A single star, careening around the monster black hole in the center of the Milky Way, has provided astronomers with new proof that Albert Einstein was right about gravity.

More than 100 years ago, Einstein’s general theory of relativity revealed that gravity is the result of matter curving the fabric of spacetime (SN: 10/17/15, p. 16). Now, in a paper published July 26 in Astronomy & Astrophysics, a team of researchers reports the observation of a hallmark of general relativity known as gravitational redshift. The measurement is the first time general relativity has been confirmed in the region near a supermassive black hole.
As light escapes a region with a strong gravitational field, its waves get stretched out, making the light redder, in a process known as gravitational redshift. The scientists, a team known as the GRAVITY collaboration, used the Very Large Telescope array, located in the Atacama Desert of Chile, to demonstrate that light from the star was redshifted by just the amount predicted by general relativity.

Scientists have observed gravitational redshift before. In fact, GPS satellites would fail to function properly if gravitational redshift weren’t taken into account. But such effects have never been seen in the vicinity of a black hole. “That’s completely new, and I think that’s what makes it exciting — doing these same experiments not on Earth or in the solar system, but near a black hole,” says physicist Clifford Will of the University of Florida in Gainesville, who was not involved with the new study.

At the Milky Way’s heart there lurks a hulking supermassive black hole, with a mass about 4 million times that of the sun. Many stars swirl around this black hole (SN Online: 1/12/18). The researchers zeroed in on one star, known as S2, which completes an elliptical orbit around the black hole every 16 years.
In May 2018, the star made its closest approach to the black hole, zipping by at 3 percent of the speed of light — extremely fast for a star. At that point, the star was just 20 billion kilometers from the black hole. That may sound far away, but it’s only about four times the distance between the sun and Neptune.
Measuring the effects of general relativity in the black hole’s neighborhood is challenging because the region is jam-packed with stars, says astrophysicist Tuan Do of UCLA, who studies S2, but was not involved with this work. If attempting to observe this region with a run-of-the-mill telescope, “you’ll just see this big blur.”

To obtain precise measurements and pinpoint individual stars in the crowd, the scientists used a technique called adaptive optics (SN Online: 7/18/18), which can counteract the distortions caused by the Earth’s atmosphere, and combined information from four telescopes in the Very Large Telescope’s array. “You can bring the light together from these four telescopes and thereby generate a super telescope … and that does the trick,” says study coauthor Reinhard Genzel, an astrophysicist at the Max Planck Institute for Extraterrestrial Physics in Garching, Germany. Genzel and colleagues have been observing this star for decades, since before its previous swing by the black hole 16 years ago.

In future work, the scientists hope to test other aspects of general relativity, including the theory’s prediction that S2’s orbit should rotate over time. A similar rotation was previously seen in Mercury’s orbit around the sun, which puzzled astronomers until Einstein’s theory explained the effect (SN Online: 4/11/18).

The GRAVITY researchers might find other stars that orbit even closer to the black hole, allowing them to better understand the black hole and further scrutinize general relativity. If that happens, Will says, “they’ll really start to explore this black hole up close and personal, and it’ll be a very cool new set of tests of Einstein’s theory.”

Hobbits took a separate evolutionary path to becoming small than did short, modern-day humans living on the same Indonesian island, a new DNA analysis suggests.

Rampasasa pygmies residing near a cave on Flores that previously yielded small-bodied hobbit fossils inherited DNA from Neandertals and Denisovans but not from any other now-extinct hominid, an international team of researchers reports in the Aug. 3 Science. The finding provides genetic backup for a fossil-based argument portraying these controversial Stone Age hominids as a separate species, Homo floresiensis, not small-bodied Homo sapiens that could have represented ancestors of Rampasasa people.
Diminutive hobbits, standing roughly a meter tall, lived on Flores from at least 100,000 to 60,000 years ago, with possible ancestors on the island dating to about 700,000 years ago (SN: 7/9/16, p. 6). Some scientists contend that hobbits were actually short humans, not an ancient hominid species (SN: 11/18/06, p. 330). So far, researchers have been unable to extract DNA from hobbit fossils. Comparing hobbit and present-day human DNA would go a long way toward clarifying the evolutionary ID of the half-size Flores hominids.
In the new study, evolutionary geneticist Serena Tucci of Princeton University and colleagues compared DNA from 32 Rampasasa individuals with that of Neandertals, Denisovans and present-day humans around the world. The Flores pygmies got smaller due in part to alterations to genes linked to height and the breakdown of fatty acids in foods, the researchers found.

Still, without hobbit DNA, it’s impossible to say with certainty that Rampasasa pygmies — on average more than 30 centimeters taller than hobbits — inherited no genes from H. floresiensis, says evolutionary geneticist and coauthor Richard Green of the University of California, Santa Cruz. “But we looked hard for [H.] floresiensis ancestry in Rampasasa people and could not find it.”

The question of how many died in the aftermath of Hurricane Maria has yet another answer.

Using vital statistics records from hard-hit Puerto Rico, researchers estimate that 1,139 more people died than expected from September 20, 2017 — the day the Category 5 hurricane made landfall — through that December.

Alexis Santos-Lozada of Penn State and Jeffrey Howard of the University of Texas at San Antonio tallied monthly death counts from January 2010 through December 2017 to determine the expected number of deaths and the variability in these numbers over time. Excess deaths spiked in September, the month the hurricane hit, and peaked in October 2017, with 564 more deaths than expected that month, the pair report online August 2 in JAMA. However, the vital statistics data don’t pinpoint individual causes of death.
The official death toll from the Puerto Rican government is 64, based on death certificates that document if a death was directly due to the hurricane. But this method doesn’t factor in deaths indirectly related to the storm, such as those from infectious-disease outbreaks or because of disruptions in electricity or medical care.
Last May, another study in the New England Journal of Medicine based on surveys of 3,299 randomly selected households across the island estimated at least 4,645 deaths from the hurricane (SN Online: 5/29/18). Survey participants were asked about deaths, delays in medical care, and electricity, water and cell phone interruptions to capture deaths indirectly related to the disaster.

Getting a good estimate on deaths after natural disasters is crucial for guiding rescue and recovery efforts, Santos-Lozada and Howard say.

Of these methods, “using excess death counts from dependable vital statistics data is probably a more solid approach,” says James Shultz, director of the Center for Disaster & Extreme Event Preparedness at the University of Miami Miller School of Medicine. But he finds it interesting that the excess deaths in the new study are concentrated in September and October, while the previous survey-based research assumes excess deaths continued past this time.

“Hopefully, future calculations will extend the analysis into the first half of 2018, when many areas of Puerto Rico still lacked power and health care services were only partially restored,” Shultz says.

Scientists now know how long it takes for a cell to tell itself it’s time to die.

Signals triggering a type of cell suicide called apoptosis move through a cell like a wave, traveling at a rate of 30 micrometers per minute, Stanford University systems biologists Xianrui Cheng and James Ferrell Jr. report in the Aug. 10 Science.

These findings resolve a debate over whether these death signals spread by diffusion, with signaling molecules working their own way across a cell, or as self-regenerating trigger waves, like toppling dominoes. The apoptosis process starts with damage that causes the release of death signal chemicals. One example is cytochrome c leaking from damaged mitochondria, the cell’s power plant. Once cytochrome c concentrations get high enough, the chemicals signal proteins called caspases to go to work. Caspases trigger other proteins to poke holes in neighboring mitochondria, releasing more cytochrome c and moving the death wave across the cell. That chain reaction happens more quickly than diffusion can, Ferrell says. In an African clawed frog egg, a trigger wave takes about a half-hour to spread across the 1.2 millimeter cell, whereas diffusion would take five hours, he says.

Like forest fires, trigger waves will keep going as long as there is fuel to feed them — in this case, the death signal chemicals and proteins, Ferrell says. He predicts that many other biological signals may move as trigger waves.

“We biologists are just waking to this idea that trigger waves are a recurring theme in biological communication,” Ferrell says.

Discovering that apoptosis travels as trigger waves in cells may give scientists clues about how to persuade cancer cells to kill themselves (SN: 1/21/17, p. 10). Or researchers may learn how to prevent cells from dying in conditions such as Alzheimer’s disease or muscular dystrophy.

Pregnant women aren’t immune to the escalating opioid epidemic.

Data on hospital deliveries in 28 U.S. states shows the rate of opioid use among pregnant women has quadrupled, from 1.5 per 1,000 women in 1999 to 6.5 per 1,000 women in 2014, the U.S. Centers for Disease Control and Prevention reports.

The highest increases in opioid use among pregnant women were in Maine, New Mexico, Vermont and West Virginia, according to the CDC study, published online August 9 in Morbidity and Mortality Weekly Report.
“This analysis is a stark reminder that the U.S. opioid crisis is taking a tremendous toll on families,” says coauthor Jean Ko, a CDC epidemiologist in Atlanta.

In this first look at opioid use during pregnancy by state, Washington, D.C. had the lowest rate in 2014, at 0.7 per 1,000 women, and Vermont had the highest, at 48.6 per 1,000. However, the data from the U.S. Health and Human Services Department represents only the 28 states that record opioid use at childbirth during the studied time frame.

“We knew the incidence was increasing” as the number of babies going through opioid withdrawal has also gone up, says Matthew Grossman, a pediatrician at Yale University. Overall, the number of U.S. deaths attributed to opioids has also been steadily rising (SN: 3/31/18, p. 18). In 2014, there were 14.7 opioid deaths per 100,000 people, up from 6.2 per 100,000 in 2000, according to the CDC.
Taking opioids during pregnancy, especially in the last trimester, increases the risk of preterm birth and stillbirth, as well as infant opioid withdrawal (SN: 6/10/17, p. 16). Pregnant women should tell their doctors if they are taking opioids, so complications can be addressed, says Alison Holmes, a pediatrician at Dartmouth-Hitchcock Medical Center in Lebanon, N.H. Mothers may be prescribed methadone, a synthetic opioid which is safer for the fetus and protects it from going through withdrawal in the womb. “What’s not safe for the child is active opioid misuse,” she says.

Only eight U.S. states require that pregnant women be tested for opioids if substance abuse is suspected, the CDC says. In Cincinnati, all pregnant women are tested at delivery, but it would be even better to test women in the first trimester, says pediatrician Scott Wexelblatt at the Cincinnati Children’s Hospital. “If we could identify a mom at 12 weeks instead of 40 weeks, then we could get her into medicated assisted treatment.”

In one published swoop, an ancient fossil fruit bat has turned into a lemur. If that transformation holds, it suggests that lemur ancestors made two tricky sea crossings from Africa to Madagascar, not one as researchers have often assumed.

A new fossil analysis finds that the ancient species Propotto leakeyi, which lived in East Africa between 23 million and 16 million years ago, was not a bat, as scientists thought, but a primate closely related to modern aye-ayes. These strange-looking lemurs are found only on Madagascar along with another closely related lemur lineage.
What’s more, Propotto teeth and jaws display key similarities with fossils of a roughly 34-million-year-old primate, Plesiopithecus teras, previously found in Egypt, researchers say. Plesiopithecus, previously suspected to have been a primate, was an ancestor of Propotto and of modern aye-ayes, they conclude. Together, the findings, published August 21 in Nature Communications, may help rewrite lemurs’ evolutionary history.

The research challenges a long-standing view that all Madagascar lemurs, including aye-ayes, evolved from a single population of African ancestors that somehow reached the island at least 54 million years ago. That estimate rests largely on genetic studies of modern lemurs and other primates. Destruction of ancient lemurs’ African habitats by global cooling around 34 million years ago left their kind isolated on Madagascar, according to this scenario.

But the survival of aye-aye ancestors in Africa millions of years after that, as suggested in the new study, raises the possibility that Propotto reached Madagascar on its own — separate from the other lemur lineage found on the island — and gave rise to present-day aye-ayes. No Propotto fossils have been found on Madagascar.

“Our identification of both Propotto and Plesiopithecus as African relatives of the aye-aye implies that [these] lemurs weren’t present on Madagascar until 30 million years or more later than previously thought,” says study coauthor paleontologist Erik Seiffert of the University of Southern California in Los Angeles.
Ancestors of the modern lemurs other than aye-ayes traveled to Madagascar sometime between around 41 million and 20 million years ago, the researchers estimate. During that period, ancestors of the only other mammal groups now inhabiting Madagascar — rodents, Malagasy mongooses and insect-eating creatures called tenrecs — also reached the island from Africa. Previous computer simulations indicated that ocean currents at that time could have carried animals stranded on storm-uprooted trees and vegetation mats from East Africa to Madagascar.

The team, which included Duke University’s Gregg Gunnell (who died in 2017), created digital reconstructions of Plesiopithecus and Propotto fossils for comparison with fossil and living primates, including aye-ayes (Daubentonia), and to closely related mammals called colugos. Evolutionary trees based on tooth and jaw analyses and available DNA data pointed to a link between the two ancient species and aye-ayes.

Plesiopithecus and Propotto might have used enlarged teeth projecting from the front of their mouths to gouge holes in trees and expose grubs’ nests, as modern aye-ayes do. Aye-ayes also poke through tree holes with long, skinny middle fingers to extract grub. But no hand fossils from either ancient creature have been found, so it’s a mystery whether they shared aye-ayes’ taste for finger food.

The discoverer of three Propotto tooth-bearing lower jaws in Kenya originally reported in 1967 that the finds belonged to a new primate species, possibly an ancestor of primate relatives of lemurs called lorises. But within the next two years, the same scientist accepted another researcher’s proposal that Propotto’s jaws and teeth more closely resembled those of a fruit bat. A 1984 report describing several more Propotto teeth unearthed in Kenya also concluded that they came from a fruit bat.

The new identification of a line of ancient African lemurs that ran from Plesiopithecus through Propotto “is an interesting discovery,” says paleoanthropologist Marc Godinot of the National Museum of Natural History in Paris. “I have thought for years that Propotto was more likely a primate than a fruit bat.”

Godinot also argued in a 2006 study that the shape and positioning of teeth at the front of Plesiopithecus’ mouth pegged it as a relative of aye-ayes, consistent with a double colonization of Madagascar by lemur ancestors.

That possibility “merits serious consideration,” but a single African origin for lemurs on Madagascar remains the simplest, most likely scenario, says evolutionary biologist Anne Yoder of Duke University. Most African mammals couldn’t manage even one colonization of the island, so attributing two of these “highly improbable” events to lemur ancestors alone demands more evidence, Yoder says.

Still, it can’t be discounted that several ancient African lines of primates might have evolved in the aye-aye lineage but only one made it to Madagascar on a sea crossing that occurred independently of other African lemurs, Yoder says. Or, in line with her own view, Madagascar may have been colonized by one group of ancient lemurs that gave rise to multiple lines of creatures, one of which was a direct ancestor of modern aye-ayes. Only further fossil discoveries can resolve this mystery, Yoder says.

For cervical cancer screening, there’s a new option in town.

Women ages 30 to 65 can opt to have human papillomavirus, or HPV, testing alone every five years, according to new recommendations from the U.S. Preventive Services Task Force.

HPV testing alone joins two other alternatives that are still endorsed: an HPV test plus a Pap test every five years, or a Pap test alone every three years. The guidelines, published online August 21 in JAMA, are the first update to the group’s cervical cancer screening recommendations since 2012.

Recent research has shown that HPV testing, which checks for the presence of the sexually transmitted virus in a sample of cervical cells, is better at catching precancerous lesions early than the traditional Pap test, which looks for those lesions in a cervical cell sample.

Globally, cervical cancer is the fourth most common cancer among women. HPV causes nearly all cervical cancers.

It’s been a year since the total solar eclipse of August 21, 2017, captured millions of imaginations as the moon briefly blotted out the sun and cast a shadow that crisscrossed the United States from Oregon to South Carolina.

“It was an epic event by all measures,” NASA astrophysicist Madhulika Guhathakurta told a meeting of the American Geophysical Union in New Orleans in December. One survey reports that 88 percent of adults in the United States — some 216 million people — viewed the eclipse either directly or electronically.
Among those were scientists and citizen scientists who turned their telescopes skyward to tackle some big scientific mysteries, solar and otherwise. Last year, Science News dove deep into the questions scientists hoped to answer using the eclipse. One year out, what have we learned?

1. The eclipse sent ripples through Earth’s atmosphere.
   Normally, the sun’s radiation splits electrons from atoms in the atmosphere, forming a charged layer called the ionosphere, which stretches from 75 to 1,000 kilometers up. But when sunlight briefly disappears during an eclipse, the electrons rejoin their atoms, creating a disturbance in the ionosphere that is detectable with receivers on the ground (SN Online: 8/13/17).

The moon’s supersonic shadow produced a bow wave of atoms piling up in the ionosphere, similar to the wave at the prow of a boat, Shun-Rong Zhang of MIT’s Haystack Observatory in Westford, Mass., reported in December. Although such bow waves were predicted in the 1960s, this was the first time they were definitively observed.
The eclipse also sent a wave traveling through the thermosphere, an uncharged layer of the atmosphere about 250 kilometers high, that was observed from as far away as Brazil nearly an hour after the eclipse ended (SN: 5/26/18, p. 14). And measurements of temperature, wind speed and sunlight intensity showed that the eclipse briefly changed the weather along the path of darkness.

2. Showing Einstein was right is not so simple.
   Physicists chased the moon’s shadow to redo the iconic experiment that showed Einstein’s theory of general relativity was correct (SN Online: 8/15/17). In Einstein’s view, the sun’s mass should warp spacetime enough that the positions of stars should appear to be slightly different during an eclipse. During the May 1919 solar eclipse, British astronomer Arthur Stanley Eddington took photographs that proved Einstein right.

During the 2017 eclipse, almost a century later, amateur astronomer Donald Bruns of San Diego made similar measurements with modern equipment and came to the same conclusion as Eddington: Stars visible during the eclipse were all askew. Bruns published his results in Classical and Quantum Gravity in March.

But astrophysicist Bradley Schaefer of Louisiana State University in Baton Rouge and others had far more difficulty reproducing the measurement with enough precision to show that Einstein was right. “‘Bummer’ is an understatement,” Schaefer says. “This all may have been for naught.”

Schaefer had enough trouble that he thinks it may have been impossible for Eddington to get the precision he claimed. The earlier astronomer may have hit upon the right answer by luck, not because he actually measured it.

3. Infrared light will help measure the corona’s magnetic field.
   Some eclipse experiments didn’t revolutionize our understanding of the sun on their own, but will enable future ones to pull back the veil. One of these was the first infrared observations of the sun’s corona, the shimmering halo of hot, diffuse plasma that is only visible in its entirety during a total solar eclipse. The shape and motion of all that plasma are guided by magnetic fields, but the corona’s magnetic field is so weak that it has never been measured directly (SN Online: 8/16/17).
   Previous studies suggested that infrared wavelengths of light might be particularly sensitive to the corona’s magnetic field. So two groups chased the August 2017 eclipse in airplanes to get some infrared observations. Amir Caspi of the Southwest Research Institute in Boulder, Colo., and his colleagues took the first infrared image of the entire corona.
   Flying in another aircraft, Jenna Samra of Harvard University measured the corona in five specific wavelengths, one of which had never been seen before. Comparing those results with observations taken from the ground in Casper, Wyo., (where I watched the eclipse) showed that those wavelengths are bright enough that a telescope now under construction in Hawaii will be able to help map the corona’s magnetism (SN Online: 5/29/18).
4. Figuring out what heats the corona will take more work.
   Almost every experiment aimed at the eclipsed sun last August had some bearing on the biggest solar mystery of all: Why is the corona so hot? The solar surface simmers at around 5500° Celsius, but the corona — despite being farther away from the solar furnace and made of much more diffuse material — rages at millions of degrees.

One year after the Great American Eclipse, scientists are still scratching their heads. Caspi’s team searched for waves rippling through the corona, which could distribute energy far from the solar surface. Those waves could also help comb out magnetic tangles in the corona and explain its well-ordered look (SN Online: 8/17/17).

In a complementary measurement, the group in Wyoming saw signs of neutral helium atoms in the corona, says solar physicist Philip Judge of the National Center for Atmospheric Research in Boulder. Those uncharged atoms probably represent cool material embedded in the corona (SN Online: 6/16/17).

Similar cool spots have been seen during earlier eclipses, although it’s hard to imagine how the cool atoms survive in the searing heat, like ice cubes remaining solid in hot soup. But collisions between charged ions and neutral atoms could help convert ordered motions, like Caspi’s waves, into coronal heat.

The results so far are interesting, but inconclusive, Caspi says. “It’s certainly possible we will get some very interesting results from this set of observations alone,” he says. But for such a big problem as coronal heating, eclipse observations may play a supporting role to more direct measurements, such as those that the recently launched Parker Solar Probe will make (SN Online: 8/12/18).

5. People are already looking to the next eclipse.
   A survey done by researchers at the University of Michigan found that eclipse watchers sought more information about eclipses and the scientific questions involved an average of 16 times in the three months following the event.

Several research groups are planning observations for the next total eclipses, visible in South America in July 2019 and December 2020 (SN: 8/5/17, p. 32). Caspi and Samra’s teams both hope to fly through those eclipses in aircraft again.

And amateurs and pros alike are preparing for the Great American Eclipse version 2.0, which will cross from Texas to Maine in 2024.

“Everybody’s eyes are on 2024,” Caspi says.

Particle accelerator technology has crested a new wave.

For the first time, scientists have shown that electrons can gain energy by surfing waves kicked up by protons shot through plasma. In the future, the technique might help produce electron beams at higher energies than currently possible, in order to investigate the inner workings of subatomic particles.

Standard particle accelerators rely on radiofrequency cavities, metallic chambers that create oscillating electromagnetic fields to push particles along. With the plasma wave demonstration, “we’re trying to develop a new kind of accelerator technology,” says physicist Allen Caldwell of the Max Planck Institute for Physics in Munich. Caldwell is a spokesperson of the AWAKE collaboration, which reported the results August 29 in Nature.
In an experiment at the particle physics lab CERN in Geneva, the researchers sent beams of high-energy protons through a plasma, a state of matter in which electrons and positively charged atoms called ions comingle. The protons set the plasma’s electrons jiggling, creating waves that accelerated additional electrons injected into the plasma. In the study, the injected electrons reached energies of up to 2 billion electron volts over a distance of 10 meters.

“It’s a beautiful result and an important first step,” says Mark Hogan, a physicist at SLAC National Accelerator Laboratory in Menlo Park, Calif., who studies plasma wave accelerators.

Previously, scientists have demonstrated the potential of plasma accelerators by speeding up electrons using waves set off by a laser or by another beam of electrons, instead of protons (SN: 5/8/10, p. 28). But proton beams can carry more energy than laser or electron beams, so electrons accelerated by protons’ plasma waves may be able to reach higher energies in a single burst of acceleration.
The new result, however, doesn’t yet match the energies produced in previous plasma accelerators. Instead, the study is just a first step, a proof of principle that shows that proton beams can be used in plasma wave accelerators.

High-energy electrons are particularly useful for particle physics because they are elementary particles — they have no smaller constituents. Protons, on the other hand, are made up of a sea of quarks, resulting in messier collisions. And because each quark carries a small part of the proton’s total energy, only a fraction of that energy goes into a collision. Electrons, however, put all their oomph into each smashup.

But electrons are hard to accelerate directly: If put in an accelerator ring, they rapidly bleed off energy as they circle, unlike protons. So AWAKE starts with accelerated protons, using them to get electrons up to speed.

Prior to the experiment, there was skepticism over whether the plasma could be controlled well enough for an effort like AWAKE to work, says physicist Wim Leemans of Lawrence Berkeley National Laboratory in California, who works on laser plasma accelerators. “This is very rewarding to see that, yes, the plasma technology has advanced.”

We now have the most precise estimates for the strength of gravity yet.

Two experiments measuring the tiny gravitational attraction between objects in a lab have measured Newton’s gravitational constant, or Big G, with an uncertainty of only about 0.00116 percent. Until now, the smallest margin of uncertainty for any G measurement has been 0.00137 percent.

The new set of G values, reported in the Aug. 30 Nature, is not the final word on G. The two values disagree slightly, and they don’t explain why previous G-measuring experiments have produced such a wide spread of estimates (SN Online: 4/30/15). Still, researchers may be able to use the new values, along with other estimates of G, to discover why measurements for this key fundamental constant are so finicky — and perhaps pin down the strength of gravity once and for all.
The exact value of G, which relates mass and distance to the force of gravity in Newton’s law of universal gravitation, has eluded scientists for centuries. That’s because the gravitational attraction between a pair of objects in a lab experiment is extremely small and susceptible to the gravitational influence of other nearby objects, often leaving researchers with high uncertainty about their measurements.
The current accepted value for G, based on measurements from the last 40 years, is 6.67408 × 10−11 meters cubed per kilogram per square second. That figure is saddled with an uncertainty of 0.0047 percent, making it thousands of times more imprecise than other fundamental constants — unchanging, universal values such as the charge of an electron or the speed of light (SN: 11/12/16, p. 24). The cloud of uncertainty surrounding G limits how well researchers can determine the masses of celestial objects and the values of other constants that are based on G (SN: 4/23/11, p. 28).
Physicist Shan-Qing Yang of Huazhong University of Science and Technology in Wuhan, China, and colleagues measured G using two instruments called torsion pendulums. Each device contains a metal-coated silica plate suspended by a thin wire and surrounded by steel spheres. The gravitational attraction between the plate and the spheres causes the plate to rotate on the wire toward the spheres.

But the two torsion pendulums had slightly differently setups to accommodate two ways of measuring G. With one torsion pendulum, the researchers measured G by monitoring the twist of the wire as the plate angled itself toward the spheres. The other torsion pendulum was rigged so that the metal plate dangled from a turntable, which spun to prevent the wire from twisting. With that torsion pendulum, the researchers measured G by tracking the turntable’s rotation.

To make their measurements as precise as possible, the researchers corrected for a long list of tiny disturbances, from slight variations in the density of materials used to make the torsion pendulums to seismic vibrations from earthquakes across the globe. “It’s amazing how much work went into this,” says Stephan Schlamminger, a physicist at the National Institute of Standards and Technology in Gaithersburg, Md., whose commentary on the study appears in the same issue of Nature. Conducting such a painstaking set of experiments “is like a piece of art.”

These torsion pendulum experiments yielded G values of 6.674184 × 10−11 and 6.674484 × 10−11 meters cubed per kilogram per square second, both with an uncertainty of about 0.00116 percent.

This record precision is “a fantastic accomplishment,” says Clive Speake, a physicist at the University of Birmingham in England not involved in the work, but the true value of G “is still a mystery.” Repeating these and other past experiments to identify previously unknown sources of uncertainty, or designing new G–measuring techniques, may help reveal why estimates for this key fundamental constant continue to disagree, he says.