A sample of 1834 leaves from 83 plants in 26 families was collected in tropical rain forest in Madagascar from three vertical strata: top of emergent trees (up to 37 m), top of trees in upper canopy (about 22 m), and shrubs and saplings at ground level. These leaves were examined for damage by seven different agents: fungi, epiphyllae, mechanical injury, galls, leaf miners, grazing insects and skeletonizing insects. Fungi affected more than 60% of the leaves and grazing insects 45–65%, with other agents each affecting 3–30%; fewer than 4% of the leaves escaped unscathed by any agent. Individual leaves were attacked by up to five agents. There was a sharp decline in proportion of leaves affected by fungi, leaf miners, epiphyllae and mechanical breakage with increasing category of severity. Grazing insects, mechanical injury, and perhaps galls, had greater impact at ground level than in the canopy, with fungi and skeletonizing insects showing the opposite pattern. Leaf miners had lower incidence in the canopy than elsewhere. The observed vertical stratification means that a tree not only needs to balance its defences to meet multiple threats in any stratum, but must adapt to a different suite of challenges during its lifetime. Attack by grazing insects and fungi are major challenges for saplings at ground level, but with increasing height above the ground fungal attack on leaves becomes more prevalent, but attack by grazing insects less so.

The horse-chestnut leaf miner, Cameraria ohridella, is an invasive alien species defoliating horse-chestnut, a popular ornamental tree in Europe. This paper presents quantitative data on mortality factors affecting larvae and pupae of the leaf miner in Switzerland and Bulgaria, both in urban and forest environments. Two sampling methods were used and compared: a cohort method, consisting of the surveying of pre-selected mines throughout their development, and a grab sampling method, consisting of single sets of leaves collected and dissected at regular intervals. The total mortality per generation varied between 14 and 99%. Mortality was caused by a variety of factors, including parasitism, host feeding, predation by birds and arthropods, plant defence reaction, leaf senescence, intra-specific competition and inter-specific competition with a fungal disease. Significant interactions were found between mortality factors and sampling methods, countries, environments and generation. No mortality factor was dominant throughout the sites, generations and methods tested. Plant defence reactions constituted the main mortality factor for the first two larval stages, whereas predation by birds and arthropods and parasitism were more important in older larvae and pupae. Mortality caused by leaf senescence was often the dominant mortality factor in the last annual generation. The cohort method detected higher mortality rates than the grab sampling method. In particular, mortality by plant defence reaction and leaf senescence were better assessed using the cohort method, which is, therefore, recommended for life table studies on leaf miners.