In order to evaluate the herbicide low rate concept, the weed flora associated to the crop was surveyed during 1995 and 1996 to estimate the importance of the different species. Between 1997 and 1999 field herbicide trials were made using mixtures of nine active ingredients.
Randomized complete block design is arranged, with untreated check plots to make easier the studies about the efficiency of the different herbicide mixtures. The analysis of variance is made. Duncan's multiple range test is used to establish which treatments are significantly different and also Dunnett method in field trials ( 1999) to compare a particular check treatment (pendimetalina 0.990 kg a.i. ha^-1) with each of the other different treatments.
The weed species with more importance in the tobacco crop were: Echinochloa crusgalli, Polygonum spp. (hydropiper, lapathifolium, persicaria), Stellaria media, Chenopodium album, Portulaca oleracea, Solanum nigrum, Sulanum physalifolium, Capsella bursa-pastoris, Anummthus hybridius y Arrhenatherum elatius subsp. bulbosum.
The mixtures (kg a.i. ha^-1) "metobromuron 0.25 + (oxadiargyl 0.06 + aclonifen 0.2) + oxyfluorfen 0.12 + propaquizafop 0.05" and ''linuron 0.225 + (oxadiargyl 0.06 + aclonifen 0.2) + oxyfluorfen 0. 12 + propaquizafop 0.05", have given the best results for all the weed species, with average efficiencies of 92.6% and 89% respectively.
Thus, we can conclude that in the former treatments there was a decrease of herbicide volume used, in comparison with the check treatment. Indeed, it has been confirmed that the use of active complementary ingredients amplifies the efficiency of this kind of treatments and prevents the selection of resistance flora. Finally, a stop of the vegetative growth has not been shown and the crop tolerance to the different herbicide mixtures has been satisfactory.