Occurrence of mold was visually monitored for 26 days on samples of major anatomical components of corn stover maintained at several storage temperatures (T) and water activities (a(w)). Glass desiccators with saturated salt solutions placed in temperature controlled chambers provided simulated storage conditions with temperatures ranging from 10 degrees C to 40 degrees C and water activities ranging from 0.11 to 0.98. Mold affected leaf, stalk skin, and stalk pith equally at water activity greater than 0.9. As expected, a combination of increased water activity greater than 0.9 and temperatures greater than 30 degrees C was conducive to mold growth. Based on material moisture content during tile initial mold growth, it was postulated that among the corn stover components the stalk pith was the least resistant to mold growth followed by stalk skin and leaf for the studied range of temperature and water activity. Mold growth models fitted well with the observation. A linear mold-free days predictions using a three-parameter regression model (T, a(w), and T x a(w)) was superior (R-2 = 0.99) to other models considered. The exponential spoilage model using two parameter T and a(w) also gave comparable performance (R-2 = 0.95). Among tile independent factors, T x a(w) product was the most significant (p = 0.0069) followed by T (p = 0.0114), and a(w) (p = 0.3140) in explaining the experimental data. The developed models can be applied to predict the safe storage period of corn stover components exposed to various temperature and moisture environmental conditions. (c) 2007 Elsevier Ltd. All rights reserved.