This thesis investigated the phenotypic and genetic variation of five groups of mid-infrared (MIR) predicted fatty acids; short-chain (4 to 10 carbons), medium-chain (12 to 16 carbons), long-chain (17 to 22 carbons), saturated, and unsaturated fatty acids. The predicted fatty acid values (g/dL of milk) were log transformed in order to improve normality. The data set included 49,127 test-day records from 10,029 first lactation Holstein cows in 810 herds. The total number of animals in the pedigree was 76,074. The random regression animal test-day model included: days in milk, herd-test date, and season-age of calving (polynomial regression) as fixed effects, and herd-year of calving, animal and permanent environment effects as random polynomial regressions, and random residual effect. Third and fourth degree of Legendre polynomial were estimated for fixed regression of season-age of calving effect and for the three random regressions, respectively. The average daily heritability estimates were 0.24, 0.32, 0.23, 0.33, and 0.21 for short-chain, medium-chain, long-chain, saturated, and unsaturated fatty acids, respectively. Estimated average daily genetic correlations were positive among all fatty acid groups and ranged from moderate to high (0.63-0.96). This study demonstrates the existence of genetic variation in milk fatty acid profile predicted by MIR spectra. Therefore, milk fatty acid composition can be improved through genetic selection. Keywords: milk fatty acid, mid-infrared, random regression model, heritability I would like to acknowledge; Participating dairy producers CanWest DHI (Guelph, ON) and Valacta (Sainte-Anne-de-Bellevue, QC) Supported by Contribution from the Dairy Research Cluster Initiative (Dairy Farmers of Canada, Agriculture and Agri‐Food Canada, the Canadian Dairy Network and the Canadian Dairy Commission) Ontario Genomics NSERC of Canada DairyGen Council of Canadian Dairy Networ