in the B. lapidarius complex using an integration by cu- Schweer, Titze, & Tanaka, 2005; Lecocq, Brasero, et al., mulation decision framework; (b) highlight the taxonomic 2015; Lecocq, Dellicour, et al., 2015). While the efficiency changes compared with an integration by congruence deci- of commonly used taxonomic diagnostic characters as sion framework; and (c) provide the description of Bombus components of integrative taxonomy approach has been bisiculus sp. n. recently assessed (Lecocq, Brasero, et al., 2015; Lecocq, Dellicour, et al., 2015), the suitability of alternative decision frameworks (i.e. integration by congruence vs. integration 2 | MATERIAL AND METHODS by cumulation) has not been investigated to date. Indeed, published studies have used an integration by congruence 2.1 | Sampling approach leading to only strongly supported species status (Bertsch, 1997; Bertsch et al., 2005; Lecocq, Brasero, et Two hundred seventy-seven specimens from the B. lapi- al., 2015; Lecocq, Dellicour, et al., 2015). However, the darius complex were previously sampled for other studies recent bumblebee evolutionary radiation (i.e. from the late (Lecocq, Dellicour, et al., 2015; Lecocq et al., 2013): B. lapi- Miocene or even during the Quaternary for some species darius lapidarius (genetic analyses [GA] = 196, CLGS anal- groups) linked to last climatic oscillations (Dellicour et al., yses [CA] = 174, morphological comparisons [MC] = 196), 2017; Duennes, Lozier, Hines, & Cameron, 2012; Hines, B. lapidarius decipiens (GA = 23, CA = 17, MC = 20), 2008; Lecocq et al., 2013) could theoretically lead to re- B. lapidarius atlanticus (GA = five, CA = zero, MC = 10), cently diverged species where complete gene lineage sort- B. caucasicus (GA = 10, CA = three, MC = 13) and decip- ing in all genomes (i.e. mtDNA and nclDNA), reciprocal iens-like from South Italy (GA = 20, CA = 35, MC = 35). monophyly for many loci or phenotypic differentiation are Specimens were killed by freezing at −20°C. Further 282 not yet observed. Therefore, taxonomic conclusions based specimens of decipiens-like were added for taxon descrip- on integration by cumulation approach could be more suit- tion purpose (see Appendix S1). For these specimens, a map able for bumblebee taxonomy, especially for most recently was prepared with Carto Fauna Flora version 2.0 (Barbier & diverged taxa. Rasmont, 2000). In this study, we investigate the taxonomic status within the Bombus lapidarius complex (Lecocq, Dellicour, et al., 2015; Lecocq et al., 2013). According to the latest taxo- 2.2 | Genetic divergence nomic revision based on a multisource approach along with an integration by congruence decision framework (Lecocq, Two gene fragments were sequenced: mitochondrial cy- Dellicour, et al., 2015), the B. lapidarius complex in- tochrome oxidase 1 (COI) and phosphoenolpyruvate car- cludes (a) two species (Bombus caucasicus Radoszkowski boxykinase (PEPCK). Three closely related species included LECOCQ ET AL. | 3 as outgroup were analysed (Cameron, Hines, & Williams, male cephalic labial gland secretions (CLGS) were analysed. 2007): Bombus