Hemlock dwarf mistletoe (Arceuthobium tsugense (Rosendahi) G.N. Jones subsp. tsugense (western hemlock race)) is the most important and widespread disease of old-growth western hemlock forests in the Pacific Northwest. Although heavy infection of dwarf mistletoe can significantly increase growth loss and mortality of host trees, the parasite is also considered an important ecological component of forests influencing stand structure, species composition, and wildlife habitat. This retrospective study was conducted in stands that sustained extensive windthrow in the past and was designed as a simulation for stand structure and disease conditions expected after use of silvicultural systems that retain overstory canopy trees in a managed forest. The specific objectives were to: 1. characterize the forest structure and dwarf mistletoe component of stands that developed after catastrophic wind storms in the late 1800's; and 2. develop and validate a model, using site, plot, and tree variables, that predicts the dwarf mistletoe rating of hemlock trees that have been exposed to the pathogen for approximately a century. Results indicated that retention of overstory infected hemlock trees in southeast Alaska forests will result in pathogen spread to and intensification on hemlock trees that have developed since the late 1800's (post-disturbance trees), but not to the devastating levels predicted by previous research in the coastal forests of British Columbia, Washington, or Oregon. On Kuiu Island, the retention of an average of 3 heavily infected overstory residual trees/plot (83 residual trees/ha) resulted in nearly 100 percent infection of post-disturbance hemlock trees within a plot. Of those trees infected, less than one-fifth had heavy infection (a dwarf mistletoe rating of 5 or 6), and approximately one-half had moderate infection (a dwarf mistletoe rating of 3 or 4). On Chichagof Island, the retention of an average of 2 heavily infected residual trees/plot (69 residual trees/ha) resulted in nearly 100 percent infection of post-disturbance hemlock trees within a plot. Of those trees infected, less than ten percent had heavy infection and approximately one-third had moderate infection. A mathematical model that predicted dwarf mistletoe rating on post-disturbance hemlock trees was developed from eight stands on Kuiu Island and the generality of the variables was tested in two stands on Chichagof Island. The Kuiu Island model predicted that the mean dwarf mistletoe rating of post-disturbance hemlock trees within a plot increased with increasing numbers and dwarf mistletoe ratings of residual and advanced regeneration trees within a plot. No other factors that described site or tree conditions within a plot appeared important in the model. Sitka spruce trees, though rare hosts of dwarf mistletoe, did not appear to increase the mean dwarf mistletoe rating of postdisturbance hemlock trees within a plot. When present in sufficient numbers, spruce trees may act as barriers to pathogen spread, although further analysis is needed to confirm a barrier effect. The Chichagof Island models, developed separately for the two stands, also predicted that the mean dwarf mistletoe rating of post-disturbance hemlock trees within a plot increased with increasing numbers and dwarf mistletoe rating of residual trees within a plot. The importance of advanced regeneration trees could not be verified in either stand on Chichagof Island since most plots lacked this tree type.