Grain sorghum (Sorghum bicolor L. moench) stands as a globally significant cereal crop but the adversity of pre-harvest sprouting (PHS) caused by reduced grain dormancy and moist conditions prior to harvest remains unsolved. Here, we identified a dormancy QTL using a Redlan×IS9530 RIL population, where parent lines are low in tannins and early flowering but otherwise contrasting in grain dormancy and plant height. We phenotyped this population in 2 years with informative PHS-related traits (grain germination index, embryo sensitivity to abscisic acid and in one year the actual natural sprouting), revealing a robust dormancy QTL in chromosome 9 (qDOR-9). This signal overlapped with associations found for plant height (caused by the dw1 locus, used for decades in sorghum improvement) and time to flowering. The effect of qDOR-9 was validated with independent near isogenic lines carrying the IS9530 "dormant" allele while maintaining the Redlan dw1 "short" allele. Additional analyses on Yellow Milo, from which the dw1 allele originated, implied that a low dormancy allele close to dw1 was introduced to Redlan-as well as to many other currently productive lines-by breeding efforts aimed at decreasing plant height, thus illustrating a new instance of genome erosion canalised by crop breeding. However, the introgression of qDOR-9 could enhance PHS tolerance in cultivated dw1-carrying backgrounds without affecting plant stature.