This study explores the effects of counting objects and reciting numbers in preschool on 1st-grade math performance. Data on 3,125 low-income preschoolers' (4-to 5-year-olds) counting abilities (orally reciting numerals in chronological order and counting blocks) were collected during a fall session (between September 15 and December 15) and were associated with 1st-grade math performance 2 years later. Children's counting abilities were organized into five ordinal categories (from lowest to highest): (1) cannot recite or count to 10, (2) can only recite to 10, (3) can count and recite to 10, (4) can recite (but not count) to 20, and (5) can count and recite to 20. Results of hierarchical regression analyses, controlling for sex, age, ethnicity, socioeconomic status, and days absent in 1st grade, yielded a significant linear relation across the five counting ability categories. Children who were able to count and recite to 20 during the first half of preschool had the highest math performance in 1st grade. Implications for the importance of these findings for early curricula math standards are discussed.