A discussion is given of the relative merits of the D-D, D-T,or D-3He mirror fusion power systems. As a basis for discussion we have used a simple approach for estimating the ratio of Q(D, He) to Q(D, T). When needed, the value of Q(D, T) is considered to be of the order of that calculated by Marx and by Kuo-Petravic, Petravic and Watson. The systems are compared at different assumed plasma temperatures in the range 200 to 1000 keV on the basis of several criteria: the investment costs in dollars per kilowatt, the ratio of waste power to power output, and the tritium and neutron production. We have concluded, based on the estimates of the ratio of Q-values, that subject to these criteria the D-T system appears the more favoured system. On the other hand, should a means be found to enhance the Q-values the D-3He system would be preferred on all but the first criteria. We consider the three-component D-T-3He system and its optimization. Finally, the possibility of further enhancing the D-T system by utilization of tritium breeding in the plasma is discussed.