Some knowledge can still be hard-won, even in the Information Age. Yesterday, a scientific expedition released pictures and videos of a lone living specimen of the world’s largest bee: the Wallace’s giant bee, Megachile pluto, which they had dug out of its nest in a termite mound up a tree in a tropical forest in Indonesia. It had giant jaws and was undeniably impressively large, for a bee—but how large was it?
The many, many stories about the bee contained one or both of two facts, or factoids, about its size. It was “the size of a human thumb,” as the New York Times put it, and it was “four times larger than a European honeybee,” as NPR wrote it in a picture caption.
Setting aside the matho-grammatical shibboleth against using “four times larger” to mean “four times the size of”—is a honeybee really one-quarter the size of a person’s thumb, in your experience? Honeybees seem smaller than that. Which, if either, of those two sizes was the bee?
The supporting visuals just made things more confusing. One tweet about the Wallace’s giant bee called it “the length of a human thumb” right above an embedded video of the bee, captive in a plastic vial that was being held in someone’s hand—a hand whose thumb was obviously longer than the bee’s body.
Meanwhile, publications kept using a composite picture that the expedition had put out, supposedly to illustrate the bee’s size in comparison to the honeybee. In that picture, the big bee was obviously many, many more than four times the size of the little bee. It would have taken two little bees just to fit into its immense head.
This is one of the great troubles with mainstream journalism about science: even more than they do in regular journalism, publications tend to receive and repeat whatever sources have told them (or whatever they think sources have told them) without stopping to ask if it makes sense. So NPR put “four times larger” underneath a picture of two things whose sizes were clearly not anywhere near being in a 4:1 ratio to each other.
But science is not supposed to be a set of unchallenged and untested pronouncements. It’s supposed to be observable and reproducible. What was the actual relationship among the size of the Wallace’s giant bee, the honeybee, and the human thumb? How big a bee was the big bee, really?
Bee dimensions, it turned out, are one more example of the kind of information that the general-purpose internet does not reliably keep in one place. Wikipedia claimed that the female Megachile pluto, which is larger than the male, is 38 millimeters long, but the citation went to an article that required an active JSTOR account to read. Honeybee sizes were a total mess—there are multiple species of honeybee, all of which are eusocial insects, divided up into different castes with different body sizes. Google’s single-answer algorithm declared that the answer to “honey bee size” was “about 15 mm,” which it had retrieved from a page run by Orkin.
Even if you wanted to take your insect data from an exterminating company, there was no obvious way to get a 4:1 relationship out of 15 mm and 38 mm. My suspicion was that hype had gotten involved. The 15 mm number was at the high end of the honeybee lengths the internet was spitting out, which were mostly in the lower teens. It would be easy enough to round down to 10 mm, and then round the Wallace’s giant bee up to 40 mm, and there: four times the size!
Except that’s not how scaling up an animal works. It’s like the old problem of how Godzilla would collapse under its own weight: every time you double the length of something, while keeping the shape the same, you’re increasing its volume by two cubed. A bee that was four times the length of a honeybee would be 64 times its size. Or, going at it the other way, a bee that was four times the size of a honeybee would only be about 1.6 times its length.
None of these were the relationship in the picture. To sort out the relative sizes, I needed to know the actual dimensions of Megachile pluto, and suitably comparable numbers for a suitably comparable honeybee—for simplicity’s sake, an Apis mellifera worker bee, the most familiar caste of the most common honeybee species.
I wrote to the Bee Lab at the University of Minnesota, asking for help, and Zachary Portman, a research scientist there, obligingly sent along two references. The first was “A Classification of the Bees of the Australian and South Pacific Region,” a 1965 paper from the Bulletin of the American Museum of Natural History by Charles D. Michener, which listed Megachile pluto as being 39 mm in length. The second, a 1984 paper from the Journal of the Kansas Entomological Society, included a photograph of a Wallace’s giant bee side-by-side with an Apis mellifera worker.
I went to compare the relative lengths of the bees in the image and discovered that somehow—after being photographed, developed, printed in the Journal of the Kansas Entomological Society, scanned, converted to pdf, downloaded, emailed, and opened in my browser on my computer screen—they had ended up life-sized. The Wallace’s giant bee was 39 mm long on a ruler held up to the screen, and the honeybee was 13 mm. Across the fattest part of the abdomen, they were 16 mm and 5 mm, respectively.
So the Wallace’s giant bee was three times the length of the honeybee, and a little more than three times the width. There was no way to measure the depth, but in all the other pictures it looked proportionate.
What did that mean about their overall sizes? Another thing that goes wrong—a lot—in science journalism is that people get so attached to the numbers they see on a calculator that they forget about significant digits and end up with spuriously precise results. I was doing a rough estimate, with the help of a ruler, so the relevant number was “three-ish.” A length of three-ish, times a width of three-ish, times a depth of three-ish equals: twenty-seven-ish? Nope. Equals “around 30.”
The Wallace’s giant bee is around 30 times the size of a honeybee, not four times.
And the human thumb? If the bee is as deep as it is wide, it’s about 10,000 cubic millimeters. The ruler said my thumb is 70 mm from the tip to the carpometacarpal joint, and 20 mm across the knuckle, and not quite as thick as it is wide. Call it 25,000 cubic millimeters. I’m six feet tall, so allowing for some overage, it seems safe to say the Wallace’s giant bee is maybe half the size of a human thumb.
How did the thumb thing get into the record? Clay Bolt, the photographer on the expedition, wrote an article about finding the bee in which he cited the bee’s discoverer, Alfred Russel Wallace, the 19th century British naturalist best known for having come up with the theory of evolution by natural selection independently of Charles Darwin:
Wallace spends little time on the bee—one of 1,000 new species he had also discovered. But he described the female that a local brought to him as, “…about as long as an adult human’s thumb,” and as “a large black wasp-like insect, with immense jaws like a stag-beetle.”
In a photograph taken within a few years of his finding the bee, Wallace (standing beside a chair, for reference) appears to be, if anything, unusually tall, and his thumbs seem to be normally proportioned. Maybe he was assuming that other people’s thumbs were half the size of his own, or maybe he was being careless. Here’s a science lesson: more than 150 years later, people are still copying down his estimate as fact, ignoring the evidence of their own thumbs.