Considerable progress has been made in the use of corpora for research purposes to describe language in use, and more recently, through a CADS (corpus assisted discourse studies) approach, to identify the discourse features of specific text genres. While the potential benefits of working with corpora in the classroom have been recognised, there has been a lag in the promulgation of guidelines for carrying out meaningful corpus work with language learners and teachers in mind. The papers in this volume aim to make a contribution toward filling that gap by providing an in-depth account of innovative corpus work, most of which has actually been carried out with real learners in the classroom. Authors provide valuable insights into ways of structuring corpus work for specific target learners, as well as suggestions for resolving problematic issues that have arisen and avoiding errors that have been made with learners and in their own research and experimentation. The transparency and honesty with which they present their methodology and results, along with the successful techniques they have developed, constitute a step forward in defining good (and bad) practice in the use of corpora in learning.