Similar comparison as in Fig. 7, but for squared visibilities |V|², with a contribution . The |V|² values are plotted against projected baseline B/λ (in cycles), and shifted vertically according to the dataset number. The are CHARA/MIRC data at the bottom, NPOI in the middle, and CHARA/VEGA at the top. Synthetic data are denoted by yellow crosses, observed data by blue error bars, and residua by red lines. A few outliers with large uncertainties, which do not contribute much to χ² anyway, were purposely removed from the plot to prevent clutter. Even though there are some systematic differences for individual segments of data, overall trends seem to be correctly matched.