Each observation in the input data set should contain necessary information such as the identification of an individual and the first and second parents of an individual. In addition, if a CLASS statement is specified, each observation should contain the generation identification; and, if a GENDER statement is specified, each observation should contain the gender of an individual. Optionally, each observation might also contain the covariance between the first and the second parents. Depending on how many statements are specified with the procedure, there should be enough variables in the input data set containing this information.

If you omit the VAR statement, then the procedure uses the first three unaddressed variables in the input data set as the names of the individual and his or her parents. Unaddressed variables in the input data set are those variables that are not referenced by the procedure in any other statements, such as CLASS, GENDER, or BY statements. If the input data set contains an unaddressed fourth variable, then the procedure uses it as the covariance variable.

If the individuals given by the variables associated with the first and second parents are not in the population, they are added to the population. However, if they are in the population, they must be defined prior to the observation that gives their progeny.

When there is a CLASS statement, the functions of defining new individuals and assigning covariances must be separated. This is necessary because the parents of any given individual are defined in the previous generation, while covariances are assigned between individuals in the current generation.

Therefore, there could be two types of observations for a multiparous population:

• one to define new individuals in the current generation whose parents have been defined in the previous generation, as in the following, where the missing value is for the covariance variable:

     Mark   George Lisa    .    M  1
     Kelly  Scott  Lisa    .    F  1
  

• one to assign covariances between two individuals in the current generation, as in the following, where the individual’s name is missing, ‘Mark’ and ‘Kelly’ are in the current generation, and the covariance coefficient between these two individuals is 0.50:

     .      Mark   Kelly  0.50  .  1
  

Note that the observations defining individuals must precede the observation assigning a covariance value between them. For example, if a covariance is to be assigned between ‘Mark’ and ‘Kelly’, then both of them should be defined prior to the assignment observation.