The impact of horizontal heterogeneities, liquid water path (LWP from AMSR-E), and cloud fraction (CF) on MODIS cloud effective radius (<i>r</i><sub>e</sub>), retrieved from the 2.1 μm (<i>r</i><sub>e2.1</sub>) and 3.8 μm (<i>r</i><sub>e3.8</sub>) channels, is investigated for warm clouds over the southeast Pacific. Values of <i>r</i><sub>e</sub> retrieved using the CERES algorithms are averaged at the CERES footprint resolution (∼20 km), while heterogeneities (<i>H</i><sub>σ</sub>) are calculated as the ratio between the standard deviation and mean 0.64 μm reflectance. The value of <i>r</i><sub>e2.1</sub> strongly depends on CF, with magnitudes up to 5 μm larger than those for overcast scenes, whereas <i>r</i><sub>e3.8</sub> remains insensitive to CF. For cloudy scenes, both <i>r</i><sub>e2.1</sub> and <i>r</i><sub>e3.8</sub> increase with <i>H</i><sub>σ</sub> for any given AMSR-E LWP, but <i>r</i><sub>e2.1</sub> changes more than for <i>r</i><sub>e3.8</sub>. Additionally, <i>r</i><sub>e3.8</sub>–<i>r</i><sub>e2.1</sub> differences are positive (<1 μm) for homogeneous scenes (<i>H</i><sub>σ</sub> < 0.2) and LWP > 45 gm<sup>−2</sup>, and negative (up to −4 μm) for larger <i>H</i><sub>σ</sub>. While <i>r</i><sub>e3.8</sub>–<i>r</i><sub>e2.1</sub> differences in homogeneous scenes are qualitatively consistent with in situ microphysical observations over the region of study, negative differences – particularly evinced in mean regional maps – are more likely to reflect the dominant bias associated with cloud heterogeneities rather than information about the cloud vertical structure. The consequences for MODIS LWP are also discussed.