Four sets of data are presented in which the relation between X and Y is not best described as linear in the original numbers. Therefore, some transformation of X, Y, or both is appropriate. Suggestions are log Y, log X, log Y and log X, 1/Y, and 1/X. Two of the sets can be described as learning curve examples. The object of the exercise is to find the best linear transformation. Best can be defined as the transformation which results in the highest r-squared and makes some sense in terms of the example. The student should interpret the coefficient of regression for the best transformation.