We also hypothesized that the improvement of habitat quality by the alternative practices was likely to outweigh potentially negative density dependent effects on hatching success. The presence or absence of prey carrying in each individual territory was studied in the most abundant species using Generalized Linear Mixed Models (GLMMs) with binomial error distribution. In the Saône Valley, the fixed factors were meadow passerine territory density in study sites and the situation of each territory either inside (at least partly) or outside the areas devoted to mowing postponement; study sites and years were used as random effects. In the Moskva Valley, the fixed factors were mowing management the preceding year (no mowing, mowing for only one year, for several consecutive years) and passerine territory density in corresponding replicates; study sites were used as a random effect.