I report four experiments designed to investigate the perceptual grouping processes by better understanding how a simple pattern of dots comes to be represented in the visual memory systems. Subjects were briefly shown a random set of point-like dots which were then masked, and subsequently reappeared with a possible added or deleted dot. Subjects detected deletions more accurately than additions and detected changes to the most peripheral dots more reliably than changes to interior dots, even those much closer to fixation. The most peripheral points appear to be objectized in a process that might be analogized as shrink wrapping. The circumscribed area is memorable and easily recalled despite brief presentations. The data are consistent with the view that the fastest of the perceptual processes is a global system which proceeds from the periphery inward toward fixation and that a separate slower local analytical sequence starts at fixation and works outward. I conjecture that proximity, one, if not the most salient of all features in the early perceptual processes that extract data from the sensory store and encode it in the visual memory systems, is not itself stored as part of the representation.