Abstract
The patterns of star creation which provide the optical spiral structure of galaxies have been attributed either to a gravitational (density-shock) or hydromagnetic origin; in the latter model gas clumping results from varying tilt of an oblique magnetic field. The predictions of the two theories (Table I) are compared with observational data; in each theory the dust, young-star and older-star arms are separated as observed. The major differences lie in the surface density and velocity distributions of cool gas; observations in M31, M33 and M101 seem to provide strong evidence against the density-shock theory. The 21-cm line profiles of H I in our Galaxy have not yet been correctly interpreted because of the neglect of velocity perturbations which rank in importance with density variations. Hence the line profiles favour neither theory; however, observed velocities V z and shear DV / dz seem to be hydromagnetic. Observations of radio synchrotron emissions from M51, M31 and our Galaxy are either marginal or contrary to the density shock theory; a shock in our Galaxy should provide clear radio evidence, but this is not observed. Observational evidence indicates that some spiral galaxies do, and all may, have magnetic halos probably of primeval origin as required by the hydromagnetic theory.