Population Distribution
Histogram of the source population each sample is drawn from.
Distribution of Sample Means
This builds up over iterations; the dashed curve represents the theoretical N(mu, sigma^2 / n) prediction from the CLT.
How it works
For a population with mean mu and variance sigma^2, the Central Limit Theorem says the sample mean X-bar for independent draws tends toward N(mu, sigma^2 / n) as n grows, even when the original population is skewed, heavy-tailed, or multi-peaked.
Increase sample size and iterations to watch the sample-mean histogram tighten and line up with the dashed theoretical curve.