Despite extensive research on phonological awareness and reading, there has been little effort to study practical questions that would assist practitioners regarding the choice and interpretation of the phonological awareness tests available to them. This study examined the relationship between decoding (real and pseudowords) and three phonological awareness tests (segmentation, blending, and manipulation) taken from the "Comprehensive Test of Phonological Processing (CTOPP)" with an unselected population of first grade (n = 67) and second grade (n = 49) students. Segmentation displayed the weakest correlation with reading and accounted for no statistical variance in reading beyond what was found in the blending test. It also failed to account for a substantial amount of variance in reading that is captured by the manipulation test. Despite its popularity in educational contexts, phonological segmentation may be less useful than phonological manipulation or blending in assessing the phonological substrates of reading at these grade levels. (Contains 3 tables, 2 figures, and 1 note.)