**2. The frugivory hypothesis**

The oldest hypothesis for the evolution of primate colour vision is that trichromacy is an adaptation to frugivory (Allen, 1879; Polyak, 1957). This hypothesis builds on the observation that most primates rely heavily on dietary fruit (Chapman & Onderdonk, 1998; Fleagle, 1999). Given that many tropical plants produce red, orange or yellowish ("colourful") fruits and that trichromatic colour vision enhances primates' abilities to detect these colours amidst green foliage and unripe fruits, trichromatic colour vision should facilitate the detection and selection of edible fruits (Regan et al., 2001). Primates in turn may provide high-quality seed dispersal services to these fruiting plants (Garber & Lambert, 1998). The ability to distinguish between ripe and unripe fruits, fruits from leaves, and edible from inedible species is a complex task that could be facilitated by trichromacy (Smith et al., 2003), and there is compelling experimental and behavioural evidence for trichromatic advantages in these tasks.

The frugivory hypothesis was originally proposed and tested for catarrhine primates (Sumner & Mollon, 2000b) because until relatively recently platyrrhines (Mollon et al., 1984) and strepsirrhines (Tan & Li, 1999) were not known to possess trichromacy. The frugivory hypothesis has since been extended to platyrrhines and tested behaviourally (Caine et al., 2003). Theoretical studies have also assessed the conspicuity of dietary fruits to trichromatic versus dichromatic platyrrhines (Riba-Hernández et al., 2005; Stoner et al., 2005). Research evaluating the frugivory hypothesis is reviewed briefly below.

Tropical fruits exhibiting a "primate dispersal syndrome" (medium-sized red, orange or yellow fruits with succulent pulp, large seeds, tough exocarps; (Gautier-Hion et al., 1985)) have been found to occupy a narrow region of colour space that is detectable to trichromatic platyrrhines (Regan et al., 2001). Furthermore, the reflectance spectra of most primate-consumed fruits in Uganda were found to show chromatic changes as they ripen, and trichromatic catarrhines should be well adapted to discriminate fruit ripeness (Sumner and Mollon, 2000b). In the New World, the peak spectral sensitivities of trichromatic platyrrhine cone pigments were found to be well suited to the detection of fruits against a background of leaves under photopic conditions (Regan et al., 2001). The spectral tuning of the L-M cone pigments in the trichromatic platyrrhine *Alouatta seniculus* (Linnaeus, 1766) was found to be optimal for detecting ripe fruits against a background of leaf "noise" (Regan et al., 1998). Among six phenotypes in one white-faced capuchin (*Cebus capucinus*, Linnaeus, 1758) population, monkeys possessing the phenotype with the most spectrally separated L-M opsin alleles showed the highest acceptance index for red fruits (Melin et al., 2009). For this same population of capuchin monkeys, colourful fruits comprise 60% of the diet and are preferred to cryptically coloured fruits (Melin, 2011).

Trichromatic vision may also afford a selective advantage by allowing trichromats to forage on fruit under a greater range of light conditions than dichromats (Yamashita et al., 2005). In two species of *Saguinus* (Hoffmannsegg, 1807), captive trichromatic individuals were more efficient than dichromats at selecting ripe fruits, both in isolation and against a background of leaves (Smith et al., 2003). In a study linking nutritional value to fruit colouration, Riba-Hernandez et al. (2005) have provided additional support for this hypothesis. Arguing that the principle reward of ripe fruits is sugar, these authors demonstrated a positive correlation between fruit glucose content and the red-green colour channel of spider monkeys (*Ateles geoffreyi*, Kuhl, 1820).

In summary, the evidence for a ripe fruit foraging advantage for trichromats is compelling. Behavioural evidence for ripe fruit foraging advantages in wild and captive trichromats, the trichromat-accessible colour space of tropical fruits, the dietary importance of colourful fruits, and a positive correlation between nutritional value and fruit colouration all indicate that a ripe fruit foraging advantage may indeed result in the evolution and persistence of trichromacy. However, the frugivory hypothesis is problematic on several accounts. In the sections below, we review problems with the frugivory hypothesis and alternate hypotheses regarding the evolution of trichromacy.
