All of the data reported in Robilotto, Khang, and Zaidi (2002) Robilotto and Zaidi (2004), and Singh and Anderson (2002) are consistent with Robilotto and Zaidi's theory that perceived transparency (or opacity) is determined by the perceived contrast of the filter region. Kasrai and Kingdom's (2001) results also appear largely consistent with the perceived-contrast theory, since background Michelson contrast did not vary greatly among their stimuli. The data reported in Albert are largely consistent with this theory too. Greater variability in opacity matches for my zero-reflectance stimuli are also to be expected, according the perceived-contrast theory, as observers would be unable to find satisfactory perceived-contrast matches in these cases. In addition, I have now conducted a variant of my Experiment 1 using precisely the same instructions as Singh and Anderson and Anderson, Singh, and Ming, and the results were very similar to my original Experiment 1. Thus, the only results that appear to be contrary to the perceived-contrast theory are those from Experiments 1 and 6, in Anderson et al. (2006, see their Figures 3, 9, and 10). I propose a dual-process model of perceived overlay transparency in which filter-to-background relative properties (e.g., relative perceived contrast and mean luminance, boundary contrast) define certain qualitative constraints and categories of perceptual transparency. This model explains all the data under discussion here.