Ecological theory predicts that habitat growth and loss will have different effects on community structure, even if they produce patches of the same size. Despite this, studies on the effects of patchiness are often performed without prior knowledge of the processes responsible for the patchiness. We manipulated artificial seagrass habitat in temperate Australia to test whether fish and crustacean assemblages differed between habitats that formed via habitat loss and habitat growth. Habitat loss treatments (originally 16 m2) and habitat growth treatments (originally 0 m2) were manipulated over 1 week until each reached a final patch size of 4 m2. At this size, each was compared through time (0-14 days after manipulation) with control patches (4 m2 throughout the experiment). Assemblages differed significantly among treatments at 0 and 1 day after manipulation, with differences between growth and loss treatments contributing to most of the dissimilarity. Immediately after the final manipulation, total abundance in habitat loss treatments was 46% and 62% higher than controls and habitat growth treatments, respectively, which suggests that animals crowded into patches after habitat loss. In contrast to terrestrial systems, crowding effects were brief (=1 day), signifying high connectivity in marine systems. Growth treatments were no different to controls, despite the lower probability of animals encountering patches during the growth phase. Our study shows that habitat growth and loss can cause short-term differences in animal abundance and assemblage structure, even if they produce patches of the same size.