The observed finite width of spectral lines is due to both internal and external causes. The latter must be ascribed to the motion of the absorbing/emitting atoms (Doppler broadening) and the random perturbations brought about by neighbouring atoms. The former is a natural consequence of the intrinsic atomic structure. In this seminar I will review some of the steps that led in the first half of the last century to the quantum-mechanical calculation of the natural width of spectral lines. In the first place I will introduce the model put forward by Lorentz (1890's) to describe the interaction atom/radiation field in terms of harmonic oscillators (Lorentz oscillators). Then a necessary account of both the theory of radiative transitions by Einstein (1917) and the equivalence between the absorbing atoms and an ensemble of Lorentz oscillators stated by Ladenburg (1921) will be given. Finally I will consider the fundamental contribution by Weisskopf and Wigner (1930), who showed that a bound level of an unperturbed atom has a finite width determined by the lifetime of the level, according to the energy-time uncertainty established by the Heisenberg's principle.