With a growing global population, rising feed costs, and an increased awareness of environmental concerns, the Canadian dairy industry is looking to include feed efficiency into a breeding objective for Holstein cattle. Since measuring individual feed intake can be difficult and expensive, the use of a predictor trait such as mid-infrared spectroscopy may be a solution by providing phenotypes for feed intake on large numbers of animals at a low cost, utilizing technology that is already used in regular milk recording. This thesis presents a preliminary evaluation of the efficacy of using mid infrared spectral data to predict a feed efficiency trait. Adding mid-infrared spectral data did not considerably improve accuracy of prediction equations from equations created using just milk yield and milk component data. Further research is needed to validate of the results found in this study using an independent dataset.