8.17: Natural selection can cause the evolution of complex traits and behaviors.

We have seen that natural selection can change allele frequencies and modify the frequency with which simple traits, such as fur color or turkey-breast size or plant height, appear in a population. But what about complex traits, such as behaviors, that involve numerous physiological and neurological systems? For instance, can natural selection improve maze-running ability in rats?

The short answer is yes. Remember, evolution by natural selection is occurring, changing the allele frequencies for traits, whenever (1) there is variation for the trait, (2) that variation is heritable, and (3) there is differential reproductive success based on that trait. These conditions can easily be satisfied for complex traits, including behaviors.

In 1954, to address this question, William Thompson trained a group of rats to run through a maze for a food reward (FIGURE 8-30). He found a huge amount of variation in the rats’ abilities: some rats learned much more quickly than others how to run the maze. Thompson then selectively bred the fast learners with each other and the slow learners with each other. Over several generations, he developed two separate populations: rats descended from a line of fast maze-learners and rats descended from a line of slow maze-learners. After only six generations, the slow learners made twice as many errors as the fast learners before mastering the maze, while the fast learners were adept at solving complex mazes that would give many humans some difficulty. Sixty years later, it’s still unclear which particular genes are responsible for maze-running behavior, yet the selection experiment still demonstrates a strong genetic component to the behavior.

Natural selection can also produce complex traits in unexpected, roundabout ways. One vexing case involves the question of how natural selection could produce an organ as complex as a fly wing, when 1% or 2% of a wing—that is, an incomplete structure—doesn’t help an insect to fly. In other words, while it is clear how natural selection can preserve and increase the frequencies of fitness-increasing traits in populations, how do these soon-to-be-useful traits increase during the early stages if they don’t increase the organism’s fitness (FIGURE 8-31)?

The key to answering this question is that 1% of a wing doesn’t actually need to function as a wing to increase an individual’s fitness. Often, structures are enhanced or elaborated on by natural selection because they enhance fitness by serving some other purpose. Experiments using models of insects demonstrated that, as expected, a small percentage of a wing, in the form of a nub or “almost wing,” confers no benefit at all when it comes to flying. (The nubs don’t even help flies keep their orientation during a “controlled fall.”) The incipient wings do help the insects address a completely different problem, though. They allow much more efficient temperature control, so that an insect can gain heat from the environment when the insect is cold and dissipate heat when the insect is hot. Experiments on heat-control efficiency, in fact, show that as small nubs become more and more pronounced, they are more and more effective, probably conferring increased fitness on the individual fly—but only up to a point.

Eventually, the thermoregulatory benefit stops increasing, even if the nub length continues to increase. But it is right around this point that the proto-wing starts to confer some aerodynamic benefits (see Figure 8-31). Consequently, natural selection may continue to increase the length of this “almost wing,” but now the fitness increase is due to a wholly different effect. Such functional shifts explain the evolution of numerous complex structures that we see today, and these shifts may be common in the evolutionary process.