Understanding micro-image configurations in quasar microlensing

The micro-arcsecond scale structure of the seemingly point-like images in lensed quasars, though unobservable, is nevertheless much studied theoretically, because it affects the observable (or macro) brightness, and through that provides clues to substructure in both source and lens. A curious feature is that, while an observable macro-image is made up of a very large number of micro-images, the macro flux is dominated by a few micro-images. Micro minima play a key role, and the well-known broad distribution of macro magnification can be decomposed into narrower distributions with 0, 1, 2, 3, ... micro minima. This paper shows how the dominant micro-images exist alongside the others, using the ideas of Fermat's principle and arrival-time surfaces, alongside simulations.