In 2 experiments, the authors examined whether spellings improve students' memory for pronunciations and meanings of new vocabulary words. Lower socioeconomic status minority 2nd graders (M = 7 years 7 months; n = 20) and 5th graders (M = 10 years 11 months; n = 32) were taught 2 sets of unfamiliar nouns and their meanings over several learning trials. The words were defined, depicted, and embedded in sentences. During study periods, students were shown written forms of 1 set but not the other set. Spellings were not present during word recall. Results of analyses of variance showed that spellings enhanced memory for pronunciations and meanings compared to no spellings (ps less than 0.01). Better readers and spellers increasingly outdistanced poorer readers and spellers in remembering pronunciations over trials when spellings accompanied learning (p less than 0.05), suggesting a Matthew effect. An explanation is that spellings activated graphophonemic connections to better secure pronunciations and meanings in memory. Results indicate that orthographic knowledge benefited vocabulary learning and diminished dependence on phonological memory. Instructional implications are that teachers should include written words as part of vocabulary instruction and that students should pronounce spellings as well as determine meanings when they encounter new vocabulary words.