A fractal universe model has been proposed elsewhere (Driessen, 1991, 1994), based on the assumption of perfect self-similarity between some definite scales, and in particular between particles and galaxies. Its motivation rested essentially on two ideas: first, if striving for unification is justified, then the fractal idea makes sense, for in this case one single theory of particles could encompass all self-similar structures at once. Secondly, it could be demonstrated that, provided some reasonable hypothesis were made about the structure of spiral galaxy cores, they came back to the same topological state after two complete revolutions (Driessen, 1991). These ideas combined suggested that spiral galaxies might correspond to electrons on the cosmic scale. The main test of the coherence of the fractal picture was then to discover if other galaxy types and quasars do find a place in this model. Now, a small group of dissident has sustained for years that QSOs (or at least a certain class of QSOs) are local objects emitted by galaxies. This class would contain very small (d ≈ 0.3 kpc), faint (Mv ≈ −12), and most often radio-quiet objects with large (non cosmological) excess redshift. The analogy between the two statements “electrons emit photons” and “galaxies emit QSOs” hinted that QSOs embody cosmic scale photons, and entails that galaxies experience quasar exchange interactions. It was claimed further that galaxy clusters reproduce atomic and molecular-like systems with one or a few central giant ellipticals (the cosmic scale nuclei) interacting via quasar exchange with their satellite spirals (the cosmic scale electrons). The flock of dwarf galaxies around large ellipticals and spirals would result from condensation of dusty gas clouds entrained by ejected QSOs.