RESUMEN
Establishing species boundaries is one of the challenges taxonomists around the world have been tackling for centuries. The relation between intraspecific and interspecific variability is still under discussion and in many taxa it remains understudied. Here the hypothesis of single versus multiple species of the crab spider Synemaglobosum (Fabricius) is tested. The wide distribution range as well as its high morphological variability makes this species an interesting candidate for re-evaluation using an integrative approach. This study combines information from barcoding, phylogenetic reconstruction based on mitochondrial CO1 and ITS2 of more than 60 specimens collected over a wide range of European localities, and morphology. The findings show deep clades with up to 6% mean pairwise distance in the CO1 barcode without any biogeographical pattern. The nuclear ITS2 gene did not support the CO1 clades. Morphological assessment of somatic and genital characters in males and females and a morphometric analysis of the male palp uncovered high intraspecific variation that does not match the CO1 or ITS2 phylogenies or biogeography either. Screening for endosymbiotic Wolbachia bacteria was conducted and only a single infected specimen was found. Several scenarios might explain these inconsistent patterns. While the deep divergences in the barcoding marker might suggest cryptic or ongoing speciation or geographical isolation in the past, the lack of congruent variation in the nuclear ITS2 gene or the studied morphological character systems, especially the male palp, indicates that S.globosum might simply be highly polymorphic both in terms of its mtDNA and morphology. Therefore, more data on ecology and behaviour and full genome sequences are necessary to ultimately resolve this taxonomically intriguing case.
RESUMEN
The integration of independent data sets could solve problems in both traditional and DNA-based taxonomy. The aim of this study is to investigate the power of CO1 sequences and of morphometrics to distinguish closely related species in the spider genus Araniella. We put special emphasis on the species pair A. cucurbitina (Clerck, 1757) and A. opisthographa (Kulczynski, 1905) since the females are morphologically difficult to distinguish and often misidentified. A total of 216 sequences of eight Araniella species from seven European countries, North America and Asia were included in the molecular analysis. The results from both maximum likelihood and Bayesian phylogenetic inference indicate successful separation of six out of eight Araniella species, including A. cucurbitina and A. opisthographa. For the same six species, we detect no overlap of intra- and interspecific genetic divergence, leading to successful species identification with a threshold approach. In addition, morphometric analysis of the epigyna of A. cucurbitina and A. opisthographa supports species separation by two best explanatory ratios: receptaculum length and distance between receptaculum and copulatory duct. Although a small overlap in the ratios exists, the species identification rate increases when combining morphometric and molecular data, which demonstrates the efficiency of integrative approaches for distinguishing closely related species. However, none of the molecular approaches was able to separate closely related A. alpica (L. Koch, 1869) and A. inconspicua (Simon, 1874) due to shared CO1 haplotypes. Considering the clear morphological separation of the males and different habitat preferences, incomplete lineage sorting or introgressive hybridization could have led to identical CO1 sequences. Therefore, DNA-barcoding must be thoroughly tested even within small homogenous genera of spiders.