The response of strongly nonlinear dynamic systems to stochastic excitation exhibits many interesting characteristics. In this paper, excitation forms like white noise and Gaussian band limited noise are applied to a bilinear system. The nonlinear characteristics cannot be represented sufficiently accurate using statistical linearization techniques. Integration techniques have to be used. The emphasis lies on frequency domain phenomena. Phenomena like multiple resonance frequencies and multiple solutions are investigated along with their origin. The stochastic response characteristics are clarified by comparing them to periodic response features of the system. Furthermore, stochastic response phenomena can provide information on the periodic response characteristics of the system.