Genes encoding polyadenylated mRNAs depend on their poly(A) signals for termination of transcription. An unsolved problem is how the poly(A) signal triggers the polymerase to terminate. A popular model is that this occurs during extrusion of the poly(A) signal, at which time it interacts with factors on the transcription complex. To test this idea we used cis-antisense inhibition in vivo to probe the temporal relationship between poly(A) signal extrusion and the commitment of the polymerase to terminate. Our rationale was to inactivate the poly(A) signal at increasing times post-extrusion to determine the point beyond which it is no longer required for termination. We found that communication with the polymerase is not temporally restricted to the time of poly(A) signal extrusion, but is ongoing and perhaps random. Some polymerases terminate almost immediately. Others have yet to receive their termination instructions from the poly(A) signal even 500 bp downstream, as indicated by the ability of an antisense at this distance to block termination. Thus, the poly(A) signal can functionally interact with the polymerase at considerable distances down the template. This is consistent with the emerging picture of a processing apparatus that assembles and matures while riding with the polymerase.