The paper develops an algorithm by which the best input equations of a T-G flip-flop can be determined. The algorithm consists of rules for plotting the 1, 0 and “don't-care” terms from the map of the application equation onto the maps of input equations. It turns out that the 0's of the Q¯ half can be plotted in two different ways. The paper shows how this property can be judiciously exploited to produce input equations with the least-cost function. A systematic procedure to achieve this has also been described.