Since the apparent health of human faces is very closely associated with attractiveness (Joneset al., 2004, 2005), the preferences discussed in this thesis may have affected, and indeed may still be affecting the sexual selection of human skin colour, driving selection for yellower and redder skin, containing more carotenoid and more
sexual selection. Such an association may lead to the conclusion that the preferences described in this thesis are unlikely to have been subject to sexual selection, as human skin is not bright red or bright yellow. This conclusion would be unwarranted for a number of reasons.
Firstly, in order for an exaggerated trait to evolve, sexual selection must initially act on a much less exaggerated form of the trait. Fisher (1958) noted that a trait such as slightly longer tail feathers in a bird species may confer on a male a survival and foraging advantage via improved agility. Females who mate with males with the longer tail feathers would therefore obtain the direct benefits of improved male
provisioning, as well as the indirect benefits of obtaining alleles for slightly longer tail feathers for her offspring. In this way, selection for female choice for an advantageous trait begins the process of the coevolution of female preference and male trait in a classic Fisherian process (Andersson, 1994; Fisher, 1958). The preference for
carotenoid colour in human faces may be analogous to the early stages of this process. Individuals who consume more carotenoids and more fruit and vegetables in the normal diet display more carotenoid in their skin and therefore have yellower skin (Section 2). Serum carotenoid levels, which correlate with skin carotenoid levels (Stahlet al., 1998), are affected by the efficiency of uptake of carotenoids from the gut after consumption of fruit and vegetables (Stahl & Sies, 1992), which can be disrupted by gut parasites (Horaket al., 2004). Increased carotenoid levels are thought to have a number of health benefits, including improved reproductive (Brief & Chew, 1985; Schweigertet al., 2003) and immune function (Hughes, 1999, 2001) and photoprotection (Matthews-Rothet al., 1974), and individuals with increased carotenoid levels may historically have been better foragers and be less affected by
gut parasites (Horaket al., 2004). Individuals who prefer to mate with individuals who have this health benefit would therefore obtain the direct benefits of improved provisioning and reduced exposure to parasites. They may also obtain indirect benefits of genes for parasite resistance and improved carotenoid uptake for their offspring. This would provide a selection pressure for sexual choice for increased carotenoid colour. Support is provided for this hypothesis by Section 5, which shows that people perceive increased carotenoid colour as healthy. Since apparent health and attractiveness are very closely related (Joneset al., 2004, 2005), this may indicate mate choice for individuals displaying more carotenoid colour.
It may be asked at this stage why human skin is not therefore bright yellow, as are bird carotenoid ornaments. As mentioned above, we may be at an early stage in the evolution of bright carotenoid ornaments. Carotenoids are deposited in the highest concentrations in the top layer of skin where they are most visible (Edwards & Duntley, 1939; La Placaet al., 2000), but we are not yet bright yellow. Additionally, sexually selected traits are limited in their size and extravagance by natural selection, by competing forms of sexual selection, and by limiting factors in the environment. In guppies, males with larger carotenoid-based orange spots are preferred by females. Why then have the spots not expanded to cover the entire organism? Beyond a certain size, the attractiveness of the orange spots levels off (Andersson, 1994; Houde, 1987). Similarly, in humans, it may be that beyond a certain level of yellowness, additional carotenoid colour is no longer preferred. This is supported by Section 5, which shows that participants do not select the highest possible amount of carotenoid colour, but rather choose a less exaggerated amount of colouration. This may be related to the preference for averageness (Langlois & Roggman, 1990) and for a slight exaggeration
of attractive traits away from average, but not for greatly exaggerated traits very far from the population mean (Perrettet al., 1994).
A balance of selection pressures has been suggested to explain the gradient of skin lightness from very dark near the equator to very light far from the equator (Jablonski & Chaplin, 2000). Under this hypothesis, dark skin is selected for its UV protective properties near the equator, whilst light skin is selected for its increased ability to synthesise vitamin D in the absence of large amounts of UV light at high latitudes. The carotenoid colour of the skin could be sexually selected by a preference for carotenoid coloured faces, while at the same time other selection pressures act to constrain the degree of carotenoid colouration in the skin. The amount of carotenoid colour in the skin is limited by the amount of carotenoids that can be obtained from the diet (Stahlet al., 1998), which is in turn constrained by the amount of carotenoid- rich foods available and the other dietary needs of the body. A diet composed
exclusively of carotenoid-rich fruits and vegetables (which would exclude meat, fish, dairy products and potatoes/grains) could lead to deficiencies of various nutrients or caloric deficiency. Indeed, vegetarians are at increased risk of a number of nutritional deficiencies (Lowiket al., 1990). Further, displaying the carotenoids in the skin means that they are unavailable for use in immune defence and the other processes of the body (Sakset al., 2003). It may be expected that selection pressures exist for an increased preference for the consumption of carotenoid-rich foods, and improved uptake of carotenoids in the gut.
Similarly, skin blood perfusion and oxygenation are known to be positively associated with aspects of health, such as physical fitness (Armstrong & Welsman, 2001;
Johnson, 1998) and an absence of certain illnesses (Charkoudian, 2003; Panzaet al., 1990). Yet here too it could be asked why human faces are not extreme in redness like the red patches of mandrills’ faces. It could be noted that the lips of human faces are extreme in redness in light skinned groups (and many shades of lipstick are even more extreme in redness; Low, 1979). As well as the limits on skin redness caused by the preference for averageness, as discussed above, the amount of oxygenated blood colour in the skin could be limited by natural selection. Vasodilation, increasing skin blood perfusion, is a mechanism for rapid heat loss during exercise or other periods of heat stress (Johnson, 1998). Having large amounts of blood in the skin could cause excessive heat loss, which would be disadvantageous in temperate environments and also at night. Temporary flushing of the skin in a greatly exaggerated manner causes a rapid drop in blood pressure and could lead to fainting. This could potentially explain why flushing associated with social embarrassment, sexual arousal or anger takes place primarily in the face (and upper body to a lesser extent), where it will have the maximal impact as a social signal, whilst not reducing blood pressure to a dangerous extent. It is possible that evolution could select for mechanisms that would allow high levels of skin blood perfusion without excessive heat loss, or a significant drop in blood pressure.