Durable wood products such as posts, poles and timbers are expected to resist decay and termite attack for several decades. Testing the durability of new types of wood products is challenging because it is impractical to conduct tests for more than a small fraction of the expected service life. There are fundamental questions about the minimum duration of testing needed to provide meaningful results, and the relationship between the durability of the small test specimens and the much larger commercial wood products. Forest Service and Mississippi State University scientists are now beginning to provide answers to these questions with research in field plots at Harrison Experimental Forest (Saucier, Mississippi). For example, scientists have shown that 3 years of testing, the minimum length of time currently required by industry standards, provides only about 50% confidence that a product will be durable in the long term. Extending the testing by only 2 more years increases that confidence to 85%. Scientists have also shown that the small test specimens degrade an average of 2.2 times faster than product-size material, although the R2 value for this correlation is only 0.69. Models are also being developed to better define the relationship between tests conducted in Mississippi and those conducted in cooler climates such as Wisconsin. This research will improve interpretation of accelerated testing results, and provide greater confidence in the long-term performance of durable wood products.