The concept of a phylotypic stage, when all vertebrate embryos show low phenotypic diversity, is an important cornerstone underlying modern developmental biology. Many theories involving patterns of development, developmental modules, mechanisms of development including developmental integration, and the action of natural selection on embryological stages have been proposed with reference to the phylotypic stage. However, the phylotypic stage has never been precisely defined, or conclusively supported or disproved by comparative quantitative data. We tested the predictions of the ‘developmental hourglass’ definition of the phylotypic stage quantitatively by looking at the pattern of developmental–timing variation across vertebrates as a whole and within mammals. For both datasets, the results using two different metrics were counter to the predictions of the definition: phenotypic variation between species was highest in the middle of the developmental sequence. This surprising degree of developmental character independence argues against the existence of a phylotypic stage in vertebrates. Instead, we hypothesize that numerous tightly delimited developmental modules exist during the mid–embryonic period. Further, the high level of timing changes (heterochrony) between these modules may be an important evolutionary mechanism giving rise to the diversity of vertebrates. The onus is now clearly on proponents of the phylotypic stage to present both a clear definition of it and quantitative data supporting its existence.