From the perspective of the dominated splitting, under certain assumptions, there existed finitely many closed sets $A_1,A_2,\dots ,A_m$ (with $A_i\ne A_j$ for $i\ne j$) such that the support of any ergodic geometric measure coincided with one of them.