Date of Award


Document Type


Degree Name

Doctor of Philosophy (PhD)


Dairy Science (Animal, Dairy, and Poultry Sciences)

First Advisor

John E. Chandler


The objective of this study was to find whether DHI data could be used to estimate sire non-return rates to replace current technician data estimates. Bull weighted least squares means for non-return rates were calculated separately for five overlapping 60-90 day service periods from each data source. Models included stud, sire, service number, and linear and quadratic form of breeding month for both data sources, service unit for technician and lactation for DHI data. Sire and lactation were not significant $(P > .10).$ Technician differences $(P < .05)$ were in service unit, stud, service number, and linear and quadratic service month in all but one service period. DHI differed $(P < .05)$ for service number, month (linear and quadratic), and stud in two service periods. Technician R-square values were 0.23 to 0.28 versus 0.94 to 0.96 for DHI. Sire estimated non-return rates were weighted using the inverted estimator standard error squared and compared. Sire, stud, data source, service period, and appropriate interactions were modeled. Weighted bull non-returns differed $(P < .01)$ in magnitude across data sources. Stud, data source by stud, and sire within stud by data source were significant $(P .10).$ Four fertility categories based on mean and standard deviation of the weighted estimates were formed within the data sources across service periods. These categories were correlated (0.5 $>$ r $>$ 0.9) and 52.9 to 87.4% congruent within data source for adjacent service periods and across data sources within service periods. With declining availability of technician data, DHI data was shown to be a reasonable substitute. Correlations and congruency of fertility categories suggest sire choices would be very similar.