Experimental Designs for Comparing Landscape Tree Cultivars Under City Conditions

Trees in urban environments encounter many unnatural stresses. The effects of restricted growing space, air pollutants, road salts, and other stresses may best be quantified by measurements of height and foliage condition. To develop efficient experimental designs for testing trees in cities, representative age class means and variances were calculated for height, foliage condition, and six other performance traits. Existing plots of 23 different landscape tree cultivars growing in 19 U.S. cities were evaluated. Sizes and arrangements of experiments needed to show that differences of various magnitudes between two cultivars are significant were computed. Several experimental designs which optimize statistical efficiency and precision are recommended. One consists of two sets of plots containing one cultivar each and randomly located in one site category (e.g. residential, business, or industrial) in a city. This design would require at least 24 trees per cultivar and four to eight trees per plot. Other designs which may be more precise are discussed. How tests of trees in cities can be implemented, extended, and augmented by nonurban tests is also considered.

Download this file:

Download this file

Details

Author(s): W. Donald Bartoe, Henry D. Gerhold

Publication: Tree Improvement and Genetics - Northeastern Forest Tree Improvement Conference - 1978