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 (Kulczyński, 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.