Systematic position of the Clicking Frog (Kassinula Laurent, 1940), the problem of chimeric sequences and the revised classification of the family Hyperoliidae

The systematics of the African frog family Hyperoliidae has undergone turbulent changes in last decades. Representatives of several genera have not been genetically investigated or with only limited data, and their phylogenetic positions are thus still not reliably known. This is the case of the De Witte's Clicking Frog (Kassinula wittei) which belongs to a monotypic genus. This miniature frog occurs in a poorly studied region, southeastern Democratic Republic of the Congo, northern Zambia, Angola. So far it is not settled whether this genus belongs to the subfamily Kassininae as a relative of the genus Kassina, or to the subfamily Hyperoliinae as a relative of the genus Afrixalus. Here we present for the first time a multilocus phylogenetic reconstruction (using five nuclear and one mitochondrial marker) of the family Hyperoliidae, including Kassinula. We demonstrate with high confidence that Kassinula is a member of Hyperoliinae belonging to a clade also containing Afrixalus (sub-Saharan Africa), Heterixalus (Madagascar) and Tachycnemis (Seychelles). We find that Kassinula represents a divergent lineage (17–25 Mya), which supports its separate genus-level status, but its exact systematic position remains uncertain. We propose to name the clade to which the above four genera belong as the tribe Tachycnemini Channing, 1989. A new taxonomy of the family Hyperoliidae was recently proposed by Dubois et al. (2021: Megataxa 5, 1–738). We demonstrate here that the new taxonomy was based on a partially erroneous phylogenetic reconstruction resulting from a supermatrix analysis of chimeric DNA sequences combining data from two families, Hyperoliidae and Arthroleptidae (the case of Cryptothylax). We therefore correct the erroneous part and propose a new, revised suprageneric taxonomy of the family Hyperoliidae. We also emphasize the importance of inspecting individual genetic markers before their concatenation or coalescent-based tree reconstructions to avoid analyses of chimeric DNA sequences producing incorrect phylogenetic reconstructions. Especially when phylogenetic reconstructions are used to propose taxonomies and systematic classifications.