Abstract: Adding multi-touch sensing to the surface of a mouse has the potential to substantially increase the number of interactions available to the user. However, harnessing this increased bandwidth is challenging, since the user must perform multi-touch interactions while holding the device and using it as a regular mouse. In this paper we describe the design challenges and formalize the design space of multi-touch mice interactions. From our design space categories we synthesize four interaction models which enable the use of both multi-touch and mouse interactions on the same device. We describe the results of a controlled user experiment evaluating the performance of these models in a 2D spatial manipulation task typical of touch-based interfaces and compare them to interacting directly on a multi-touch screen and with a regular mouse. We observed that our multi-touch mouse interactions were overall slower than the chosen baselines; however, techniques providing a single focus of interaction and explicit touch activation yielded better performance and higher preferences from our participants. Our results expose the difficulties in designing multi-touch mice interactions and define the problem space for future research in making these devices effective.