The joint Haumea J ₂–Hi’iaka mass posterior distribution. Instead of displaying the density of sampled points as in Figure 1, we show the maximum fit quality (as measured by reduced χ ²) in a small bin. Bins without any sampled points, indicating extremely poor-quality fits, were set to the minimum bin value, although the true value is likely much worse. We find that the best fit with ﹩{\chi }_{\nu }^{2}\sim 1.1﹩ has a p-value of 0.23, meaning there is a 23% chance that random chance would produce a worse fit. A ﹩{\chi }_{\nu }^{2}\sim 1.25﹩ corresponds to a p-value of 0.05. Dashed red lines show the expected J ₂ values from different internal density models. The undifferentiated model assumes a homogeneous interior along with the occultation-derived shape model (Ortiz et al. 2017). The differentiated model is a two-layer model proposed by Dunham et al. (2019). The posterior shows that both models are consistent with the data, although the differentiated model is slightly disfavored.