BOSTON — Some aspects of speech are as Southern as pecan pie. Consider the vowel shift that makes the word pie sound more like “pah.” While that pronunciation is found from Florida to Texas, a new study reveals a surprising diversity in Southern vowel pronunciation that’s linked to a speaker’s age, social class, gender, race and geography.

The research, presented June 29 at a meeting of the Acoustical Society of America, could help software developers create better speech recognition tools for smartphones and other devices.
To understand the medley of southern vowel sounds, linguist Margaret Renwick of the University of Georgia in Athens dove into the Digital Archive of Southern Speech. The archive comprises almost 400 hours of interviews with 64 native Southerners representing a mix of ethnicities, social classes, education levels and ages.

Renwick’s analysis of more than 300,000 vowel sounds finds, for example, that Southern upper middle class women are often at the extreme end of variation in pronunciation. While Southern men and women are equally likely to shift the vowel in bet to “bay-ut,” upper middle class Southern women are more likely to stretch the vowel sound in bit to “bee-ut.” They are also most likely to pronounce bait as bite. The finding that women are more inclined to draw a sound out into two syllables — or change it entirely — is in line with other research suggesting that women are linguistic innovators, and less likely to adhere to the norms of standard American English, Renwick said.

With a final rip, an iceberg roughly the size of Delaware has broken off Antarctica’s Larsen C ice shelf. Anticipated for weeks, the fracture is one of the largest calving events ever recorded.

On July 12, satellite images confirmed a nearly 5,800-square-kilometer, 1-trillion-metric-ton chunk of ice, equivalent to 12 percent of Larsen C’s total area, split from the ice shelf. “[We] have been surprised how long it took for the rift to break through the final few kilometers of ice,” Adrian Luckman, a glaciologist at Swansea University in Wales, said in a blogpost for Project MIDAS, which has been tracking the effects of a warming climate on the ice shelf. Now the focus will shift to the stability of the remaining ice shelf and the fate of the giant iceberg.
Scientists had been monitoring Larsen C since 2014, when they noticed that a crack in the ice shelf had grown roughly 20 kilometers in less than nine months (SN: 7/25/15, p. 8). After a relative lull in 2015, the crack grew another 40 kilometers in 2016 and then 10 more in the first half of January 2017, bringing its total length to 175 kilometers. At that point, its tip was 20 kilometers from the Weddell Sea.
The crack grew another 17 kilometers between May 25 and May 31 — at times traveling parallel to the edge and ultimately putting it within 13 kilometers of the ice front. Then, in late June, the outer part of the ice shelf picked up speed, putting new pressure on the crack and the entire shelf. “It won’t be long now,” Project Midas tweeted June 30. Added Luckman, also in a tweet: “The remaining ice is strained near to breaking point.”

Yet the vigil lasted nearly two more weeks. By July 6, the crack had come within 5 kilometers of the edge of the ice. Then, sometime between July 10 and July 12, it finally reached the water, allowing the huge hunk of ice to splinter off into the sea.
The ice loss dramatically alters the landscape of Larsen C, Luckman notes. “Maps will need to be redrawn.” And that could be the least of the trouble ahead, says Adam Booth, a geophysicist at the University of Leeds in England also with Project MIDAS. “The calving event is significant because it is likely a precursor to something much bigger, potentially the collapse of the whole Larsen C ice shelf,” Booth says. The same thing happened to the neighboring Larsen B ice shelf in 2002, after it calved a Rhode Island-sized iceberg (SN: 3/30/02, p. 197).

“Glaciologists are keen to see how Larsen C will react,” says Luckman.

A complete collapse of Larsen C could have implications for sea level rise. Antarctica’s ice shelves act as buttresses, helping to slow the flow of the continent’s ice into the ocean. Since these shelves float on the water, calving icebergs don’t directly raise sea level. But calving or the collapse of an ice shelf allows glaciers and ice streams further inland to flow into the ocean, which can contribute to sea level rise.

Calving of icebergs is common, and over several decades, the shelves usually recover to their original size. But in the last two decades, ice shelves have instead continued to lose ice until collapsing, probably as a result of rising temperatures due to climate change, researchers suspect. In 2014, researchers concluded the collapse of Larsen B was the result of warming (SN: 10/18/14, p. 9).

Some computer simulations suggest Larsen C could suffer the same fate, possibly within a few years to decades, Luckman says. Still, the calving event that feeds a potential collapse may be hard to pin on climate change. “Not all ice-related stories have a clear global warming origin,” Luckman notes. Larsen C’s calving, he says, “may simply be a natural event that would have happened regardless of human activity.”
Not everyone is convinced that Larsen C will fall apart completely. Researchers from Europe predict major changes to the shelf would happen only if it loses 55 percent of its ice. At that point, a significant amount of ice could ooze from glaciers into the ocean. Still, understanding what allowed the recent rift to grow and calve will “give us insight regarding other fractures or rifts on the shelf,” says geoscientist Dan McGrath of Colorado State University in Fort Collins. While McGrath says a collapse is “very unlikely,” he adds that “these other dormant rifts are in locations where if they reinitiated, the subsequent calving event would be worrisome for the shelf’s stability.”

Discrepancies in the predictions of Larsen C’s fate raise an important point, says Richard Alley, a geologist at Penn State. Researchers don’t understand ice shelf calving and collapsing enough predict how any one individual ice shelf will behave after a break.

“The Larsen C ice shelf is, of course, just one small part of Antarctica,” Booth says. “What is worrying is that we’re seeing trends in several ice shelves that tend towards decreasing stability. Should they continue along these trends, we could be seeing the start of increased mass loss from the Antarctic continent.”

A genetic-engineering tool designed to spread through a population like wildfire — eradicating disease and even whole invasive species — might be more easily thwarted than thought.

Resistance to the tools, called CRISPR gene drives, arose at high rates in experiments with Drosophila melanogaster fruit flies, researchers at Cornell University report July 20 in PLOS Genetics. Rates of resistance varied among strains of fruit flies collected around the world, from a low of about 4 percent in embryos from an Ithaca, N.Y., strain to a high of about 56 percent in Tasmanian fruit fly embryos.
“At these rates, the constructs would not start spreading in the population,” says coauthor Philipp Messer, a population geneticist. “It might require quite a bit more work to get a gene drive that works at all.”

Gene drives are basically genetic copy-and-paste machines. These self-perpetuating machines are inherited by more than 50 percent of offspring of an individual carrying a gene drive. Working perfectly, they could transmit to 100 percent of offspring.

In its simplest form, a CRISPR gene drive consists of a piece of DNA that encodes both an enzyme called Cas9, which acts as molecular scissors, and a guide RNA that tells the Cas9 enzyme where to cut. That cutting may disrupt important genes. Researchers are experimenting with this as a way to sterilize malaria-carrying mosquitoes (SN Online: 12/7/15).

Some gene drives also carry a genetic payload. For instance, another approach to fighting malaria is to develop drives that carry genes to “vaccinate” mosquitoes against the disease (SN: 12/26/15, p. 6). Other drives might carry genes that make fluorescent proteins to indicate the gene drive’s presence; Messer and colleagues used such markers to follow two gene drives in fruit flies bred in the lab.
When an organism carrying the tool mates with one that doesn’t, gene drives go to work. Inside the fertilized egg, guide RNAs shepherd Cas9 produced by the engineered mate to a spot where it cuts the other mate’s chromosome.

If everything works correctly, cells repair that break by copying the gene drive onto the cut chromosome. But the slice can also be fixed by gluing the cut ends back together. That regluing sometimes leads to mistakes that destroy Cas9’s cutting site, creating a chromosome that is resistant to the gene drive’s insertion.

In the fruit fly experiments, some mistakes created resistance during or before fertilization. Others took place in early embryos because cells produced Cas9 for too long, allowing the enzyme to chop chromosomes again and again, Messer and colleagues discovered. That was especially a problem when females produced Cas9, they found.

Some uses of gene drives, such as those that would sterilize or kill mosquitoes, can’t tolerate any amount of resistance no matter when it arises, Messer says. Because those types of gene drives damage the organism’s fertility or viability, mosquitoes carrying resistance would have an advantage and quickly outcompete insects vulnerable to the drives.

In a separate study posted June 14 at BioRxiv.org, Messer and colleagues tested several approaches to overcoming gene drive resistance. They found that using multiple guide RNAs and turning on Cas9 only in males could reduce resistance rates.

“This is a very important and elegant set of experiments,” says MIT evolutionary engineer Kevin Esvelt.

But the conclusions aren’t news to most gene drive researchers.

“We’re aware of all these problems, and the essence of how to deal with them hasn’t been changed by these studies,” says geneticist Ethan Bier of the University of California, San Diego. Bier and lab colleague Valentino Gantz created the first gene drive in fruit flies in 2015, and have worked with other researchers to develop gene drives that would prevent mosquitoes from carrying malaria (SN: 12/12/15, p. 16).

Messer’s group is, however, the first to experimentally confirm predictions about resistance and how to avoid it, Esvelt says. “They show what’s been apparent to some people in the field for a very long time.”

Some people might think that high rates of resistance mean that gene drives are safe to release because they won’t spread easily in the wild. But that notion is misguided, says Bier. Even if a gene drive is able to affect only a small percentage of a local pest population, it could still spread around the world, Esvelt adds. “It could still screw us all over in the current form.”

Researchers should continue to conduct gene drive experiments under tight containment, he and Bier caution.

For a glue that holds up inside the body, turn to the humble slug, Arion subfuscus. A new super-sticky material mimics slug slime’s ability to stick on slick wet surfaces and could lead to more effective medical adhesives.

The material has two parts: a sticky layer that attaches to a surface, and a shock-absorbing layer that reduces strain. That makes the adhesive less likely to snap off.

Researchers tested the material as a surgical adhesive in a number of different scenarios: It stuck to pig skin and liver. It latched on to a beating pig’s heart, even when the surface was coated in blood. It sealed up a heart defect, preventing liquid from leaking even when the organ was inflated and deflated tens of thousands of times. And it was less toxic in the body than a commonly used commercialized tissue adhesive, researchers report July 28 in Science.

The researchers hope the material could someday be used in surgical procedures in place of invasive sutures and staples.

The company mice keep can change their behavior. In some ways, genetically normal littermates behave like mice that carry an autism-related mutation, despite not having the mutation themselves, scientists report.

The results, published July 31 in eNeuro, suggest that the social environment influences behavior in complex and important ways, says neuroscientist Alice Luo Clayton of the Simons Foundation Autism Research Initiative in New York City. The finding comes from looking past the mutated mice to their nonmutated littermates, which are usually not a subject of scrutiny. “People almost never look at it from that direction,” says Clayton, who wasn’t involved in the study.
Researchers initially planned to investigate the social behavior of mice that carried a mutation found in some people with autism. Studying nonmutated mice wasn’t part of the plan. “We stumbled into this,” says study coauthor Stéphane Baudouin, a neurobiologist at Cardiff University in Wales.

Baudouin and colleagues studied groups of mice that had been genetically modified to lack neuroligin-3, a gene that is mutated in some people with autism. Without the gene, the mice didn’t have Neuroligin-3 in their brains, a protein that helps nerve cells communicate. Along with other behavioral quirks, these mice didn’t show interest in sniffing other mice, as expected. But Baudouin noticed that the behavior of the nonmutated control mice who lived with the neuroligin-3 mutants also seemed off. He suspected that the behavior of the mutated mice might be to blame.

Experiments confirmed this hunch. Usually, mice form strong social hierarchies, with the most aggressive and vocal males at the top. But in mixed groups of mutated and genetically normal male mice, there was no social hierarchy. “It’s flat,” Baudouin says.

Raised and housed together, the mutated and nonmutated mice all had less testosterone than nonmutated mice raised in genetically similar groups. The testosterone levels in both types of mice were comparable to those found in females — “one of the strongest and most surprising results,” Baudouin says.

The mice’s social curiosity was lacking, too. Usually, mice are interested in the smells of others, and will spend lots of time sniffing a cotton swab that had been swiped across the bedding of unfamiliar mice. But when given a choice of strange mouse scent or banana scent, the nonmutated littermates spent just as much time sniffing banana as did the mutant mice.
When Baudouin and colleagues added back the missing Neuroligin-3 protein to parts of the mutant mice’s brains, aspects of their behavior normalized. The mice became interested in the odor from another mouse’s bedding, for instance. These behaviors also shifted in the mice’s nonmutated littermates. That experiment suggests that the missing protein — and the resulting abnormal behavior of the mutants — was to blame for their littermates’ abnormal actions.

Still, it’s hard to tease apart the mice’s roles, says behavioral neuroscientist Mu Yang of Columbia University. “It is a shared environment, and there is no sure way to tell who is influencing whom, or whether both parties are being impacted.”

Female mice that completely lacked the neuroligin-3 gene also influenced the behaviors of littermates that carried one mutated version of the gene, other behavior tests revealed. More experiments are needed to determine whether the social environment affects male and female mice differently, and if so, whether those differences relate to autism, says Luo Clayton.

For the first time in the United States, researchers have used gene editing to repair a mutation in human embryos.

Molecular scissors known as CRISPR/Cas9 corrected a gene defect that can lead to heart failure. The gene editor fixed the mutation in about 72 percent of tested embryos, researchers report August 2 in Nature. That repair rate is much higher than expected. Work with skin cells reprogrammed to mimic embryos had suggested the mutation would be repaired in fewer than 30 percent of cells.
In addition, the researchers discovered a technical advance that may limit the production of patchwork embryos that aren’t fully edited. That’s important if CRISPR/Cas9 will ever be used to prevent genetic diseases, says study coauthor Shoukhrat Mitalipov, a reproductive and developmental biologist at Oregon Health & Science University in Portland. If even one cell in an early embryo is unedited, “that’s going to screw up the whole process,” says Mitalipov. He worked with colleagues in Oregon, California, Korea and China to develop the embryo-editing methods.

Researchers in other countries have edited human embryos to learn more about early human development or to answer other basic research questions (SN: 4/15/17, p. 16). But Mitalipov and colleagues explicitly conducted the experiments to improve the safety and efficiency of gene editing for eventual clinical trials, which would involve implanting edited embryos into women’s uteruses to establish pregnancy.
In the United States, such clinical trials are effectively banned by a rule that prevents the Food and Drug Administration from reviewing applications for any procedure that would introduce heritable changes in human embryos. Such tinkering with embryo DNA, called germline editing, is controversial because of fears that the technology will be used to create so-called designer babies.

“This paper is not announcing the dawn of the designer baby era,” says R. Alta Charo, a lawyer and bioethicist at the University of Wisconsin Law School in Madison. The researchers have not attempted to add any new genes or change traits, only to correct a disease-causing version of a gene.

In the study, sperm from a man who carries a mutation in the MYBPC3 gene was injected into eggs from women with healthy copies of that gene. Carrying just one mutant copy of the gene causes an inherited heart problem called hypertrophic cardiomyopathy (SN: 9/17/16, p. 8). That condition, which strikes about one in every 500 people worldwide, can cause sudden heart failure. Mutations in the MYBPC3 gene are responsible for about 40 percent of cases. Doctors can treat symptoms of the condition, but there is no cure.

Along with the man’s sperm, researchers injected into the egg the DNA-cutting enzyme Cas9 and a piece of RNA to direct the enzyme to snip the mutant copy of the gene. Another piece of DNA was also injected into the egg. That hunk of DNA was supposed to be a template that the fertilized egg could use to repair the breach made by Cas9. Instead, embryos used the mother’s healthy copy of the gene to repair the cut.

Embryos’ self-healing DNA came as a surprise, because gene editing in other types of cells usually requires an external template, Mitalipov says. The discovery could mean that it will be difficult for researchers to fix mutations in embryos if neither parent has a healthy copy of the gene. But the finding could be good news for those concerned about designer babies, because embryos may reject attempts to add new traits.

Timing the addition of CRISPR/Cas9 is important, the researchers also discovered. In their first experiments, the team added the gene editor a day after fertilizing the eggs. Of 54 injected embryos, 13 were patchwork, or mosaic, embryos with some repaired and some unrepaired cells. Such mosaic embryos probably arise when the fertilized egg copies its DNA before researchers add Cas9, Mitalipov says.

Injecting Cas9 along with the sperm — before an egg had a chance to replicate its DNA — produced only one patchwork embryo. That embryo had repaired the mutation in all its cells, but some cells used the mother’s DNA for repair while others used the template supplied by the researchers.

None of the tested embryos showed any signs that Cas9 was cutting where it shouldn’t be. “Off-target” cutting has been a safety concern with the gene editor because of the possibility of creating new DNA errors.

The study makes progress toward using gene editing to prevent genetic diseases, but there’s still has a long way to go before clinical testing can begin, says Janet Rossant, a developmental biologist at the Hospital for Sick Children and the University of Toronto. “We need to be sure this can be done reproducibly and effectively.”

Famously sneaky particles have been caught behaving in a new way.

For the first time, scientists have detected neutrinos scattering off the nucleus of an atom. The process, predicted more than four decades ago, provides a new way to test fundamental physics. It will also help scientists to better characterize the neutrino, a misfit particle that has a tiny mass and interacts so feebly with matter that it can easily sail through the entire Earth.
The detection, reported online August 3 in Science, “has really big implications,” says physicist Janet Conrad of MIT, who was not involved with the research. It fills in a missing piece of the standard model, the theory that explains how particles behave: The model predicts that neutrinos interact with nuclei. And, says Conrad, the discovery “opens up a whole new area of measurements” to further test the standard model’s predictions.

Scientists typically spot neutrinos when they interact with a single proton or neutron. But the new study measures “coherent” scattering, in which a low-energy neutrino interacts with an entire atomic nucleus at once, ricocheting away and causing the nucleus to recoil slightly in response.

“It’s exciting to measure it for the first time,” says physicist Kate Scholberg of Duke University, spokesperson for the collaboration — named COHERENT — that made the new finding.

In the past, neutrino hunters have built enormous detectors to boost their chances of catching a glimpse of the particles — a necessity because the aloof particles interact so rarely. While still rare, coherent neutrino scattering occurs more often than previously detected types of neutrino interactions. That means detectors can be smaller and still catch enough interactions to detect the process. COHERENT’s detector, a crystal of cesium and iodine, weighs only about 15 kilograms. “It’s the first handheld neutrino detector; you can just carry it around,” says physicist Juan Collar of the University of Chicago.

Collar, Scholberg and colleagues installed their detector at the Spallation Neutron Source at Oak Ridge National Laboratory in Tennessee. The facility generates bursts of neutrons and, as a by-product, produces neutrinos at energies that COHERENT’s detector can spot. When a nucleus in the crystal recoils due to a scattering neutrino, a flash of light appears and is captured by a light sensor. The signal of the recoiling nucleus is incredibly subtle — like detecting the motion of a bowling ball when hit by a ping-pong ball — which is why the effect remained undetected until now.
The amount of scattering the researchers saw agreed with the standard model. But such tests are still in their early stages, says physicist Leo Stodolsky of the Max Planck Institute for Physics in Munich, who was not involved with the research. “We’re looking forward to more detailed studies to see if it really is accurately in agreement with the expectations.” Physicists hope to find a place where the standard model breaks down, which could reveal new secrets of the universe. More precise tests may reveal discrepancies, he says. “That would be extremely interesting.”

Measuring coherent neutrino scattering could help scientists understand the processes that occur within exploding stars, or supernovas, which emit huge numbers of neutrinos (SN: 02/18/17, p. 24). The process could be used to detect supernovas as well — if a supernova explodes nearby, scientists could spot its neutrinos scattering off nuclei in their detectors.

Similar scattering might also help scientists detect dark matter, an invisible source of mass that pervades the universe. Dark matter particles could scatter off atomic nuclei just as neutrinos do, causing a recoil. The study indicates that such recoils are detectable — good news since several dark matter experiments are currently attempting to measure recoils of nuclei (SN: 11/12/16, p. 14). But it also suggests a looming problem: As dark matter detectors become more sensitive, neutrinos bouncing off the nuclei will swamp any signs of dark matter.

Coherent neutrino scattering detectors could lead to practical applications as well: Small-scale neutrino detectors could eventually detect neutrinos produced in nuclear reactors to monitor for attempts to develop nuclear weapons, for example.

Physicist Daniel Freedman of MIT, who predicted in 1974 that neutrinos would scatter off nuclei, is pleased that his prediction has finally been confirmed. “It’s a thrill.”

I heard it for the first time a few days ago: “She’s copying me!” my 4-year-old wailed in a righteous complaint about her little sister. And she most certainly was copying, repeating the same nonsense word over and over. While it was distressing to my older kid, I thought it was funny that it took her so long to realize her sister copies almost everything she does.

This egregious violation occurred just after I had read about an experiment that pitted young kids against bonobos in a test to see who might copy other individuals more. I’ll get right to the punch line: Kids won, by a long shot. The results, published online July 24 in Child Development, show that despite imitation annoying older siblings everywhere, it’s actually really important.

“Imitation is one of the most essential skills for being human,” says study coauthor Zanna Clay, a comparative psychologist at the University of Birmingham and Durham University, both in England. Learning how to talk, operating the latest iPhone and figuring out how to buy bulk goods at the local co-op — these skills all rely on imitation. Not only that, but imitation is also important for cementing social relationships. My daughter notwithstanding, “Humans like to be imitated, and we like those who imitate us,” Clay says.
Clay and her colleague Claudio Tennie tested just how strong the urge to imitate is in 77 children ages 3 to 5 and a group of 46 bonobos ages 3 to 29. In one-on-one trials, the researchers sat next to the kids and bonobos with a small wooden box about the size of a hand. Inside was a treat: a sticker for the kids and a bit of apple for the bonobos.

Before opening the box, the researcher performed nonsensical actions over it, either rubbing the box with the back of the hand and doing a wrist twist in the air or tracing a cross into the top of the box and then tracing the edges.

These hand motions were totally irrelevant to the actual opening of the box. Nonetheless, after seeing the gestures, the vast majority of the kids made the same motions before trying to open their own box. Not a single bonobo, though, copied the irrelevant actions.
What the bonobos did — not copying the meaningless gestures — “is the rational thing to do,” says Clay. “Yet the irrational thing that the kids did is part of the reason why human cultures have evolved so rapidly and so diversely.”

Such excessive imitation, called overimitation, is a special form of copying in which people perform actions that clearly serve no purpose. It may be behind rituals, social norms and language that keep our societies running smoothly.

And it may be unique to humans: Other studies have failed to spot overimitation among chimpanzees and orangutans. These findings hint that our powerful urge to imitate even nonsensical gestures may be one of the things that separate humans from other apes.

Speed has its limits — on the open road and the Serengeti. Midsize animals tend to be the speedsters, even though, in theory, the biggest animals should be the fastest. A new analysis that relates speed and body size in 474 species shows that the pattern holds for animals whether they run, fly or swim (see graphs below) and suggests how size becomes a liability.

This relationship between speed and size has long stumped scientists. Big animals have longer legs or flippers to get from point A to point B. And bigger bodies have higher metabolic rates and more fast-twitch muscle cells, needed to convert chemical energy into mechanical energy and rapidly accelerate. So, why aren’t wildebeests faster than cheetahs?
The make-or-break factor is the time it takes an animal to accelerate to its top theoretical speed, an upper limit based on mass and metabolic rate, researchers report July 17 in Nature Ecology & Evolution. Fast-twitch muscle cells provide the power for acceleration but tire quickly. When an animal gets too big, it takes too long to accelerate, and these cells use up their energy before hitting top speeds. More modestly built critters need less time to accelerate to those speeds.

The researchers gathered speed and size data from past lab and field studies. The animals (some shown as icons in the slideshow below) ranged in mass from 30-microgram Spanish mites to a blue whale weighing 108 metric tons.

Maybe it’s more than reptile fashion. The high percentage of citified sea snakes wearing black might be a sign that pollution is an evolutionary force.

Off the coasts of Australia and New Caledonia, some turtle-headed sea snakes (Emydocephalus annulatus) sport pale bands on their dark skins. Others go all black. In 15 places surveyed, the all-black form was more likely to predominate in waters near cities, military sites or industrial zones than along reefs near less built-up coastlines, says evolutionary ecologist Rick Shine of the University of Sydney.
That trend plus some analysis of trace elements in snakes’ skin suggests that the abundant dark forms could turn out to be an example of industrial melanism, Shine and his colleagues propose August 10 in Current Biology.

The most famous example of this evolutionary phenomenon comes from a dark form of peppered moth that overtook pale populations in 19th century England (SN: 6/25/16, p. 6). Dark wings created better camouflage from hungry birds in the grimy industrializing landscape.
Shine doesn’t think the sea snakes are going for camouflage, though. Instead, the snakes could be more like the dark-feathered pigeons of Paris. The melanin that gives that city’s feral birds their urban chic also does a great job of binding traces of toxic metals such as zinc, explains evolutionary ecologist Marion Chatelain of the University of Warsaw. When birds molt, getting rid of darker feathers lets them unload more of the unhealthful urban pollutants, she and colleagues have reported.
This could explain why marine biologist and study coauthor Claire Goiran has so many dark turtle-headed sea snakes in a lagoon not far from her campus, the University of New Caledonia in Nouméa. Earlier studies had found only downsides to dark coloration: Seaweed spores preferentially settle on dark snakes and sprout fuzz that can cut swimming speed by 20 percent and cause a snake to shed its skin more often than normal.
To test a scenario of industrial melanism, or darkening due to pollution, the researchers collected data on skin colors for a total of about 1,450 snakes, both live and museum specimens, from 15 sites in New Caledonia and Australia. Higher percentages of all-dark snakes wriggled around the nine polluted sites surveyed. At one, a remote Australian reef that the military had long used as a bombing range, all 13 specimens were dark.

To test shed skins for trace metals, Goiran and Shine enlisted Paco Bustamante of the University of La Rochelle in France, who studies trace metal contamination in marine life.

Researchers managed to collect sloughed skins from 17 turtle-headed snakes, which inconveniently shed their skin underwater. To compare light and dark patches, the scientists turned to two local species of sea kraits, which have banded skin and visit land to shed it.
Overall, skins held concentrations of trace elements higher than those that can cause health problems in birds and mammals, the researchers report. In the krait skins, dark zones had slightly more of some contaminants, such as zinc and arsenic, than the pale bluish-white bands did.

The idea that polluted water favors melanized sea snakes “is a reasonable hypothesis based on what we know,” Chatelain says. Definitive tests will require more data and different approaches. Genetic testing, for example, would clarify whether dark populations arose instead from small groups of pioneers that happened to have a lot of black snakes.

That testing could be a long way off. Sea snakes are evolutionary cousins of cobras and mambas, and some of the species swimming around Australia and New Caledonia are “bowel-looseningly large,” Shine says. At least the little turtle-headed ones, which eat eggs of small reef fishes, have venom glands that have atrophied and “probably couldn’t fit a human finger in their mouths.” But until someone figures out how to keep them alive in captivity for more than a few days, Shine isn’t expecting definitive genetics.