Fifty years ago at a conference on the structure of galaxies almost everything would have been ascribed to gravity. A decade later turbulence became the new thing - partly to solve angular momentum problems. Again a decade later magnetic fields followed: synchrotron radiation was observed, angular momentum could be efficiently transported and a variety of bizarre extragalactic shapes was thought to be related to magnetic fields. In the sixties, however, the situation was reversed when it turned out that spiral structure could perhaps be understood on a gravitational picture with some simple hydrodynamics and that some of the most intriguing shapes could be reproduced in simulations of the gravitational interaction between galaxies. As a consequence magnetic fields again diminished in importance. The next decade saw a much increased role for the thermodynamics of the interstellar medium. At the end of this meeting it is clear that while gravity, hydrodynamics and thermodynamics are important, magnetic fields cannot be neglected and should be integrated into the overall picture.