Dorazus went away, taking his papers and his diagrams with him, though he left the exquisitely detailed scale model behind. Psellus sat staring at it for several minutes after he'd gone; then, quite tentatively, he reached out his hand and scratched gently at the edge of the miniature embankment. A few flakes of green-painted plaster fell away, leaving a white scar.
He sighed, and opened the book. He'd marked the page with a scrap of paper torn off the corner of some report.
To cross a flooded ditch, he reminded himself, the usual method is to construct a siege mound directly overlooking the ditch at its narrowest point, or, where more expedient, at the point at which it is most advantageous to cross. Sappers then undermine the base of the mound, causing it to subside into the ditch, filling it. Where necessary, firm standing can be provided to cross the filled section by nailing planks to long ropes, which are then rolled up. A special machine (see Appendix Twenty-Six) is used to roll them out again.
A siege mound. He did some calculations, using brass counters on a chequerboard. No, they wouldn't have time for that; and besides, if that was what they had in mind, why hadn't they started it already? Instead, as far as he could make out from the reports, they were deepening the wide, broad trenches, which suggested an attempt to drain the ditch and carry the water away into the slight dip at the base of the ridge. That, however, would be impossible; it would take the sappers far too long, they'd be vulnerable to mangonel and scorpion fire, and even if they succeeded, they'd be swept away and drowned when the water broke through; surely even the savages weren't fanatical enough for that. So, Dorazus and his subcommittee were quite right. There was no way it could be done. If he'd been at the meeting, he'd have had no choice but to vote in support of the motion.
He picked up his pen, dipped it in ink and wrote on a piece of scrap paper:
Allow three days for breaching/crossing/filling in the flooded ditch.
Because he knew they'd do it, somehow; and then they'd face the embankment itself; which they'd sap and undermine. Allow days for sapping the embankment. He frowned at the space he'd left blank, then turned back through the pages of the book.
There wouldn't be time. Surely they understood that. Undoubtedly they had some plan for coping with the flooded ditch; their preparations told him that, even though he couldn't figure out what they were planning to do. But once they were past the ditch, they had to get through, or over, or under the embankment before they could get at the walls, and there was no way they could achieve that in the time available to them…
A thought occurred to him, and he made way for it politely. Reports said that the allies were plundering farms in the Cure Doce country, on the pretext of punishing them for the attack on Duke Valens. Wagons loaded with food and hay had been seen coming down the border road. Would these extra supplies extend the time available to them to any significant degree? His fingers paddled the brass counters across the chequerboard, and he frowned. Answer: no. The simple fact remained that the allied army presently camped on and behind the ridge consumed in a single week more than the three adjoining Cure Doce counties were capable of producing in a year.
Assault, he said to himself; a straightforward, brutal attack, with scaling ladders, siege towers, similar primitive equipment. Suppose they sent a hundred thousand men (he couldn't visualise that many people all in one place) to carry the embankment by storm. He found the place in the book where the author set out a clearly tabulated ready-reckoner. To carry a defended position by storm. (His finger traced along the line.) Light or heavy defences; well, say light, for the sake of argument. That gave him a multiplication factor of forty, reduced by thirty per cent (untrained or poorly trained defending army). He did the calculation. No, not possible. According to the tables, twenty thousand defenders entrenched on the embankment, poorly trained but equipped with short-and medium-range artillery, should be able to resist an assault by up to two hundred and thirty thousand attackers; projected casualties for an army of a hundred thousand, assuming the assault wasn't abandoned until the losses reached the point of critical perceived failure (there was a complicated equation to find this, but as a basic rule of thumb, say fifteen per cent losses in eight hours or less), between twenty and twenty-six thousand for the attackers, no more than fifteen hundred for the defence…
Psellus closed his eyes. He didn't for one moment doubt the accuracy of the tables, but how on earth did the book's author know these things? It could only be that, at some time in the past, so long ago that nobody remembered them any more, there had been sieges of great cities; so frequent and so commonplace that scholarly investigators had been able to collate the data-troop numbers, casualty figures-and work out these ratios, qualified by variables, verified by controls. The cities, the men who'd lived in them, had been forgotten for so long that nobody even suspected they'd ever existed. The only mark they'd left behind was the implication of their existence, to be inferred from the statistical analyses in a manual of best city-killing practice. Extraordinary thought. There were people who held that certain kinds of stone weren't stone at all, but the compressed bones of innumerable billions of fish, crushed into solid blocks by the weight of the sea. He didn't actually believe that; but suppose the book was the only residue left by the death of thousands of cities, each one of them as huge and arrogant in its day as the Perpetual Republic-the Eternal City of this, the Everlasting Kingdom of that, squashed down by time and oblivion into a set of mathematical constants for predicting the deaths of men in battle.
So what, he thought; all that told him was that the real enemy wasn't General Daurenja or the Aram Chantat, but war itself, a truth so profound as to be completely useless. Even so, something had snagged in his mind, like a bramble on a sleeve, and he wondered what it could be. Curious, he glanced over the tables once more, until suddenly he saw it, and was immediately paralysed by its implications.
Light and medium artillery; and in the earlier chapters of the book there were detailed descriptions of each class of engine-the light field engines, such as scorpions, springalls and torsion rock-throwers; the medium engines, mangonels, onagers, the heavy springalls and lithobales; the heavy engines, such as the trebuchet. When he'd first read it, he'd been pleased and impressed-the engines in the book are just like the ones we use now, he'd told himself, so the data in the book is still useful and relevant. Now, it stunned him that he could have missed such a devastating point so completely; because the engines described in the book weren't just similar to the Guilds' approved types. Apart from a few inconsequential details, they were the same. But the book wasn't Mezentine; it had been translated by a Mezentine, two hundred years ago, from a very old manuscript, written by some foreigner belonging to a city and a race that had completely disappeared. In which case, the designs, the specifications, were hundreds of years old, quite possibly thousands; in which case, the Guilds hadn't created them, they'd simply copied them from somewhere else. True, there was a special dispensation for military equipment; but that wasn't enough to prop up the gaping sap that suddenly threatened to undermine his entire world. The Guilds hadn't created these specifications; they were the work of foreigners, savages, who'd achieved perfection at some point in the obscure past, long before the Mezentines had even left the Old Country. In which case…
He was having trouble breathing. In which case, we aren't the authors of Specification. We're just thieves, like Ziani Vaatzes, who stole designs from our betters, which is the greatest sin. In which case…