Molecular dating techniques require the use of calibrations, which are usually fossil or geological vicariance-based. Fossil calibrations have been criticised because they result only in minimum age estimates. Based on a historical biogeographic perspective, I suggest that vicariance-based calibrations are more dangerous. Almost all analytical methods in historical biogeography are strongly biased towards inferring vicariance, hence vicariance identified through such methods is unreliable. Other studies, especially of groups found on Gondwanan fragments, have simply assumed vicariance. Although it was previously believed that vicariance was the predominant mode of speciation, mounting evidence now indicates that speciation by dispersal is common, dominating vicariance in several groups. Moreover, the possibility of speciation having occurred before the said geological event cannot be precluded. Thus, geological calibrations can under- or overestimate times, whereas fossil calibrations always result in minimum estimates. Another major drawback of vicariant calibrations is the problem of circular reasoning when the resulting estimates are used to infer ages of biogeographic events. I argue that fossil-based dating is a superior alternative to vicariance, primarily because the strongest assumption in the latter, that speciation was caused by the said geological process, is more often than not the most tenuous. When authors prefer to use a combination of fossil and vicariant calibrations, one suggestion is to report results both with and without inclusion of the geological constraints. Relying solely on vicariant calibrations should be strictly avoided [Current Zoology 57 (1): 116–124, 2011].