Then, starting in the 1970s, a tiny aphid-like insect known as hemlock woolly adelgid, originally from Japan, unleashed the tree version of a pandemic in American hemlock forests. The adelgid, recognizable by cotton-like fuzz it produces while feeding on hemlock needles, has killed millions of trees and upended ecosystems throughout the eastern United States. Having turned much of Appalachia and New England into tree graveyards, the insect reached the Eastern Shore of Lake Michigan by 2016 and threatens to continue its death march through the upper Midwest.

Many scientists and foresters wrote off the hemlock as a lost cause. But a few wondered whether rare combinations of adelgid-resistance genes might lurk in the trees. Those scientists sought, propagated and planted cuttings from trees that remained green when their neighbors had become gray ghosts.

Researchers have reported that these trees survive better and grow faster than nonresistant ones. The result could mark a steppingstone toward a potential hemlock comeback.

When hemlocks started dying in large numbers in the late 1970s and early 1980s, forest experts knew they had a problem. The tree hosts dozens of insects and birds such as the blue-headed vireo and hermit thrush, and its year-round shade keeps mountain streams cool enough for trout. Scientists started looking for ways to keep hemlocks around.

One strategy involved looking for rare hemlocks that appeared to tolerate the adelgid. In the mid-2000s, a New Jersey state entomologist surveying near Delaware Water Gap found a stand of lush, green hemlocks amid gray skeletons. University scientists cloned some cuttings from what they called “bulletproof” trees and, in 2015, planted them in test plots near other hemlocks that were infested with adelgids. Four years later, the researchers returned to each plot and assessed the trees.

Ninety-six percent of the clones from the cuttings had survived, compared with 48 percent for other hemlocks. Bulletproof clones were taller and had more foliage and insect-resisting chemicals called terpenes, the researchers reported in a paper published in May in the journal Forests. That’s even more impressive, says Evan Preisser, a University of Rhode Island ecologist who led the research, because the scientists did nothing to help the trees during those four years.

Because of the study’s small sample size — just eight resistant and four nonresistant trees at each site — Preisser calls it a “proof of concept” that adelgid-resistant trees can be found and propagated.

But some scientists are skeptical that genetics can rescue the hemlock.

Rusty Rhea, an entomologist with the U.S. Forest Service in Asheville, N.C., believes the bulletproof trees may have fared better because of environmental factors where they were growing, and that Preisser’s study period was too short to prove the trees could survive in the forest.

“I’m a little hesitant to give people hope that there is resistance based on . . . propagules that are only four or five years old,” Rhea says.

Rhea’s agency, the main source of funding for forest-related research in the United States, has taken a different approach.

The service has backed a long-term effort to identify predatory insects that could limit adelgid numbers on a permanent basis, similar to, for example, how wolves control elk populations in Yellowstone National Park. This “biological control” strategy has gone awry before, but when done carefully, it can reduce populations of damaging pests — for example, introduced insects have aided California’s citrus industry. But biocontrol has rarely succeeded in natural forests.

Starting in the 1980s, scientists searched Japan, where the adelgid came from, and the West Coast of North America, where adelgids also live, for insects that munch on them. Researchers have spent more than 20 years testing a beetle called Laricobius nigrinus, native to British Columbia. In April, scientists reported that in sites in the eastern United States where they released Laricobius, the beetles damaged around a third of the adelgid egg sacs laid in the winter. As a result of the feeding, fewer adelgids emerged in the spring at these sites.

But that’s hardly cause for celebration, because the adelgid has two life cycles per year. A follow-up paper published in June by the same authors reported that adelgid numbers at the study sites rebounded in the summer, thanks to the insect’s high reproductive rate; one adult female can produce up to 500 offspring.

Biocontrol researchers are now pinning hope on two tiny silverflies, both native to the western United States, to control the summer adelgid generation. The flies have been shown to reduce adelgid numbers in bags tied onto tree branches. But when released in the open, the flies have tended to disappear rather than attain large enough numbers to control adelgids. Still, Rhea says, the flies are “the best bet we know of right now.”

Forest Service researchers have also found that hemlocks in full sun grow faster than in shade, even with adelgids feeding on them. Rhea hopes that growing hemlocks in forest clearings and deploying beetles and flies will eventually allow trees to reach maturity and reproduce before the adelgid kills them.

To Preisser, however, the ongoing struggle to establish sustainable adelgid predators population in the forest suggests biocontrol may be a dead end. He advocates seeking stands of adelgid-resistant trees adapted to different locations, growing them in large-scale demonstration plots to determine which genes provide the most adelgid resistance and breeding the best performing trees together, an effort that Preisser concedes will be expensive and time-consuming.

Even with modern practices, a hemlock may need to grow for seven to 10 years before its seeds can be collected, says Ben Smith, a forestry researcher at North Carolina State University’s Mountain Research Station in Waynesville who is developing a hemlock breeding program there. “To crop breeders, that’s an eternity,” Smith says. “For tree breeders, it’s not terrible.”

Preisser is not sticking around to find out. Frustrated by difficulties in persuading funders to support his research, he is moving on to other things. “It’s become clear to me over the past 10 years that science doesn’t make a difference,” he says.

The idea that beleaguered trees’ genomes could hold the keys to their own salvation has indeed encountered resistance from scientists, says Jennifer Koch, a biologist at the Forest Service’s research station in Delaware, Ohio.

It was “a huge hurdle to get people to believe resistance existed,” she says. “Particularly with hemlock, that’s one that people really struggled to believe.”

Research by Preisser and others has laid the groundwork for breeders to cross partially resistant trees to produce ones that can survive to reproductive age in the forest, Koch says. She is working to launch a hemlock breeding program with the nonprofit conservation organization American Forests; it will be hosted at Holden Arboretum in Kirtland, Ohio, and will also include ash and beech trees. New Jersey is setting up its own hemlock breeding program.

Success won’t require finding or producing a completely resistant tree, Koch says. Beetles and flies can also play a role by helping to keep adelgid numbers down and lengthen trees’ lives.

“We’re not looking for immunity. We’re looking for that balance. [The tree] can still be a host to this insect, but . . . the insect population won’t reach a level that’s going to kill the tree,” she says. “You mimic what happens in nature.”