Variation is everywhere in biology. Structural variation is present at molecular, cellular, organismal, and population levels, and functional variation occurs in processes from metabolism to development to behavior. In spite of this, we often describe biology in typological terms, and this is often a source of confusion.
Some variation is crucial; for example, evolution is dependent on genetic variation, and behavioral variation within ant and bee colonies ensures that all the necessary jobs get done. Much variation, though, is simply biological noise, an unavoidable consequence of the mostly analogue nature of living systems. In extreme cases, variation of this sort can complicate and even derail development, but in general development is remarkably robust. A variety of regulatory mechanisms prevent small amounts of variation early in development from being amplified into large variations in adults.
Pierre Haas and colleagues have posted a preprint to arXiv describing variation in the developmental process of inversion in Volvox globator. Facultatively sexual organisms such as Volvox are great for studying non-genetic sources of variation, because it’s pretty simple to produce millions of genetically identical individuals. When they are raised in identical conditions, variation due to environmental differences is minimized, and most of the observed variation is stochastic.