Investigation 1: Artificial Selection

Pick the tallest plants in a population and breed only them, and the offspring generation starts out taller on average than the parent generation did, because tall parents pass on the alleles that made them tall. This investigation uses Wisconsin Fast Plants and a trait called trichome number, the tiny hairs on the stem and leaves. You grow a parent generation, count trichomes on every plant, and plot the distribution: a bell-ish curve centered on some mean. Instead of breeding randomly, you take only the top sliver, say the highest 10 percent, and let those interbreed.

Draw both generations on the same trichome axis and you see the real story. The parent distribution has its peak somewhere in the middle. The selected breeders sit far out on the high tail. The offspring distribution is shifted: its peak sits closer to the breeders' mean than the original parent mean did, though usually not all the way there, because heritability is never a perfect 1. The breeder's equation captures this: R = h times S, where S is the selection differential, the gap between the selected breeders' mean and the whole parent population's mean, and h is heritability, the fraction of trait variation due to genes rather than environment. A bigger gap or higher heritability means a bigger generational shift.

This connects to natural selection and to Hardy-Weinberg ideas elsewhere in evolution: artificial selection is the same mechanism, just with a human choosing who breeds instead of the environment. It also sets up the misconception question about limits: selection cannot push a trait upward forever, because the process itself uses up the genetic variation it depends on.