Cooperative spectrum sensing improves reliability and the detection performance of sensing. However, the fully cooperative scenario may not be realistic to assume in many cases. We consider a cognitive radio network with heavy traffic users and light traffic users and analyze their behavior towards sensing using the concept of mixed strategy Nash equilibrium. Further, we design a distributed game using evolutionary game theory such that as long as light traffic users are present, they share the sensing responsibility while heavy traffic users get a free ride. Finally, the evolution dynamics is shown to converge to the evolutionarily stable strategy for the case of multiple users.