Techniques for reducing edge bias are discussed and a new approach, termed the linear expansion method, is presented. This method is applicable to a wide variety of plot shapes and sizes and is unbiased under very general assumptions about the forest tree spatial pattern. The accuracy and efficiency of the linear expansion method were compared with (i) no edge bias correction and (ii) corrections by plot image translation techniques for several forest stands and spatial patterns. Results indicated translation and the linear expansion method behaved equally well on square plots 0.08 ha in size. As the plot size diminished or its shape deviated from square, the linear expansion method provided greater accuracy and lower bias than translation techniques. The linear expansion method also provided useful accuracy on circular plots where translation was not feasible.