The notochord is derived from the embryonic mesoderm and consists of an inner core of vacuolated cells filled with glycoproteins, covered by two helical collagen-elastin sheaths.

The notochord provides a directional reference to the surrounding tissue as a midline structure during embryonic development, acts as a precursor for vertebrae and a primitive axial endoskeleton.

[5] The angle between these fibers determines whether increased pressure in the core will result in shortening and thickening versus lengthening and thinning.

[6] Alternating contraction of muscle fibers attached to each side of the notochord result in a side-to-side motion resembling stern sculling, which allows locomotion.

Initially, it exists between the neural tube and the endoderm of the yolk-sac; soon, the notochord becomes separated from them by the mesoderm, which grows medially and surrounds it.

From the mesoderm surrounding the neural tube and notochord, the skull, vertebral column, and the membranes of the brain and medulla spinalis are developed.

The hypochord is a transient structure ventral to the notochord, and is primarily responsible for correct development of the dorsal aorta.

[14][15][better source needed] By the age of 4, all notochord residue is replaced by a population of chondrocyte-like cells of unclear origin.

[20] The notochord secretes a protein called sonic hedgehog (SHH), a key morphogen regulating organogenesis and having a critical role in signaling the development of motor neurons.

The Ordovician oceans included many diverse species of Agnatha and early Gnathostomata which possessed notochords, either with attached bony elements or without, most notably the conodonts,[23] placoderms,[24] and ostracoderms.

Even after the evolution of the vertebral column in chondrichthyes and osteichthyes, these taxa remained common and are well represented in the fossils record.

[25] They point out that, although many of these ideas have not been well supported by advances in molecular phylogenetics and developmental genetics, two of them have actually been revived under the stimulus of modern molecular approaches (the first proposes that the notochord evolved de novo in chordates, and the second derives it from a homologous structure, the axochord, that was present in annelid-like ancestors of the chordates).

Deciding between these two scenarios (or possibly another yet to be proposed) should be facilitated by much more thorough studies of gene regulatory networks in a wide spectrum of animals.