2021MNRAS.508..125S

Scintillation spectra of some pulsars have suggested the existence of <= AU scale density structures in the ionized interstellar medium, whose astrophysical correspondence is still a mystery. The detailed study of Brisken et al. suggested two possible morphologies for these structures: A parallel set of filaments or sheets (the 'parallel stripes model'), or a filament broken up into denser knots (the 'threaded beads model'). Here, we propose a straightforward test that can distinguish these two morphologies: Whether the apex of the main parabolic arc created by the scattered images deviates from the origin of the scintillation spectrum or not. In the 'parallel stripes' model, the scattered images move along the stripes as the relative position of the pulsar moves. As a result, the pulsar is always co-linear with the scattered images, and thus, the apex of the main parabolic arc stays at the origin of the scintillation spectrum. In the 'threaded beads' model, the scattered images remain at almost fixed positions relative to the density structures, and thus, the pulsar is not co-linear with the scattered images at most times, leading to an offset between the apex and the origin. Looking for this possible offset in a large sample of pulsar scintillation spectra, or monitoring the evolution of parabolic arcs will help pin down the morphology of these tiny density structures and constrain their astrophysical origin.