One of the key issues in Case-Based Reasoning (CBR) is the efficient retrieval of cases when the case base is huge. In this paper we propose a case memory organization in two steps: 1) the case memory is organized using an unsupervised clustering technique, and 2) explanations for each cluster are constructed using all the cases associated to each one. The role of the explanations is twofold. On one hand they index the memory and allow CBR to do a selective retrieval. On the other hand, the explanation provide to the user additional information about why the cases have been both clustered together and retrieved.