• MacLean, David A.
  Canadian Forestry Service
  
• Ostaff, Donald P.
  Canadian Forestry Service
  

Data on spruce budworm (Choristoneura fumiferana (Clem)) caused tree mortality and stand characteristics collected between 1977 and 1981 from 25 plots on the Cape Breton Highlands, Nova Scotia, were used to derive relationships between sample size and precision level. Comparison of three plot types (prism point samples with BAF - 2.3, 0.01-ha circular fixed area plots, and 0.05-ha fixed area plots) indicated that using larger numbers of small plots (either prism or fixed area) within each sampled stand would be the best method for estimating tree mortality. Curves relating required number of plots to desired precision level (specified as a bound on the morality estimate) were presented for estimating both annual mortality and cumulative (total) mortality. These relationships will assist in choosing sample size in future studies determining stand characteristics and budworm-caused tree mortality. As a rough approximation, sampling about 15 prism points within each stand should result in estimation of cumulative mortality to about ±10% precision.