Many soybean (Glycine max) producers consider white-tailed deer (Odocoileus virginianus) to be costly crop pests, and chemical repellants are an option for reducing deer damage. To cost-effectively utilize repellants, it is crucial to know the spatial and temporal aspects of deer-browsing patterns. During 2003During -2006 in Delaware, USA, we investigated where and when browsing was most intense within crop fields and subsequently had the greatest effect on yield of soybeans at an estimated deer density of 21 deer/km 2 . We systematically placed plots (n ¼ 1,340) at 10-m intervals from the forest edge. We designed protection treatments to mimic the effect of chemical repellents. We systematically assigned types of protection treatments to plots and protected plots using fencing. Protection did not affect the yield of double-crop soybeans in any year. In 2 years of the study (2003 and 2006), protection did not affect full-season soybean yield. In the other 2 years (2004 and 2005), plots protected from deer browse yielded less than plots that were not protected. Yield tended to be the lowest in the first 20 m from the forest edge, but there was no interaction effect between deer browsing and distance from the forest edge for any year or crop type. We found that deer browsing increased yield, possibly because browsing increases branching of individual plants, and thus the number of bean pods. We also found that the distance from the forest edge affected yield, but independently of deer browsing. Deer damage to soybeans, at a deer density of 21 deer/km 2 , may be more of a perceptual issue and the use of a chemical repellant would not be justified. Ó