The Evidence Below

Scaithe keeps its inspection forms in a tin box more carefully protected from damp than the mine plan. Ilse Ostrum, the Dwarven structural inspector, laid both before me on arrival and explained what my Collegium credentials entitled me to sign. The mine’s rules accepted a ventilation certificate from a holder of a Royal Collegium chair in natural philosophy; I could find no provision requiring that the holder had ever descended a mine, a qualification which the Collegium might reasonably have hesitated to confer by appointment. My certificate was required before the resident manager, Mara Tennick, a Dwarf, could authorise an inspection party to enter the abandoned level, while each idle day added to the mine’s bill for pump hire. I had also given my colleague Professor Osric Cradock, of the Royal Collegium at Tarnis, a written promise of a catalogue of inspection customs that could safely be discarded. He considered their investigation beneath scholarship; I consider a practice which governs where men put their bodies a tolerable subject for intellectual exertion, even when its practitioners have neglected to publish.

The plan showed a separate intake supplying the maintained approach to the abandoned level’s gate, and a drainage sump isolated behind masonry. I consequently expected scant communication between the pump drawing water from that sump and the current through the gate. Scaithe’s hired pump had run for six weeks and operated throughout every maintenance visit, without reported ill effects. Those visits remained on our side of the gate with the shutter closed. The reopening proposal contemplated intermittent pumping.

Ilse, who has spent seventy years surveying mines, explained that the inspection crew would “give the level its silence” while “taking the gate’s breath”. The silence was twelve minutes of listening for movement in the roof and supports. Threads were to remain at several heights across the narrowly opened shutter while the pump stopped; the other nearby machinery was already idle. If the lowest threads turned outwards, from the abandoned level towards us, the shutter must close and everyone withdraw. Such a current might bring air deficient in oxygen, or charged with mine gases, into the approach; we had no measurement of what lay beyond the gate.

At the opened shutter my anemometer registered an inward current of forty-two feet per minute, and I expected it to show any change during the stoppage. I recommended omitting the threads and, if the reading remained satisfactory and entry were authorised, proceeding to the accessible supports just beyond the gate. Ilse refused. “Then you may record the redundancy,” she said, and I agreed with the satisfaction of a man offered assistance with his own argument.

Tom Ellard, the Human ventilation attendant, recited a verse asking departed masters to steady the threads, then hung them across the opening, where they inclined inwards. While he signalled the operator, I entered the reading and my proposed alteration in my notebook, drawing an inward arrow labelled “pump running”. The pump’s knocking slowed to a final stroke. Within the first minute of silence, the lowest threads hung slack for a moment before lifting outwards towards us. I kept the anemometer near the top, where I had taken the first reading, intending to compare like with like; its needle now fell to zero. Without the threads below it, I should have taken that reading to mean that the current through the shutter had ceased, and waited there for it to resume. Before I could lower the anemometer into the returning current, Tom closed the shutter, and at Ilse’s order we withdrew along the passage. Under the proposed intermittent pumping, each stoppage might reproduce what we were leaving, a consequence I had failed to find in the proposal itself.

I wanted the relation between the stoppage and the reversal tested. Masonry drawn upon a plan proves very little about the present integrity of a wall, and a line representing an intake says nothing about what has happened to the passage since the survey. I privately suspected an unrecorded opening behind the masonry: a falling water level might draw air towards the sump, and recovering water displace it again. The threads had shown one reversal after one stoppage, at one opening; they told us neither what the returning air contained nor how far it would travel into the approach. A fine thread has very little work to do before becoming informative. My instrument had measured at the height where I held it; the threads occupied several heights, and I had mistaken the precision of one reading for the extent of an examination.

My wish to lower the instrument before withdrawing recalled the Bellfounder workings with an exactness I should have preferred in the instrument itself. There my headache had begun while my Dwarven companions still seemed comfortable, and I had accepted their comfort as reassurance before Mistress Hadrik ordered the gallery cleared. I argued for time to complete my readings, and she removed my case, accurately judging which inducement would move me. The crew afterwards found a jammed ventilation door. I wrote the lesson down myself. A Dwarf’s comfort could establish nothing about the air’s fitness for Human lungs; at Scaithe, Tom would bear the consequences of forgetting it.

In the outer approach I proposed comparing pump states under a fixed shutter opening, then repeating the trials without the recitation. Ilse asked where Tom would stand during these comparisons. I indicated the gate, then looked at him as he wound the threads upon their board. “That is the opening we have just left,” she said. I proposed observations from farther back, with nobody stationed at the shutter. She agreed that comparisons so arranged deserved consideration, but insisted that the inherited sequence had found what six weeks of visits had missed. “Keep the threads, then,” I conceded. “Their distribution has an advantage. The rest remains open to examination.”

I asked whether twelve minutes of silence deserved preservation when today’s had ended within its first minute. She objected that the silence had its own purpose: a roof does not oblige an inspector by making its sounds immediately. I granted this readily; I had spent enough time waiting for lecturers to reach their subject to appreciate the general principle. When I asked what comparison justified twelve minutes, Ilse acknowledged that she knew of none. I wished to hear what experienced inspectors listened for, and how often those sounds were absent where defects were subsequently found. She could offer no count of silent failures, and said so without attempting to substitute her seventy years for the missing number.

Tom held that the masters had answered: to steady the threads, he explained, was to make them faithful to the air, and a faithful thread must turn when the air turned. I called the appeal a local embellishment, which offended him. He told me that his teacher, a Dwarven ventilation master, had learnt the procedure through a succession which included Human and Aesir instructors. I accepted the correction and disputed the inference. Apparitions are documented. Whether the dead persist remains unproved, and this movement of thread supplied no evidence of their participation. “You cannot prove they kept away,” he said. “Certainly,” I replied. “I also cannot enter their attendance upon your evidence.” My unrecorded opening had no better claim to admission as a fact, though its advantage was that a survey could find it, or oblige me to abandon it. Tom remained unconvinced.

Mara Tennick heard Ilse’s account without interruption, then opened the pump-hire ledger, whose entries she had made herself. Another week would consume money intended for the first working shifts; the promised work was already uncertain. She asked what condition I could certify. Her question deserved better than an airy declaration that safety has no price. Delay had a price, entered beside the names of people expecting wages, and I wanted to discover how work could proceed. The running pump offered a possibility: every uneventful visit took place during its operation, while our warning followed its cessation.

I began drafting my report on the inspection form and made continuous pumping a requirement in the section reserved for certification of the air. I considered including my advice to dispense with the threads, then decided that a withdrawn proposal need not encumber the report. Ilse prohibited entry while the pump was stopped, entirely properly within the limits of her evidence. I was enjoying the prospect of establishing conditions for reopening, and found no difficulty in attributing this improvement to the breadth of my education, as I wrote that, provided the pump remained in operation, the air could be certified fit for the proposed inspection. Ilse read that sentence over my shoulder. “How far beyond the gate?” she asked.

I could picture Tom at the shutter because I had watched him there. How far beyond it could I certify the air fit to breathe? I did not know. My notebook’s inward arrow carried its limiting condition plainly: “pump running”. The same reading had supported my earlier prescription to omit the threads. Under my proposed certificate he could be inside the abandoned level when the pump failed, with that air between him and the way out. An hour earlier I had recalled Hadrik carrying off my case; I now recognised how readily I made another person’s warning wait upon whichever occupation best pleased my scholarly pride. I struck out the certification and wrote into the report that my advice to discard the thread test was mistaken and was withdrawn.

I supported Ilse’s recommendation to withhold approval, without claiming that the level could never be made safe, and requested a ventilation survey incorporating controlled comparisons of pump state and recitation, observed through a telescope from beyond the affected approach with the shutter operated by a line. I also requested that, when safe access allowed the supports to be examined, the listening interval be tested against timed records of sounds and the defects actually found. Withholding the ventilation certificate left Mara paying hire from dwindling funds and the crew waiting for wages. Cradock could fairly reply that I promised a catalogue of dispensable customs and began by proposing the removal of a useful test. Ilse and Tom already knew my advice. Concealing it from Cradock would chiefly have protected me from the scrutiny I invited when I made my promise, while denying him a considerable pleasure; he could hardly conclude that examining the custom was useless, since we now had a specific reason to question the proposed intermittent pumping. My error followed the direction of his own contempt.

An outward arrow joined the first in my notebook, showing the direction of the lowest threads and labelled “pump stopped”. I made two identical fair copies of the corrected report, one for the mine and one for Cradock at Tarnis, and reproduced the drawing on each. The mine’s copy lay beside Ilse’s findings, and the other was folded for the post.