Subtraction combinations are particularly challenging for children to learn (Kraner, 1980; Smith, 1921; see Cowan, 2003, for a review). This study examines whether the group of children receiving the "experimental subtraction-as-addition" training outperform the "control" group, which received training on a different reasoning strategy involving 8s or 9s, on both practiced and unpracticed subtraction combinations and (2) a group that receiving "unstructured subtraction practice" on at least the unpracticed subtraction items. A total of 75 first graders (6.1 to 7.6 years old, mean = 6.6) from five schools in two school districts serving a mid-sized mid-western community participated in the study. Regarding the delayed posttest results with the "practiced subtraction" combinations, the structured subtraction-as-addition group, which received supplemental subtraction instruction/practice, significantly outperformed the use-a-ten group, which received regular classroom subtraction instruction/practice. Moreover, as indicated by a small, but appreciable, effect size (Cohen, 1992), the structured subtraction group outgained the unstructured-practice group, which received supplemental subtraction practice but not training on how subtraction is related to addition. The unstructured practice group also significantly out performed the use-ten group. For the "subtraction combinations not practiced" by any group, the structured subtraction out gained the other groups at a marginally significant level and with a medium effect size indicative of instructional effectiveness (IES, 2011). The other groups did not differ on the unpracticed subtraction items. The pattern of these results provides evidence of the efficacy of the structured subtraction intervention. Although either structured "or" unstructured supplemental practice is more effective than typical classroom training in promoting fluency with practiced subtraction items, the structured training provided additional benefit over haphazard supplemental practice with such items. (Contains 3 tables.)