The eeccentricity of Mars is variable on Gyr timescales, and the current eccentricity of Mars is close to the most likely long-term value (Laskar, 2008). Therefore, if Mars were to start out with a very low (or very high) value of eccentricity, chaotic diffusion is likely to, over time, bring Martian eccentricity to its present value. While Cuk and Nesvorny (2008) explored the effects of chaotic Martian eccentricity on the current population of Hungarias, here we are interested in the possible effects of early changes of Martian eccentricity on a primordial proto-Hungaria population (Cuk, 2012), which is essentially the same concept as the E-belt proposed by Bottke et al (2012). A steadily rising eccentricity of Mars over the first Gyr of Solar System history could lead to a delayed release of impactors from the E-belt region. While this impactor flux would not be large or sudden enough to produce a classical “lunar cataclysm”, it would be compatible with the currently accepted picture of a more prolonged bombardment (Fassett and Minton, 2013, Morbidelli et a, 2018). It would also be compatible with evidence that the impactors were not main-belt asteroids (Minton et al, 2015, Nesvorny et al, 2017), in which case different (large-end) size-distributions for E-belt objects and the asteroids would be required. In order to model this process accurately, we have modified SWIFT-rmvs4 to include relativistic precession which significantly affects the secular dynamics of the inner planets. At the meeting we will present our results on the plausible initial conditions for the orbit of Mars, and the consequences for the stability of proto-Hungarias and the resulting bombardment of the inner planets.