"I have heard," said the man, "a large relieving force bound for Ar's Station departed from Ar weeks ago."

"I heard that, too," I said. I knew that it was true. I also knew that Ar, inexplicably, to my mind, had literally invested the bulk of its land power in that very expedition, and had done so with the main forces of Cos not in the north but in the vicinity of Torcadino. This seemed to me a military mistake of almost unbelievable dimension. I had been in Torcadino several weeks ago, indeed, at the very moment when the city, housing Cosian siege engines and supplies, serving as a depot and staging area for the eastward advance of Cos, had, in a daring stratagem, been seized by Dietrich of Tarnburg with no more than a few thousand mercenaries. These had entered the city through aquaducts, literally over the heads of unsuspecting Cosian armies camped about the city. This act had stalled the invasion. I expected Dietrich to be able to hold Torcadino through the winter, but little longer. I had borne letters from Dietrich to Ar germane to these matters.

In the intrigues of the time, and to divert suspicion, Gnieus Lelius, high councilor, and first minister of Ar, he who was acting as regent in the absence of Marlenus, Ubar of the city, had even had me brought to the Central Cylinder under guard, as though I might have been arrested, and was to be examined on some charge. There, personally and at length, I had spoken to him. I had urged him to march to Torcadino and confront the main body of Cosian forces. But the troops of Ar had not been recalled, nor diverted to Torcadino. They had continued to march northward, as though the major danger lay at Ar's Station. This, in effect, seemed to negate the bold stroke of Dietrich, to slow the Cosian advance, and give Ar time to organize, to arm and march. Ar had not moved against the Cosians at Torcadino. She had marched north, presumably to relieve Ar's Station. Gnieus Lelius had listened to me thoughtfully and patiently. But he would, it seemed, trust to the judgment of his officers.

I had then been kept in Ar for weeks, a guest in the Central Cylinder, waiting and waiting. Then at last I had been given a sealed letter for the commander of Ar's Station, whose name was Aemilianus. That was all. That very night, on tarnback, I had streaked northward from Ar. I had sold the tarn only two days ago, to proceed on foot. The skies had seemed heavily patrolled. I had little doubt they would become more so as I proceeded farther northward. It seemed to me that my chances of successfully delivering the message to Aemilianus, whatever might be its contents, might be improved if it were borne not by tarnsman but by one afoot, one who might, say, among mercenaries, or civilians, mix inconspicuously. This speculation was further encouraged by the fact that Ar's station would surely have its tarn wire strung and the skies about it, as nearly as I had determined, were currently controlled by Cos. "But," said the man, "such a force has not passed this point." "I do not know its location," I said. I had stayed at certain inns in the south, past which it had taken its march, taking five days to pass given points. Then, moving northward, I had stayed at inns, also on, or near, the Vitkel Aria somewhere north of Venna.

"It cannot have just disappeared," he said.

"It is a mystery to us," I said, "but doubtless to those with access to the proper intelligence network, its movements and position are well known." I had encountered refugees from Ar's Station and its environs even south of Venna. Some told me they had seen the army pass. Some had even told me that men and women they knew had followed the army northward, as though confident of its victory and returning to their homes. What puzzled me most was that the Viktel Aria was the most direct route, for hundreds of pasangs, to Ar's Station. Indeed, Ar's Station, in effect, secured the northern terminus of the Viktel Aria, or Vosk Road, at the Vosk.

The Viktel Aria was a military toad, one laid out by military engineers as a military route. It sped almost directly from Ar to the Vosk. It made few concessions to towns or communities. Its primary purpose was to provide a reliable, nearly indestructable surface for the rapid movement of armed men. this being the case, however, why had the army of Ar not kept to it, on its presumed journey to raise the siege of Ar's Station? The most likely hypothesis seemed to me to be that it was making its way not to Ar's Station but to Brundisium, where, months ago, the Cosians had landed. This suggested that either Ar's Station was to be sacrificed in these harsh games, or that it was the thinking of Ar's commanders that a move to Brundisium would lift the siege of Ar's Station, the Cosians there perhaps then being withdrawn to protect Brundisium. Such a move, of course, might isolate the Cosian main forces, both depriving them from their fellows at Ar's Station. I did not doubt, incidentally, that the military might which Ar now had in the north, if it were what it was said to be, would be sufficient to take Brundisium. The objections to this strategy, of course, were obvious. Ar's bastion on the Vosk, Ar's Station, was being treated as expendable, which it was not, if Ar wished to maintain its power in the Vosk Basin. Even if Brundisium should fall, this would not be likely to keep open her lines of communication and supply. Similarly, Ar, lacking a sizable navy, had no way to follow up the capture of Brundisium, either by interdicting the coast or attempting an invasion of Cos.

The major objection, of course, was that this move exposed Ar herself to the main force of Cosians, which was in the vicinity of Torcadino. It was almost as though the officers of Ar were content to exchange Ar for a port, and one which, strictly, was not even a Cosian port. If this were the case, however, that Ar was advancing on Brundisium, I had, interestingly enough, heard nothing of it. By now, in the normal course of events, given Ar's start, and the typical marches of armies, she would have had time to reach not only Ar's Station but even Brundisium, much farther away.

I did not know where the main force of Ar was. In this sense I was confronted with a mystery, at least as far as my own limited information went. Perhaps, for some reason, the forces of Ar were intending to relieve Ar's Station from the west, thus interposing themselves between the siege forces of Cos and their likely routes of escape, either substantially west by southwest to Brundisium or more to the southwest, toward Torcadino. If this were the case, however, it seemed that we should, by nor, have heard something to this effect. Indeed, if this were true, it seems that Ar, by now, should have appeared on the western flank of the Cosians.

"I fear for Ar's Station," said the porter.

"How is that?" I asked.

"I do not think she can long hold out," he said. "The attackers are numerous. The defenders are thinned. The walls are weakened. New breaches are made daily. In places they are being mined. Fires have occurred in the city, from saboteurs, from fire javelins, from flame baskets catapulted over the walls. There is starvation in the city. If the forces of Ar do not soon raise the siege, I think she must succumb."

"I see," I said. "Too," said he, "the fighting, in which civilians have participated, has been lengthy and bitter. The men of Cos expected an easier time of it. Their losses have been heavy. They will not be pleased."

I nodded.

"I would not care to be there when the gate gives way," he said.

"It is late," I said.

He then opened the door in the interior gate. "The keeper's desk, and the paga room," said he, "are in the building to the right."

I looked out through the door, into the court of the inn. I was soaked to the skin. It was still raining heavily. It was dry, at least, in the covered, shedlike entrance way, between the gates. The inn itself, aside from certain ancillary buildings, was built of heavy logs, and in two parts, or structures, with a common, peaked roof, and an open space, covered from above by the roof, between the two parts. Each part, or structure, contained perhaps three or four floors, possibly joined by ladders. It was about a hundred feet between the door in the interior gateway, where I stood, and, to the right, the covered way between the separate parts of the inn. The flooring of the court was formed largely, leveled and carved, from the natural stone of the plateau. Narrow drainage channels had been cut in it. Through these water now flowed under the palisade, down the moat. It also flowed, doubtless by design, midway here and there, between the palisade's anchor post wells and bracing recesses, cut in the stone, sealed about with tar. Water was running from the long roof of the two-part structure, perhaps two hundred feet in length, falling some thirty or forty feet down to the court.