We subsequently evaluated the effect of felling month,
species and the presence of shelter by relating these variables
to the estimated half-life and lower asymptotic moisture
content (g) using a generalized linear models technique (the
GLM procedure in SAS ver. 9.2).