As a spatio-temporal data-management problem, taxi ridesharing has received a lot of attention recently in the database literature. The broader scientific community, and the commercial world have also addressed the issue through services such as UberPool and Lyftline. The issues addressed have been efficient matching of passengers and taxis, fares, and savings from ridesharing. However, ridesharing fairness has not been addressed so far. Ridesharing fairness is a new problem that we formally define in this paper. We also propose a method of combining the benefits of fair and optimal ridesharing, and of efficiently executing fair and optimal ridesharing queries.