Holstein (n = 19) and Jersey (n = 18) cows were used to study effects of two feeding systems on fatty acid composition of milk. Confinement cows were fed a total mixed ration with corn silage and alfalfa silage and pastured cows grazed a crabgrass (90%) and clover (10%) pasture and were allowed 5.5 kg of grain per head daily. Two milk samples were collected from each cow at morning and afternoon milkings 1 d each week for four consecutive weeks in June and July 1998. One set of milk samples was analyzed to determine fatty acid composition, and the second set was used for crude protein and total fat analyses. Data were analyzed by the general linear models procedure of SAS, using a split-plot model with breed, treatment, and breed x treatment as main effects and time of sampling and week as subplot effects along with appropriate interactions. Milk from pastured cows was higher than milk from confinement cows for the cis-9, trans-11 octadecadienoic acid isomer of conjugated linoleic acid (CLA). Also, milk from Holsteins was higher than milk from Jerseys for C16:1, C18:1, and CLA and lower than Jerseys for C6:0, C8:0, C10:0, C12:0, and C14:0. Several treatment x week interactions existed, but main effects were still important; for example, proportions of CLA in milk of grazed cows were relatively constant across weeks (0.66, 0.64, 0.64, and 0.69% +/- 0.02%, respectively), but the CLA in milk of confinement cows increased in wk 4 (0.35, 0.31, 0.31, and 0.48% +/- 0.02% for wk 1 to 4, respectively). There are potentially important differences in fatty acid composition of milk from cows consuming a warm season pasture species compared with milk from cows consuming a total mixed ration, as well as differences between Holstein and Jersey breeds.