When honey bees began dying en masse in late 2006, one of the early suspects was a class of insecticides called neonicotinoids. These chemicals are often applied to seeds before planting, so that the poison permeates the entire plant as it grows, including its pollen and nectar. The European Union placed a moratorium on the chemicals even as research results were mixed: at the doses honey bees might experience on a farm, neonicotinoids seemed to cause disorientation and a weakening of the immune system, but nothing to explain the die-offs.

Now, two studies published in Nature indicate the neonicotinoids are a problem for pollinators, though not in the way many first assumed. In the first study, researchers at Lund University looked at 16 fields of oilseed rape, a major source of vegetable oil, in southern Sweden. Half were planted with seeds coated in a neonicotinoid and a fungicide; the other half, the control, had seeds coated with only fungicide. The researchers found that placing bumblebees near the neonicotinoid-treated fields impeded the ability of colonies to grow and reproduce. Solitary bees also failed to reproduce near treated fields. The honey bee colonies, however, showed no signs of negative effects from the insecticide.

"We saw significant influence from treatments on wild bees," Maj Rundlof, the lead author on the study, said at a press briefing. The population density of wild bees — that is, bumblebees and the solitary Osmia bicornis bee — fell by about half near the fields treated with neonicotinoids. Bumblebee colonies near the treated fields grew more slowly and produced fewer queens than colonies near the control fields. At six of the eight control fields, Osmia bees built brood cells to hatch the next generation, but they failed to do so at any of the treated fields.

"To some extent it’s not surprising that an insecticide has negative consequences on insects," said Henrik Smith, another author on the paper. It’s unclear exactly how the neonicotinoids — or neonics, as they’re often called — harmed the bees, but past studies offer clues. Maybe the treatments impaired the Osmia bees’ ability to navigate, the researchers speculate, an effect seen in previous studies of neonicotinoids. The insecticide could also impair the bumblebees’ ability to forage for pollen and care for their brood.

The finding that honey bees didn’t show signs of harm shouldn’t be a complete surprise either. Honey bees live in such large colonies and regenerate so quickly that thousands of bees can die without significantly effecting the hive. Their resilience is part of the reason they’re used for migratory crop pollination, rather than bumblebees, which live in colonies of tens or hundreds, or solitary bees. Previous work suggests honey bees may also be better at detoxifying after neonicotinoid exposure than bumblebees. The researchers also caution that their study shouldn’t be taken as a sign that honey bees aren’t harmed at all: because so much can vary in a field study, anything less than a 20 percent decline in colony size wouldn’t be detected, and there could be longer-term effects.

"This is certainly the most decisive work we’ve seen on neonics at field-relevant doses," says Dennis vanEngelsdorp, an entomologist at the University of Maryland bee laboratory.

Van Engelsdorp was also one of the first researchers to study Colony Collapse Disorder after it emerged in 2006, and when neonics were a prime suspect. Honey bees have continued to die at higher than normal rates since then — about a third of the managed colonies in the country each year — though the mass vanishing symptomatic of CCD is less common. (The total bee population has stayed about the same, a testament to both the resilience of honey bee colonies and to beekeepers’ efforts.) Researchers now believe that a mix of factors are combining to kill the bees, namely a lack of flowering plants due to monocrop agriculture, disease, parasites, and exposure to pesticides. If neonicotinoids are playing a role, in that mix, vanEngelsdorp says, the study suggests that it’s probably minimal. "The really important finding is that levels that aren’t affecting honey bees are affecting native bees."

This finding is troubling in its own right, especially given that honey bees are the model species for measuring potential impacts of new pesticides on pollinators. Neonics represent more than one-fifth of the world’s insecticide market and are common on crops like corn, canola, cotton, and soy. If honey bees are better able to handle pesticide exposure, there’s a risk that bumblebees and other native bees have been dying undetected.

There’s no comprehensive record of non-honey bee populations, but what studies exist show disturbing declines. Nearly a third of the bumblebee species in the United States are declining, according to the Xerces Society. As with the honey bees, it’s not just pesticides: disease and habitat loss also likely play major roles. These bees also pollinate crops and native plants, often more efficiently than honey bees, having evolved in tandem with them.

The second study looks at whether bees can detect, and therefore avoid, neonicotinoid-treated plants. It turns out the bees appear to prefer treated plants.

The study came out of an earlier experiment by Erin Jo Tiedeken at Trinity College, Dublin, designed to measure whether bumblebees can detect toxins in sucrose syrup, a stand-in for nectar. She placed a bumblebee in a plastic box along with two tubes of sucrose syrup, one of which was laced with a neonicotinoid. After waiting 24 hours, the researchers weighed the tubes to see which the bees had eaten from, expecting to find either no difference, if the bees couldn’t detect the toxin, or a preference for the pure syrup, if they could. Instead they found the bee preferred the toxic syrup. "We were both a bit shocked," said Geraldine Wright, another author on the study.

As part of the more comprehensive experiment, the researchers presented the bees with neonicotinoid-laced syrup and found that the neurons on their proboscises didn’t respond as they would if presented with other types of toxins. "I don’t think they can taste it at all," Wright said. Yet the second experiment found that both bumblebees and honey bees prefer syrup with levels of neonicotinoids comparable to what they’d experience in a seed-treated field.

It’s unclear why bees would prefer neonics, especially if they don't seem to taste it, but vanEnglesdorp cautions that it could be dilution itself that creates the attraction, because natural nectar isn’t 100 percent pure sucrose. He agrees that the study shows bees aren’t deterred from consuming neonics.

Another possible explanation for their preference is neurological: maybe, Wright reasoned at the briefing, bees are experiencing low doses of neonicotinoids as rewarding on some level and returning to the toxic food tube. "A little bit is medicine, a lot’s a toxin," Wright said, comparing the insecticide-laced syrup to cigarettes. After all, nicotine is produced by plants as a defense against insects, and neonicotinoids are derived from that; maybe small doses of neonicotinoids function similarly to nicotine in humans, with low doses having a pharmacological effect. Wright said that further research would be needed to find out whether bees actually get addicted to the insecticide.