This paper reports on the second of two studies designed to assess the effectiveness of a metacognitive approach to teaching word identification and reading comprehension skills to upper primary poor readers, and to investigate effective methods for implementing the program in the regular classroom. To improve word identification skills, experimental subjects were given metacognitive training in the analysis and monitoring of word identification strategies. Reciprocal teaching procedures, incorporating the word identification strategies, were used for comprehension training. Subjects in control conditions received either (a) normal classroom reading activities, or (b) normal classroom activities plus phonics-based remedial instruction. To facilitate implementation of the program in the regular classroom a three phase model was used during which responsibility for instruction was gradually passed from the experimenter to the class teacher. Measures of (a) word identification, (b) metacognitive awareness of word identification cues, and (c) comprehension were taken on several occasions during the study. Results of repeated measures analysis of variance showed significantly greater improvements for subjects in the experimental condition. However, most of the improvements took place in the experimenter-led, rather than the teacher-led phase of the Study. The implications of these findings for classroom practice are discussed in the light of current research.