In this study we chronicle the development of the Australian-Antarctic discordance (AAD), the crenelated portion of the Southeast Indian Ridge between ~120¿ and 128 ¿E, since anomaly 6y time (19 Ma). We reconstruct satellite-derived marine gravity fields and depth anomalies at selected times by first removing anomalies overlying seafloor younger than the selected age, and then rotating the remaining anomalies through improved finite rotations based on a very detailed set of magnetic anomaly identifications. Our gravity field reconstructions reveal that the overall length of the Australian-Antarctic plate boundary within the AAD has been increasing since 19 Ma. Concomitantly, the number of propagating rifts and fracture zones in the vicinity of the discordance has increased dramatically in recent times, effectively dividing it into its present-day configuration of five distinct spreading corridors (B1--B5) that are offset alternately to the north and south and exhibit varying degrees of asymmetric spreading. Our bathymetric reconstructions show that the regional, arcuate-shaped, negative depth anomaly (deeper than predicted by normal lithospheric cooling models) presently centered on the discordance began migrating westward before anomaly 5ad time (~14.4 Ma), and that a localized depth anomaly low, which at time 5ad lay on the ridge axis in spreading corridor B5, has been split apart by subsequent seafloor spreading. The magnetic anomaly patterns suggest that the depth anomaly is not always associated with a particularly contorted plate boundary geometry. Although the plate boundary within the AAD has been getting progressively more crenelated with time, this effect shows little to no migration along the ridge axis since 19 Ma. Thus any geodynamic models of the evolution of the discordance must account for the following observations: (1) the crenelation of the plate boundary within the AAD has increased with time, (2) the center of the crenelated zone does not appear to have migrated along the ridge crest, and (3) both the depth anomaly and the isotopic boundary between Pacific and Indian mantle have been migrating westward along the ridge axis but at apparently different rates. We suggest that both along-axis migration of the depth anomaly and isotopic boundary, as well as temporal variation in the upwelling mantle material beneath the AAD, and local tectonic effects are required in order to explain these observations. ¿ 1999 American Geophysical Union