In recent years behavioral genetic studies have provided conclusive evidence that reading disability and related learning disorders, such as mathematics disability, are due at least in part to heritable factors (DeFries et al. 1987; Alarcón et al. 1997). Although the observed relationship between performance in these areas also may be due substantially to genetic influences (Light and DeFries 1995; Thompson et al. 1991), relatively few studies have examined the genetic and environmental etiology of this covariation in a multivariate framework. In the present study data from 196 identical (monozygotic; MZ) and 155 same-sex fraternal (dizygotic; DZ) twin pairs in which at least one member of each pair evidenced reading problems in school (reading disabled) were subjected to a multivariate behavioral genetic analysis. Structural equation models were fitted to twin data for verbal IQ (VIQ), phonological decoding ability (PHON), reading performance (READ), and mathematics performance (MATH) to assess the extent to which VIQ and PHON mediate the observed covariation between READ and MATH. Results suggest that VIQ and PHON account for most of the covariation between READ and MATH. Moreover, approximately 82% of the observed correlation between READ and MATH was due to genetic factors that also influence VIQ and PHON. When data from 132 MZ and 91 same-sex DZ control twin pairs in which neither twin had a history of reading problems were subjected to the same analyses, the covariation between READ and MATH was found to be due to both genetic and shared environmental influences. Thus genetic factors that influence VIQ and PHON also contribute to the observed covariation between READ and MATH in both a reading-disabled and a control twin sample.