Experiment designs to evaluate inter- and intra-specific interactions in mixed plantings of forest trees

A review of three mixed-species trials reveals the utility of competition indices for evaluating inter- and intra-specific interactions between trees, the desirability of experiments that span a range of tree spacing and composition to inform calibration of these competition indices, the need for extremes of species composition and stand density to calibrate response surfaces, and the far-reaching impact of edge-effects. Experiment layouts commonly used for mixed-species trials in forestry (such as replacement series) rarely provide a strong basis to calibrate competition indices and response surfaces. Alternative designs involving systematic changes in species composition may offer a better basis for calibrating response surfaces, especially when used in conjunction with designs that vary tree spacing systematically. Systematic spacing designs (such as Pudden clinal plots, Marynen plaids and Nelder fans) are well established, but designs that vary composition systematically are less common. The Goelz triangle and an alternative are contrasted to illustrate how both may offer inspiration when designing trials for mixed-species research in forestry.