Cichlid fishes are one of the most important model systems for evolutionary biology. Unfortunately, however, the timeline of cichlid diversification is still insufficiently known and limits our understanding of the mechanisms that generated their spectacular diversity. The uncertainty regarding this timeline stems from a decades-old controversy surrounding the phylogeographic history of cichlid fishes. Did cichlid subfamilies diversify as the result of Gondwanan vicariance, as supported by their distribution on former Gondwanan landmasses? Or did they diverge much more recently through oceanic dispersal, as suggested by the fossil record? While a large number of studies have already addressed this question with molecular-clock analyses, no single conclusion has emerged from these investigations. Here, I review the molecular evidence for Gondwanan vicariance or trans-Atlantic dispersal resulting from these studies. I discuss the weaknesses and strengths of each study, aiming to promote the formation of consensus on the matter and to prevent the repetition of previously made mistakes. I find that after accounting for inappropriate calibration strategies and saturation in mitochondrial datasets, the molecular evidence points to trans-Atlantic dispersal long after continental separation, probably around 75–60 Ma.