This study involved the close observation of the reading of 59 adults in literacy tuition and 25 twelve-year old children in remedial classes, and the recording of their oral reading errors and ability in comprehension, short term memory, word reading and definitions measures. Fundamental differences were found in the structures of the two samples and their approach to the reading task. Differences between subjects in the adult sample stemmed from the presence or not of semantic acceptability among errors coupled with ability in areas like comprehension, definitions skills and metalanguage. The measures used by which the adults were grouped by the Literacy Scheme were found to be inadequate and alternative adult groupings were discovered in the corpus of the data. Suggestions were made for incorporating meaning related tasks into the teaching of reading to adults.