1 Experimental tests of whether predation pressure on insects is sometimes restricted to particular forest microhabitats have been carried out only in one or two vegetation periods and described for only a few predators. In the present study, we describe the seasonal dynamics of a wide spectrum of insect predators among forest microhabitats. 2 We also examine the impact of weather conditions on insect predation, and predict that forest openness would influence the predation trends among forest microhabitats. 3 The design of our experiments enabled direct measurement of relative predation pressure on bait (larvae of the blowfly Calliphora vicina) pinned onto selected microhabitats (the base, trunk and leaves of trees) within a temperate floodplain forest (Czech Republic). 4 The most parsimonious generalized additive model showed significant trends in the predation rate among the forest microhabitats. The highest predation rate for bait was at the base of trees and the lowest predation rate was on leaves. 5 We also observed significant differences in the species structure of predators in various microhabitats. The most common source of predation on trunks was from birds, whereas wasps were the most common predator on leaves and ants were the most common at the base of trees.in biomass partitioning among forest layers can influence the abundance of predators, with higher densities tending to be found in layers with the greatest amount of biomass (De Dijn, 2003;Vance et al ., 2007). This information can be useful when examining whether certain safe refuges for prey exist. For insects, spatial refuges can be located on different scales: (i) various habitats within a plant; (ii) different layers within the forest; and (iii) different types of forest habitat.Studies from tropical and temperate forests imply that predation activity differs across forest layers (Jeanne, 1979;Olson, 1992;Loiselle & Farji-Brener, 2002). The main disadvantages of recent predation studies are that they: (i) focus only on a few predators (mostly ants and birds); (ii) are carried out only in one or two growing seasons; (iii) do not examine diurnal or seasonal variability in the predation rate in different forest layers; and (iv) do not examine the effect of