In terms of EPR-chameleon models and local and causal measures, the Bell's argument is reanalyzed. Contrary to Bell, it is shown that the nontriviality of the joint probability measure does not always imply the nonlocality. It is analyzed that under what conditions correlations of distant particles are obtained which are different from the standard correlations. The protocol for the correlations of distant particles admits nontrivial probability measures respecting the locality