Animal diets crucially affect fitness, yet many aspects of their ultimate determinants are unknown. The distribution and extent of herbivory in lizards, its evolutionary history, and ecological factors that may favour it are discussed. Most lizards are exclusively or primarily carnivorous, yet many species eat some plants and a few are almost exclusively herbivorous. Based on a literature survey of diets of over 450 lizard species, the distribution and degree of omnivory and herbivory are described. Some plants occur in the diets of slightly over half of lizard species, and plants formed 10% or more of the dietary volume of 12.1% of species, and 90% or more of the diet of 0.8% of species. The greatest percentage of omnivorous species (> 10% plant diet), over 30% in each, and highest mean percentage plant matter in the diet are in Iguanidae, Corytophanidae, Gerrhosauridae, Agamidae, Xantusiidae, and Tropiduridae. Numerous other omnivores occur in Lacertidae and Scincidae and fewer in several additional families. Herbivorous lizards (> 90% plant volume) tend to be folivorous and to possess adaptations for processing leaves, including specialized dentition for cutting or reducing leaves, elongated intestines, colic valves that slow passage of food, and intestinal flora that digest cellulose. Omnivorous lizards lacking such specializations may eat some leaves, but consume much more fruit, flowers, and seeds, plant parts that are easy to digest, likely to be very abundant seasonally, and may be highly nutritious. Some lizards eat nectar and pollen; even sap is eaten by at least one gecko. Ontogenetic increase in plant consumption and decrease in prey consumption is known, but its generality has been controversial. Such ontogeny has been demonstrated in three iguanid species, a skink, a lacertid, two tropidurids, a phrynosomatid, and two corytophanids, but it does not occur in some other species. The importance of specific foods may vary with age. Omnivory and/or herbivory have originated in many lizard families, with at least nine origins in Iguania and 23 in Scleroglossa. Origins have been rare in Gekkonoidea and Anguimorpha and common in Scincomorpha, especially in Lacertidae and Scincidae. Losses of omnivory have been much less frequent than gains. Only a few origins can account for all the herbivory in lizards. Concentrated changes tests show that there is a significant association in Lacertidae, Lacertiformes, Lacertoidea, Scincidae, and Scleroglossa between insularity and omnivory. Insular lizards may broaden their diets to compensate for limited availability of prey. Addition of other factors that reduce availability of prey, i.e. extreme aridity and cave-dwelling, to insularity, strengthened the relationship to omnivory in Lacertidae and Lacertoidea. We were unable to demonstrate a role of aridity independent of insularity, but present anecdotal evidence suggests that it may promote evolution of plant consumption. Large body size in lizards has long been associated with herbivory, and more recently, with omnivory in lacertid lizards. Using a conventional regression approach in which each species is considered to supply an independent data point, this relationship was confirmed for all lizards. Although larger species have diets with more plants, plant consumption accounts for only 9% of the variation in body length, which is not surprising given that other factors such as predation, competition, and sexual selection affect body size. The frequency of transitions body size associated with transitions to omnivory or carnivory was also examined. In Iguania, Scleroglossa, and all lizards, transitions supporting the hypothesis that omnivory favours increase in body size were significantly more frequent than non-supporting transitions. This suggests that substantial plant consumption favours evolution of larger size, probably because of the energetic considerations first presented by Pough (1973). Because actively foraging lizards move widely through the habitat to locate prey and tongue-flick to locate prey by chemical cues, we hypothesized that they may be more likely to evolve omnivory than ambush foragers, which wait motionless for prey and do not tongue-flick to locate or identify prey. The basis of this prediction is that the wider seaching of active foragers predisposes them to contact with a greater variety and quantity of plants and that chemosensory tongue-flicking used by omnivores to identify plant food might be easier to evolve in active foragers that already use pre-chemical discrimination. The prediction is supported by a significantly greater per species frequency of origins of omnivory by active foragers than by ambushers. A scenario for the progressive evolution of omnivory and herbivory from ancestrally carnivorous lizards is discussed.