The distribution of the light-variation amplitudes, A(a), in addition to determining the number of undiscovered contact binary systems falling below photometric detection thresholds and thus lost to statistics, can serve as a tool in determination of the mass-ratio distribution, Q(q), which is very important for understanding of the evolution of contact binaries. Calculations of the expected A(a) show that it tends to converge to a mass-ratio dependent constant value for a->0. Strong dependence of A(a) on Q(q) can be used to determine the latter distribution, but the technique is limited by the presence of unresolved visual companions and by blending in crowded areas of the sky. The bright-star sample to 7.5 magnitude is too small for an application of the technique while the the Baade's Window sample from the OGLE project may suffer stronger blending; thus the present results are preliminary and illustrative only. Estimates based on the Baade's Window data from the OGLE project, for amplitudes a>0.3 mag. where the statistics appear to be complete allowing determination of Q(q) over 0.120. The mass-ratio distribution can be approximated by a power law, either Q(q)~(1-q)^a1 with a1=6+/-2 or Q(q)~q^b1, with b1=-2+/-0.5, with a slight preference for the former form. Both forms must be modified by the theoretically expected cut-off caused by a tidal instability at about q_min 0.07-0.1. An expected maximum in Q(q), is expected to be mapped into a local maximum in A(a) around 0.2-0.25 mag.

AASTeX5, 12 figures, 5 tables, accepted by AJ, Aug.2001