Back to the past: a new take on the timing of flowering plant diversification.

‘In the case of plants, an adequate fossil record does not exist . . .’ (Boulter et al., 1972) ‘Molecular clocks are like Santa Claus: everyone wants to believe in them, but no one really does.’ (H. Brad Shaffer, pers. comm.) The second most interesting thing about Magall on et al.’s new analysis of the timing of angiosperm diversification in this issue of New Phytologist (pp. 437–453) is their compendium of molecular clock estimates for the age of angiosperms (their fig. 1). Since 1972, when Ramshaw and colleagues (in this journal) used some of the first amino acid sequence data from plants to estimate an age of angiosperms at 300–400 million years ago (Ma), molecular-based estimates have ranged wildly. Also since then many things have happened: genomics, phylogenetic biology, statistically sophisticated inference of divergence times, and a much better understanding of the early angiosperm fossil record, to name the most relevant. Why then are estimates of the ‘crown group’ age of angiosperms (see Fig. 1) typically 50–150 million years older than the earliest angiosperm fossils from the Early Cretaceous (c. 135Ma); and why have the error bars on estimated ages not gotten much narrower?