Hyptis marrubioides were collected in March 2003 at the mature vegetative stage from their natural habitat; 20 randomised individual plants at the same age representing the local population were collected as homogenous samples from each locality: (A) Lavras (21° 14′ S/44° 59′ W), at an altitude of 919 m; (B) Tiradentes (21° 6′ S/44° 10 m W), 927 m.

The results were submitted to Principal Component and Cluster analysis which allowed three groups of oils to be distinguished with respect to sampling site and post-harvested process: cluster I (fresh leaves and fresh or dried stems from Lavras site) with high percentage of caryophylla-4(14),8(15)-dien-5beta-ol (16.7%) and eudesma-4(15),7-dien-1beta-ol (12.8%); cluster II (dried leaves and stems from Tiradentes site) with epi-longipinanol (16.2%) rich oil, and cluster III (dried leaves from Lavras) containing a high content of beta-caryophyllene (17.4%) and alpha-copaene (10.1%). Canonical discriminant analysis showed that is possible to accurately predict 100% well-classification in the original clusters using beta-caryophyllene, epi-longipinanol and caryophylla-4(14),8(15)-dien-5beta-ol as predictor variables. The whole or sliced plant materials resulted in similar chemical composition.

Plants from Horto de Plantas Medicinais of Universidade Federal de Lavras (UFLA) were cultivated in the form of a randomized block with six replications in an experimental field at Setor de Horticultura, UFLA, in Lavras city (S 21° 14′, W 44° 59′, 920 m). Samples were collected four times at 3-month intervals from March to December 2004.

There was a significant difference for oil contents in the different seasons. The lowest oil content was obtained in the summer (about 50% inferior to the other seasons). Methyl (E)-cinnamate was obtained as the major compound (86-90%) of the 14 identified components.

Aerial parts of P. dysenterica were collected during the flowering stage from two different locations in Greece in August 2002. Sample A: Katara (Perfecture Trikala). Sample B: Arahova (Perfecture Viotia).

Fifty-four components were identified representing 80.5% (sample A) and 72.6% (sample B) of the total oils. The main components in sample A were (Z)-nerolidol (11.2%), caryophyllene oxide (9.1%) and (E)-nerolidol (6.6%), while those of sample B were beta-caryophyllene (12.8%), caryophyllene oxide (12.8%) and (E)-nerolidol (6.9%).

The aerial parts of T. chamaedrys were collected at the flowering stage in June 2004 near Corti, Corsica, France and near Oristano, Sardinia, Italy

The Corsican and Sardinian oils of T. chamaedrys investigated in this study were qualitatively similar but they differed by the amount of their major components. The major components were beta-caryophyllene (29.0% and 27.4%, respectively) and germacrene D (19.4% and 13.5%, respectively), followed by alpha-humulene (6.8%) and delta-cadinene (5.4%) in the Corsican oil and by caryophyllene oxide (12.3%) and alpha-humulene (6.5%) in the Sardinian oil. These quantitative differences are also noticeable on the amounts of the different class compounds. Especially, the monoterpene hydrocarbons amounted for 10.3% and 4.1% in Sardinian and Corsican oils respectively and the oxygenated sesquiterpenes amounted for 18.9% and only 7.4% in both oils, respectively. Both oils were qualitatively rather similar in comparison with those reported in the literature from various geographic regions. However, among the 87 components identified in this study, 47 minor components (< 0.6%) reported were identified for the first time in T. chamaedrys oil. This study confirms the quantitative variability of the major components according to the plant origin.