Hitherto we have been treating the case of a _judicium populi_ consequent on the magistrate’s recognising the limitations on his power. But there is a possibility of his refusing this recognition, and in this case the matter can be brought to the people only by means of an appeal (_provocatio_) lodged by the accused. This contingency was, in the middle and later Republic, unusual but not unknown, for the jurisdiction of the _duumviri perduellionis_ was, as we know from the case of Rabirius,[1217] regulated at times in such a manner that an appeal to the people was an essential part of the procedure. In such a case there were two magisterial investigations instead of one. The first was the _quaestio_, as a result of which the magistrate had pronounced the appellable sentence; the second was the _anquisitio_ before the people preceding the decision in the _comitia_. It must occasionally have happened that different magistrates conducted these two stages of procedure; for if an individual appealed against the decision of a magistrate in a province or in the field, this magistrate might himself be unable to conduct the case at Rome. The people is represented from a very early time as rescinding its own sentences.[1218] This rescission was simply the repeal of a law, and was perhaps not regarded originally as the revision of its own sentence by a court. No provision was made that the particular assembly which had pronounced the sentence should repeal it. This was, indeed, sometimes the case. Popilius, for instance, who had been held responsible for the judicial murders following the fate of Ti. Gracchus, was both banished and restored by _plebiscita_.[1219] But, on the other hand, Metellus, “interdicted” by a consular bill, which must have been passed at the _comitia centuriata_,[1220] was restored by the _rogatio_ of a tribune,[1221] while Cicero himself, banished by a tribunician enactment, was recalled from exile by a consular law passed at the _comitia centuriata_.[1222]