The first American dogs were different, and arrived a very long time ago

New genetic research using buried dogs found at two sites in Illinois suggests that the first American dogs came over from Siberia over 16,000 years ago, and were genetically distinct from European dogs.

In the 1960s and 1970s, archaeologists excavated two sites in western Illinois, where ancient hunter-gatherers collected shellfish from a nearby river and stalked deer in surrounding forests. These people also appear to have buried their dogs: One was found at a site known as Stilwell II, and four at a site called Koster, curled up in individual gravelike pits.

Radiocarbon analysis of the bones reveals that they are around 10,000 years old, making these canines the oldest dogs known in the Americas, researchers report on the bioRxiv server. It also makes these the oldest solo dog burials anywhere in the world. The Stilwell II dog was about the size of an English setter, whereas the Koster dogs were smaller and slenderer, says the study’s lead author, Angela Perri, a zooarchaeologist at Durham University in the United Kingdom. “It wouldn’t be surprising if they were all used as hunting dogs.” But where did they come from in the first place?

A second study, published today in Science, may have the answer. A large, international team of researchers sequenced DNA from the mitochondria, or cellular power plants, of 71 North American and Siberian dog bones—including from one of the Koster dogs—dated from about 10,000 to 1000 years ago. When they compared this material, which is passed down only by the mother, to that of 145 modern and ancient dogs, they discovered that the ancient American dogs have a genetic signature not found in any other canines.

There is much uncertainty still about the dates, but not about the genetics. The dogs were larger, resembles wolves, and even howled instead of barked. They were wiped out after the arrival of the Europeans in American, probably because of disease.

North American mountains get 3Xs the snow previously estimated

The uncertainty of science: A new analysis based on computer models suggests that the mountains of North America get three times more snow each year than scientists has previously estimated.

Those figures come thanks to a new analysis in which researchers used computer simulations to estimate the typical annual snowfall in each of 11 North American mountain ranges. After supercomputer simulations of regional climate that would have taken 50 years on the average laptop, the team found that those mountain ranges receive about 3018 cubic kilometers of snow a year. Although those ranges together cover only about 25% of the area stretching from the Arctic Ocean down to Mexico’s southern border, they get about 60% of its snow, the researchers report in Geophysical Research Letters. That’s nearly three times the estimate for mountain snow from one previous study, the team notes.

First, this is based on computer simulations, not actual data in the field. I wouldn’t put much money on it. Second, it does show us how little climate scientists really know about the climate, as this simulation is still using all the knowledge they have, and it comes up with a conclusion that confounds them. Third, I was astonished the article didn’t try to push the idea that this larger estimate should be blamed on human-caused global warming. It didn’t, which I suppose is a sign of some progress.