I meant to cover this paper when Tyler Cowen and Kevin Drum discussed this, but, now that the NY Times has raised the topic, I can be topically again (it just goes to show that the internet repeats itself). Anyway, from the NY Times (boldface mine):

A first warning sign comes from looking at other animals. Our pets have been getting fatter along with us. In 2012, some 58.3 percent of cats were, literally, fat cats. That is taken from a survey by the Association for Pet Obesity Prevention. (The very existence of this organization is telling.) Pet obesity, however, can easily be tied to human behavior: a culture that eats more probably feeds its animals more, too.

And yet, a study by a group of biostatisticians in the Proceedings of the Royal Society challenges this interpretation. They collected data from animals raised in captivity: macaques, marmosets, chimpanzees, vervets, lab rats and mice. The data came from labs and centers and spanned several decades. These captive animals are also becoming fatter: weight gain for female lab mice, for example, came out to 11.8 percent a decade from 1982 to 2003.

But this weight gain is harder to explain. Captive animals are fed carefully controlled diets, which the researchers argue have not changed for decades. Animal obesity cannot be explained through eating behavior alone. We must look to some other — biological — driver.

As you might guess, the microbiome is implicated as a culprit. Look, I would love it if there were a probiotic that would let me eat lots of junk food (who wouldn’t?). But the reasons for the increases–note the plural–are probably different, even though that doesn’t make for a very good story.

Regarding the lab mice, a decade is ~33 generations, and that’s more than enough time for selection to increase body size. It would be relatively trivial to increase the average weight of a population of dogs from 60 lbs to 68 lbs in thirty plus generations (this is what dog breeders do). So then the issue is what would select for larger mice? Several possibilities come to mind (#3 seems to be the most straightforward):

1) Mice in captivity (especially males) can be incredibly aggressive towards each other. That’s before you get to the infanticide. Being bigger certainly would help.

2) Larger body size often is correlated with higher fecundity. Depending on how one propagates the colony, this could be an issue.

3) Researchers probably won’t pick the ‘runty’ ones, which skews the distribution. For mouse colony propagation, ‘stud’ mice (males) are typically chosen. If there is a researcher bias towards picking the big ones…

This isn’t to say the microbiome couldn’t be involved (hard to rule out and not mutually exclusive), but we have many examples across a wide range of animals that moderate phenotypic change can occur in ‘observable’ time, when there’s strong selection–there are subdisciplines in microbiology and insect biology that focus on this very question. Considering animal breeders have bred significant changes over short time spans, it seems likely that the lab setting might be doing it accidentally.

I’m not sure how this could apply to people though, so I’m inclined to think that lab animal and human increases are merely coincidental.