In 1924, Bob Marshall and his Harvard University advisor Richard Fisher developed an integrated historical approach for forest reconstruction to address ecological questions and provide insights to forest management. Their approach utilized complementary methods and data: dendrochronology, diverse historical records (e.g., deeds, town and oral histories, and census and lumber mill records), and intensive field mapping and sampling of the site including stumps and uproot mounds, and forest composition and structure. Marshall applied this approach broadly to a 61-ha Tsuga canadensis (eastern hemlock)-Pinus strobus (eastern white pine) forest on sandy dry soils and intensively to a 0.15-ha sub-plot. He sought to test his hypothesis that pine and hemlock displayed compositional resilience to disturbance and to understand the life-history and growth characteristics of the two species that enabled co-dominance. Marshall concluded that, in contrast with the successional tendencies of white pine across New England's mesic uplands, pine and hemlock had dominated on these dry, sandy soils since at least the early 1700s and through multiple episodes of logging. Based on his interpretation of the contrasting growth rates and shade tolerance of these species, he developed a simple model of forest development and silviculture to guide management on dry sites for pine and hemlock timber production. Fisher, Marshall, and Harvard Forest colleagues used these historical insights to plan a suite of harvesting experiments on the site in 1924–25 to perpetuate hemlock and pine. Unfortunately, the hurricane of 1938 and subsequent salvage logging terminated the experiment. In 2007, we tested Marshall's interpretations and prediction of resilience in this forest by examining its long-term response to the series of intense disturbances. We synthesized decades of observations, photographs, and data, and relocated Marshall's plot to measure forest age and size structure, composition, and site features including decaying stumps, pits, mounds, and bent and sprouting individuals. The current hemlock and pine forest is strikingly similar in structure and composition to that of 1924. These results reinforce Marshall's conclusion that hemlock and white pine forests on well-drained sandy soils can be remarkably resilient in composition to intense disturbance. The work highlights the development and application of an integrated approach to forest reconstruction by Marshall and Fisher and underscores the contribution of historical insights to addressing basic ecological questions, designing large long-term field experiments, and guiding forest conservation and management.