Chapter 379: An Answer Unasked

The sealed system — a watch post.

The new reader no longer had to ask where the interval cards were kept. She had one beside the third instrument and checked it whenever she called a value, even when the scale had barely moved.

The first reader kept her old station. Across the room, the duty keeper attended the second line while the recorder prepared the next sending entry. Its scheduled time was already written down; the actual time would come from the sending keeper. Between the expected windows, they continued the closer observations Tamar had ordered.

The first instrument registered a mark.

The reader called its time and value. The recorder’s pen paused over the page.

"That’s the scheduled time. I haven’t got a sending time."

"It’s when the mark appeared." She repeated it. "Write it there."

He entered the observation. The mark’s strength fell within the line’s posted range, and its shape resembled an ordinary checking return.

The duty keeper called a mark on the second line. The new reader watched her scale until the third appeared, then supplied its time and value. All three were within their respective strength limits. Their arrival times differed, as observations through separate paths could.

"Is that the next return already?" the recorder asked.

Tamar had left the desk. She stood behind the first reader’s stool, looking between the instrument and the unfinished sending entry.

"Leave it unmatched. We’ll see what follows the working."

The recorder entered it for immediate review while the duty keeper requested the sending details.

The sending keeper’s confirmation came through the duty keeper with the actual release time. When the first mark appeared, he had not yet sent the working. The recorded separation exceeded the uncertainty of the clocks. Against the confirmed sending time, the other two marks were also too early for their expected windows.

They kept watching. A further return registered within each line’s window. The recorder grouped those three observations beneath the sending entry and left the earlier event separate.

The recorder turned the ledger toward the first reader.

"I thought you were reading off the card."

"I was reading the instrument."

"I know that now."

He left the book where she could check his entry before taking it to Tamar. She opened the earlier pages beside the unmatched observation.

The recorder checked back through the sending entries covered by the earlier review, then forward to the latest completed check. Each had a matched return. None remained outstanding, though an additional echo of an answered working was still possible.

The sending record and the keeper’s account agreed about the first mark. For a controlled test, Tamar wanted someone watching the release itself.

"Put another keeper at the sending station," she told the duty keeper. "I want an observation of the input as well as the returns."

They repeated the equipment checks one line at a time, keeping the others attended. Each instrument read the reference input within its limits. They checked the clocks against the separate time standard and recorded the results with the settings used.

"Three paths agreed," the duty keeper said when they finished.

"They all connect to the same structure," Tamar said. "A fault there could reach all three."

He looked back at the first instrument, then pulled his interval card closer.

Tamar checked the prescribed behavior before planning the test. This particular response was supposed to follow the checking input through the designated section. The checking mechanism had no scheduled output of its own. Continuous containment had its own activity, maintained by the keepers; withholding a check would not mean leaving the prison untended.

She proposed delaying the next input while its form, strength and section remained unchanged.

"Watch through the old time and its return windows," she said. "Then send the working. We’ll have separate times to compare."

The recorder drew a new field beside the scheduled time.

"When?"

Tamar gave him a proposed release time. The new reader checked her card and looked up.

"That’s beyond the first window. It isn’t beyond mine."

Tamar came to her station.

"Show me."

The reader put her finger beneath the third line’s closing time.

"If we send then, my old window is still open. I want it closed before anything new goes out."

Tamar compared all three cards and extended the delay. For the second test she chose a longer interval, still allowing the deferred return to be observed before the following ordinary slot. The duty keeper sent both proposed intervals to the keeper responsible for that section.

Permission came before they began. Both delays were allowed; the continuous containment work would remain unchanged.

The recorder fitted the approved times into the row, then looked at the space left for observations.

"If we move it again, I’m using the next page."

"Use it now," Tamar said. "Keep this one attached."

He turned the page and ruled wider columns. At the first instrument, the reader set her pencil against the printed scheduled time, hesitated, then wrote the deferred time beside it. She labeled both and placed the card back where she could see it.

At the sending station, a second keeper watched the assigned keeper prepare the prescribed working. He would confirm that no checking input was released during the hold, then record the actual sending time.

The duty keeper remained at the second instrument. The first reader took the first, and Tamar joined the new reader at the third. For this observation, they disconnected the third instrument from the structure and supplied a steady reference input. Its setting and the changeover time went into the record.

The original scheduled time came.

The first line marked. The reader called its arrival time and strength without looking toward the desk. The duty keeper supplied the second observation when it appeared on his line.

At the third instrument, the reference reading held steady.

"No change here," the new reader said. Tamar checked the scale beside her and confirmed it.

The recorder entered the two connected readings as a further unmatched event, with their actual times. The keeper at the sending station confirmed that the checking input remained withheld.

They watched until all three original windows had closed. No further mark appeared. The new reader then restored the third instrument’s ordinary connection, checked its setting and allowed it to settle. Tamar confirmed it was ready before the deferred sending time.

The assigned keeper released the working at the approved time. The observer confirmed the release, and the recorder entered the actual time beside the planned one.

A return followed within the windows calculated from that sending. Each instrument registered it within its own timing and strength limits.

"Two events," Tamar said. "Keep them apart on the page."

The recorder showed her the line he had already drawn between them.

She read both entries before returning to the duty keeper. The instrument on the reference had remained steady while those connected to the structure registered the earlier event. That narrowed the equipment question, but it did not exclude a fault shared by the connected paths or the structure itself.

"Then what produced the first one?" he asked.

"We haven’t found that. The sending observer saw no checking input."

He passed her the confirmation from the sending station. She put it with the readings and asked the new reader to prepare the cards for the second test.

For that run, they used the longer approved delay and moved the reference control to the second instrument. The first and third remained connected. The sending observer stayed in place through the original scheduled time and all the original return windows.

The connected instruments again registered an unmatched event around the old scheduled phase, each at its own observed time. The second instrument held its steady reference reading. After the old windows closed, the duty keeper restored its connection and checked the setting before the delayed input was released.

That working produced a return within all three recalculated windows. The recorder read the entries aloud, and each observer confirmed the times and values from their station.

Tamar took the ledger to the desk. There were three unmatched events now: the first observation and one from each controlled deferral. Both deferred workings also had accounted-for returns.

Her report stated that return-shaped events had recurred at the original scheduled phase during two authorized delays, while the prescribed checking input was observed withheld. The later workings had produced returns within their own expected windows. Reference checks had passed, and the instruments used as disconnected controls had remained steady during the unmatched events. A fault in the shared paths or structure remained possible; the cause was unestablished.

She attached checked copies of the initial observation, the controlled tests, the equipment checks and the sending-side confirmations. The original ledger stayed at the post. The report went through the keepers’ chain to the others responsible for containment.

"The next check goes at its ordinary time," she told the duty keeper. "Confirm that with the sending station."

He sent the instruction and brought her the acknowledgment. Both temporary deferrals were complete. The third instrument remained connected, and the closer observations continued under the existing unmatched-response rule.

The new reader called the next observation time. Tamar checked the card beside the first instrument and took her station.