Artificial damage is a popular method in plant-herbivore studies, because the use of real herbivores is often laborious and because it may be virtually impossible to use herbivores in many experimental setups. We made a literature search of studies that tested whether natural and artificial damage have similar effects on plants. Of 46 studies found, 33 (72 %) reported a significant difference between responses to artificial and natural herbivore damage in at least one of the statistical tests included. The studies contained 280 statistical tests, of which 99 (35 %) showed a significant difference between artificial and natural damage. Phytochemical responses to artificial and natural damage were different in 41 % of the statistical tests and 75 % of the studies found at least one significant difference. Plant resistance, measured as secondary damage, herbivore performance, fungal growth in damaged tissue or plant attractivity to parasitoids of herbivores, differed in 60 % of the statistical tests and 85 % of the studies had significant differences. Growth, reproduction and physiological responses to artificial and natural damage differed in 20-30 % of statistical tests and 50-83 % of studies had significant differences. Thus, studies on plant tolerance (growth and reproduction after damage) more often showed similar effects for artificial and natural damage than studies on plant resistance to herbivory, but even in tolerance studies artificial and natural damage often have different effects. Some studies indicated that application of herbivore saliva and careful imitation of timing and spatial pattern of damage helped in reaching the same effect with simulations and natural damage.
IntroductionIn many experimental settings, artificial damage has several practical benefits over the use of real herbivores (Hjältén, Chap. 12, this Vol.). The extent of damage and the location of damaged parts can easily be controlled, and collateral damage to other than target tissues can be minimized. The removed biomass can be collected and measured. Furthermore, there is no need to collect and rear herbivores. The use of artificial damage enables efficient experimental designs, with balanced sample sizes of experimental groups and a low variation of treatment intensity within each experimental group. For these reasons, artificial damage is used more often in herbivory research than real herbivores.Artificial damage does not, however, always adequately mimic natural damage (Baldwin 1990; Hjältén, Chap. 12, this Vol.). Many types of herbivory are not applicable for simulations. Damage by stem borers, miners, galling insects, root feeders or sucking insects is seldom tried to simulate. However, even when the apparent damage pattern is easy to reproduce, several characteristics of natural herbivory may be difficult to simulate, such as the timing of damage, herbivore host choice, location of damage within a plant, and subtle details of damage by herbivore mouth parts. Herbivore saliva can also play a role in plant responses (Walling ...