The slime fits into both your hands, if you let it spread.
You do not. You hold it cupped above the assay tray while the master turns a shaded lamp over it. The little body gathers itself against your palms: cool, green-brown, heavier than rainwater and lighter than mud. Its skin takes the light without shine. Grit travels through it in slow currents. At one end, a feeding mouth worries the saltless oat paste she has put on the glass.
"Weight?" she asks.
You read the balance arm. "Four and three-quarter marks."
"Again."
You settle your wrists. Outside the assay room, water knocks somewhere inside the reservoir wall, a thin, hurried sound that stops and starts again.
"Four and three-quarter."
The master touches a reed fiber to the slime's flank. It draws in from the irritation, then relaxes and resumes feeding. She places a blue assay slip beneath the drop it leaves. The paper yellows at the expected rate.
"Mouth ordinary," she says. "Acid ordinary. Startle ordinary. No hard inclusion under the lamp. No second sac."
"It made the pattern."
"This one?"
"It was on the wall with the others."
"That wasn't my question."
The sluice warden stands in the doorway with rain stippling the shoulders of her oilskin. Behind her, a bell rings twice from the upper gauge walk: not alarm, only a reading passed down. Two strokes, a pause, one. The reservoir has risen another hand.
"How ordinary?" the warden asks.
The assay-master does not look up. "Entirely, so far."
The wall had seemed ordinary too, until several dozen drain slimes spread themselves over its wet mica render. They had gone opaque along one branch, glass-clear along another, and cold blue at the fork, and you had stood in the gallery watching them.
Then the masonry split on the same fork. The crack was wide enough to take the corner of your gauge card, and it ran from the old side spill toward the gallery floor. By the time you pulled the alarm cord, water had already begun to bead through it.
"They warned us," you say. "The blue ran right along the line where it split."
The warden stays in the doorway. "When did the colors come up? When did you pull the cord?"
"I wasn't sure what I was seeing at first. The light was low under the cloud and I thought it might be that, so I waited to see if the colors moved with it, and they didn't. And the reed keeper knows them, so I went down to the beds for him first, because if he said it was nothing—"
"Minutes."
"The stair clock said a quarter past when the blue came up. The clerk's board has the cord at twenty-two past." You make yourself say the rest. "Seven minutes."
"And the crack?"
"Five. Maybe six."
The little slime presses its mouth against a crease in your palm, finding no food there.
"I saw it first," you say.
"You did." The warden steps inside. "The salt scour goes into the drain before crest. If the drain chokes, the lower houses flood. Salt kills everything on that wall, and whatever you saw goes with it. Give me a test that changes that choice."
The assay-master lifts the lamp. "How long do I have?"
"Till the next red mark. Less if the rain thickens."
From the passage, the reed keeper says, "You won't learn anything from one. I've tried."
He carries two shallow clay trays under one arm and leaves wet heel prints behind him. His trousers are stained to the knee with the black-green water of the winter beds. In his other hand, an iron bucket shifts with the soft, sticky weight of more slimes.
The warden glances at the floor. "Cloth."
He puts the trays down. "After. They don't do it on their own."
"You told me they change color every wet season," you say. "You said when it rains long enough they go blue."
"After a long rain they bank up on the mica. Some go blue, some years, when the beds are full." He looks into your hands, not at you. "Take one off the wall and it's green again before you reach the shed. I told you that part too."
You had heard the part that made your claim larger. The rest had seemed like mud around it.
The assay-master draws a clean wet line of mica clay around the first tray. "Put eleven on the film. If you have salt on your hands, wash it off first."
"Salt burns them," the reed keeper says. "They ball up and you get nothing."
"I washed before I came up."
"Then rinse again." The master holds out her hand, and the reed keeper passes her the bucket. She sets it at your elbow.
The words are ordinary and irritating. You obey them.
One by one, you lift the slimes from his bucket. Each behaves as the one in your palm did: contracting at first touch, spreading toward wet clay, finding the oat paste. Eleven become an untidy ring with no interest in your urgency. Their skins remain the cloudy green of pond glass.
"Nothing," the warden says.
"It was dozens on the wall," the reed keeper says. "That's eleven."
You set the examined slime into the last open space. Its edge touches two others. A pulse moves through the ring, not outward from the new arrival but both ways at once. Three bodies turn opaque. Four clear until you can see grit suspended inside them. Along the tray's lower edge, a cold blue travels through the rest and branches around the oat paste.
You lean in over the tray. "There. That's it, that's the warn—"
"Drip," the assay-master says.
The eave above the open shutter is leaking. Each drop strikes the sill, and the vibration reaches the table through the wall.
The warden leans out and closes the shutter. "Again."
The master lifts the slime you examined clear on a reed spatula. After a slow count of twenty the colors thin and swim, and the ring settles back to green. She sets the slime back into the same gap. The pulse comes again, fainter: two bodies cloud, three clear, and a blue edge runs partway along the lower rim without branching.
"It came back," you say. "It's the twelfth. Every time a twelfth goes in, they start."
"It started when a twelfth touched this group, on this clay, once under a dripping eave and once, weaker, without it." The assay-master gives you a wax pencil. "Write that."
You write it, pressing hard enough that the wax skids on the card.
The warden taps the red mark on her gauge card. "Two comparisons. Five minutes. If one matters, we go to the wall."
The reed keeper has already begun coating the second tray to match the first. "Bucket water do?"
"Yes, if each cup gets the same measure," the master says.
You lift the one you examined out and set it apart in a clay bowl with a smear of wet mica and paste. Then you divide the rest, twelve to each tray, adding more from the bucket. The first tray receives one continuous skin of wet mica. On the second, the master draws a dry wooden blade twice across the clay, leaving narrow gaps from rim to rim. Food goes at the center of each. Water goes into two stoppered cups.
"Ready," the reed keeper says.
The master pours one cup against the upper rim of each tray. Water presses through the mica in a shallow fan. On the continuous film, the slimes nearest the load go opaque. Those below clear. Cold blue forms along both edges of the wet fan. On the broken film, the same change begins and stops at the first dry cut. Beyond it, the slimes continue eating.
"They're showing where it'll go," you say. "Look, it runs down past the blue, same as the crack on the wall."
The reed keeper turns the continuous tray a quarter circle. "Let's see what they follow."
The water already in the clay now lies against a different edge. The blue branch frays, disappears, and reforms downhill, along the new edge of the load, in a shape you have not seen before.
The assay-master slides a flat weight beneath one side of the tray, changing the slope without adding water. Opaque bodies gather above the pressure; clear ones stretch below it. Blue holds at the boundary between.
You draw the three states on the back of the gauge card. "So it's the pushing. They go cloudy where it pushes hardest, clear where it pushes less, and blue where the two meet."
"That is what these twelve did," the assay-master says, and writes under your drawing: relative water stress, connected wet film. "If the film is broken, I expect nothing to cross the break. That tray agrees."
"And it takes twelve," you say.
"Twelve did it here," the reed keeper says. "Don't marry it yet. Three winters back I had forty on the upper face after a long rain, and they stayed green till the beds filled."
A runner's boots slap down the passage. The impacts jar both tables. Blue leaps in broken flecks across the continuous tray. Several slimes go opaque at once, then clear without relation to the slope.
The master waits for the runner to pass. "That was his boots. Irregular shaking gives noise."
The warden takes the pencil from you and writes that herself. "Will the ones on the wall show me what's pushing on that crack? Yes or no."
"If the film is still connected," the master says. "If feet and tools and rain don't drown the response. If the wall changes when a gate changes."
"That's three ifs."
"Yes."
"Useful?"
You nearly answer for her. The reed keeper lifts the first tray instead.
"Enough to carry down to the wall," he says. "I wouldn't stop the salt on two trays, though."
The warden points. "Wall. Bring the mallet. Trough stays covered till I say."
You take the mallet in one hand and the clay bowl in the other.
The service gallery smells of wet stone and the sharp ash brine waiting in its covered trough. Rainwater comes down the stair treads in threads. Along the reservoir wall, the crack has darkened. The colony still spans it, bodies touching across a sheen of mica clay. Opaque branches crowd above the split. Clear bodies spread below. Cold blue points sideways in a fork toward the bricked face of the old spill.
The pattern is so close to what you remember that certainty returns before caution can catch it.
"It's pointing at the old spill," you say. "There's something behind those bricks, and they're showing us where."
The assay-master kneels without touching the wall. "Or the crack is loading them from the side, and they are answering that. Mallet."
You pass it to her. She raps the masonry well away from the crack: three uneven blows. The colony shivers into noise. Blue spots flash everywhere. Clear skins cloud. The fork disappears.
"Stop feet," the warden calls. "Stop tools. Nobody on the upper walk."
The order travels away in shorter voices. For half a minute the gallery holds only water sounds. The colony settles. The fork returns, less cleanly now, toward the bricked spill.
"Gate one notch," says the warden.
The gate tender at the wheel looks down. "Main gate?"
"Main. One. Hold."
Iron teeth engage. The wheel turns; chain takes weight; the main sluice lifts the width of two fingers. Water changes pitch inside the wall. The blue fork weakens near the main channel, but along the brick face it deepens. A clear line opens below the old spill.
The reed keeper presses two fingers to the mortar. "Here."
A bead of water swells where there had been only damp. Then another. It carries pale clay, not the black silt of the main drain.
"There's a pocket behind the bricks," you say. "That's what they were pointing at."
"Evidence of one," the master says. "If there is water trapped there, it is separate enough not to drop with the main gate."
The warden watches the upper gauge runner's raised board. "If I scour, what happens to that pocket?"
No one answers at once. The covered ash-salt trough sits along the far wall.
The reed keeper rubs clay between finger and thumb. "You'll wash this face clean, slimes and clay and all. Whatever's behind those bricks will still be pushing after."
"Can you prove it?" the warden asks.
"Not before red, no."
The assay-master rises carefully. "If you scour, we lose the only indicator on this wall. It may shift the load behind the bricks too, and I can't say which way."
"The old spill runs out over the winter reed beds," the warden says. "Break it now and they're gone."
The reed keeper looks at her. "Drowned before cutting. No dry stalk, no roof bundles, no matting this year. If the water strips them to the roots, next planting too."
You had thought of the beds as an empty place for unwanted water. In the rain, you can smell the cut reeds stored above them, grassy and sweet through the stone vents.
"Then just a handspan," you say. "Open the old spill a handspan and watch the wall and the seep and the gauge together, and if the pocket drops we keep opening, and if it doesn't we stop, and nobody's lost more than—"
The warden looks at the bricks. "Opening crew needs eight minutes."
"Six," says the reed keeper. "The lower left corner's been soft since last winter. It's on your sheet."
The warden takes the inspection sheet from her oilskin and runs a thumb down to the third line. She folds it away. "Six, then. Tools. Two pry bars, rope, eye shields. Trough stays covered. Clear the spill run."
Then, to the reed keeper: "How much bed at a handspan?"
"First terrace. Maybe half the second."
"At full?"
He looks through the rain vent. "All of it."
The warden waits one beat. "Count the loss. Help them open."
Work makes speech smaller. Bar under. Hold. Rope tight. Not that brick, the shoulder brick. The assay-master asks for chalk, then for a dry cloth because the first will not mark wet stone. You take readings at the crack, the old spill, and the upper gauge. Twice you start to call a blue change before she makes you wait for the boot noise to fade.
The first brick comes free with a gush no wider than your wrist. Muddy water spits across the gallery and runs into the old channel. The colony blanches into disorder.
"They've gone to pieces," you say. "That's just the knocking, isn't it?"
"Some of it," the master says. "Some of it may be the load moving, and I can't separate the two yet. We wait."
"Gauge?" calls the warden.
"Still rising," comes the shout above.
They widen the opening to a handspan. Water drives through with a deep, hollow cough. The old channel catches it and carries it toward the reed beds. Behind the bricks, something settles with a sound you feel through your soles.
"It's slowing," you say. "The seep's slowing. It's working."
The master bends close. "Call the card, not the meaning."
You put the gauge edge to the bead. "Three drops in ten. Was nine."
"Upper gauge down one finger," the runner calls.
On the wall, the cold-blue branch beside the crack dims. Near the opened spill it strengthens, then thins as water finds the channel. Opaque bodies relax to green-brown. Three measures changing together. None of them says safe.
Then the runner's board goes up again: the crest is coming in, and the upper gauge climbs.
The warden turns to the reed keeper. Rain shows silver through the vent behind him.
He does not wait for her question. "They go blue by that spill every long rain once the beds are full. That tells me nothing. The beds I know. Two handspans open, the water's over the first two terraces. Four, and the lower bank goes. Past one, with this much coming, the crop's lost today."
"Can we hold one?"
He studies the flow, not the slimes. "Not through crest."
The upper bell strikes the red mark.
The warden's next words come one by one. "Full old spill. Main gate holds one notch. No scour. Move everyone off the lower walk."
The reed keeper grips the second pry bar. "Open the bed cuts first or you tear the bank."
The warden changes the order without defending the first. "Bed cuts first. Two people with him. Then full spill. Go."
You remain at the gauges with the assay-master while the gallery fills with boots, rope calls, brick dust, and rain. The wall pattern becomes useless. Mallet shocks throw blue everywhere; passing bodies send opaque ripples through the colony. When the storm strikes the outer face in earnest, even the wet mica runs.
"No read," you say.
"Good."
You look at her.
"Good call," she says, and points to the gauge. "Now use that."
So you do. You call heights and drop rates. You mark the crack at fixed points. The old spill opens brick by brick, and brown water takes the first terrace, the second, then the whole reed bed. Stalks flatten in shining rows. The lower bank holds because the reed keeper is out there in waist-deep runoff, shouting where to cut the overflow before the full release reaches him.
The reservoir gauge stops climbing.
It holds through three readings. Then it drops.
At the crack, the beads slow and cease. The masonry does not heal; the fork remains, black and undeniable, but it does not widen. No ash-salt goes into the drain. The covered trough stays covered.
By dusk, the storm has moved past the ridge. Water still pours through the sacrificed beds, but the houses below the reservoir show cooking lamps instead of evacuation lanterns.
On first repair watch, you sit at a trestle table in the gallery with the warden's wet record book between you. The crack has been braced. The old spill remains open under guard. Every quarter hour, someone measures the wall and reads the gauge aloud.
The warden writes: alarm delayed seven minutes after first observed color pattern.
You watch the nib. "Five before the crack. Two after."
She inserts the figures. "Reason?"
The answer you want is that no junior keeper should be blamed for failing to understand several dozen common slimes. The answer she can use is shorter.
"I stood and watched the colors. Then I went for the reed keeper before I went for the cord."
She writes that too. "Crop loss?"
Across the table, the reed keeper names terraces, planting days, bundles expected, and the bank section to inspect at morning light. He does not make the loss noble. The warden does not ask him to.
The assay-master leaves three cards beneath a stone: continuous film response; dry break; controlled gate response. Under them she adds limits in her square hand. Twelve observed, threshold not established. Relative water stress only. Irregular vibration and rain corrupt. No speech, prediction, identity, or safety claim supported.
"And the slime?" you ask.
She snaps her case shut and does not answer.
The slime rests in the same clay bowl beside your elbow, feeding on the last of the oat paste. Ordinary weight. Ordinary mouth. Ordinary acid. When you lift the bowl, it draws away from the vibration of your thumb.
"I thought you'd keep it for the assay room."
"I have its weight and its acid," she says. "In my room it would dry out inside a week. Put it where you took it."
The colony has gathered on the surviving wet mica below the brace. You rinse your hands, tip the bowl, and let the slime find the film by its own slow movement. At the twelfth contact—or perhaps the thirteenth; two bodies overlap beyond your count—a cold-blue edge passes through the group.
You write: blue after contact on wet film, old-spill load active. Relative pressure response; no directional use.
Behind you, the warden turns a page. "Gate order for midnight?"
"Old spill holds full," you say. "Main gate back down a notch if the upper gauge drops another finger."
The reed keeper is already coiling the measuring rope. "And the bed cut?"
"Check at second bell."
"First bell. Bank's soft."
The warden looks to you.
"First bell," you say. "I'll take the wall reading while he checks the bank."
"Record both," she says.
The gauge clerk calls the next number from above: four fingers under red. You set the bowl upside down to dry and write it down.