The Samoan moss spider, the world’s smallest arachnid at a third of a millimeter, is nearly invisible to the human eye. The largest spider in the world is the goliath bird eater tarantula, which weighs 142 grams and is about the size of a dinner plate. For reference, that is about the same difference in scale between that same tarantula and a bottlenose dolphin.
And yet the bigger spider does not act in more complex ways than its tiny counterpart. “Insects and spiders and the like—in terms of absolute size—have among the tiniest brains we’ve come across,” says William Wcislo, a scientist at the Smithsonian Tropical Research Institute in Panama City. “But their behavior, as far as we can see, is as sophisticated as things that have relatively large brains. So then there’s the question: How do they do that?”
No one would argue that a tarantula is as smart as a dolphin or that having a really big brain is not an excellent way to perform complicated tasks. But a growing number of scientists are asking whether it is the only way. Do you need a big brain to hunt elusive prey, design complicated structures or produce complex social dynamics?
For generations scientists have wondered how intelligent creatures developed large brains to perform complicated tasks. But Wcislo is part of a small community of scientists less interested in how brains have grown than how they have shrunk and yet shockingly still perform tasks as well as or better than similar species much larger in size. In other words, it is what scientists call brain miniaturization, not unlike the scaling down in size of the transistors in a computer chip. This research, in fact, may hold clues to innovative design strategies that engineers might incorporate in future generations of computers.
Scientists interested in brain miniaturization often refer to something called Haller’s rule, proposed by German neuroscientist Bernhard Rensch and named for the 18th-century father of physiology, Albrecht von Haller. It holds that smaller creatures will have smaller brains but that the ratio of brain to body size will actually go up. And what is amazing is that few if any creatures on earth violate this rule. “It’s extremely general, and it’s been known for a long time. And there seem to be no good ideas as to why in the world it’s true,” says William Eberhard, a spider researcher and frequent collaborator with Wcislo, who also works at the Tropical Research Institute.
Imagine packing for a trip with a massive suitcase and then learning that the plane will accept only luggage half that size. The trip is the same, but the space just got tight, so you will have to be more efficient, and your bag might be bursting at the seams. The same thing happens to some of Eberhard’s smaller spiders. “Their brains were not staying in the right parts of their body. In the tiny ones they were going into the legs, and the sternum was bulging out, and it was full of brain. Their bodies were being deformed by these brains,” he says.
The comparison of scale in this spider world boggles the mind. Take Eberhard’s favorite group of creatures, orb weaver spiders. The largest he has worked with weighs around three grams, whereas the smallest weighs 0.005 milligram—roughly 600,000 times as small as its cousin. For perspective, imagine a normal adult man standing next to a giant who stood 400 kilometers tall and weighed more than 300 blue whales. The giant’s brain alone would weigh 910,000 kilograms.
So would such a giant be more intelligent than a human? If the scaling principles hold from the world of spiders, the answer is no, as can be seen by looking closely at the webs they spin.
As a spider constructs a web, it must continually make decisions, finding the most efficient places to attach each thread. And although they are exceptional architects, they do make mistakes—and those mistakes are pretty consistent over time. So Eberhard used these web-making mistakes as a proxy for cognitive capacity. Knowing the incredible costs of having a tiny body and thus an outsize brain, he expected to see that cost reflected in their webs. The smaller spiders should make more mistakes.
Shockingly, they do not. In fact, species to species and even within the species, the number of mistakes was exactly the same. Then a student of Eberhard’s tested the little critters, forcing them to build in a constrained environment—inside a piece of tubing about the diameter of a large air-rifle BB. Again, the spiders made the same number of miscalculations, even as newly born nymphs. The same seems to be true for parasitic wasps, which span from the massive tarantula hawk to a fairy wasp that is smaller than a single-celled paramecium. The latter have truly minuscule brains but are equally as adept at locating and ambushing prey. “We haven’t yet found any behavioral costs of having a totally tiny brain,” Wcislo says.
How could such a tiny brain perform as well as a bigger one? Through vicious, cutthroat evolutionary efficiency. Some tiny creatures actually have shrunken brain cells with dramatically shorter connecting axons, the wirelike extensions from neurons. But even then, there is a lower limit—a cell cannot get smaller than its nucleus (although some beetles may simply jettison the nucleus altogether). And if axons get too short, they start interfering with one another like tangled electrical cabling.
So having a halfway-decent brain is a tough job for small invertebrates. What does this mean for us larger creatures? It turns out that Haller’s rule does not care if you are a spider, wasp, bird or even a human. As animals evolve to become smaller because of a change in climate or other selective pressures, their brain demands an ever higher percentage of energy and real estate in their body.
One species of salamander that, like insects, can vary wildly in size has evolved a thinner skull to make room for its brain. And although it is not yet clear how all this applies to humans, we do know that human brains have shrunk over the past 10,000 years. Perhaps rather than becoming less intelligent, our ancestors’ brains were just becoming more efficient.
Diego Ocampo, a biologist currently finishing his Ph.D. at the University of Miami, took a survey of more than 70 bird species and found that they perfectly follow Haller’s rule, with the smallest ones having proportionally larger brains. But when he looked at individual groups, he noticed hummingbirds had their own supercharged version of the rule. Take two species of hummingbird. The violet sabrewing, a sizable bird at 12 grams, is about 2.4 percent brain. Meanwhile the striped-throated hermit, which is a fifth the size, is 4.8 percent brain. Compared with other creatures, these numbers are oddly low. Far bigger birds that he sampled, such as thornbills, have a brain that takes up an ungainly 7 percent of their body.
It is as if the hummingbirds as a group have come up with a far more efficient type of brain than other birds—a slight bending of Haller’s rule. And if that was not enough, the hermit, far from being a simpleton, actually demonstrates the most complex behaviors. Whereas the sabrewing tends to sit and guard a single plant, the hermit memorizes complex lines to follow through the forest to find food.
What if birds have unlocked some kind of ultraefficient brain design that allows them to do more with less? Certainly this would explain some of the stupendous abilities observed in, say, African grey parrots, which can identify shapes and even count, as well as corvids, which have an equivalent number of neurons to some primates and, it is suggested, may even be self-aware. Do not forget octopuses, which have very primitive brains and yet perform tasks that rival those of dogs.
Lars Chittka, who studies bee behavior and intelligence at Queen Mary University of London, flips these questions about animal smarts on their head. It is not that they require large brains to do complicated things, he says, it is that complicated behavior really does not require much brainpower. “The task that requires a large brain hasn’t been discovered yet,” he says. “You can do a whole lot with very little brains.” Some wasps, he says, are able to recognize the faces of every other wasp in their communities. But when he looks at their brain, there is nothing to explain such an impressive ability. Chittka suggests facial recognition may have evolved from simpler abilities, such as recognizing food sources. And given that bees have complex social interactions, symbolic language and excellent spatial memory, there is not really much to separate their intelligence from that of, say, a rodent.
Still, it stretches credibility to compare two species from vastly different parts of the animal kingdom and even harder to understand how physiology corresponds to specific behaviors. But, Eberhard says, any animal that has been pushed “up against the wall of Haller’s rule” by evolving to a smaller size while maintaining complicated behaviors is bound to have come up with a few interesting ways to streamline its brain.
Wcislo compares large animals such as whales and perhaps humans, with the large Apple IIe computers that sat on so many desks in the 1980s and revolutionized personal computing. They were powerful tools, but there was lots of wasted space and excess heat production. Now compare that with modern iPhones, and you see the power of miniaturization.
So maybe it is not surprising that Wcislo’s work has attracted the attention of Silicon Valley. His oldest and most devoted funder is Frank Levinson, a venture capitalist and founder of the fiber-optics giant Finisar. To explain why he started investing in bug research, Levinson describes the time he watched a pair of male butterflies near his home compete for a female’s attention, ducking and weaving around a bush. “The best chip out of Intel can’t fly, can’t dance, can’t romance a woman, can’t dogfight,” he says. “I don’t know anything in silicon that could do anything remotely as complex as this.”
If tiny animals have learned to do more with less, what is stopping electronics from doing the same?
Levinson says electronics companies today are obsessed with artificial intelligence—how to make machines more humanlike—at the same time that the increase in computing speeds seem to be slowing down for the first time since the 1970s. So, Levinson says, there is a huge need to both understand how intelligence works and make circuits smaller and more efficient. In other words, more insectlike.
Insects provide plenty of examples of high-performing computational machines. Take Wcislo’s latest obsession, nocturnal sweat bees that live under a jungle canopy with 10 to 20 times less light than on a moonless night. It is so dark that the laws of physics say there are not enough photons to distinguish a visual signal from background noise. “How the hell do they see?” Wcislo says. “They should not be able to see.” It seems their tiny brain acts as a filter for the image, like night-vision goggles, extracting an image out of the surrounding darkness. He is also training ants to walk through mazes and then comparing their brain with those of other ants living less intellectually challenging lives. These are the kinds of questions that may suggest cutting-edge materials and designs to allow computers to shrink as fast as animal brains have.
At the end of the day, insect brains offer more than just incredible efficiency—they also offer simplicity. Investigations into artificial human intelligence are tricky, partly because the human brain is inordinately complex. But as these scientists are finding, there is much you can do with a very small, efficient brain. Perhaps there is more programmers can learn from them as well.
“Silicon Valley is always looking for those new niches,” Levinson says. “One interesting place to look is with [Wcislo] and the guys studying something as simple as ants and bees and spiders—and see what they can tell us about thought processes and learning.”