Time course of grouping of shape by perceptual closure as a function of spatial proximity and collinearity between the closure-inducing fragments was examined in three experiments using primed-matching. When only closure was available, early priming of the global shape was observed for spatially close fragments, but not for spatially distant fragments. When closure and collinearity were available, the global shape of both spatially close and spatially distant fragments was primed at brief exposures. These results indicate that spatial proximity is critical for the rapid grouping of shape by perceptual closure in the absence of collinearity, but collinearity facilitates the rapid grouping of shape when the closure-inducing line segments are spatially distant. These findings suggest a rapid computation of collinearity between closure-inducing line segments that is insensitive to spatial proximity within a certain range. This fast-occurring mechanism enables efficient image descriptions and apparently is crucial for a fast, reliable interpretation of the visual scene. The results also showed, however, that stable priming effects of the global shape over time were observed only when the closure-inducing fragments were collinear and spatially close, suggesting that maintaining a stable representation of shape beyond the first stages of visual processing depends both on spatial proximity and collinearity between the closure-inducing fragments.