Technical Abstract: Five apple orchards in West Virginia, USA, were managed with different pest management practices ranging from unsprayed to conventional insecticides. Data were collected on aphid, predaceous insect, and spider abundance every 2 weeks from May to July along one or two randomly selected, 15 consecutive tree transects. A total of 892 individual predaceous insects were observed: 32% were Aphidoletes aphidimyza, 21% were Coccinella septempunctata, and the rest were from 13 other taxa. The most important predator controlling aphid populations was C. septempunctata. Of the 32 sample transects with predators, 72% showed an edge effect in which either the standard error or mean of predator abundance differed from the edge towards the center of the orchard. This edge effect extended a mean of 7 trees (30 m) into the orchard. Only 40% of the 32 transects that had spiders showed an edge effect, extending into the orchard a mean of 8 trees (33m). However, 75% of the transects with more than one spider per tree had an edge effect of 10 trees (43m). We recommend a minimum border of 7 trees before sampling for aphid predators and 10 trees before sampling for spiders. A recommended experimental plot size, for semi-dwarf apple orchards, to eliminate edge effects is 1.25ha for aphidophaga, 1.5ha for spiders; a minimum plot size of 0.6ha for aphidophaga and 0.75ha for spiders would minimize edge effects.