The more ancient statements of the Greeks carry us much further back than the reasoning of Trogus Pompeius and his successors. If we set Pliny aside, who asserts "that the kind of Magism established by Zoroaster was many years older than that taught by Moses," Hermippus of Smyrna puts Zoroaster 5000 years before the Trojan war. Even before Hermippus, Theopompus of Chios, and Hermodorus, the pupil of Plato, had ascribed the same date to him. Eudoxus of Cnidus, the contemporary of Plato, placed him still higher; he thought that Zoroaster lived 6000 years before the death of Plato. According to Pliny, Aristotle ascribed to him the same antiquity, and, as we learn from Diogenes Laertius, maintained that the Magians were older than the Egyptians. And even in the fifth century B.C., Xanthus the Lydian is said to have written that from the time when Zoroaster lived to the march of Xerxes against Hellas a period of 6000 years had elapsed.[143] Through these statements there runs, beyond all doubt, a system, the knowledge of which began in the fifth century B.C. among the Greeks and continued beyond the time of Alexander. Whether we take 5000 years before the Trojan war, or 6000 years before Plato's death, we are equally brought back into the seventh millennium B.C. If the later statements of the West, which make Zoroaster a contemporary of Ninus and Semiramis, are the results of combining the most prominent name in Bactria with the conquest of Bactria by the founder of the Assyrian power, as related in the Medo-Persian Epos, the fixing of Zoroaster's date so many thousand years previously must have been taken by the Greeks from the Persians. In these dates we seem to be dealing with certain cyclic periods. We learn from Theopompus of Chios, that according to the doctrine of the Magians, one of the two gods Oromazdes and Areimanius would reign and the other be subject for 3000 years; for another 3000 years they would be in conflict, and one destroy the works of the other, until at length Areimanius would succumb and men become happy.[144] From this we may with certainty conclude that periods of 3000 years were in use among the priests of Iran to denote certain spaces of time, and that these cycles form the base of the statements of the older Greeks, if we can prove the use of such periods in the Avesta or in the books of the Parsees.