An overarching question guided this study:What is mathematics difficulty (MD) independent of reading difficulty (RD)? The sample included 203 children whom the researchers followed from kindergarten to third grade. The researchers used latent growth modeling to investigate the relationship between MD and measures of working memory, short-term memory, cognitive processing speed, early numerical skills, and phonological processing, independent of reading. Two main findings emerged: (a) deficits in early numerical skills were a defining feature of MD with or without RD, and (b) early numerical skills and phonological processing influenced growth in mathematics from kindergarten to third grade. These results indicate that future MD research should focus on numerical and language-based skills, including whether the mathematics disorder in MD--specific or otherwise--reflects underlying language and/or numerical deficits. (Contains 6 tables and 1 figure.)