Predicting yield loss due to interference from two weed species using early observations of relative weed leaf area

Experiments were conducted on linseed ( Linum usiãissimum L.) and autumn-sown field bean ( Vicia faha L.) to test how well the relative leaf area-based yield loss prediction model of Kropff & Spitters could estimate yield loss due to interference from Stellaria media L. (common chick-weed) or barley ( Hordeum vulgare L.) alone or a combination of both S. media and barley . Damage coefficients were calculated for all comparisons. Generally, the model provided a better estimate of yield loss due to interference from barley than from S. media because of both the variability in crop yield response stemming from S. media's plasticity and the generally minor crop yield response to S. media interference. The addition to the model of a parameter accounting for asymptotic yield loss was, generally, not warranted for either S. media or barley in the single-weed species data sets. When both S. media and barley were present as weed species in the crop, the model that fitted the data best for six out of 10 data sets was the one in which a separate damage coefficient was included for both species. There was no evidence that the presence of S. media influenced damage coefficient values for barley. However, in the presence of barley a parameter accounting for asymptotic yield loss was warranted for S. media in the two-weed species model. The combination of the addition of this parameter and the presence or barley affected the values of the damage coefficients for S. media in the two-weed species compared with the single-weed species models. Consequently, it was shown for these two weed species that two-weed species yield loss prediction models parameterized using data from single-weed species experiments generally resulted in different estimates of yield loss in comparison with models parameterized using data from two-weed species experiments.