There is a well-known relation between spatial ability and mathematics dating back to the work of early twentieth century factor analysts. This connection is a ripe opportunity for educators, who might use spatial training to improve math learning. However, a closer look at the literature reveals gaps that impede direct application. The primary problem is that although this relation is well established in older children and adults, its emergence in early development and subsequent developmental interactions are not well documented. Moreover, there is a need for more mechanistic explanations that might be leveraged to improve math education. In this chapter, we attempt to address these issues by reviewing the existing literature to identify instances where answers are available and others where further research is needed.