Bat Milk: Ranking mammals’ milk

Generally speaking, adult humans are reluctant to consume bodily secretions. I hesitate to over-generalise, but I feel confident in saying that most people don’t hanker after a glass of saliva or spread ear wax on their toast. But we do, in most parts of the world, drink vast quantities of milk. Or eat it as cheese. Or ice cream! We’ve found a thousand different ways of consuming the stuff. Not human milk, of course. Once we’re weaned we’re not supposed to go back to the stuff. And even for children, Western cultures are very picky these days about whose milk you’re allowed to drink. Women from many cultures will quite happily breastfeed any hungry child (of appropriate age), but in Western societies many of us are squeamish about the idea. We’re a bit suspicious of milk. Cow milk is definitely ok, as is sheep and goat, usually, along with a few others. But we mostly don’t like to think too much about how we get the stuff or where it comes from. I think we all feel a little icky about the whole secretion-drinking idea, so we’ve made a lot of unwritten rules about it so we can reassure ourselves that it’s all ok.

I thought it would be interesting to delve into this a little bit, so I employed some of my (extremely rudimentary) web development skills to make a website which would collect some data. I give you www.batmilk.net.

Here’s how it works: Participants are presented with the image of two glasses of milk, each labelled as coming from a different, randomly chosen mammal. They are told they must drink one of the glasses of milk, but they can choose which. Participants choose a glass, and then the website presents them with two more random mammals’ milk to choose between. When participants grow tired of drinking hypothetical glasses of milk, they can go to a page which shows tabulated results of all previous votes. Milks are ranked in order of the proportion of matched contests that they win. For example, llama milk wins 85% of its match-ups, while horse milk wins 71%.

This methodology was inspired by Fivethirtyeight’s “Halloween Candy Power Ranking”, which is also a fascinating study. I shared my work on Twitter, and garnered a few thousand votes (and several disgusted reactions) over the next couple of weeks. The methodology doesn’t allow me to easily tell how many individual people participated, by my estimates are that it was one or two hundred people.

Here’s the chart of milk results.

As you might expect, some strong preferences emerged. Llama, goat, sheep, cow, and camel were the top ranked milks, winning between 82% and 85% of their match-ups. Goat, sheep, and cow are good traditional picks for the mostly white, western audience I’d expect to reach from my Twitter account. Llama and camel are a bit more surprising, especially llama, which was the top-scoring milk. I can only speculate, but I believe that llamas thread the needle between oddity and novelty. It’s different enough that participants whose choices were motivated by a desire to “try something new” might pick it over, say, cow or goat, but it’s also less unusual than many of the other mammals featured, which means that people with more traditional tastes will reliably pick it more often than those others as well.

At the other end of the scale, bats, mice, weasels and especially rats were considered to have especially undesirable milk. These results are less surprising, perhaps. As well as being traditionally considered unclean, these are all very small animals, which makes the thought of extracting a whole glass of milk out of them a bit disturbing.

The big surprise for me on this list was actually human milk, which rated near the bottom at only 29%. That’s only slightly better than mole milk, and worse than platypus milk, which doesn’t even come from nipples but rather is secreted from sweat glands on the chest. Given that almost every one of us has drunk our share of human milk in our early life, it seems bizarre to be put off by drinking the one kind of milk that’s actually intended for us. When I talked to some participants, their answers about this question were interesting to me. When considering drinking human milk, they were suddenly much more concerned with the specific individual that the milk came from, and the circumstances under which it was extracted.

Another comment from one of the participants struck me. She said “I always vote for the larger animal”. That struck me as significant. To what extent do the results reflect simply the size of the animal? I plotted the relationship between the mammals’ scores, and their (approximate) adult weight. I excluded whales and elephants since they were large outliers in this data set. I also ended up taking the natural logarithm of the weights, which worked better for comparing a cow with a mouse while also preserving the difference between a mouse and a squirrel. That produced this chart.

As you can see, there’s a relationship. It’s not a strong correlation, but it is notable. Cows, as one of the heaviest mammals, also have one of the most popular milks. Rats, one of the smallest, have the least popular milks. Measuring this relationship allows me to construct a “weight adjusted score”, a measure of the difference between the mammal’s actual score, and the score we’d expect them to get, given their weight.

This gives some interesting results. At the low end humans, dolphins and hyenas feature. We’ve already discussed human milk, but dolphins are interesting — as a marine mammal, it makes sense that their milk is considered less palatable than a terrestrial mammal of equivalent weight. Hyenas have an unsavoury reputation, which seems to flavour their milk. Dominating the top end of the field though is the humble squirrel. For reasons I don’t entirely understand, squirrel milk is vastly more popular than you’d expect, given their size. Maybe it’s that squirrels are generally viewed pretty positively — happy-go-lucky tree dwellers — especially compared to the rats and weasels which are a similar weight. Or maybe it’s simply squirrels’ diets — squirrel milk is just nut milk with extra steps.

So what have I learned? I feel like the theme that emerged in this data is the relationship between the perceived palatability of a mammal’s milk, and the perceived “cleanliness” of that mammal. Squirrels, rabbits, otters and kangaroos, which all have surprisingly desirable milk, are also animals which, at least in Western cultures, we tend to think of as “good” animals — friendly and cute. More “unclean” animals, like weasels and rats, marine mammals, like seals and dolphins, and oddballs like pangolins and anteaters also have less desirable milk. Perhaps it shouldn’t be surprising that the animals whose milk we prefer to drink are also the animals we’d like to cuddle. Probably the biggest take-away from this analysis is that there’s a huge and currently untapped market for squirrel milk. I welcome offers of VC funding.