In this longitudinal study, Norwegian children received daily training sessions in word analysis with the aim of stimulating the children to discover and attend to the phonological structure of language. Ten first-grade classes received one of three treatments during the fall semester. In the three classes (60 children) assigned to the phoneme isolation or positional treatment, teachers taught the children to attend to individual phonemes and identify those phonemes in initial, medial, or final position in target words. In the three classes (52 children) assigned to the phoneme segmentation or sequential treatment, teachers taught the children to identify the phonemes in a word in the correct sequence and to blend them. Finally, four classes (100 children) were assigned to a control treatment in which they looked at illustrations and discussed them. Metaphonological skills were measured at various points during the training period, and long-term effects of the training on the pupils’ progress in reading and spelling were assessed at the end of Grades 1 and 2. Both forms of phonological training had a facilitating effect on reading and spelling. At the end of Grade 1, students who had received the sequential phoneme segmentation training scored significantly higher in spelling than students who had received the positional (phoneme isolation) training; however, this difference disappeared by the end of Grade 2. A significant interaction between type of treatment and intelligence suggested that students of lower ability profited the most from the phonological training.