Completely Automated Public Turing tests to tell Computers and Humans Apart (captcha) are challenge-response tests used on the web to distinguish human users from automated bots. Mobile devices such as smartphones and tablets have become a primary means of accessing online resources for many users, however most existing captchas do not properly fit mobile devices and may lead users to abandon tasks. Captchas have become sufficiently hard for users to solve that some web sites refrain deploying them and others are actively looking at alternatives. For users of smartphones, the reduced screen size can lead to typing mistakes and loss of position. In addition, environmental context and device orientation also have an impact on the user experience. In this thesis, our research revolves around three primary, inter-related questions: How can we effectively assess usability issues of captchas accessed on smartphones? What are the most prevalent usability issues of captchas accessed on smartphones? How can we improve captchas for smartphone usage? We conducted lab and heuris- tic evaluations on existing and prototype captcha schemes, and identified areas for improvement. We developed, refined and tested a set of domain specific heuristics to evaluate captcha schemes on smartphones. We designed and tested four captcha prototypes to assess the viability of different input methods. From the empirical work, we identified design strategies for the development of new captcha schemes for smartphones.