The etiology of the observed relationship between reading and mathematics performance was examined by analyzing data from samples of same-sex twin pairs tested in the Colorado Learning Disabilities Research Center. Bivariate phenotypic and genetic structural equation models were fitted to data from 526 twin pairs selected for reading deficits (290 identical and 236 same-sex fraternal) and 355 control pairs (220 identical and 135 same-sex fraternal). Subtests of the Peabody Individual Achievement Test (PIAT; Reading Recognition, Reading Comprehension, and Spelling) were used as measures of reading performance, and scores from the Wechsler Intelligence Scale for Children-Revised (WISC-R) or Wechsler Adult Intelligence Scale-Revised (WAIS-R) Arithmetic subtest, the Wide Range Achievement Test Arithmetic subtest, and the PIAT Math subtest were used as indices for mathematics performance. The results of these confirmatory factor analyses indicate that genetic and environmental covariances between reading and math latent factors do not differ significantly for twin pairs in the proband and control groups. Estimates of heritability for reading performance in the proband and control samples were 0.81 and 0.69, respectively, and those for math performance were 0.88 and 0.67, respectively. Moreover, genetic influences accounted for 83% of the covariation between the reading and math factors in the proband group and for 58% of the covariation between these two latent variables in the control group; in contrast, shared environmental influences did not contribute significantly to the relationship between the reading and math latent factors nor to their independent variation.