Wondering if somebody has quantitative data for observations on ants attacking herbivores and/or pollinators (as # observations/period of time) without any sort of manipulation. In other words, a person sits, observes, and records a natural predation event involving insects. There is interesting research on predation by (e.g.) spiders or ants, but it seems obvious that conclusions are mostly based on the results of removing the predator (either naturally not present in some plants, but present in others or experimentally removed) and measuring the effects of this onto a variety of plant features (involving vegetative growth or reproduction). Is this all what we have? Can anecdotal observations only be counted with one hand?