Changes of terpene content in milk from cows grazing natural diversified upland pasture were examined in this observational trial. A homogeneous plot divided into 2 subplots was used from May 31 to July 1, 2003 (first growth) and again from October 1 to October 7, 2003 (vegetative regrowth). Each subplot was grazed by 6 dairy cows in 2 ways: strip grazing (SG), with new allocations of pasture strips at 2-d intervals, and paddock grazing (PG). The PG subplot was divided into 3 paddocks and the cows were moved to a new paddock on June 13 and June 24, 2003. Milk from the 6 cows was collected twice a week, pooled, and used for terpene analyses by dynamic headspace gas chromatography-mass spectrometry system. Twenty mono- and 23 sesquiterpenes desorbing from the milk fat were separated. The most abundant monoterpenes were beta-pinene, alpha-pinene, gamma-terpinene, limonene, alpha-tujene, terpinolene, and alpha-phellandrene. The most abundant sesquiterpenes were beta-caryophyllene, alpha-copaene, beta-cedrene, transmuurola-4-(14)-5-diene, beta-bisabolene, and delta-cadinene. Both mono- and sesquiterpenes in SG milk increased across time with an 8-fold increase in total terpenes in milk from the beginning to the end of June. In parallel, dicotyledons, including the main terpene-rich plants, increased from 17 to 31% of total biomass of the vegetation and the development of Dactylis glomerata progressed from boot to ripening stage. The terpenes in PG milk were equivalent to those in SG milk for the first paddock at the beginning of June and remained constant or doubled for the sum of mono- and sesquiterpenes, respectively. The lower variability of the PG milk terpene content could be related to the opportunity that PG cows had to choose ingested herbage, whereas the SG cows had limited choice within the smaller allocated pasture strips. Milk from cows grazing regrowth pastures in October contained low levels of terpenes, and values were similar for PG and SG milks. Change in the sum of monoterpenes is representative of changes for most monoterpenes (except for limonene) and, conversely, the change in the sum of sesquiterpenes mainly reflected changes in beta-caryophyllene and alpha-copaene. In addition to effects of botanical composition of pasture, it appears that terpene content in milk may vary according to factors linked to grazing management that need more intensive study. Nevertheless, current results raise questions about the precision of terpenes as feed tracers.