you are able to read, even though ordinary reading requires you to integrate words that are widely sperated on the page. How should we explain this fact? The obvious suggestion is that you’re relying on some form of memory. Working memory is the memory you use for information that you are actively working on. Working memory holds information in an easily accessible form, so that the information is, so to speak, at your fintertips, instantly available when you need it. How can we ensure that working memory’s content will be instantly available, always ready when you need them? In part, the answer lies in working memory’s size: working memory is hypothesized to have only a small capacity, and so, with only a few items held in this store, there will never be a problem locating just the item you want. Can we test this proposal? One way to measure working memory’s capacity is via a span test. In this test, we read someone a list of, say, four items. This person has to report these back, immediately, in sequence. People start making errors whit sequences of seven or eight letters. This not only confirms our hypothesis that working memory is limited in size, but, just as important, also provides a simple example of how we can learn about this memory’s properties by seeing how this memory influences observable performance.