Abstract : Five experiments are described that study the relationship between measures of working memory and reading comprehension. Two experiments investigated whether the complex span measure must be similar to the reading comprehension task to be predictive of comprehension. The correlation found between reading comprehension and two reading-related complex spans was similar to those found between two arithmetic-related complex spans and comprehension. The relationship remained significant when quantitative skills were factored out. The simple digit and word spans (measured without a background task) did NOT correlate with reading comprehension. The complex span/comprehension correlations were a function of the difficulty of the background task. When the difficulty level of the reading-related or arithmetic-related background tasks was moderate, the span/comprehension correlations were higher in magnitude than when the background tasks were simple or very difficult. The third experiment showed that if serial recall was required in the span tasks, simple word span did significantly predict reading comprehension but not as well as the sentence span. The fourth experiment showed that the ordering of list lengths in the span tasks had little influence on the correlation between span scores and comprehension. The fifth experiment is the first in a series investigating variables whether variables that influence simple word span also influence the sentence word span. This study demonstrated that the word length has the same effect on the sentence span task as on the simple word span.