Pollination in Viola results both from sternotribic and nototribic pollinator responses, the frequencies of which vary according to the plant species. Syndromes of floral characters determine these frequencies, and each is adaptive to a group of pollinators, of variable taxonomic heterogeneity, characterised by particular morphology and behavior. Some syndromes elicit equal frequencies of sterno- and nototribic responses, while others elicit chiefly, or exclusively, one type, and coevolution with a single insect genus may be in progress. One effect of the differing syndromes is the partitioning of pollinator resources among simultaneously flowering species. An evolutionary sequence for the floral syndromes is proposed which, in broad agreement with systematic conclusions, considers yellow-flowered, sternotribic Chamaemelanium violets as the most ancient and blue-flowered sterno- or nototribic violets of the sections Rostellatae and Plagiostigma as the most recently evolved. Species of the section Melanium (the pansies), in which cleistogamy is almost absent, appear to be an early offshoot leading to multicolored, nototribic flowers. The "generalist" syndromes were probably vital to the success of the genus in (a) invading the northern and southern temperate zones and (b) adapting to widespread habitat disturbance by man. The basic structure of the violet flower is adaptive to a temperate pollinator complex to be found over very large geographic areas. Floral variations demonstrate adaptive radiation to many ecological conditions. One of the interesting fields remaining in pollination biology is the creation of a synthetic theory of evolution that links the processes of differentiation of pollination systems as observed in living plant populations with the major trends of coevolution over very long periods of time. The evolution of pollination systems in individual angiosperm families has been the subject of several elegant studies such as Lewis and Raven (1961), Grant and Grant (1965), and Eyde and Morgan (1973). In most cases the sequencial stages in evolution have been typological; the data indicating only major changes in mode of pollination from one group of animals to another, for example from insects to birds. The actual processes of differentiation are still very poorly understood. The purpose of this paper is to suggest processes of change in the pollination systems of Viola and then to examine their operation both from contemporary and historical perspectives. Therefore, discussion will focus first on data from living populations and second on how this data may be applied to a phylogeny of pollination systems in the genus.