Since Ed Lorenz’s seminal contribution on empirical orthogonal functions (EOFs) in 1956, their use has become ubiquitous in atmospheric science. As well as the basic technique, known more widely as principal component analysis, many variations on the theme have been introduced. The definition of EOFs is simple, but their use involves a number of tricky decisions. A number of myths and misconceptions exist regarding when and how to use them and their variants. These myths are discussed and it is noted that a number of questions about the use of EOFs are still not completely resolved.