Background and aims – Keraunea is a genus recently described in Convolvulaceae, though it has sat uncomfortably in this family. A recent molecular phylogenetic study suggests that its two morphologically almost identical species actually belong to different families, Malpighiaceae (Superrosids) and Ehretiaceae (Superasterids), although with little-to-no morphological evidence to support it.

Material and methods – Sequences of matK, rbcL, and ITS for all the 77 currently accepted genera of Malpighiaceae, K. brasiliensis and Elatinaceae (outgroup) were compiled from Genbank and analysed with Maximum Likelihood and Bayesian Inference criteria for nuclear, plastid and combined datasets. Additional database and herbarium studies were performed to locate and analyse all duplicates of the holotype of K. brasiliensis to check for misidentified or contaminated materials.

Key results – Our examination of expanded DNA datasets and herbarium sheets of all K. brasiliensis isotypes revealed that an error in tissue sampling was, in fact, what led to this species being placed in Malpighiaceae. Kew’s isotype had a leaf of Malpighiaceae (likely from Mascagnia cordifolia) stored in the fragment capsule, which was unfortunately sampled and sequenced instead of the actual leaves of K. brasiliensis.

Conclusions – DNA sequences can be helpful in classifying taxa when morphology is conflicting or of a doubtful interpretation, with molecular phylogenetic placement becoming a popular tool that potentially accelerates the discovery of systematic relationships. However, good knowledge of plant morphology is essential for formulating the phylogenetic hypotheses to be tested and for a critical re-interpretation of the results in the context of biological information of the species or families. Thus, these techniques are, much like any others, prone to methodological errors. We highlight the crucial need to observe plant morphology alongside molecular phylogenetic results, particularly when the new hypotheses are in disagreement with the existing classification and at risk of incurring gross taxonomic mistakes.