The history of isospin is described, starting from the introduction by Heisenberg (1932) of the concept (ρ-spin), associated with the idea that proton and neutron are two states of the same entity (later to be called "nucleon"). It is stressed that from the start Heisenberg had the picture of the transmission of proton-neutron interaction by a charged agent (field) in mind. It is shown that, in contrast, Majorana's modification of Heisenberg's interaction scheme rejected the isospin formalism and moved away from a field picture. The revival of the isospin idea by Cassen and Condon (1934) on the basis of the "charge-independent" description of nucleon-nucleon interaction and the formal "iso-invariance" of that interaction is discussed. The consequent wide application of the isospin idea in nuclear structure theory is described. In particular the introduction (Wigner, 1937) of the total isospin vector of a system of nucleons, of isospin multiplet classification of states and of applications to nuclear reactions, β-decay, etc. The introduction of isospin into field theories of nucleon-nucleon interaction is described : the charge independent extension first of the Fermi (β, ν) field (Kemmer, 1937) and then of the Yukawa field (Kemmer, 1938). A brief account is given of the consolidation of the "symmetric theory" of the isospin of the nucleon-pion system, following the post-war experimental proof of the existence of the pion isotriplet and the improvement in the understanding of field theories in terms of the renormalisation concept. The application of this iso-invariance of the (N-π) system in the description of a wide range of phenomena (weak interactions) is sketched. The first step into non-abelian gauge theory - the Yang-Mills iso-vector gauge field - is briefly described, concluding with brief remarks on the status of isospin within the framework of modern particle theory.