SUMMARY Biologists measure developmental time by dividing development into arbitrary time blocks called “stages.” This is a reasonable approach, provided that developmental timing is precisely controlled within a species. However, the degree of this precision is unknown. This is unfortunate because precision in developmental timing at the population level is a central issue to the whole research program of heterochrony. To examine this issue, we apply Ontogenetic Sequence Analysis to 261 embryos of the Lake Victoria cichlid Haplochromis piceatus. The result of our analysis can be mapped as a complex web of 26,880 equally parsimonious developmental sequences. This topology reflects timing polymorphism (intraspecific heterochrony) among embryos of this species. Because of this timing polymorphism, it is not possible to define discrete “stages” in this cichlid (although there is sufficient sequence signal to assess the maturity of embryos). More generally, we show that sequence polymorphism creates uncertainty about how a given embryo will develop implying that the mechanisms controlling developmental timing in embryos lack precision. For this reason, it is imperative to consider patterns of embryonic variability when measuring developmental time.