An Observatory on Borrowed Time

That leaves, by Albright’s count, 38 fields to photograph in six months, on a hill where winter offers perhaps six usable nights a month. Nine of the 38 lie in parts of the sky that are well placed only on summer nights, and they will not climb high enough in darkness before the money runs out. “So we know already that we can’t finish,” said Miriam Hale, the observatory’s director, who is 58 and one of its two employees. “What we can do is finish everything that can be finished. That turns out to be a surprisingly motivating distinction.”

A century of decades

The survey was the idea of Agnes Penhale, the observatory’s first director, who in 1931 was 36 years old and had, by her own account in that year’s annual report, “more telescope than purpose.” Brandeth Hill had been built six years earlier with money left by an Asterby mill owner who wanted the valley to have an observatory as fine as any city’s. Penhale proposed to give it something no city observatory had the patience for: the same 212 fields, photographed on several nights in each of 10 consecutive decades, so that astronomers in 2031 could compare the sky with what it had been a hundred years before. She knew she would not see the end of it. Her handwriting is on the paper sleeve of the survey’s first plate, exposed on the night of March 3, 1931: “Field 1, epoch 1. For whoever looks next.”

Over 77 years the telescope exposed 41,300 glass plates, about a quarter of them for the survey and the rest for programs that came and went with the directors who devised them: comet hunts, a search for faint companions of nearby stars, a decade spent photographing the Moon for a mapping project nobody now remembers. They live in the basement in 140 oak cabinets, each plate in a paper sleeve marked with the date, the field, the length of the exposure and often a word or two about the night: “haze,” “dew on lens,” “fox in dome.”2

Since 2015 a rota of 19 volunteers has been scanning them. They work in a converted storeroom beside the cabinets, with two scanners built for aerial photographs, a light table and a box of cotton gloves. Each plate takes about six minutes and becomes a file of nearly a gigabyte: some 450 million shades of gray, in which every star is a small dark smudge whose size and density record how bright it was on the night the plate was exposed. They have scanned 28,900 plates so far, including every plate in the survey. Even with a third evening a week added this autumn, about 70 plates an evening, they will not finish the rest before March. After that the cabinets are to go to a university storage depot in Corran, where the plates will be safe and, for practical purposes, out of reach.

What the old glass knows

Astronomers have photographed the sky digitally, in depth and in bulk, only since the late 1990s. For almost everything before that, the plates are the record. They are imperfect instruments; a plate responds to light less predictably than a modern sensor, and its measurements of a star’s brightness are good to about 10 percent. But they have one quality no new telescope can buy, which is time. Many stars change over days or weeks, and modern surveys catch those easily. Some change over decades. A few do something dramatic once in a human lifetime. The only way to see those is to have been watching for a very long time, or to own the glass of someone who was.

In the spring of 2017, Gordon Pye, a retired dentist from Little Varne, the village at the foot of the hill,3 was checking a batch of scans when the comparison software flagged a star in field 88 that was 40 percent fainter on a run of plates from 1979 and 1980 than on any others. Field 88 had been photographed far more often than the survey required, because it contains a star cluster that two later directors used for testing new batches of plates. Pye assumed a flaw in the emulsion. Then he pulled the field’s plates from every other decade and found the same dimming beginning in the summer of 1938 and lasting about 14 months. “I was a dentist for 38 years,” Pye said. “You learn the difference between a stain and a cavity. This was a cavity.”

The two dimmings began 41 years and four months apart. Hale and Pye worked out what kind of object could produce them: most likely a pair of stars, one of them wrapped in a vast, opaque disk of dust that once every 41 years swings across the face of its companion and blocks part of its light. Systems like it are known, but only a handful, because an eclipse that comes round twice a century is almost impossible to catch on purpose. If the interval held, the next one would begin in early 2021. The observatory photographed field 88 on every clear night from the autumn of 2020. On February 11, 2021, nine days after the predicted date, the star began to fade.^*

The paper that followed is, by some distance, the most cited thing the observatory has produced since Penhale’s day, and it rests on 312 plates, the oldest exposed when Pye’s parents were children. Hale puts its moral plainly: “No telescope built this century could have found that star. Not because they aren’t better. Because they weren’t there in 1938.” The archive holds quieter discoveries, too: a nova that flared on six plates in 1953 and went unnoticed at the time, and a handful of stars that have brightened steadily for 60 years for reasons nobody yet understands. Hale suspects that most of what the plates know has not yet been asked.

Nine fields short

The observatory has been running out of money, in one sense or another, for most of its life. The mill owner’s bequest was exhausted by 1970, when the University of Corran took the site over; since then the university has paid for two salaries, the electricity and the heating of a building with a large hole in its roof. In February it announced that it would consolidate its astronomy in a new building in Corran and end its support for Brandeth Hill at the close of its financial year. Hale calls the decision “perfectly reasonable on paper,” and the university points out that its students now do their observing remotely, on large telescopes abroad, and that a century-old refractor is not where modern careers are made. A local group, the Friends of Brandeth Hill, has raised about a third of what it estimates it would need to keep the dome open for five more years. Its treasurer keeps a hand-drawn fundraising thermometer in the window of the village shop.

Meanwhile the work has been reorganized around the deadline. Albright, who is 34 and the observatory’s technician, night assistant and, when necessary, roofer, has built a schedule that ranks the remaining fields by how soon each will sink into the evening twilight, and checks the forecast for the hill three times a day. On clear nights the telescope runs from dusk until dawn, and Albright naps on a camp bed in the plate room between exposures. Downstairs, the volunteers have moved the plates in the worst condition—fungus spreading across the gelatin, cracked glass taped at the corners sometime in the 1960s—to the front of the queue. Pye now comes in three evenings a week, which his wife describes as a phase.

In a chest freezer in the darkroom, under a layer of ice packs, the observatory keeps the last of its photographic plates: 46 unexposed sheets, bought in 2002 as the last suppliers stopped making them and kept cold ever since to slow the fogging that ruins old emulsion. Hale intends to use one on the survey’s final clear night, on whichever field is still above the horizon, and to develop it with chemicals she has been hoarding for the purpose. The plate will be foggy and of little scientific use. “It’s not for the data,” Albright said. “It’s for the cabinet.” Hale has already written its sleeve, in pencil, leaving the date blank. It reads: “Epoch 10, last plate. For whoever looks next.”