Perfetti's representational quality hypothesis suggests that orthographic learning in developing readers is associated with two important changes to the orthographic lexicon that drive continued word reading development, namely, addition of word-specific representations and acquisition of complex context-dependent decoding skills. Perfetti further hypothesizes that these two changes are mutually facilitative suggesting a bidirectional relationship. To test this hypothesis, cross-lagged panel models were used to model the relationship between word reading and nonword reading longitudinally across Grades 1-4 in a diverse sample (N = 433) of developing readers. Overall results revealed the codevelopmental pattern between word and nonword reading to be bidirectional in nature such that word reading impacted later nonword reading and vice versa. These bidirectional relations did not differ significantly between those who qualify for free and reduced lunch and those who do not. However, differences were observed between those at-risk for dyslexia and those not at-risk.