A multi-time-period method of stock reconstruction was used to estimate harvest and stock interception rates by area and time period, exploitation rates and total run size by stock. The data required for these reconstructions were catch by time and area, daily escapement by population (stock), the residence time of each population in each harvest area, and the routing of the populations. Migration routes for each stock were defined using information from the 1982-85 north coast tagging studies. The initial set of migration routing parameters were adjusted until the run reconstruction results approximated those from the tagging study years. Two different sets of migration parameters were required to fit the 1982 and 1983 interception rates for sockeye. While the 1983 set provided the best fit to the interception rates for Alaskan fisheries based on scale data for 1984-95, year to year variability can be substantial. Consequently, we incorporated all the available stock composition estimates for Alaskan fisheries into our sockeye reconstruction analysis.