Though confined within the walls of his capital, Croesus had still good reason for hoping to hold out until the arrival of his allies, to whom he sent pressing envoys of acceleration. For Sardis was considered impregnable--and one assault had already been repulsed, and the Persians would have been reduced to the slow process of blockade. But on the fourteenth day of the siege, accident did for the besiegers that which they could not have accomplished either by skill or force. Sardis was situated on an outlying peak of the northern side of Tmolus; it was well fortified everywhere except toward the mountain; and on that side the rock was so precipitous and inaccessible, that fortifications were thought unnecessary, nor did the inhabitants believe assault to be possible in that quarter. But Hyroeades, a Persian soldier, having accidentally seen one of the garrison descending this precipi tous rock to pick up his helmet which had rolled down, watched his opportunity, tried to climb up, and found it not impracticable; others followed his example, the stronghold was thus seized first, and the whole city speedily taken by storm. Cyrus had given especial orders to spare the life of Croesus, who was accordingly made prisoner. But preparations were made for a solemn and terrible spectacle; the captive king was destined to be burned in chains, together with fourteen Lydian youths, on a vast pile of wood. We are even told that the pile was already kindled and the victim beyond the reach of human aid, when Apollo sent a miraculous rain to preserve him. As to the general fact of supernatural interposition, in one way or another, Herodotus and Ctesias both agree, though they described differently the particular miracles wrought. It is certain that Croesus, after some time, was released and well treated by his conqueror, and lived to become the confidential adviser of the latter as well as of his son Cambyses: Ctesias also acquaints us that a considerable town and territory near Ekbatana, called Barene, was assigned to him, according to a practice which we shall find not infrequent with the Persian kings.