Primates are noted for their varied and complex pelage and bare skin coloration but the significance of this diverse coloration remains opaque. Using new updated information, novel scoring of coat and skin coloration, and controlling for shared ancestry, we reexamined and extended findings from previous studies across the whole order and the five major clades within it. Across primates we found (i) direct and indirect evidence for pelage coloration being driven by protective coloration strategies including background matching, countershading, disruptive coloration and aposematism, (ii) diurnal primates being more colorful, and (iii) the possibility that pelage color diversity is negatively associated with female trichromatic vision; while (iv) reaffirming avoidance of hybridization driving head coloration in males, (v) darker species living in warm, humid conditions (Gloger’s rule), and (vi) advertising to multiple mating partners favoring red genitalia in females. Nonetheless, the importance of these drivers varies greatly across clades. In strepsirrhines and cercopithecoids, countershading is important; greater color diversity may be important for conspecific signaling in more diurnal and social strepsirrhines; lack of female color vision may be associated with colorful strepsirrhines and platyrrhines; whereas cercopithecoids obey Gloger’s rule. Haplorrhines show background matching, aposematism, character displacement, and red female genitalia where several mating partners are available. Our findings emphasize several evolutionary drivers of coloration in this extraordinarily colorful order. Throughout, we used coarse but rigorous measures of coloration, and our ability to replicate findings from earlier studies opens up opportunities for classifying coloration of large numbers of species at a macroevolutionary scale.

Torso pattern as 0 = Not black-and-white or may be all black or all white; 1 = ‘Salt and pepper’ like splotches of black-and-white that appear conspicuous; 2 = Transverse stripes of black-and-white; 3 = Longitudinal stripes of black-and-white; 4 = A solid block of black/dark brown/dark grey on the dorsum with a solid block of white/light yellow/light grey on ventrum; 5 = A solid block of white/light yellow/light grey on dorsum with a solid block of black/dark brown/dark grey on ventrum; 6 = Irregular blocks of black-and-white; 7 = Disordered stripes of black-and-white, often resembling contor lines; 8 = Large spots of black on white background or vice versa.

Tail color proportions were scored as 0 = Not black or white (may be grey, brown or red); 1 = All black; 2 = Black more than 50% and white less than 50%; 3 = 50% of tail black and 50% white; 4 = White more than 50% and black less than 50%; 5 = All white; 999 = No tail. In tail analyses we reformulated these categories as having any white fur (#2-5) or not.

Total number of distinct color and separately shades (Figure S2) were documented on each part of the body (with increasing numbers here termed as being ‘more colorful’). Species were additionally categorized as having either red (yes or no) and/or dark (yes or no) torso coloration (as described in the legends SOM Figures 1 and 2).

Male and female genitalia scores were 0 = not different from surrounding pelage, 1 = white, 2 = red or pink, 3 = black, 4 = yellow, 5 = orange, 6 = blue. For females, we focused on category 2 in analyses as the vast majority of colorful female genitals are red or pink. Male genitalia are often multicolored so for males we combined categories 1-6 in analyses.

Activity cycle: 1 = Nocturnal, 2 = Crepuscular, 3 = Diurnal, 4 = Cathemeral, 12 = Nocturnal/Crepuscular, 13 = Diurnal/Crepuscular’ 0 = absence of data. We combined categories 2 and 4 into a new category diurnality.

Group size: 1 = 1 individual (solitary); 2 = 2 - 5 individuals; 3 = 6 - 25 individuals; 4 = 26 - 100 individuals; 5 = >100 individuals.

Social organization: (i) pair bonded; (ii) multimale; (iii) fission-fusion; (iv) multimale and multifemale, or fission-fusion; (v) multimale, one female, (vi) one male and multifemale, and (vii) solitary foraging, (viii) variable, i,e., reported as (i) or (ii), or (i) or (vi), or (ii) or (vi), or a  ‘a few males’. We subsequently lumped (iii) and (iv) together as ‘fission-fusion societies’. We combined (ii), (iii), (iv) and (v) together as ‘multimale societies’ irrespective of female number.

Multilevel societies from Table S2 in Grueter et al (2015).

Mating system: (i) monogamous, (ii) polygynous, (iii) polyandrous and (iv) promiscuous (also termed polygynandrous). Some species displayed more than one mating system ((v) monogamous and polygynous, (vi) polygynous and promiscuous, (vii) monogamous and polyandrous, (viii) polyandrous and promiscuous, and (ix) monogamous and promiscuous) and in these cases, we recorded both. We lumped categories (ii), (v) and (vi) together as ‘polygynous’; and categories (iv), (vi), (viii) and (ix) together as ‘promiscuous’.

Hair: “long” or “shaggy” (1) or else as short/absent (0). Fur thickness was scored as dense if reported as “dense”, “shaggy” or “thick” (1) or as not dense (0). These two variables were subsequently combined to generate a new variable thick hair, where species that had either long or thick hair were scored as 1, and those that had scores of 0 for both were scored as 0.

Species range overlap. Species were scored as either having sympatric congeners (1) or not (0).

Shade scores: see methods.

Visual system: monochromatic, dichromatic, trichromatic or polychromatic (i.e., sex-linked alleles in females). If a species was trichromatic or polymorphic (for females), it was recorded as a 1. If a species was dichromatic or monochromatic, it was recorded as a 0. Trichromatic males were scored as 1, otherwise 0.

But please see Methods in the paper

Contact Tim Caro or Theodore Stankowich for details