The principal components are uncorrelated and orthogonal to one another They are ordered in such a way that the first component explains the largest amount of variance in the original dataset, the second explains the largest amount of variance of the original dataset among all components that are orthogonal to the first principal component, and so on and so forth.