The general purpose of any breeding program is to produce some change in the genetic structure of the population in question. The existence of control populations in which attempts are made to prevent any changes in the genetic structure indicates the value of some device for measuring progress due to selection. The experimental design discussed in this paper due to Goodwin, Dickerson and Lamoureaux provides 3 measures of genetic changes. All 3 depend essentially on the use of matings which are repeated indentically during 2 successive breeding seasons. The 1st of these measures the difference between the means of 2 groups of birds of different generations but reared in the same year. The 2nd method measures a combination of changes in genetic and joint maternal-year effects and is, therefore, of doubtful value. The 3rd measures the regression of response due to selection on years. This last method has the disadvantage that the proper error is difficult, though not impossible, to estimate. The number of degrees of freedom available for this estimate is also very limited. Fully detailed procedures for estimating the errors of the methods are given and illustrated with a numerical example.