This article reports on an analysis of a small corpus of fiction written for children, extracted from the BNC. Quantitative analyses of most frequent words and sequences of words, and of parts-of-speech, were conducted, and compared with their equivalents in two other sub-corpora of the BNC, of adult fiction and of newspaper texts. The main findings point to some characteristics of both the fiction corpora which are very similar, and which contrast markedly with the news texts. However, more nuanced comparison of concordance lines in which the frequent items occur reveal subtle but telling differences between their use in context in adult fiction and in fiction written for children.