Viking images of the auroral distribution have been used to investigate the relevance of various ionospheric coordinate systems. An important aspect of the large-scale auroral shape is its dependence on asymmetries of the Earth's internal field. Model predictions of where the aurora occurs, using the equatorial plane's volume current density, agree with observations and imply that the internal field plays a more important role than generally believed. Historically, the belief that the internal field has only small effects seems to stem from the widespread use of the corrected geomagnetic and invariant coordinate systems. These system's involve the mapping of field lines and have advantages in statistical studies and comparisons; less sophisticated systems such as the eccentric dipole coordinate system should be used in individual studies and in studies involving the differentiation or integration of some observational parameters. Observations of the auroral distribution are given to illustrate the universal time, tilt angle, and KP variability in different coordinate systems and demonstrate that the dominant variability of the aurora is due to internal field asymmetries. A new set of coordinate systems are briefly developed as examples of how to incorporate external field models into studies of auroral images. It is proposed that one of these coordinate systems can be used as a test of how well an external field model can match observed auroral distributions.