Cluster analysis principles were used to allot bulls to be used as Multiple Service sires (MS). Expected Progeny Differences (EPDs) on pre- and postweaning traits were used to calculate distances between bulls. EPDs were standardized and weighed to form a final selection index. The criteria to form lots was based on minimizing the sum of all standardized distances for all possible pairs of bulls. The program was tested on a set of 158 top bulls from a Nelore herd. A set of 4,740 breeding cows was used to evaluate three breeding strategies with the goal of producing genetically superior individuals: (1) random mating; (2) single sire under designed matings by a specific program (PAD); and (3) lots of multiple sires under designed matings using PAD. Principal components was used to obtain the genetic biotypes existing in these populations. Extreme values in EPDs amongst MS lots averages and overall were very similar, indicating the program's ability to preserve total variance. Variances of EPDs from calf crops obtained under the use of PAD were increased three fold. Using designed matings by PAD allowed an increase of 70% in the number of animals which would be recognized with official and fiscal benefits in comparison with random mating. First principal components for cows, indicated that most genetic variability is accounted for preweaning traits and visual scores at postweaning; the second indicated animals can be contrasted as good weaners and poor performer at yearling and vice-versa; the third component showed diferences in sexual precocity. For single and multiple sires, despite the stronger selection pressure behind them, most of the remaining genetic variability is associated with differences in precocity (traits associated with early finishing).
animal breeding; beef cattle; expected progeny differences; fortran; precocity; selection indexes
