The inner ear is a complex sensory organ responsible for balance and sound detection in vertebrates. It originates from a transient embryonic structure, the otic vesicle, that contains all of the information to develop autonomously into the mature inner ear. We review here the development of the otic vesicle, bringing together classical embryological experiments and recent genetic and molecular data. The specification of the prospective ectoderm and its commitment to the otic fate are very early events and can be related to the expression of genes with restricted expression domains. A combinatorial gene expression model for placode specification and diversification, based on classical embryological evidence and gene expression patterns, is discussed. The formation of the otic vesicle is dependent on inducing signals from endoderm, mesoderm and neuroectoderm. Ear induction consists of a sequence of discrete instructions from those tissues that confer its final identity on the otic field, rather than a single all-or-none process. The important role of the neural tube in otic development is highlighted by the abnormalities observed in mouse mutants for the Hoxa1, kreisler and fgf3 genes and those reported in retinoic acid-deficient quails. Still, the nature of the relation between the neural tube and otic development remains unclear. Gene targeting experiments in the mouse have provided evidence for genes potentially involved in regional and cell-fate specification in the inner ear. The disruption of the mouse Brn3.1 gene identifies the first mutation affecting sensory hair-cell specification, and mutants for Pax2 and Nkx5.1 genes show their requirement for the development of specific regions of the otic vesicle. Several growth-factors contribute to the patterned cell proliferation of the otic vesicle. Among these, IGF-I and FGF-2 are expressed in the otic vesicle and may act in an autocrine manner. Finally, little is known about early mechanisms involved in guiding ear innervation. However, targeted disruption of genes coding for neurotrophins and Trk receptors have shown that once synaptic contacts are established, they depend on specific trophic interactions that involve these two gene families. The accessibility of new cellular and molecular approaches are opening new perspectives in vertebrate development and are also starting to be applied to ear development. This will allow this classical and attractive model system to see a rapid progress in the near future.