Selection Pressures (Skills)

Most of evolution is too slow to watch. Guppies are the exception. They breed fast enough that a change in predation shows up in a couple of years, which is why John Endler was able to do something unusual — set natural selection and sexual selection against each other in ten ponds and measure which one won.

Why guppies are such a good model

Guppies are small fish found in the mountain forest streams of Trinidad and Tobago. Three features make them almost perfect for studying selection.

• They vary a lot. There is wide variation in colour and in the size and number of spots, and there is strong sexual dimorphism — males are brightly coloured while females are dull.
• Their spots do two jobs at once. The spots mimic the gravel of the streambed, giving camouflage from predators. But bright, obvious spots are also what females look for in a mate.
• The streams differ. Some streambeds are fine and sandy, others are coarser and more pebble-like. Some stretches have plenty of hiding places from predators; others have almost none.

So the same trait — spottiness — is pulled in opposite directions by two different pressures, and the balance point depends on the stream. That is the whole experiment in one sentence.

Endler’s hypotheses

In the 1970s Endler noticed that guppy colour patterning in Trinidad changed with predation pressure. He then set out to test three ideas in the lab and in the field:

• When predators are present, the substrate type of the streambed affects survival, which changes spot brightness.
• When the predation rate is low, spot brightness on male guppies increases, because sexual selection is unopposed.
• As predation increases, spot brightness decreases.

The laboratory experiment

Endler used a greenhouse to recreate a tropical environment. Before the experiment began, the guppies lived in large tanks and bred freely for six months with no predation, which gave a starting population with a wide range of spot sizes and numbers. He then counted and measured the spots on every fish, so he had baseline data.

The guppies were then randomly assigned to ten ponds inside the greenhouse. Five ponds had coarse gravel and five had fine gravel, and the ponds were exposed to three levels of predation.

Once the predators were introduced, the experiment ran for five months. Endler then counted and measured the spots on all of the guppies again. The experiment was left to run for a further nine months, after which more data was collected. The weak predator was a killifish; the dangerous one was a pike cichlid.

Predictions and results

That last row is the most interesting result in the whole study, and it is worth sitting with for a moment. Take the predators away and the rule reverses. The gravel still matters — but now guppies are selected to clash with it rather than blend into it, because being noticed is suddenly an advantage instead of a death sentence.

The field experiment

The greenhouse result could be dismissed as an artefact of artificial ponds, so Endler repeated the idea in the wild. He took a number of dull male guppies from an area of high predation and moved them to an area of low predation. He left them for 15 guppy generations — about two years — then returned.

The male guppies now had more colourful patterning. With strong predation removed, sexual selection was free to push colour upwards, exactly as the lab ponds had suggested.

Interpreting the results

The overall conclusion is not “predators reduce spots”. It is that guppy evolution is a dynamic balance between two selection pressures acting on the same trait:

• Natural selection acts through competition for food and avoidance of predators. Predators spot brightly coloured males more easily, so bright males survive less well. This selects for less colourful, less spotted individuals.
• Sexual selection acts through mate choice. Males with brighter and bigger spots are more likely to get a mate, reproduce, and pass on the alleles for those spots. This selects for more colourful individuals.

Where the balance sits depends on how dangerous the stream is. Where predation is high, brightly coloured males are less likely to survive regardless of how attractive they are — and a dead guppy has no reproductive advantage at all.

The exam skills this topic tests

Variables

• Independent variables (what Endler changed): predation level and gravel type.
• Dependent variables (what he measured): number, size and brightness of spots per male.
• Controlled variables (what he kept the same): temperature, light, food supply, pond size, and the starting population the fish were drawn from.

Why the design choices matter

• Random assignment to ponds prevents bias — without it, spottier fish might end up concentrated in one treatment by accident.
• Two ponds per combination is replication. If both ponds show the same trend, the result is unlikely to be a fluke of one pond.
• Baseline measurements before the predators arrived let him show that any later difference was caused by the treatment, not present from the start.
• Six months of free breeding first made sure there was plenty of variation for selection to act on.

Reading the graphs

Endler’s published data comes as several small bar charts, one per colour of spot, with shaded bars for artificial ponds and unshaded bars for natural ones, and error bars showing two standard errors. Two habits will save you:

• Read the key first, then the axes, then the bars. Never start with the bars.
• Use the error bars. If the error bars of two groups overlap a lot, you cannot claim a real difference between them. If they are clearly separated, the difference is likely to be genuine.

Worked examples