Jonathan Weiner
Guppies
[from The Beak of the Finch]


What the Grants are to Darwin’s finches, Endler is to guppies. His guppies are not the variety they sell in pet stores (he considers those trash fish). His guppies live in northeastern South America, in the small streams that zigzag down the mountains of Venezuela, Margarita Island, Trinidad, and Tobago, flashing through steep, undisturbed green forests and then the broad spreads of the old cacao and coffee plantations, on their way to the Caribbean Sea and the Atlantic.

The male guppies wear black, red, blue, yellow, green, and iridescent spots in varying sizes, shapes, hues, and combinations. In fact their spots vary so much that they are like fingerprints: no two guppies are alike.

These spots, like the beaks of Darwin’s finches, are heritable. Although the exact placement and arrangement of spots is unique, each guppy inherits its particular palette of colors, and also the general size and brightness of the ensemble, from its parents. The spots only show up on the males (they can be made to appear on the female guppies with testosterone treatments).

Like minute variations in the beaks of finches, the spots on a guppy are the sorts of details that one might imagine are beneath the notice of natural selection. Nature may scrutinize the slightest variation, but there are some things even Darwin’s process cannot see. Design could not possibly govern a thing so small.

In the 1970s, while Peter and Rosemary Grant were watching the finches of the Galapagos, Endler began watching the guppies of Venezuela’s Paria Peninsula, and Trinidad’s Northern Range. There the streams run down the mountains roughly parallel, as if in a series of vertical stripes. The streams are clear, swift, and clean, deeply shaded by tropical evergreens and punctuated by waterfalls. Their beds are lined with brilliant, many-colored gravel, much like the floors of the fish tanks in the pet stores.

It is obvious to anyone who has ever tried to watch a school of these guppies against the parti-colored sands and pebbles of a streambed that the spots are excellent camouflage. In fact you could watch one of these clear streams for quite a while before you noticed the guppies at all, because they tend to swim close to the gravel while the sun is out.

The fish need this camouflage because they have seven enemies: six species of fish and one freshwater prawn. All seven of these enemies hunt guppies from dawn till dusk. The most dangerous is Crenicichla alta, a cichlid fish, which eats about three guppies an hour; the least dangerous is Rivulus hardi, which eats one guppy in about five hours.

Endler found guppies and at least a few of their enemies in almost every section of almost every stream, from the headwaters near the summit of each mountain to the plains and plantations below. Neither the guppies nor the guppy eaters can swim up a waterfall, and the population of each section of stream tends to stay put. (Sometimes a few fish get swept downstream, but none of them can get back up.)

High up near the headwaters of each stream, the only enemy the guppies have is the comparatively mild-mannered Rivulus hartii. But moving downstream, section by section, the population of guppies lives and dies in the company of more and more of its enemies, until down near the base of each mountain, the stream is loaded with all seven of the guppy eaters. So a graph of risk and danger runs with the current. For the guppies, the higher in the stream, the lower the risk; and the lower in the stream, the higher the risk. In stream after stream the intensity of natural selection is graduated in the same way: gentle pressure among the guppies at the top, violent pressure among the guppies at the bottom.

Endler saw that the streams would make a wonderful natural laboratory for the study of natural selection. He developed standardized methods of measuring guppy spots, as careful and ritualized as the Grants’ methods with Darwin’s finches. He learned to anesthetize and photograph each guppy he caught. (Like Darwin’s finches, the guppies have met very few human beings, so they are easy to catch.) From the photographs he recorded the color and position of each spot of each and every male guppy, dividing each guppy into dozens of sectors to make a standardized guppy map that is easy to read, to tally, and to enter into a computer.

When Endler analyzed his surveys he discovered a pattern. The spots on each guppy look chaotic, but the spots of all of the populations of guppies in a stream, taken together, from the headwaters down to the base, have a kind of order. The spots on each population of guppies bear a simple relationship to the number of guppy eaters in their part of the stream. The more numerous the guppies’ enemies, the smaller and fainter the guppies’ spots. The fewer their enemies, the larger and brighter their spots.

The lucky guppies in the headwaters wear sporty coats of many colors, and each color is represented by big clownish splotches. Many of their spots are blue. These blue spots are iridescent, like the Day-Glo patches that cyclists wear; they flash as the fish swim, and they can be seen a great distance through the clear water.

Meanwhile the guppies downstream tend to wear conservative pin dots of black and red. The spots are almost vanishingly small. Most wear only a tiny amount of blue.

Endler looked at his data from stream after stream. In every one of them, the size and number of spots ran steeply downhill. And Endler drew the same sort of conclusion that Lack did when he noticed the patterns of beaks in the Galapagos. Endler thought he could see the hand of natural selection at work among the guppies. The greater the pressure from their predators, the more camouflage they wear; the less the pressure, the slighter the camouflage.

Of course, that interpretation did not explain why guppies are colorful at all. If they are in some danger everywhere, even in the headwaters, then why doesn’t natural selection favor the best-camouflaged guppy everywhere?

The answer is that a male guppy has more to do in life than merely survive. It also has to mate. To survive it has to hide among the colored gravel at the bottom of its stream and among the other guppies of its school. But to mate it has to stand out from the gravel and stand out from the school. It has to elude the eyes of the cichlid or the prawn while catching the eyes of the female guppy.

The gaudier the male, the better his sex life. He is more popular among females, and he gets many chances to pass on his gaudy genes as long as he lives. In a quiet spot near the headwaters of the stream his life is likely to be long and happy and he may father innumerable gaudy children. But in a spot near the base of the stream he may not father a single guppy before he vanishes down the gullet of a cichlid.

The quieter the colors of a male, the less luck he has in courting females. On the other hand he is likely to have more time to try, because the less he stands out among his own kind, the less he stands out among his enemies.

This is not just a problem for Trinidadian guppies. Wherever males court females, or females court males, whether the signals are a bright splash of color, as in guppies or red-winged blackbirds, or loud far­carrying songs, as in frogs and crickets, their broadcasts are always in danger of being intercepted by the enemy. Strong colors or loud calls can attract a mate from one side and a predator from the other. Every bullfrog calling in the night is in the dangerous spot of Romeo calling out beneath the balcony of the house of Capulet. A few species have found ways to finesse this problem. Among fish, some wrasses change color only very briefly, to flash a sexual signal in dangerous waters – the equivalent of a sexy whisper, pssst!

Looking at his guppy data, Endler read into them a struggle between two contending forces. Everywhere in the stream the gaudier fish produce gaudier young – pushing the next generation toward loud colors and self-advertisement. And everywhere in the stream the quieter fish produce quieter young, pushing the next generation toward modesty. In the relative safety of the headwaters the gaudier guppies live long enough to win many females before they are eaten, so the population evolves in the direction of greater and greater gaudiness, and almost every male wears a coat of many colors. But in the dangerous waters at the base of the mountain the gaudy guppies live such a short time that they are out-reproduced by the modest guppies. So the whole population evolves in the direction of greater and greater drabness. Males court females at distances of 2 to 4 centimeters, and from there the little spots are visible; but from farther away the males blend into the gravel. So the small-spotted guppies can blend into the background in the eyes of their predators, Endler says, “yet still be visible and stimulating to females.”

When Endler first began studying the guppy streams, he was in the same position as David Lack after the Galipagos. Endler could see pat­terns that strongly suggested the forces of selection at work. He did not actually see selection shaping the patterns, but the closer he looked the more he was sure that the hand that shaped the patterns really was the hand of natural selection. Within the broad patterns he kept find­ing curious subsidiary patterns. For instance in a few of the headwaters there are prawns. In these headwaters the guppies favor red spots. This red shift makes sense because although guppies and other fish see more or less the same colors that humans do, prawns and shrimp are red­blind – they cannot see the last band in our rainbows. So in those particular headwaters, male guppies with big red spots can show off to female guppies while hiding from the prawns.

Back in the 1940s, Lack made his selectionist argument about Darwin’s finches without trying to measure it in the field to see if he was right. But Endler went the extra step: he decided to test the predic­tions of his theory by trying to detect these processes in action. He built ten ponds in a greenhouse at Princeton University. Four of the ponds were about as wide, deep, and long as the low-water territories of Crenicichla alta. The other six ponds were about the size of the headwater streams with the comparatively mild-mannered Rivulus hartii. Endler put black, white, green, blue, red, and yellow gravel in the bottom of his artificial ponds and pumped water through them to give them a current, like the streams in the wild.

Meanwhile, Endler collected guppies from up and down a dozen streams in Trinidad and Venezuela. In some places he took guppies that lived with just one predator, in some places guppies that lived with two predators, and so on up to the maximum, seven. He wanted stocks of wild guppies that had evolved under the whole spectrum of guppy menace, that were coping in the wild with every level of dan­ger. He bred each stock in a separate aquarium.

When the artificial streams were ready for his guppies, Endler took five pairs at random from each stock and put them all together into two of his ponds to let them breed and mingle and get used to their new homes. Guppies can give birth at the age of five or six weeks, and a female guppy can spawn a lot of baby guppies, so it did not take long for the populations to double. After a month he took guppies from those two ponds and used them to stock two more ponds. A month after that, he had enough guppies to seed each of his ten ponds with two hundred fish per pond.

What he had done, in effect, was to shuffle and reshuffle the deck. He now had a highly heterogeneous assortment of guppies. They had all kinds of spots, and their spots were completely random with respect to the gravel at the bottom of their homes.

He let these guppies breed in their new streambeds for months. Then he added a few of their natural enemies to the streams, according to a careful plan. The evolutionary experiment had begun.

According to his prediction, the guppies should now evolve rap­idly. The guppies in each tank should begin to look more like guppies that live with that same set of predators in the wild, they should come to look more like the gravel in their particular stream, and those in the most dangerous tanks should come to mimic the gravel more closely than those in the safer tanks.

After five months, Endler took his first census. He drained each stream, counting every male’s spots and noting their position, anesthe­tizing them, photographing them, as he had done in the wild, and then starting up the stream again. Nine months later he took a second census. By that time nine or ten generations had passed in the lives of his guppies.

Some of the guppies were safe, with no enemies. These guppies got gaudier between the foundation of the colony and the first census, and they got gaudier still by the time of the second census. The males evolved more and more spots, bigger and bigger spots, wilder and wilder palettes of spots.

Meanwhile males in tanks with the dangerous cichlids evolved fewer and fewer, smaller and smaller spots. They were still visible to fe­males, but they got less and less visible to cichlids, who strike from 20 to 40 centimeters away. These guppies mostly dropped the blue and the iridescent spots, their Day-Glo patches, just like guppies that live with cichlids in the wild. Endler measured these differences as metic­ulously as the Grants measure finch beaks. "Spot height, spot area, to­tal area, and total spot area relative to body area also decreased significantly with increased predation intensity" he reports. The fish themselves changed size, too. Full-grown guppies in the dangerous tanks were smaller, while mature guppies in the safe waters were larger – again, just as in the wild.

Each tank had a different bottom: different mixes of gravel colors and different gravel sizes. In the pools with no predators the guppies did not change their spots to match the gravel – the opposite. Their spots evolved to be smaller than the big gravel and larger than the small gravel, making the males easier and easier to see, like chameleons in reverse. They carried more iridescent spots, and a wider palette of colors per fish, and generation after generation they looked less and less like their background, all of which is just what we would expect if they were competing for attention. Sexual selection was operating to make males as different from the gravel bottom as possible.

If only one force or the other had been operating, just natural selection or just sexual selection, the guppies would not have evolved in this remarkable way. Without natural selection all of the fish would have gotten gaudier. Without sexual selection none of them would have gotten gaudier. But the safe ones did get much more colorful, adding, in particular, blue spots. It is probably not a coincidence that guppies’ retinas are exquisitely sensitive to blue. Almost all males carry some blue somewhere, even in the most dangerous waters – it may be a sine qua non of courtship.

The fish had evolved in Endler’s greenhouse until they replicated the patterns that they display in nature, and they had done so in a very short time. Of course, Endler’s streams were artificial. He had not seen natural selection in the wild. A skeptic could still argue that Endler was wrong about his explanation for the pattern in the wild. So Endler figured out a way to run the same sort of evolutionary experiment in nature.

Early in his fieldwork he had found a Trinidadian stream that con­tained the guppy eater Rivulus bartii, but no guppies. About 2 kilometers away was a second stream that contained both guppy eaters and guppies. Endler took a random sample of about two hundred guppies from one of the high-danger zones in the second river. He measured each and every one, as usual, and then he transferred them to the safe place in the first river. He took a sample of their descendants more than a year later, after a passage of fifteen generations.

The males in the safe stream were now much gaudier than their immediate ancestors, who were still living in the stream next door and coping with many enemies. The immigrant males wore bigger spots, and more of them, and each male sported a wider assortment of colors. Natural selection had acted just as predicted. Evolution had run as fast in the wild as in the greenhouse.

Everywhere in those streams, daily and hourly, natural selection in the form of cichlids and prawns is not just metaphorically but literally scrutinizing the male guppies. The result of enemy predations on each generation keeps pushing the males to blend in with the stream bot­tom. At the same time, daily and hourly, sexual selection in the form of female guppies is scrutinizing those same males. The result of their choices is that generation after generation of males is pushed to stand out.

Now it is clear why there is such virtually infinite range of varia­tion in the way, each individual male guppy is spotted. Many different random patterns of splotches will be equally good camouflage, because the streambed patterns are random too. It would not help the guppies to sport the same pattern as all the others, and in fact it would hurt them. If the guppy males all looked alike, their enemies could develop a search image – an inner template. They would search for that pat­tern, as we search for the face of a friend in a crowd. The rare misfit would have a great advantage. Meanwhile the females would go for the unusual males, too, and that would drive more and more diversity of patterns. So in this respect natural selection and sexual selection cease to oppose each other and push in the same direction: toward al­most infinite diversity

We see here an example of what Darwin saw in the wide world. He understood that his simple process can lead to the most bewilder­ing and chaotic-looking diversification and variety – but underneath, the driver is as simple and plain and commonsense as ever, “small consequences of one general law leading to the advancement of all organic beings, – namely, multiply, vary, let the strongest live and the weakest die.” The guppy experiments suggested to Endler what Darwin’s finches were suggesting at about the same time to the Galapagos finch watchers: that natural selection can be swift and sure. The process is flowing along, all around us, much faster than Darwin ever dreamed.

Endler’s study is leading him into deeper and deeper waters. He now suspects that the guppies’ spots, their mating habits, and their color vision are all evolving simultaneously, with change in any one of these factors driving change in all the others. To measure variations in the guppies’ retinas, Endler is collaborating with physiologists. These “hard science” types often remind him how “soft” the science of evolution is perceived to be by the outside world – even by biologists. “I was talking with someone in vision physiology the other day” Endler says, “and he told me, ‘Wow, I had no idea that the subject was so rigorous. I had no idea that you actually did experiments.’

“We have a serious public-relations problem,” Endler says. “People don’t realize this is real science.”