Deer Mice on Lava: Selection or Wishful Thinking?
A biologist studies deer mice in a desert preserve where fur color is controlled mainly by one gene with two alleles: D (dark) and d (light). D is dominant to d. On pale sand, light mice are camouflaged; on dark lava, dark mice are camouflaged. In 1990, on a large pale-sand area, a sample of 1,000 mice contained 150 dark and 850 light. In 2005, a lava flow covered about 30% of that area. By 2020, a sample of 300 mice from the new dark lava patch contained 220 dark and 80 light. In the same year, a sample of 1,000 mice from the remaining pale sand contained 220 dark and 780 light. Hawks and coyotes are the main predators and hunt by sight. The lava patch is contiguous with the sand; mice can move between both surfaces, though mostly over short distances. A student claims the mice on the lava 'turned dark because they needed to blend in.'
1. Using the data, evaluate whether natural selection has occurred in the lava patch. Compare the 1990 sand population, the 2020 sand population, and the 2020 lava population, and explain what evidence would be needed to support the claim.
Read the worked answer
The data are consistent with natural selection, but they do not prove it by themselves. In 1990, only 15% of the sand population was dark. In 2020, the sand population was 22% dark, while the lava population was 73% dark (220/300). The lava population is much darker than both the historical sand population and the contemporary sand population, which serves as a nearby control. If dark fur is heritable, the shift toward dark mice on lava is what visual predation would produce: light mice stand out on dark rock and are removed before reproducing, while dark mice survive and pass on the D allele.
To support the claim more strongly, the biologist would need evidence that fur color is inherited, that light and dark mice differ in survival or reproduction on lava, and that the difference is caused by predation rather than drift or migration. Rough allele-frequency estimates also suggest change: in 1990 sand, light mice are dd, so q^2 = 0.85, q = 0.92, p = 0.08; in 2020 lava, q^2 = 0.27, q = 0.52, p = 0.48. Those estimates assume Hardy-Weinberg conditions, so they are only approximate.
2. Explain why the student's statement is not an accurate description of natural selection. Include what actually changes across generations and what does not.
Read the worked answer
The statement implies that individual mice changed color in response to a need. That is not how natural selection works. Individual mice do not rearrange their pigment genes because the rock is dark, and a color change acquired during life would not be passed to offspring.
What actually happens is that the population already contains variation in fur color. On dark lava, light mice are more visible to hawks and coyotes and are less likely to survive and reproduce. Dark mice survive better and leave more offspring, so the frequency of the dark allele increases in the population over generations. The population changes; the individual does not. The student's wording reverses the cause: the need does not create the trait, and selection acts on traits that are already present.
3. Design a field experiment to test whether predation is the cause of the pattern. Describe the setup, a control, and the predicted result.
Read the worked answer
One design is a clay-model predation experiment. Make equal numbers of dark and light clay mice that are the same size and shape. Place 100 dark and 100 light models on the dark lava and 100 of each on the pale sand, in paired grids. Record the number of models with bite or peck marks after two weeks. To control for the effect of predators, place an identical set of models inside predator exclosures on each surface, or use a nearby area where predators are excluded.
Prediction: on lava, light models should be attacked more often than dark models; on sand, dark models should be attacked more often than light models. Inside exclosures, attack marks should be low and similar for both colors. That pattern would support visual predation as the mechanism. A complementary design is to use enclosed mouse populations with and without predators and measure survival by color over several generations.