The incidence of species-level paraphyly in animals: A re-assessment

[Display omitted] •Incidence of species-level gene-tree paraphyly confirmed at ∼20% using an independent dataset.•Approximately 10% of cases of paraphyly may be incorrect due to operational reasons.•Strong effect of sampling effort on detection of paraphyly.•Animal species on average diverged 2–3 Ne generations in the past. Species-level paraphyly was found by Funk and Omland (2003) to occur in 23% of animal species on the basis of a meta-analysis of published mitochondrial gene trees. Given the potential for bias in the selection of study organisms and the subsequent publication of their gene trees, I re-estimated the incidence of paraphyly in an independent dataset of publicly accessible COI sequences from the Barcode of Life Data System. Among 7368 animal species represented by two or more sequences, 19% were paraphyletic, slightly less than in the previous study. Rates within major taxonomic groups mirrored, but were slightly lower than, that observed earlier. Tests were made for operational factors that could inflate, and sampling effects that could underestimate, the rate of paraphyly. Overall the previous findings are confirmed. The observed incidence suggests that on average animal species diverged 2–3 Ne generations in the past, far short of the predicted 5 Ne generations required for complete monophyly.