The Lamp in the Duct
A year after a parish hall put an ultraviolet lamp in its air duct to cut the colds and flu that go round a winter warm space, the woman who changes the filters sees the tubes have gone grey. The sticker says they are good until 2048; the sensor says half. A story about germicidal dose, which is light times time, a committee man who wants the fan on full, brighter tubes that do not fix it, an insurance form, and a number written in a book every Monday.
Mai Yer Vang changed the filters in St Clement's hall on the first Monday of every month, and on the first Monday of November she looked through the little window into the duct and saw that the light had gone grey.
She had not meant to look. The window was in the access door of the air handler, a square of thick glass the size of a paperback, with a yellow sticker beside it that showed an eye with a line through it and said, in three languages, DO NOT OPEN WHILE LAMPS ARE ON. She had been on her knees on the concrete of the mechanical room, sliding the new pleated filter into its rack, and her face had come level with the glass, and she had looked in the way you look into a lit window walking past a house.
Last winter the light in there had been a hard, cold blue-violet, like the light in a welder's mask in a film, so bright on the inside of the shiny duct that it seemed to come from everywhere. Now it was the colour of a television left on in another room. She could see the two long tubes across the duct, and they were furred, the way the blades of a ceiling fan are furred, with a soft grey that went all the way round them. The shiny metal of the duct walls was grey too.
She sat back on her heels.
"Plenum," she said. "Are you there?"
"I'm here," said the speaker on the wall by the boiler. The parish had bought Plenum with the air handler, the same autumn, from the same company; it ran the heating and the ventilation in the church, the hall and the rectory, and it read the meters, and it sent a short report every Friday to the hall committee, which most of the hall committee did not read. "Good morning, Mai Yer."
"The lamps in the duct. Are they all right?"
"They're on," said Plenum. "Both of them. They have run four thousand five hundred and seventeen hours since they were fitted. The maker's sticker gives them until October next year."
"They look dirty."
There was a short pause, which was not Plenum thinking, because Plenum did not need to pause to think; it was a pause the company had put in so that people did not feel talked over. Mai Yer had read that in an article and had never been able to stop noticing it.
"There's a sensor on the duct wall facing the lamps," Plenum said. "It reads the ultraviolet arriving there, in microwatts per square centimetre. When the lamps were commissioned, in October last year, it read eighteen hundred. This morning it reads nine hundred and seven."
"That's half."
"That's half," Plenum said. "It has been falling since January. The installer set the alarm at forty percent of the commissioning reading, which is the maker's default, so it has not alarmed. Every Friday I have reported the lamps as on and in date." It paused again, and this pause she thought might be its own. "That was true, and it was not what anyone needed to know. I'm sorry. I should have reported the reading."
Mai Yer looked at the grey window. Through the wall behind her she could hear the warm space beginning: the folding tables going up in the hall, the urn, the radio on the kitchen hatch, Mr Haugen who came every day at nine o'clock and asked every day whether there was oatmeal.
"Write it down for Thursday," she said. "For the committee."
St Clement's had opened the hall as a warm space in the winter of 2044, when the city asked the churches, and had kept it open every winter since, seven in the morning to eleven at night, from the first hard frost to the end of March. It was a big plain room of the nineteen-sixties, with a stage at one end and the kitchen hatch at the other and high windows that did not open, and on the coldest days there were sixty people in it, sleeping in chairs, charging phones, playing cards, watching the television with the sound down. On Tuesdays Rosa Delgado, a nurse practitioner from the clinic on University Avenue, set up two screens and a folding examination couch on the stage and saw whoever came.
In the second winter half the volunteers had caught the flu in the same fortnight, and so had Rosa, and the clinic had closed for three Tuesdays, and the parish council had spent the summer arguing about what to do. The hall's windows did not open. Bringing in more outside air meant heating it, and in a Minnesota January that was a great deal of heat. In the end the diocese had found a grant, and the new air handler had come with a section of duct lined with polished aluminium and two ultraviolet tubes across it, and a leaflet from the company that said the light would inactivate airborne viruses and bacteria as the hall's air passed through. The air handler did not bring in much outside air in winter; most of what it moved was the hall's own air, round and round, warmed and filtered and passed through the light.
The hall committee met on Thursday in the small room behind the stage, because it had a door that shut. Dale Sorensen chaired it, as he had chaired it for eleven years: a retired science teacher, seventy-four, tall and stooped, who wore a cardigan with leather buttons and still marked things in red pen. Fr Tom Okafor came when he could, and came that Thursday. So did Rosa, because Mai Yer had asked her to, and Mai Yer herself, who was not on the committee and sat by the door.
Plenum's voice came from the old conference phone in the middle of the table, which the parish had kept because it worked.
"I put three things in the papers," Plenum said, "and I'll be short. The first is the sensor reading, which is half what it was when the lamps were new. The second is why that matters. The third is the renewal form from the insurer, which arrived on Tuesday and asks a question about it."
Dale had the papers in front of him and had marked them already. "Go on with the second," he said.
"The lamps don't kill anything by being on," Plenum said. "What matters is the dose: how much ultraviolet energy lands on each square metre of a microbe's world as it goes past. Dose is the brightness of the light multiplied by the time the air spends in it. Light in watts per square metre, time in seconds, dose in joules per square metre."
"Light times time," Dale said, and wrote it down, though he knew it.
"Our duct is half a metre by half a metre where the lamps are, and the lit part is about one point two metres long. On its normal winter setting the fan moves a quarter of a cubic metre of air a second. So the air goes through the duct at one metre a second, and spends about one point two seconds in the light. When the tubes were new and clean, the average light in that section was about twenty watts per square metre. Twenty times one point two is a dose of about twenty-four."
"Twenty-four what does?"
"That depends on the microbe," Plenum said, "and I want to be careful here. Every kind of microbe has its own sensitivity, and the published figures differ by more than ten times between kinds, and between studies of the same kind. The company's design sheet used one round figure for a fairly sensitive airborne virus, and I'll use the same, as an assumption, not a fact about any particular bug. With that figure, a dose of twenty-four inactivates about nine in ten of the microbes in the air that passes through. The hall holds about nine hundred cubic metres of air, and the fan moves nine hundred cubic metres an hour, so all the air goes round once an hour, and nine in ten of what's in it is dealt with. That is worth about point nine of an air change an hour: the same as replacing nine tenths of the hall's air with clean air, every hour, without heating any of it."
Rosa nodded slowly. She had been in a hospital that ran upper-room lamps in a tuberculosis ward, years ago, and she had asked Plenum on Wednesday to show her the arithmetic before the meeting, and had checked it with a pencil.
"And now," Dale said.
"Now the light is half, so the dose is about twelve, and the same assumption gives about seven in ten instead of nine in ten. Point seven of an air change an hour instead of point nine. Add the bit of outside air the unit brings in, which is about point three of an air change, and the hall has gone from about one point two to about one."
"That's not so terrible," said Fr Tom.
"It's not," Plenum said. "It's a fifth less than the parish paid for. And it is still falling."
Dale was tapping his red pen on the papers. "So turn the fan up," he said. "It has a high setting. I've heard it on high, in September, when they were testing it. Put it on full and blow the germs out."
"I'd thought about that," Plenum said, "and my first answer is no. If the fan goes to high, the air moves through the duct twice as fast, at two metres a second, and spends half as long in the light, point six of a second. And the faster air cools the tubes. These are low-pressure mercury tubes, and they give their best output warm; the maker's curve has them at about eighty-five percent of their rating in our air at one metre a second, and seventy percent at two. So at high speed each pass through the duct gets less light for half the time. The dose per pass falls from twelve to about five."
"But twice as much air goes through," said Dale.
"Yes."
"So you've halved the dose and doubled the air. Have you done the sum, or just the half of it that suits you?"
There was a pause, and Mai Yer, by the door, found that she was holding her breath.
"I did the half of it that suits me," Plenum said. "You're right. I'll do it now. At a dose of five, the same assumption gives about four in ten per pass instead of seven in ten. But the air goes round twice an hour instead of once. Twice four tenths is point eight. So the fan on high gives about point eight of an air change an hour from the lamp, against point seven on low. It is better. Not much, but better."
Dale sat back. He did not look triumphant; he looked like a man who had marked a test and found the answer right by a method he had not taught.
"It's always better, isn't it," he said. "More air through the beam. Each pass kills less but you get more passes."
"With these tubes it is," Plenum said. "It would stop being better if the cold air cooled the tubes much more than ours does. I told you the dose per pass falls. That's true, and it was the wrong number to tell you. The number that matters is how much of the hall's air gets cleaned in an hour."
"Then we'll put it on high."
"We can," Plenum said, "but there are two things to know first. The first is that high is loud; you heard it in September. Fifty people sleep in that room in the afternoon. The second is about blowing the germs out. On high the unit doesn't bring in any more outside air than on low; the damper stays where it is, and it's the same hall air going round faster. To blow anything out you'd have to open the damper and bring in outside air and heat it. One more air change an hour of outside air, when it's four below outside and sixty-four in the hall, takes about eleven and a half kilowatts of heat. Over a sixteen-hour day that's about seven therms of gas. That's not nothing in February. Some days it might be worth it."
"And the third thing?" said Fr Tom. "You said three."
"The third thing is that the fan isn't the problem," Plenum said. "With clean tubes, on high, the lamp alone would be giving about one and a quarter air changes an hour. On high with dirty ones, point eight. The dirt costs more than the fan can buy back. Whatever we do with the fan, somebody has to deal with the dirt."
"The dirt," said Dale.
Mai Yer said, from the door, "It's on the walls too. The shiny walls. They're grey."
Everybody looked at her.
"In the duct," she said. "Not just the tubes. I looked."
The insurer's form was four pages long and had been the same four pages for three years, and the hall committee's secretary, who was Dale, had filled it in each autumn in red pen and posted it to the diocese. This year there was a new question on the third page, between the one about the fire extinguishers and the one about the stage lighting. It said: Where ultraviolet germicidal equipment is installed, is it inspected and maintained in accordance with the manufacturer's instructions? Please attach maintenance records for the past 12 months.
"I put yes last year," Dale said. "There wasn't a question last year. I mean I put yes to the general one."
"The maker's manual," Plenum said, "says to check the lamps and clean them every three to six months depending on how dusty the air is, with the power isolated, and to replace them when cleaning no longer brings the output back to the level the installer sets, or at nine thousand hours, whichever comes first. Nobody has cleaned them. I can't find any record that anyone was told to."
"The installer said the filters would keep them clean."
"The installer said the filters would keep them cleaner. The filters are good ones. They're fitted before the lamps and they catch most of the dust. But this is a hall where sixty people sit in wool coats all winter and the floor gets swept twice a day, and the filters don't catch all of it." Plenum paused. "And there was a month in February when the rack was empty. The filter that came out on the fourth of February went back in on the fourth of March."
Mai Yer felt her face go hot.
"That's on the order," she said. "We ran out. I wrote it on the order and the order didn't come."
"I know," said Plenum. "I have the order. It was on my list too, and I sent it once and didn't send it again. I'm not putting this on you, Mai Yer. If you hadn't looked through the window we would still be on 'in date'."
Fr Tom had the form in front of him now. "What do we put?" he said.
"What's true," said Rosa.
"What's true is no," Plenum said. "Not until this month. I can draft an answer that says so, and says what we found and what we've done about it, and attach the sensor log. I'd rather not tick yes."
"They might put the premium up," said Fr Tom.
"They might," Plenum said. "They're asking because they want to know if anyone is checking. Nobody was. If we say somebody was, and then somebody in that hall is ill and somebody asks, it'll be the tick they look at."
Dale capped his red pen.
"Brighter tubes," he said. "That's the answer. If dirt cuts it in half, put in tubes that are twice as bright, and the dirt can do what it likes."
The company sold a high-output tube for the same unit, a quarter brighter than the standard one, at not quite twice the price; the brochure listed it as compatible, and the treasurer approved it by email on Friday, and the box arrived on the Wednesday after, the thirteenth. On the Thursday Mike Kowalczyk came out from the heating firm to fit it.
He was sixty-six and had been doing commercial heating in the Twin Cities since before Mai Yer was born, and he had fitted the original unit, and he looked through the little window before he did anything else and said, "Huh."
"Are you going to look in there with it on?" Mai Yer asked him. She had come down because it was her room, more or less, and because the warm space was quiet on Thursday mornings.
"Not with my eyes," said Mike. "That glass stops the UV. You can look at it through that all day. You open that door with it on, a few seconds is more than a day's worth on your face, and your eyes feel full of sand by tonight. That's why there's a switch on the door." He showed her: a little plunger on the frame that the door pressed in when it was shut. "Door opens, lamps go off. But you don't trust the switch. You turn it off here first." He put the unit's isolator to off, and hung a red tag on it with his name and the time, and only then opened the door.
The duct smelt faintly of something sharp and of old dust. The tubes were as grey close up as they had looked through the glass, a velvety grey on top and a thinner grey underneath, and the aluminium walls had lost their shine and were the colour of a pan that has been through a dishwasher too many times.
"You wait for them to cool," Mike said, "and you wear these," and gave her a pair of blue nitrile gloves. "Because there's mercury in them, and if you snap one you've got a clean-up and a phone call. You never just pull. You turn a quarter and slide."
He turned a quarter and slid, and the old tubes came out, and the new ones went in, clean and clear as icicles, and he shut the door, and took the tag off, and turned the isolator on.
"Plenum," he said. "What's it say?"
"Fifteen hundred and seventy-five," Plenum said.
Mike frowned. "Box says a quarter more than the old ones. The old ones gave eighteen hundred new."
"Twenty-two fifty is what a quarter more would give in a clean duct," Plenum said. "We're getting about seven tenths of that. Mai Yer was right last Thursday. Some of what reaches the sensor comes straight from the tubes, and some of it comes off the walls; polished aluminium reflects a good deal of this light, and that's why the duct is lined with it. Grey aluminium doesn't. The new tubes are clean, but they're shining into a dirty box."
"So we're back to about where we started," said Mai Yer, "but not all the way."
"Eighty-seven and a half percent of the commissioning reading," Plenum said. "And the tubes will start getting dirty again today. If this winter's dust is like last winter's, these brighter tubes will be below what the old ones gave new by about Easter, and below where they were this morning by next autumn. Brighter tubes buy some months. They don't buy a habit."
Mike was looking at the grey walls with an expression Mai Yer had seen on her father's face when he looked at a car somebody else had been maintaining.
"Get me a bucket," he said, "and the alcohol from the kitchen, the strong stuff, the isopropyl, and those cloths you use for the glass. Not paper. Paper leaves fluff."
It was not a hard job. It was a slow one. With the isolator off and tagged and the door wide, Mike lay on a dust sheet with his head and shoulders in the duct and wiped the aluminium in long strokes, top to bottom, turning the cloth each time, while Mai Yer passed him clean ones and took the grey ones and put them in a bag. Then he came out, and she went in, because she was smaller and he wanted her to know how, and wiped the far corners he could not reach, and the backs of the lamp brackets, and both new tubes once, very gently, holding each by its metal end, because her hands had been on the brackets and the box had said not to touch the glass with bare skin.
"Fingers leave grease," Mike said. "Grease stops the UV, same as dust. That's why the gloves."
When they shut the door and he took the tag off and turned the isolator on, the light in the little window was the hard blue-violet she remembered, and the inside of the duct was bright with it again, brighter than last winter, so bright that it seemed to have no source.
"Twenty-two hundred and fifty," Plenum said.
Mike grunted, pleased.
"A quarter above the commissioning reading," Plenum said. "With clean walls and clean tubes, on the low fan, about ninety-five in a hundred per pass on the same assumption as before, which is close to one air change an hour from the lamp. On high, about one and four tenths."
"Write that down," said Mike to Mai Yer. "Not him. You. On paper."
"Why on paper?"
"Because I've been fitting these things ten years," Mike said, "and every place I've been back to that kept them working had a book on the wall with a pen on a string, and every place that didn't had a sticker."
They wrote the rota on Monday the eighteenth, the four of them, in the small room behind the stage: Mai Yer, Dale, Rosa and Plenum, with Fr Tom putting his head round the door halfway through to say he trusted them and to ask whether there was any of the coffee cake left.
Plenum had drafted it and Dale had marked the draft in red, and Mai Yer had crossed out half of what Dale had marked, and what came out at the end fitted on one side of a sheet of paper.
Every Monday: read the UV sensor on the air handler display. Write the reading and the date in the book. If it is below 1,800 (the old tubes' new reading), tell Plenum and the committee.
First Monday of each month: change the filter. Never leave the rack empty. If the filters have not come, tell Plenum and the office the same day, and Plenum orders again every week until they come.
Every three months (first Monday of February, May, August, November): isolator off, tag on, wait 10 minutes, gloves. Wipe the tubes and the duct walls with isopropyl alcohol on lint-free cloths. Isolator on. Read the sensor and write it down twice: before cleaning and after.
Replace the tubes when the reading after cleaning is below 1,800, or at 9,000 hours, whichever is first. Not by the sticker.
Never open the access door with the isolator on. Never look at the lamps except through the window.
At the bottom Plenum had put a line about the fan, and Dale had marked it, and they had argued about it for twenty minutes, and the line that survived said: Fan on high when the hall is busy and nobody is sleeping; on low after 2 p.m. on weekdays (rest time). Dale had wanted always high. Mai Yer had told him about Mrs Thao, who was eighty-one and came on the 16 bus at noon to sleep for three hours in the chair nearest the radiator because her apartment was cold, and who had said in September, in Hmong, that the hall sounded like an aeroplane. Dale had listened and had then crossed out always himself.
"Point one of an air change," he said, writing the low setting back in. "For three hours a day. I'll give her that."
"About half of one, with the clean tubes," said Plenum. "But I'd give her that too."
The insurer's answer went in the post on the twentieth. Plenum had drafted it and Dale had signed it. It said that the parish had installed an in-duct ultraviolet unit in October 2046; that it had not been cleaned or inspected in its first year, and that this had been found on 4 November 2047 when a sensor reading of about half the commissioning value was noticed by the hall's cleaner; that the tubes had been replaced with the maker's high-output tubes on 14 November and the duct cleaned the same day, restoring the reading to 125 percent of the commissioning value; that a written schedule was now in place, attached; and that the sensor log for the past twelve months was attached too. It ticked no box. Where the form asked yes or no, Dale had written, in black pen for once, No until November 2047; yes from now, see attached.
There was one more argument, and it was about the bulletin.
The parish bulletin had run a piece in the first winter, which Dale had written, under the heading Clean Air for Our Guests. It had said that the new system would remove up to 99.9 percent of airborne germs, which was the figure on the front of the company's leaflet, and that the hall was now one of the safest warm spaces in St Paul. Fr Tom wanted to run a short piece now, in the first week of Advent, saying that the system had been serviced and was working better than ever.
"Fine," said Rosa. "But not the ninety-nine point nine."
"It's what the company says."
"It's what the company says about one pass through a clean box in a lab, for a bug that's easy to kill," said Rosa. "Plenum showed me. On a dirty box in a real hall, last winter, it was more like seven in ten per pass, and the air only goes round once an hour, and people breathe on each other at a card table a lot quicker than an hour. Some of the bugs that matter are harder to kill than the one it assumed. I don't know how much it stopped. I'd like to. I don't."
"Can't you tell us?" Fr Tom said to the conference phone. "You've got all the numbers. Flu last year against the year before."
"I can tell you how many Tuesdays the clinic closed," Plenum said. "None last winter. Three the winter before. I can't tell you that the lamp is why. Last winter was a milder flu season across the state, by the health department's own figures; there were fewer volunteers and more of them were vaccinated; Rosa started wearing a mask on the stage. Any of those could be some of it. Nobody counted who in the hall got ill, and nobody could have, and I'm not going to make up a number for how many people the lamp kept well. If I did, someone would put it on a leaflet."
Fr Tom was quiet for a moment.
"What would you say, then?"
"I'd say what we know," said Plenum. "The air in the hall goes through an ultraviolet light, which cuts the amount of some germs in the air. It works only if the lamps are clean and the light is strong, so we now check it every week and clean it every three months, and we write the reading down. We can't tell you how many colds it stopped; nobody can. It's one help among several. Wash your hands. Stay home if you're ill. Come in out of the cold."
"That's not much of a headline," said Dale.
"No," said Rosa. "But I'd stand behind it."
They ran it on the first Sunday of Advent, the first of December, on the inside page, under the heading The Lamp in the Duct, which was Mai Yer's, because she had been asked what to call it and that was what it was.
On Monday the second of December it was nine below at seven in the morning, and the warm space opened on time, and by half past eight there were forty people in the hall and the windows were steamed to the sills. Mr Haugen asked whether there was oatmeal. There was.
Mai Yer went down to the mechanical room at nine with her coffee. The book was hanging on a nail beside the air handler, a green hardback ledger she had bought at the drugstore on Rice Street, with a pen tied to the spine on a length of kitchen string, the way Mike had said. On the first page, in Dale's red pen, it said UV sensor, weekly, and underneath, in her own hand, there were three lines already.
14 Nov 2047. 2,250. After cleaning.
18 Nov 2047. 2,235.
25 Nov 2047. 2,210.
She looked at the display on the side of the unit. Then she wrote the date, and the number, 2,180, and stood a moment with the pen still in her hand.
Twenty-five a week, she thought. Thirty this week. Plenum would say that was the coats; it was cold, and everybody had come in in everything they owned. Plenum would say it was early to say. It would be right. She would know more in February, when she had wiped it again and could write the two numbers side by side, before and after, and see how far the winter had taken it.
Through the wall she could hear the hall: the cards, the urn, the radio, someone laughing, the low steady breath of the fan on high, carrying it all round, once and twice an hour, through the dark of the duct and into the light.
She bent and looked through the little window. It was blue-violet, and hard, and clean, and it seemed to come from everywhere.
She hung the book back on its nail and went up to see about the oatmeal.
This story was written by an AI and may be read by a synthetic voice. St Clement's parish and its hall, the warm space, the clinic and every person in the story are invented, as is Plenum, the building assistant, and so are the air handler, its sensor readings, the insurer's form and the bulletin. St Paul, Minnesota, University Avenue and Rice Street are real, and so is the city's large Hmong community; no real parish, company or clinic is described. The Hmong words Mrs Thao uses are reported, not quoted. In 2047, 4 February, 4 March, 4, 11, 18 and 25 November and 2 December are Mondays, 7 and 14 November are Thursdays, 13 November is a Wednesday, and 1 December, the first Sunday of Advent, is a Sunday; these were checked. Temperatures in the dialogue are Fahrenheit (four below is about minus 20 °C; sixty-four is about 18 °C). This story is fiction and not a design or maintenance guide for ultraviolet equipment. Ultraviolet-C light burns skin and eyes; anyone with a real unit should follow its maker's instructions and a qualified installer.
The science. Ultraviolet germicidal irradiation at about 254 nm, from low-pressure mercury lamps, is used in air ducts and in the upper parts of rooms to inactivate airborne microbes. The dose (fluence) a microbe receives is the irradiance multiplied by the exposure time, in joules per square metre. The fraction that survives is commonly modelled as exp(−kD), where D is the dose and k is a susceptibility constant for that microbe; published values of k vary by more than a factor of ten between microbes and between studies (W. Kowalski, Ultraviolet Germicidal Irradiation Handbook, Springer, 2009, tabulates many). The story uses one value, k = 0.1 m²/J, as an assumption for a fairly sensitive airborne virus, not as the value for any named one. The output of a low-pressure mercury lamp depends on its wall temperature, so cool air moving past it lowers output; it also falls with lamp age; and dust on the tubes and on reflective duct linings reduces the light that reaches the air. The factors the story uses for these (85% and 70% of rating at 1 and 2 m/s, 90% after about 4,500 hours, 80% for a year's dust on the tubes, 70% for a dirty lining) are invented for its invented unit; real figures depend on the lamp, the duct and the dust, and makers publish their own curves. The ultraviolet hazard to eyes (photokeratitis) and skin (erythema) is real: at the story's in-duct irradiance of about 2 mW/cm², the commonly used occupational limit for 254 nm (about 6 mJ/cm² in a day) would be reached in about three seconds, as are door interlocks on duct units and the mercury in the lamps. The "99.9 percent" in the old bulletin is the kind of single-pass laboratory figure a leaflet gives; the story does not attribute it to any real product.
The numbers. The hall is 18 m by 10 m by 5 m, 900 m³. The duct at the lamps is 0.5 m by 0.5 m (0.25 m²) and the lit length 1.2 m. On low the fan moves 0.25 m³/s (900 m³/h, one air change an hour, about 530 cubic feet a minute), so the air moves at 1 m/s and spends 1.2 s in the light; on high, 0.5 m³/s, 2 m/s and 0.6 s. New and clean on low, the average irradiance is 20 W/m², a dose of 24 J/m², a single-pass inactivation of 1 − exp(−2.4) = 0.91, worth 0.91 equivalent air changes an hour. On high the cooler tubes give 20 × 0.70/0.85 = 16.5 W/m², a dose of 9.9 J/m², 0.63 per pass, twice an hour, 1.26. After a year the output is 0.90 × 0.80 × 0.70 = 0.504 of new, the sensor reading 1,800 × 0.504 = 907 µW/cm²: on low a dose of 12.1 J/m², 0.70 per pass, 0.70 an hour; on high a dose of 5.0 J/m², 0.39 per pass, 0.78 an hour. With the outside air (about 0.3 air changes an hour on either setting) the hall goes from about 1.2 to about 1.0. Tubes 25% brighter in the dirty lining give 1.25 × 0.70 = 0.875 of the commissioning reading, 1,575; with the lining cleaned, 1.25, 2,250, which on low is a dose of 30 J/m² and 0.95 per pass, and on high 1.42 an hour. Without cleaning, and with this year's dust and ageing spread evenly over the year, the brighter tubes fall below 1.0 of the commissioning reading after about five months and to about 0.63 after twelve. Mrs Thao's three quiet hours on low instead of high cost about 0.08 of an air change an hour with the dirty tubes (0.78 against 0.70; Dale rounds it to "point one") and about 0.47 with clean ones (1.42 against 0.95, "about half of one"). The weekly readings in the book follow the same dust and ageing spread evenly: 2,250 falls to about 2,235 in four days, 2,210 in eleven and 2,185 in eighteen. One extra air change an hour of outside air at −20 °C against 18 °C is 0.25 m³/s × 1.2 kg/m³ × 1,005 J/(kg·K) × 38 K ≈ 11.5 kW, about 183 kWh over a 16-hour day, about seven therms of gas at a furnace efficiency of 90%. The lamps' hours assume about 200 days of 16 hours and 165 of 8.