All Chinook salmon Oncorhynchus tshawytscha stocked in Lake Huron during 2000–2003 were marked with oxytetracycline (OTC), fin clips, or a combination thereof to determine the relative contributions of natural reproduction and stocking to recruitment. Oxytetracycline administered in feed proved to be an effective, low‐cost method of marking spring fingerling Chinook salmon. Vertebrae proved to be suitable tissue for OTC detection, meeting standards for accuracy and reproducibility. Vertebrae were not suitable as aging structures, perhaps due to the slow somatic growth of Lake Huron's Chinook salmon during this study. Based on lakewide recoveries from recreational and commercial fisheries, we estimated that nearly 80% of Chinook salmon from the marked year‐classes were from natural reproduction. Wild‐origin Chinook salmon constituted 96% of open‐water samples from Georgian Bay; 82, 80, and 68% of those from the central, southern, and northern portions of the main basin; and 64% of those from the North Channel. The percentage of wild‐origin Chinook salmon in some tributaries in the Ontario watershed of Lake Huron approached 100%. There was less variation in wild recruitment between years than between lake areas. This Chinook salmon population was introduced in the late 1960s and until recently had been widely viewed as hatchery dependent; our findings suggest otherwise. A self‐sustaining predator population would represent a paradigm shift from an era when predation rates could be manipulated by adjusting salmonid stocking (i.e., 1968 to about 1995). These findings should help managers to critically review the effects of stocked Chinook salmon on Lake Huron's fisheries and ecosystem, including prey consumption demand, and to assess future needs for propagated Chinook salmon in Lake Huron.