A large body of work has developed over the last decade examining the relation between the approximate number system (ANS) and mathematical performance across a wide range of ages, but particularly for preschool-age children. Largely, the evidence is mixed and suggests that a small relation exists that is dependent on a number of child-related or measurement-related factors. In contrast, little work has focused on understanding the stability and predictors of the ANS. These issues were examined by assessing 113 preschool children in the fall and spring of the preschool year on mathematical and cognitive assessments. Mixed-effect regressions indicated fall ANS performance was the strong predictor of spring ANS performance, suggesting moderate stability of this variable during preschool. However, cardinality and response inhibition were also significant predictors, and school-level variance was high. These findings indicate that the ANS may not be as foundational for mathematics development as previously suggested. (author abstract)