Those gadgets at the hardware store advertised as "10 tools in one" have their place -- when packing for a camping trip, for example -- but none of the components will do its job as well a single knife, screwdriver, or pliers would perform it. It seems to be a law that the more tasks an implement is asked to take on, the worse it does any of them.

Looking at nature, scientists have discovered the same trade-off between versatility and performance when the organs of living things have evolved to do more than one job. In fact, a body structure that creates a survival advantage in some situations may be downright harmful in other environments. That explains why some disadvantageous traits haven't been eliminated by natural selection and instead persist.

Scientists invoke the "principle of frustration" in these cases. All organisms are under a variety of selection pressures at once, and they evolve structures or behavior that inevitably are compromise solutions to the conflicting needs. Sometimes there are several such solutions successful enough to give a survival edge.

One recent experiment used computer simulations to show how the principle of frustration can cause the evolution of several solutions. Karl Niklas, a scientist at Cornell University, started with the forms of the earliest known land plants. He let them evolve along several paths, aimed at adapting to different selection pressures. One was reproductive success, and these plants would need many seeds as high off the ground as possible. For harvesting light, the production of many non-overlapping horizontal branches would be the best solution. For stability and strength, the trees would need a limited number of branches and few long, horizontal ones.

In each of these cases where only one selective force played on the plants' evolution, they developed very different solutions. But then Niklas did a simulation with all three pressures operating at once. The result? A much greater range of equally optimal compromises, and many of them looked a lot like trees we're familiar with today.