A 3 x 3 Latin Square experiment was designed to compare 2 ways of bypassing the effects of the rumen with olive oil fatty acids in 'Manchega' dairy ewes. Treatments were a control diet, CaOFA (control diet plus 45 g of olive fatty acids as calcium soap), and OO (control plus 45 g/d of olive fatty acids as olive oil emulsified in skim milk) and bottle-fed to animals trained to maintain the reticular groove reflex). No differences were found in milk, protein, and lactose yields, but fat yield and milk fat content were greater in treatments with added fat (CaOFA and OO). Content of short- and medium-chain fatty acids in milk fat was greater for control treatment than for the other 2 groups, the yield of these fatty acids being similar for all 3 diets, except that of C12:0, which was greater for the control treatment. Content and yield of C18:0 and isomers of C18:1 others than oleic acid were greater in milk from the CaOFA diet than from the other 2 diets. Oleic acid content and yield were greater in milk after OO treatment (23.9% and 16.8 g/d, respectively), intermediate after CaOFA treatment (19.2% and 13.8 g/d, respectively), and lower after control diet (10.7% and 6.52 g/d, respectively). Linoleic acid yield and content were greater in ewes fed the OO diet than in those on the other 2 diets, both of which showed similar data. All these changes indicated that the "protected" olive fatty acids (as calcium soap) were severely affected by the rumen environment and that the use of the reticular groove reflex seems to be a more effective way of bypassing the rumen in adult lactating dairy ewes.