Old-growth forests have long been the focus of much interest and attention. Even so, most of what we think we know about them is not well-tested. This is because they are 'slow systems'. Appropriate time-frames for studying the properties of these systems are measured in decades and centuries - longer than the careers of researchers. For example, old-growth forests are often treated as 'equilibrial' - maintaining more or less constant composition and structure over long periods of time through interactions among their constituent species. However, this assumption has rarely been tested effectively because appropriately long-term quantitative observation and measurement have been unavailable. Similar problems apply with regard to many other questions about the workings of old-growth, or late-successional forests. My project uses data collected over many decades (up to 70+ years) from permanent study plots in old-growth forests in Michigan. While these data sets begin to give appropriate temporal perspective for testing hypotheses about these forest communities, they are difficult to work with because they are very large and complicated. Data have been collected by different researchers, using different approaches and at variable intervals for several hundred study plots. Fortunately, these problems have become more tractable with the development of sophisticated analytic approaches and modeling tools used at NCEAS. I hope my work with this unusual data-set will lead to collaborative and parallel studies using similar data-sets for other 'slow systems' and, ultimately, to a fuller, synthetic understanding of the properties of forest ecosystems over time and space.