It is currently considered imperative to introduce reading instruction as early as possible. This proposition was tested by assigning groups of kindergarteners (N = 256) to two conditions differing in their emphasis on prevention. In the first, teacher-implemented research-based interventions were implemented during kindergarten and first grade. In the second, only the first-grade intervention was implemented. Analyses were conducted separately for students not at risk, at low risk, and at high risk of reading problems. In the short term, the kindergarten intervention was highly effective for both low-risk and high-risk students, but it is only for the latter group that early gains translated into better reading skills at the end of first grade. Not-at-risk students did not benefit from the early introduction of reading instruction.