Graph
at left show no general clustering of points around any of the three
vectors which represent test sites. If the entire test population was
treated equal there would be no preference to any clone relative to
site.
Further
analysis of trial stratifying by region of collection shows a distinct
lack of red clones collected 56 degrees latitude performing well
at any site except on site 10.Local is assumed best, however the dark green clones, which in all cases have been moved north, are the best performance
