We have described ongoing research on applying multiple types of learning within a case-based reasoning context, focusing on how case-based reasoners can learn to apply cases more effectively, both by learning how to adapt prior cases to new situations and by learning similarity by estimating which response plan cases are easiest to adapt. Preliminary trials of the effects of adaptation learning on our system are encouraging for decreasing memory search cost, but more thorough tests are needed, both to study how the process ``scales up'' when large numbers of adaptations are learned and to determine effects on the quality of the response plans generated. Tests are also needed to examine how well current estimates of adaptation cost predict the difficulty of future adaptations. Future work will include refining our model in preparation for more extensive tests of the system as a whole and the effects of its multiple forms of learning.