If lithium (or something) is responsible for obesity, we should see low levels near lean cities and high levels near obese cities
In the least obese communities, we found little evidence of lithium contamination. We were able to find drinking water measurements for five of these communities, and in all five, the level of lithium in their drinking water was quite low, with averages around 2 ng/mL.
In the most obese communities, we found abundant evidence of lithium contamination
Coal plants maintain groundwater testing wells to monitor contaminants, including lithium, and many of the most obese communities were downriver of plants with testing wells showing lithium levels above 100 ng/mL
We're still surprised at how well this lines up. We didn't expect cities to matter very much — but based on what we've found, there looks like a pretty strong relationship between obesity rates and lithium contamination in the local water