The traditional framework for the description of arthropod development takes the molt-to-molt interval as the fundamental unit of periodization, which is similar to the morphological picture of the main body axis as a series of segments. Developmental time is described as the subdivision into a few major stages of one or more instars each, which is similar to the subdivision of the main body axis into regions of one to many segments each. Parallel to recent criticisms to the segment as the fundamental building block of arthropod anatomy, we argue that, while a firm subdivision of development in stages is useful for describing arthropod ontogeny, this is limiting as a starting point for studying its evolution. Evolutionary change affects the association between different developmental processes, some of which are continuous in time whereas others are linked to the molting cycle. Events occurring but once in life (hatching; first achieving sexual maturity) are traditionally used to establish boundaries between major units of arthropod developmental time, but these boundaries are quite labile. The presence of embryonic molts, the 'gray zone' of development accompanying hatching (with the frequent delivery of an immature whose qualification as 'free-embryo' or ordinary postembryonic stage is arbitrary), and the frequent decoupling of growth and molting suggest a different view. Beyond the simple comparison of developmental schedules in terms of heterochrony, the flexible canvas we suggest for the analysis of arthropod development opens new vistas into its evolution. Examples are provided as to the origin of holometaboly and hypermetaboly within the insects.