This book investigates the structure, ontogeny, and evolution of the Old English poem Beowulf through the use of ?lexomic? methods of computer-assisted statistical analysis. Our results are,This book investigates the structure, ontogeny, and evolution of the Old English poem Beowulf through the use of ?lexomic? methods of computer-assisted statistical analysis. Our results are, to date, the highest profile success of these methods, which have produced the first truly new evidence about the text of Beowulf in nearly a century. The cluster analyses employed in this investigation have identified patterns of vocabulary distribution that are perhaps otherwise too hidden, distributed, or subtle to be detected by the unaided eye and mind, but which allow us to calculate and represent visually the relationships of similarity and difference among different segments of the poem. The information thus gleaned is then integrated with that gained from traditional philological and literary techniques, and the resulting synthesis significantly improves our understanding of Beowulf. Among the most striking results are the distinctive difference in vocabulary between the Unferth episode and the rest of the poem, and the surprising similarity between the beginning of the poem and the Finnsburg episode. Even beyond its effect on Beowulf studies, the investigation is valuable in its demonstration of the ways that methods of hierarchical agglomerative clustering can be used to analyze a widely studied, canonical literary text that is at the foundation of the English canon.