On the basis of a substantial, robust, and well-replicated cognitive science research literature, interleaving has been recommended across a range of professional pedagogical works as an evidence-based strategy for mathematics teachers. This report first notes the conflation of interleaving and spacing effects underpinning those recommendations and the limitations of some laboratory-based research as grounding for practice, particularly given that no direct interleaving research involves mathematical concepts with school students. It then reports on two large classroom experiments aimed at replicating the direct interleaving-versus-blocking effect, alongside a more common teaching approach, for learning well-defined mathematical categories (angle relations). Repeated failure to detect an interleaving-versus-blocking effect raises the question of whether this effect is relevant to learning defined mathematical concepts.