The weed community was composed by the follow family and species: Amaranthaceae: Alternantera tenella Colla, Amaranthus deflexus L.; Asteraceae: Acanthospermum hispidum DC., Acanthospermum australe (Lofl.) Kuntze., Bidens pilosa L., Blainvillea rhomboidea Cass., Emilia sonchifolia DC.; Commelinaceae: Commelina benhalensis L.; Convolvulaceae: Ipomoea acuminata Roem. et Schult; Cyperaceae: Cyperus rotundus L.; Euphorbiaceae: Chamaesyce hirta (L.) Millsp., Chamaesyce hissopifolia (L.) Small., Plyllanthus tenellus Roxb.; Malvaceae: Sida rhombifolia L.; Mimosaceae: Aeschynomense rudis Benth.; Poaceae: Brachiaria decumbens Stapf., B. plantaginea (Link.) Hitch; Cenchrus echinatus L., Cynodon dactylon (L.) Pers., Digitaria horizontalis Willd., Eleusine indica (L.) Gaertn., Panicum maximum Jacq.; Portulacalaceae: Portulaca oleraceae L.; Rubiaceae: Richardia brasiliensis Gómez, Spermacoce latifola Aubl.; Família Solanaceae: Physalis angulata L., Solanum americanum Will.

The dicots predominated, corresponding to 75% of all species found. The most numerous dicot families were Asteraceae, with five species, and Euphorbiaceae, with three species. Among the monocots, the major family was Poaceae, with seven species, which among all weed families present in the experimental area, was the most numerous. The species from Poaceae corresponded to 26% of all species found.

According to Clark (1971), the different species composing an infesting community can affect, in part, the degree of interference, since the competitive ability varies among the species. It is hypothesized that the more physiologically similar two species are, and often these are related to the taxa classification, the closer their needs will be and the more intense the competition for the limited factors of the common ecosystem.

It was observed that in terms of number of individuals (Fig. 1), after the 15th day of planting the infesting community was established, reaching its maximum estimated density on day 30, with 565 plants/m². Later the population density decreased until day 87, when it reached 67 plants/m². This reduction in density was probably due to the interference between the individuals, which caused plant mortality in the infesting community. Dekker and Meggit (1983), in studies of population dynamics, observed that the increase in population density of Datura stramonium L. (jimsonweed), associated to plant growth, induced an increase in the mortality rate. In the present study, after the stabilization of the interference effect, an increase in the weed density was observed (Fig. 1), probably resulting from a new emergence outbreak, reaching 235 individuals/m², on day 105 after peanut planting.

The increase in weed density, until 30 days after peanut seeding, could have affected the crop. Direct competition for the medium resources may have affected the crop. Otherwise, the return of a population increase, at the end of the growing season, also could have affected the crop, prevailing the indirect interference, obstructing the harvest process, in operational terms as well as in maturation time for the peanut branches.

Bhan et al. (1971) observed, in a study done in India, that 75% of the components of an infesting community of the peanut crop could emerge in the first 30 days of the growing season. In Brazil, Pitelli (1980) observed that in a “rainy season” peanut crop the maximum number of individuals was established at 20 days. In a “dry season”, Pitelli et al. (1984) observed that this establishment occurred after 14 days.

The weed community accumulated dry biomass linearly, until 66 days after peanut seeding, when a maximum accumulation of dry matter (1049 g/m²) was observed (Fig. 2). Dry matter accumulated decreased at 105 days, reaching a value of 630 g/m², due probability to natural senescence. Making a correlation between the data on infesting community density and dry matter accumulation, one can observe that, although there was plant mortality as a consequence of interference among individuals of the infesting community, the plants remaining in the area developed and accumulated dry biomass.

Among the species of the infesting community the three most important were Brachiaria plantaginea (Link) Hitchc. (alexander grass), Digitaria horizontalis Willd. (crabgrass) and Bidens pilosa L. (hairy beggarticks) (Fig. 2). In terms of population density, D. horizontalis was the most numerous species in most evaluations, followed by B. pilosa and B. plantaginea (Fig. 3). At 30 days after peanut seeding, when the infesting community reached its maximum density, D. horizontalis contributed with 22% (123 plants/m²) of this infestation, B. pilosa with 11% (62 plants/m²) and B. plantaginea with 5% (28 plants/m²). At 105 days, the major weed responsible for the new population density increase was B. pilosa with 17% (36 plants/m²), followed by D. horizontalis with 12% (26 plants/m²) and by B. plantaginea with 6% (12 plants/m²).

Digitaria horizontalis was also the weed that presented the greatest dry biomass accumulated in most evaluations, reaching, at 75 days, the maximum accumulation of 240.0 g/m² (Fig. 4). Only when the dry biomass accumulation by the infesting community was maximum, at 60 days after planting, B. plantaginea reached greater value of dry biomass accumulated than D. horizontalis, accumulating 287.2 g/m². At 60 days after planting, the population of B. plantaginea represented 26% of the total accumulated by the infesting community and the populations of D. horizontalis and B. pilosa represented 15% (161.4 g/m²) and 8% (87.9 g/m²), respectively.

When the weight of 100 seeds is compared between peanut cultivars that had grown free of weed interference (weed-free) (Fig. 5), one observes that the cultivar that presented the greatest weight was ‘Caiapó’ (61.9 g), followed by Runner ‘Tégua’ (60.2 g), ‘IAC-22’ (56.3 g), ‘ST-Tatu’ (45.5 g) and ‘IAC-1075’ (42.9 g). Comparing the weights of 100 seeds from the treatments maintained weed-free with those maintained weed-infested during the whole growing season, the weed interference caused a 15% reduction on the weight of 100 seeds of cultivar ‘Caiapó’, 14% of Runner ‘Tégua’, 31% of ‘IAC-22’, 8% of ‘ST-Tatu’ and 15% of ‘IAC-1075’. Thus, the cultivar most sensitive to weed interference, in relation to the weight of 100 seeds, was ‘IAC-22’ and the least sensitive ‘ST-Tatu’.

Considering the weed-infested periods, the weight of 100 seeds for all cultivars adjusted well to the linear regression (Fig. 5 and Table 1), where, according to the equations obtained, the reductions on the weight of 100 seeds were of 0.06 g/day for cultivar ‘Caiapó’, 0.07 g/day for Runner ‘Tégua’, 0.13 g/day for ‘IAC-22’, 0.03 g/day for ‘ST-Tatu’ and 0.06 g/day for ‘IAC-1075’. The increase on the period that the crop was kept weed-free resulted in a positive effect on the weight of 100 seeds for all cultivars, with the data adjusted to the Boltzman sigmoidal equation.

Pitelli (1980), Bianco (1978) and Ishag (1971) also observed negative effects of the infesting community presence on the average weight of 100 peanut seeds. Pitelli (1980) mentioned that, due to the great non-uniformity of the infesting community in the crop areas, peanut plants are subjected to the most diverse conditions of shading, competition, of soil thermal and hydric variations, of allelopathic processes and of other factors that could direct or indirectly, change flowering. Thus, with the non-uniformity of flowering and, fruit set, by harvest there would be mature seeds side by side with growing seeds. This would make the average size of the seeds smaller and reflect on the weight of 100 seeds. However, Dajos (1983) stated that, generally, the competition between plants leads to a reduction on the individual weight of a plant and on the number of seeds, but the weight of each seed tends to be constant. Gavioli (1985) observed that in the commercial classification of the seeds as well as in the average size, the seeds did not differ as a consequence of interference by an infesting community, in which D. horizontalis predominated. That author attributed this result to the specific condition of that study, where the infestation was uniform, differently from what was reported by Pitelli (1980).

In the absence of weed interference, the most productive cultivar was ‘Caiapó’ (3490 kg/ha), followed by Runner ‘Tégua’ (2640 kg/ha), ‘IAC-22’ (2370 kg/ha), ‘ST-Tatu’ (2180 kg/ha) and ‘IAC-1075’ (1280 kg/ha) (Fig. 6). Weed interference during the whole growing season caused a 88% yield reduction in cultivar ‘Caiapó’, 86% in Runner ‘Tégua’, 92% in ‘IAC-22’, 74% in ‘ST-Tatu’ and 88% in ‘IAC-1075’. Thus, similarly to what was observed for the weight of 100 seeds, the most sensitive cultivar to interference, in relation to yield, was ‘IAC-22’ and the least sensitive ‘ST-Tatu’, both of them with an erect growth habit.

Previous studies indicated that maximum peanut loss was 63% in Brazil (Bianco, 1978; Pacheco, 1980; Pitelli et al., 1981; Pitelli et al., 1984; Gavioli, 1985; Barbosa and Pitelli, 1990; Martins and Pitelli, 1994; Kasai et al., 1997). For the time of weed removal, the critical periods in the current study were smaller than those previously published. Therefore, in the present study, weeds became to interfere with crop growth earlier in the season, perhaps due to elevated weed population density observed in the beginning of the growing season (Fig. 1). Critical weed-free period in this study were longer than those previously published in Brazil, which could may be attributed to wider row spacing (90 cm), relative to previous studies (<70 cm). The broader spacing impedes the soil shading by the crop, allowing the weeds to germinate for a longer period and, in turn, longer control periods could be needed.

Analyzing the yield of the cultivars as a function of weed-infested or weed-free periods indicated that the Boltzman sigmoidal equation best fit all data (Fig. 6 and Table 2). When the cultivars were maintained weed-infested for increasing periods, there was a more intense yield reduction between 15 and 75 days after planting. As a function of the length of weed-free period with, more intense increase on yield occurred especially when the weed control was done in the first 30 days. Longer control periods did not provide great increases in peanut yield.

Accepting a maximum yield loss of 5% the critical time of weed removal was 13, 7, 12, 16, and 7 days and the critical weed-free period was 47, 65, 31, 26 and 61 days for the cultivars ‘Caiapó’, Runner ‘Tégua’, ‘ST-Tatu’, ‘IAC-1075’ and ‘IAC-22’, respectively (Fig. 6 and Table 3). Consequently, the critical periods of weed control were 13 to 47, 7 to 65, 12 to 31, 16 to 26, and 7 to 61 days for these cultivars, respectively, with no observations of behavior patterns for either growth habit, runner or erect. For the present results the critical period for weed control for all cultivars should begin at 7 DAP and continue until a maximum of 65 DAP.

In the literature review from Brazilian experiments until 2005, no long critical periods of weed control were found. When this period was evaluated, it was determined that only one weed control practice would be enough to ensure high crop yield (Bianco, 1978; Pacheco, 1980; Pitelli et al., 1981; Pitelli et al., 1984; Gavioli, 1985; Barbosa and Pitelli, 1990; Martins and Pitelli, 1994; Kasai et al., 1997). In a comercial area of peanut with Cyperus rotundus, Commelina benghalensis, and Euphorbia heterophylla the productivity was reduced in 76%, with a critical period of weed control of 34 to 90 days after peanut emergency, with a great difficult for mechanical harvest (Nepomuceno et al., 2005). In a weed community composed by Xanthium strumarium, Cenchrus echinatus, Indigofera hirsuta, Portulaca oleraceae, and Digitaria horizontalis, Dias et al. (2005) founded critical periods of weed control of 30 to 88 days after peanut emergency for row spacing of 80 cm and 38 to 101 days for row spacing of 90 cm, with a yield loss more than 80%. All long critical periods of weed control were found outside Brazil. Rodrigues Marquina et al. (1974) determined a critical period of weed control of 30 to 60 days after planting. Drennan and Jennings (1977) determined such a period between six and ten weeks after peanut emergence and Hill and Santelmann (1969) between three and six weeks.