How do we know?
Observation: Developmental processes have been meticulously documented for centuries. A contemporary of Charles Darwin, Ernst Haeckel, was a noted illustrator and natural historian and recorded embryonic development of numerous species (unfortunately, with a great deal of creative licence). Visual records provide an important foundation for work on the biological mechanisms of development. Such work continues today, with 'virtual embryo' projects providing digital images of different stages of animals' development.
Cell labelling: More detailed information can be obtained by experiments in which cells are labelled and the fate of their descendants is tracked. This technique can be used to track cell migration and also to identify a cell's descendants or ancestors.
Tissue explants: Early embryological research relied heavily on moving parts of embryos, either around the same embryo or between embryos. Such work identified key areas of the embryo that controlled the development of neighbouring regions (organising centres).
Genetics: Genetic approaches aim to identify the genes controlling developmental processes. Classically, the aim is to identify a genetic change (mutation) associated with a specific phenotypic change. The nature of the phenotypic defect provides a clue to the function of the unmutated gene - if a mutation stops a head forming, it can be assumed the gene is normally involved in head formation.
From the 1980s onwards, genetic approaches have revealed much about the development of model organisms such as the fruit fly, nematode worm, zebrafish and mouse. Particularly important has been the ability to 'knock out' (eliminate) the function of specific genes (and, more recently, to 'knock in' new genes). Lately, it has even been possible to knock out genes at particular times in development or in specific tissues.
Comparative studies: The common evolutionary origins of different organisms means that results in one animal can be compared with those from another. The function of a human gene implicated in an inherited disease, for example, can be analysed in mice or zebrafish.
Functional studies: Genetic studies can identify genes involved in developmental processes; their sequences may provide clues to their function. Generally, though, cellular and biochemical studies need to be carried out to find out what role a gene has in the cell.
Mathematical modelling: Mathematical approaches to development were proposed by computing pioneer Alan Turing in the 1950s. The goal is to develop mathematical representations that model the genetic and molecular control networks that create different morphologies during development. The growth of systems biology - looking at how many different components act together - is stimulating much activity in this area.
Image: Chromosomes; Kate Whitley, Wellcome Images