ABSTRACT
   Since man learned to domesticate and grow plants in the Neolithic period, resulting in the birth of agriculture, has needed to maximize crops production. To increase yields, different techniques are employed: fertilization, irrigation and pests (insects, fungal pathogens and weeds) control. The development of agrochemical products in the 1940s increased agricultural productivity, based on the employment of chemical fertilizers and pesticides and the use of genetically modified crops. The overuse of agrochemicals involved negative consequences: the excess use of synthetic herbicides caused resistant weeds and severe contamination to the environment, due to accumulation in soils and ground water, with negative impacts in li-ving organism and human health.
   Since the 1990s, agrochemical products regulation is becoming more restrictive, according with the requirements of a society concerned about negative impacts of its widespread use. Other techniques based on natural products to control weeds are being developed, because of its safety for environment and human health. Allelopathy, based in plant interactions (including microorganisms) through release of secondary metabolites by volatilization, exudation and lixiviation from vegetal tissues, offers new possibilities for weed control. Allelochemicals that suppress plant species near the source plant have received special attention due to its potential as natural herbicides.
   The main objective of this work is to test phytotoxic activity of essential oils and aqueous extracts from mediterranean plants against weeds germination and seedling growth, to select the more actives in order to develop them as natural herbicides. Species that elaborate essential oils were selected based on the existing literature about the biological activities of their secondary metabolites or of that from species taxonomically closed: Lantana camara L., Eucalyptus camaldulensis Dehnh., Eriocephalus africanus L., Cistus ladanifer L., Artemisia gallica Willd., Artemisia annua L., Lavandula angustifolia Mill., Rosmarinus officinalis L., Thymus capitatus (L.) Hoffmanns. et Link, Tagetes lemmonii A. Gray, Pelargonium odoratissimum (L.) L’Hér., Thymus vulgaris L. and Origanum vulgare L. and their essential oils and aqueous extracts were obtained. As weedy species were selected Amaranthus hybridus L., Portulaca oleracea L. and Chenopodium album L., because they are considered serious and problematic weeds in many crops around the world, and another two ruderal weeds recently found in crops due to changes in agricultural techniques, as no-tillage practices, Conyza canadensis (L.) Cronq. and Parietaria judaica L.
   Essential oils were analyzed by means of GC and GC-MS and aqueous extracts by HPLC-MS.
   In vitro assays were performed in germination chambers to assess the effects of essential oils and aqueous extracts on weeds germination and seedling growth. Essential oils and aqueous extracts from L. camara, E. camaldulensis, E. africanus, C. ladanifer and A. gallica were tested against all selected weeds. The other essential oils were tested against both P. oleracea and C. canadensis. All essential oils were phyto-toxic against at least one weed species, being their activity dose-dependent or not showing significant differences between concentrations. In general, essential oils revealed to be more phytotoxic than aqueous extracts obtained from the same species, acting against more weeds and causing more severe damages. A complementary activity was observed between essential oils and aqueous extracts from L. camara and E. africanus, as one was effective against species where the other did not show any effect and vice versa. Essential oils from E. camaldulensis, with spathulenol as the main compound, and T. capitatus, which contained high levels of carvacrol, were the most effective treatments. A. hybridus, C. canadensis and P. judaica germination was very sensitive to the majority of essential oils, whereas P. oleracea and C. album showed more resistance. None of the essential oils tested inhibited completely C. album germination. In this species, aqueous extracts manifested greater phytotoxic activity, being L. camara aqueous extract the best treatment to control it. C. canadensis was the most sensitive species to aqueous extracts effects. 
   Herbicidal potential of L. camara, E. camaldulensis, E. africanus and C. ladanifer aqueous extracts was tested under greenhouse conditions. All except C. ladanifer were also evaluated under field conditions. E. camaldulensis and E. africanus showed more activity. Their effect, after only one application, was maintained during six weeks, under greenhouse conditions. In field conditions, three applications every fifteen days extended their effect for sixteen weeks. Likewise, the herbicidal potential of E. camaldulensis and E. africanus essential oils, which showed phytotoxic activity in vitro, was tested under greenhouse conditions, displaying no effect at the doses employed (0.5 µl/ml).
   The results obtained from in vitro bioassays revealed that essential oils and aqueous extracts tested showed selective phytotoxic activity, being their activity dependent of the species against they act, so they could be developed as natural selective herbicides. However, future experiments, involving greenhouse and field conditions must be performed in order to verify their herbicidal potential. Aqueous extracts maintained their activity under greenhouse conditions but essential oils lost it, due to their high volatility. Application of essential oils must be improved with other formulations, as microencapsulation, which would simplify their handling, and increase their effectiveness by reducing their volatilization and slow down their degradation in the environment.