THE USE OF INDIVIDUAL TREE STATISTICS TO QUANTIFY EFFECTS IN AN ACID MIST EXPERIMENT WITH MATURE TREES

Mature trees of a single clone of Sitka spruce (Picea sitchensis (Bong .) Carr.) growing in the field were exposed to simulated acid mist at pH 2.5 over three successive years. The growth of each tree in the aci d mist treatment, in terms of the stem area increment was related to t he pattern of growth of the untreated trees at the site during 1991 an d 1992. In 1993, the effects of stopping the acid mist treatment on ha lf of the treated trees was investigated by comparing the actual growt h against that predicted relative to the untreated trees on the basis of the previous years' data. The prediction of growth from one year to the next was consistently too small for the smallest acid-misted tree s, suggesting that the growth rates relative to the untreated trees we re increasing with time. Relative to the trees which continued to rece ive acid mist, however, the smaller trees which received no further ac id mist in 1993 grew less than predicted, but there was a significantl y greater growth than predicted for the largest 'recovering' trees. Re sults from the individual tree modelling approach showed that a 10% ch ange in relative stem area increment would be detectable at the 5% lev el, using a block of 4 trees as a treatment unit. The statistical tech niques used are alternatives to conventional ANOVA calculations in. de termining statistically significant treatment effects where numbers of replicates are small.