The model proposed in the first paper in this series predicts that in mixtures of plant species with different rooting depths there will be an inverse correlation between the relative crowding coefficient of the deep rooting species with respect to the shallow rooting one and the frequency of the deep rooting plants. Two field experiments are reported in which this phenomenon was observed. The first experiment involved Plantago lanceolata L. and Lolium perenne L. and the second involved Plantago lanceolata L. and Anthoxanthum odoratum L. In a third competition experiment between Plantago lanceolata L. and Anthoxanthum odoratum L. it was tested whether this observed phenomenon was indeed caused by the differences between the rooting depths of these species. The experiment was divided into one series where different rooting depths were possible and one series where gauze with a narrow mesh width prevented differences between rooting depths. In the series without gauze the predicted frequency-dependence of the relative crowding coefficient was observed, whereas in the series with gauze the relative crowding coefficient was independent of plant frequency. A comparison between the yields in the series with and without gauze suggests that in mixture Plantago is forced to utilize the nutrients from deeper soil layers, whereas this is not the case in monoculture. Although such phenomena complicate the description of the process investigated, it is concluded that the model proposed provides an approximate prediction of most results of the experiments which were carried out.