Deviation

A deviation describes a single observation rather than the whole data set, and you write it as $x_i - \bar{x}$, where $\bar{x}$ (x-bar) is the mean. A positive deviation puts the value above the mean and a negative deviation puts it below. In the data 2, 4, 6, 8 the mean is 5, so the four deviations are -3, -1, 1, and 3. They add to zero, which is true for every data set and is exactly why the variance squares the deviations before adding them up.