Glossary
Stationarity
A series is stationary when its mean, variance and dependence over time do not change. Prices usually are not; returns are closer to it.
A series is (weakly) stationary if its mean and variance are constant and the autocorrelation between two points depends only on the gap between them. Most statistical tools assume it. A random-walk price is not stationary, because its spread grows with time; its returns, the differences, are much closer to it.
A series with a unit root (an AR(1) coefficient of 1) is the standard non-stationary case.
Example: with a daily volatility of 1%, a random-walk price has a standard deviation of change of 10% after 100 days and 20% after 400, so its variance is not constant.