Literature that describes the prevalence of inherited red-green color deficiency in different populations is reviewed. Large random population surveys show that the prevalence of deficiency in European Caucasians is about 8% in men and about 0.4% in women and between 4% and 6.5% in men of Chinese and Japanese ethnicity. However, the male: female prevalence ratio is markedly different in Europeans and Asians. Recent surveys suggest that the prevalence is rising in men of African ethnicity and in geographic areas that have been settled by incoming migrants. It is proposed that founder events and genetic drift, rather than natural selection, are the cause of these differences.
© 2012 Optical Society of America
Congenital red-green color vision deficiency is the most common -linked inherited abnormality in human populations. There are two types, protan and deutan, and differences in severity. Dichromats (protanopes and deuteranopes) have severe deficiency and are able to match all spectral hues using two color matching variables. Anomalous trichromats (protanomalous and deuteranomalous) form two heterogenic groups that require three variables but make color matches that are unacceptable to normal trichromats. Anomalous trichromats may have slight, moderate, or severe deficiency. All color deficient people see fewer colors in the environment and some colors that are easily distinguished by people with normal color vision look the same if there is no perceived luminance contrast. Color confusions are mainly between red, yellow, and green hues and between blue-greens, greys, and purples. Dichromats and severe anomalous trichromats confuse fully saturated hues but slight/moderate anomalous trichromats, who form the majority of color deficient people, only confuse pale or desaturated colors.
A large number of surveys have been made to establish the prevalence of deficiency in different populations. Published summaries of all data available prior to 1960 showed that the male frequency is about 8% in European Caucasians, 5% in Asians, 4% in Africans and less than 2% in indigenous Americans, Australians, and Polynesians [1,2,3,4,5,6,7] However individual surveys vary in precision, depending on the number of people examined and in accuracy due to the efficiency of the screening method. The characteristics of the population are not usually taken into consideration. In large randomly mating populations the prevalence of nonfatal inherited abnormalities, such as red-green deficiency, remains constant from one generation to another (The Hardy–Weinberg Law). A different prevalence may be found in small populations that are isolated geographically or by religious faith because the gene pool is restricted and marriage is between individuals who share a common ancestor. The gene pool is also reduced if marriage is limited to near neighbors with a similar frequency of color deficiency. These populations are not in Hardy–Weinberg equilibrium and surveys provide a “snap shot” of the situation at a particular time. A higher prevalence will be found in later years if migrants, with a higher prevalence of deficiency, become fully integrated with the original population.
A large number of people must be examined to obtain a precise prevalence figure within narrow confidence limits. If the estimated prevalence is greater than 5% this can be determined using standard deviations The standard deviation (sd) is given by the square root of where is the percentage of affected people found, is the percentage of people testing normal, and is the total number examined. The 95% confidence limit is within of the figure obtained (Table 1). These figures show that more than 5000 men must be examined to obtain a precise estimate of the frequency of red-green deficiency. The prevalence in women is too small to be assessed using standard deviations and is compared with that found for the corresponding male population. It is necessary to know the ratio of different types of red-green deficiency in the male population to do this effectively since the expression of deficiency in women depends on paired chromosomes that program the same type of deficiency (protan or deutan).
A large random population is needed to avoid bias. Examination of all school leavers entering employment or military service is ideal [8,9]. Data obtained in a single school are not ideal because a large number of siblings and cousins may be included. A representative group of color deficient subjects, that includes protans and deutans with different severities of deficiency, must be included in the survey in order to obtain accurate results. A low prevalence will be found if volunteers are recruited because symptomatic people with severe deficiency will come forward whereas asymptomatic people with slight deficiency may not. Conversely symptomatic people with severe deficiency may not volunteer for examination in Japan because abnormal color vision is considered to be a social stigma. People with severe deficiency are also likely to be under represented if data is collected from applicants for employment in an occupation known to require normal color vision  An average figure obtained from a large number of separate studies is inherently inaccurate because each survey may be subject to bias. This is particularly the case if the individual groups are small and some examiners are inexperienced.
The screening test used in a survey must have a proved capacity to identify protans and deutans with slight deficiency as well as severe deficiency. Comparative studies show that the Ishihara plates are the most efficient pseudoisochromatic screening test [11,12,13]. Background knowledge of the design format, administration, and optimum pass/fail criteria are needed to obtain maximum screening efficiency . Only the transformation and vanishing numeral designs are needed for screening. The Ishihara pathway designs are intended for nonverbal subjects but the longer viewing time needed is likely to reduce screening accuracy. The american optical company Hardy, Rand, and Rittler (HRR) test has relatively poor screening efficiency [15,16]. Women with normal color vision are specifically disadvantaged on this test and fail to interpret a greater number of low threshold red-green screening designs than men . Reliable protan:deutan classification is not always obtained with pseudoisochromatic designs and dichromats are not distinguished from anomalous trichromats with severe deficiency. A spectral anomaloscope, such as the Nagel anomaloscope, is the “gold standard” reference test for identifying, classifying and diagnosing different types of red-green deficiency. Unfortunately this instrument is rarely available outside Europe.
A theory based on natural selection was proposed in the 1960’s to explain the global differences in the frequency of color deficiency found prior to this date . The theory proposed that selection against color deficiency occurred in primitive hunter-gatherer tribes that relied on good hue discrimination to find food and avoid predators in a natural environment. Selection was then relaxed in people that moved away from equatorial regions in the African subcontinent to form settled agricultural communities allowing the prevalence of deficiency to rise. Further relaxation of selection then took place in northern urban populations with the longest history of cultural development allowing the prevalence to increase again.
The aim of the present paper is to review the original data that forms the basis for this theory and to consider alternative explanations for the differences found.
A literature search was made to access published data for the frequency of red-green deficiency in different populations. Forty-seven original publications were found. It was not always possible to obtain original data published before 1950 and some data are reproduced from ten of these papers that included summaries of other findings [1,2,3,4,5,6,7,19,20,21]. The emphasis is on results obtained in random surveys that included more than 5000 men and 5000 women in the same population. Results obtained in less precise surveys that included more than 1000 men are also reported. Data obtained from smaller populations and isolated groups are reported for comparison. Priority is given to accurate surveys made with the Ishihara plates for identification and with the Nagel anomaloscope for diagnosis of the type of deficiency.
Results obtained for the five ethnic groups distinguished by anthropologists are reported separately. Each of these groups contains a number of different populations that vary in life style and cultural development.
The five groups are the following:
- (1) European: Europeans are frequently described as Caucasians and are predominantly white-skinned. Populations range from Northern Scandinavia to Southern Europe, the Eastern Mediterranean, the Near East (including Turkey), and coastal areas of North Africa.
- (2) Asian: The populations of the Indian–Pakistan subcontinent, Mongoloid people, Chinese, and Japanese.
- (3) African: People from the African subcontinent who are predominantly black skinned, including Negroes and Black-Americans.
- (4) Native American: The indigenous people of North and South America.
- (5) Pacific: Indigenous Native Australians and Pacific Island populations including Melanesians and Polynesians.
A. Prevalence of Red-Green Deficiency in European Caucasians
Results obtained in five random surveys, with the Ishihara plates, that included more than 5000 men are shown in Table 2. A male prevalence of about 8% was found in all these surveys. The precision of measurement is slightly better than (Table 1). The prevalence of deficiency in German men did not change between 1936 and 1979 or after two generations [7,8,22,23].
A similar prevalence was obtained in five other random surveys that included more than 1000 men (Table 3) [24,25,26,27,28]. The precision of measurement is approximately (Table 1). A higher prevalence of nearly 9% was found in a survey that relied on recruiting volunteers at a public science exhibition .
Prevalence data was obtained in six surveys that included more than 3000 women (Tables 2 and 3). The female prevalence varied between 0.36% and 5% in these studies rising in step with a slight increase in the corresponding male prevalence (Table 2). The mean prevalence was 0.43%. A male prevalence of 9.25% in 1343 men in the former USSR using the Rabkin plates is reported by Kherumian and Pickford (1959)  However it was also reported that the frequency varied between 1% and 10% in groups of fewer than 500 Russian men with different ethnic backgrounds.
A prevalence of 8.2% was found by Miles (1929), with the Ishihara plates, for 1286 American men of European Caucasian origin . Miles also supervised three large surveys at public science exhibitions in which volunteers “walked by” a static display of Ishihara plates and recorded their own results . Later surveys in the USA were made with the HRR plates. Thuline (1964) examined 5263 school boys and 5075 girls and found the frequency of color deficiency to be 6.1% and 0.5%, respectively  The nationwide multicenter survey of school children, aged between 12 and 17 years, reported by the US Department of Health in 1974 found a mean prevalence of 7.7% in boys and 0.7% in girls . However, the male prevalence varied between 6.3% and 9.2% and the female prevalence between 0.1% and 1.5% in different geographic regions. A number of combined protan/deutan and tritan/”tetartan” classifications were recorded for both boys and girls. A nationwide survey consisting of volunteers for service in the Royal Navy and Air Force examined in 70 recruitment centers in the United Kingdom during World War II, found a mean prevalence of 7.2% using a selection of pseudoisochromatic plates taken from different tests. Volunteers were aware that good color vision was required. The frequency obtained for 6000 men examined in different centers varied between 5.4% and 9.5% [7,10].
Two recent surveys have been made in the Near East using the Ishihara plates. A survey in 1996 found a prevalence of 8.2% in 1136 Iranian men and 0.47% in 992 women . The second survey obtained a prevalence of 7.3% in 941 Turkish soldiers in 2005 .
Classification of the type of deficiency was made with the Nagel anomaloscope in six large random surveys (Table 4). These data show that the frequency of different types of red-green deficiency is in the approximate ratio:1 Protanope (P): 1Protanomalous trichromat (PA): 1 Deuteranope (D): 5 Deuteranomalous trichromats (DA) [5,8,22,25,26,33]. The deutan (): protan () ratio is therefore about . These data are used to explain the prevalence of deficiency found in European women (Tables 2 and 3).
If red-green deficiency results from an abnormality at a single locus on the chromosome the expected female prevalence would be 0.64%, the square of the male frequency. The lower frequency of about 0.4% was interpreted by Kalmus (1965) in terms of a “two locus hypothesis” . The hypothesis assumed that there were two separate genes for protan and deutan deficiency at different positions on the chromosome with multiple allelic forms that programmed normal red-green discrimination, anomalous trichromatism, and dichromatism in rank order of dominance. Heterozygotes and mixed heterozygotes, with a gene for protan deficiency on one chromosome and a gene for deutan deficiency on the other, are expected to have normal hue discrimination because a dominant normal gene is present at one locus on each female chromosome (Table 5). The prevalence of different types of deficiency in men shows that approximately 0.4% of women are homozygous for either deutan or protan deficiency and 0.24% of women are mixed hereozygotes. A total of 0.36% of women are homozygous for deutan deficiency and are expected to express this type of deficiency. Similarly 0.04% of women are homozygous for protan deficiency which is expressed. If anomalous trichromatism is dominant over dichromatism, the ratio of different types of deficiency in females would be 1 P: 3 PA: 1 D: 35 DA or in a deutan: protan ratio of about . The prevalence ratio in women found in two surveys with the Nagel anomaloscope is in close agreement with these figures [2,23]. For example, Waaler (1927) identified 3 PA, 1 D and 36 DA in a group of 40 color deficient women and Koliopoulos et al. (1976) identified 1 P, 3 PA, 2 D and 31 DA in a group of 37 women [3,22].
B. Prevalence of Red-Green Color Deficiency in Asian Populations
Summaries of results obtained in large Asian populations are listed in several reviews [3,4,6,7,19,20,21]. Some original data are difficult to access and other reports merely state the prevalence found. Most surveys mention that the Ishihara plates were used but do not describe the pass/fail criterion or whether the numeral or pathway designs were utilized. The results obtained in surveys that included more than 1000 men and 1000 women in the same population showed that the male prevalence varied between 4% and 6.5% in different regions of China (Table 6). A male prevalence of 5% was obtained in Chengtu in 1950 compared to a prevalence of 6.3% in 1934. The precision of measurement is better than 0.2% in the four surveys that included more than 7000 men. The female prevalence varied between 0.4% and 1.7% in these surveys and was always greater than the corresponding male frequency squared.
A higher male prevalence was found for Korean men in 1989 than in 1967 but the increase, after only one generation, might be due to the different screening tests used [19,21] (Table 7). The data reported by Ahmad and Chai (1980) were obtained at a medical examination given to all men born in Singapore in 1962 conscripted for military service . The precision is better than (Table 1)
A mean prevalence of about 4% was obtained in four Japanese surveys that included more than 6000 men prior to 1950 . The precision of measurement is better than (Table 1). A smaller survey in Nagoya in 1969 found a prevalence of nearly 6% (Table 8). The female prevalence varied between 0.2% and 0.6% in these surveys and always exceeds the corresponding male frequency squared.
Sato (1935) listed results from 19 other regional surveys in Japan with the Ishihara plates that included more than 5000 men and 3000 women in the same population and eight that included more than 1000 men and 1000 women. The male prevalence varied between 3.1% and 7.4% and the female prevalence between 0.2% and 0.9% in these studies .
Three surveys in populations of more than 1000 men have been made with the Ishihara plates in the Indian subcontinent. A male prevalence of 3.8% was found for 1306 Jat Sikhs in Patiala City in 1995 . A frequency of 2.5% was found in 1155 men from nine tribes in Andhra Pradesh and 6.5% in 569 nontribal boys using the pathway designs . A prevalence of 3.69% was obtained in a cross-sectional study of 1023 men attending an ophthalmic out-patient clinic in Benares in 1963 . Benares is a Holy City and the population is derived from all parts of the Indian subcontinent. Data collected for more than 10000 male and more than 3000 female students in 92 further education colleges throughout India found a mean prevalence of 3.6% in men and 0.2% for women .
Classification of the type of deficiency, with the Nagel anomaloscope was made in two small Japanese studies. Iinuma and Handa (1972) examined 838 boys and 747 girls: 50 boys (5.97%) and 10 girls (1.34%) were identified as color deficient . The classifications were 6 P, 9 PA, 8 D, and 27 DA for boys and 3 PA and 7 DA for girls, the deutan: protan ratio is therefore about in both groups. Ichikawa and Najima (1974) examined 738 male and 700 female students: 46 men (6.83%) and 5 women (0.80%) were identified as color deficient . The classifications were 4 P, 6 PA, 5 D, and 31 DA in men giving a deutan: protan ratio of about . Classifications were not reported for women.
C. Prevalence of Red-Green Color Deficiency in Africans and African-Americans
Kheruminan and Pickford (1959) list results derived from color naming mistakes that suggested that the prevalence varied between 1.72%, 2.5%, and 2.7% in 1000 men from three different African Tribes . Two modern surveys with the Ishihara plates in Lagos, Nigeria have found very different figures. Williams et al (1998) examined 5580 boys and 5405 school girls and found frequencies of 3.60% and 0.81%, respectively . The deutan: protan ratio was reported to be the same in both groups with approximately 50% classified as deutan, 40% as protan, and 10% unclassified. A previous study of 750 male and 491 female students in 1986 obtained frequencies of 7.1% and 0.61%, respectively. The prevalence of 7.98% was found for men with mixed Nigerian and European ancestry .
Clements (1930) examined 200 African-American men in the Southern United States with the Ishihara plates and obtained a prevalence of 4.24% . A prevalence of 3.91% was found by Crooks (1936) for 2019 men in the state of Virginia . The deutan: protan ratio was estimated to be about in this study. Examination of 1137 school boys, between the ages of 6 and 15 years, made 30 years later initially reported a prevalence of 6.77% but the results were clearly age-related and reexamination with the Nagel anomaloscope reduced the figure to 2.99% . The 34 color deficient boys identified were classified as 9 PA, 3 D, and 22 DA, a deutan: protan ratio of approximately . The survey made by Thuline (1964), using the Ishihara pathway designs with the HRR plates “for confirmation”, found a prevalence of 5.7% for 439 “nonwhite” boys . The national survey reported by the US Health Department in 1974 found a mean prevalence of 6.35% in African-American boys but large regional variations were found . The report concluded that the difference in the mean prevalence found for African-American and Caucasian-American boys (7.7%) was not statistically significant.
D. Prevalence of Red-Green Color Deficiency in Indigenous Native American and Pacific Populations
The prevalence of color deficiency found in isolated groups of Indigenous Americans and in Polynesian islanders using “ad hoc” screening methods, are reported by Post (1962) and is extremely variable . However Cruz–Coke (1970) interpreted these differences in terms of cultural development and pointed out that in South America the frequency was less than 2% in primitive nomadic tribes living at high altitudes, about 5% in settled agrarian communities neared to the coast and 8% in ports visited by European traders that had intermarried with the local population in the 17th century . Garth (1933) examined isolated populations in Central America and reported a male prevalence of 2.3% in 571 Mexicans, 2.5% in 535 American Indians from mixed tribal backgrounds, and 1.1% in 535 Navajo Indians. He found the highest prevalence (5.5%) in Indians that had intermarried with incoming migrants .
The survey made with the Ishihara plates in Native Australians and the indigenous people of Papua, New Guinea in 1965 found the prevalence to be 2% in the 4455 men examined . This survey provides the strongest evidence of a low prevalence of red-green color deficiency in an isolated indigenous population. An examination of 1258 male high-school students in Manila in 2010 obtained a prevalence of 5.17% .
Precise accurate surveys in European Caucasian population all find that the prevalence of red-green color deficiency is approximately 8% in men and about 0.4% in women (Tables 2 and 3). Similar figures are obtained in countries that are widely separated geographically and in Turkey and Iran on the border of the European region [33,34]. The ratio of types of red-green deficiency is also remarkably consistent (Table 4). The European area has been subject to mass migration and settlement by people of diverse backgrounds for over 3000 years and the general population conforms to the Hardy–Weinberg Law. However local populations may not be in Hardy–Weinberg equilibrium if marriage has only been within the community and with near neighbors. For example different figures were obtained in geographic areas of the United Kingdom in the 1940’s and in regions of the USA in 1974 [7,10,32] Regional variations in the frequency of color deficiency are still found in populations, such as Italian fishing communities which have preserved a strong social and religious identity . The characteristics of the population and the date that a survey is made therefore have a bearing on Hardy–Weinberg equilibrium and influence the results obtained.
The prevalence of deficiency found in large widely separated Chinese populations before 1959 found that the male prevalence varied between 4% and 6.5% and the female prevalence between 0.7% and 1.7% (Table 6) [4,5,6] Regions of China may have been in limited Hardy-–Weinberg equilibrium during this period because travel was difficult over long distances and most marriages would have been with near neighbors. The mean male prevalence derived from the large number of Chinese surveys reported by Sato (1935) was 4.0% . The largest and most precise survey in men of Chinese ethnicity found a prevalence of 4.6% (Table 7) . These data show that the prevalence of red-green deficiency in Chinese men is between 4% and 5%. Large surveys in Japan before 1960 obtained a mean prevalence of 4% in men (Table 8) [3,19]. However a prevalence of nearly 6% was found in 3434 men in Nagoya in 1969. The number of color deficient men examined in two studies with the Nagel anomaloscope is too small to obtain an accurate figure for the ratio of different types of deficiency. One study found a deutan: protan ratio of about and the other a ratio of . If the deutan: protan ratio is expected to be the same as in European populations, a deutan: ratio of suggests that some deutans have not been identified [40,41].
The prevalence of red-green deficiency in African-American men is usually described as 4% but this figure is derived from two small surveys made in the 1930’s [1,44]. Surveys made after 1960 using different screening tests show that the prevalence has risen to between 6% and 7.2% [31,32]. This rise occurs after one generation but could be due to intermarriage. “Snapshot” surveys in small isolated populations in India, the African subcontinent and Central America report a male prevalence between 2% and 4%. However frequencies of about 5% are found in groups that have intermarried with different tribes and a prevalence of 8% has been reported in coastal areas visited by Europeans in historical times [18,46]. A prevalence of 7.1% was found for 750 Nigerian men in Lagos in 1986 but was nearly 8% in men of mixed Nigerian and European ancestry . Evidence that the prevalence of deficiency in indigenous people was obtained in the large survey, made with the Ishihara plates, in 4455 Native Australians and men from Papua New Guinea. “Snap shot” surveys in small groups of Polynesian Islanders suggest that the prevalence is also very low . A survey in Manila obtained a male prevalence of 5% similar to that found in men of Malay ethnicity [9,47].
Advances in molecular biology during the past 25 years have revolutionized the understanding of the underlying genetic causes of color deficiency and a genetic explanation has been sought to explain racial differences in prevalence. It is now known that all types of red-green deficiency are caused by loss of a gene necessary for normal color vision and that most types of deficiency result from paired genes, positioned at a single locus on the chromosome which program abnormal peak wavelength sensitivity. The basic assumption underlying “the two locus hypothesis” is therefore incorrect. Peak wavelength separation determines the phenotype rather than dominant genes. The long wavelength () and medium wavelength () sensitive genes are positioned in a head-to-tail array that were formed by duplication of a single ancestral chromosome gene about 30 million years ago followed by point mutations that changed peak wavelength sensitivity [50,51] Some ethnic differences in gene structure have been found. Polymorphism at site 180 of the gene is common. Serine at this site is found in 84% of Japanese, 78% of African-Americans but in only 60% of Europeans and produces an approximate 4 nm shift in peak wavelength sensitivity. The gene is normally second in the chromosome array and further duplication events add more or -like genes that are not expressed. The average number of genes is greater in Europeans than other races suggesting that more duplication events have occurred. Only 25% of Caucasians have a single gene compared with 50% of Japanese and 42% of African-Americans [52,53]. One parental chromosome array with at least genes is needed to produce a protanomalous phenotype but there is no evidence that the prevalence of protan deficiency is lower in Asian and African-American men. The photopigment identity of individual cones in the central retinal mosaic is genetically determined but the mechanism is not fully understood. The cone ratio varies individually but the mean ratio varies in different racial groups and the percentage of cones in the central mosaic is slightly lower in African-Americans than in Europeans . Most intriguingly, a polymorphism in the promoter of the gene second in the array has been found to produce a deuteranomalous phenotype in Japanese but not in other racial groups . The same study failed to identify a genetic cause in 15% of deuteranomalous trichromats. These data show that additional genetic causes of red-green deficiency are still to be discovered.
Explanation for the high prevalence of deficiency in Asian and African women is challenging. Sampling error can be discounted in very large population surveys but inadequate screening methods have to be considered. If the same screening technique is used for men and women in the same population both prevalence figures will be equally affected and the male: female prevalence ratio will be the same in surveys made by different examiners in widely separated geographic areas. One possibility is that the examination method was changed due to poor female education, when the survey was made, requiring that the Ishihara pathway designs were used because no verbal response could be made to the numeral designs. On average women have poorer hue discrimination ability than men and make more errors on HRR screening plates that require good blue-green, gray, and purple discrimination [13,56]. On average Caucasian heterozygotes are also found to make slightly more “misreadings” than men on Ishihara screening plates based on red-yellow-green confusion [57,58]. This would not result in false positive identification of red-green deficiency if the instructions provided with the test that describe pass/ fail criteria are followed carefully. Paired chromosomes in women contain duplicate genes that program different physical characteristics and suppression must occur at an early stage of development. It is not known whether there are racial differences in this mechanism. If inactivation in heterozygotes is not totally biased in favor of the male or female chromosome, a proportion of either or dedicated cones will contain two photopigments with different peak wavelength sensitivity. The uneven red-green response derived from three medium wavelength photopigments would result in reduced hue discrimination ability in this spectral range and lead to a greater number of “misreadings” or some errors on the Ishihara plates that would be interpreted as slight color deficiency.
The theory of relaxation of natural selection to explain racial differences in the prevalence of red-green deficiency was proposed very tentatively using all the available data available at the time without considering differences in precision and accuracy [7,18,21]. It is therefore surprising that the theory has been universally accepted and repeated in contemporary literature. The theory was strongly criticized by ophthalmologists at the time because “color blind” people are not totally without hue discrimination and hunter-gathers would not be so disadvantaged nutritionally that they died before reaching reproductive age . Red-green deficiency has no impact on survival fitness and therefore is neutral in terms of natural selection. In any case only 2% of our remote male ancestors, who were dichromats, would have experienced difficulty finding nutritious ripe fruits and leaves by color but would have been able to utilize perceived lightness difference, position, and pattern to be successful as they do today. In modern society people with severe deficiency experience more problems with denotative and connotative color coding in the work environment than in agricultural activities .
Cruz–Coke (1970) described data that showed that the prevalence of deficiency was higher in populations in coastal areas, particularly in South America that had a long history of contact with Europeans and suggested that this was because the genes for color deficiency and alcoholism are linked . This theory provoked widespread adverse criticism in the medical journalss of the day. However Cruz–Coke published a world map with contours delineating the average prevalence found in different global regions that appeared to show a relationship with geographic latitude. This led to a theory that the higher prevalence of deficiency in latitudes further from the equator was related to the longer hours of twilight  Support was provided by anecdotal reports that deutans have superior luminance discrimination in mesopic light levels than normal trichromats and that natural selection would have acted to increase the prevalence of this type of deficiency. Although complex interaction between luminance and color discrimination is evident at low light levels there is no evidence that color deficient people are advantaged in this way. There are also a number of surveys reported in this review that show that the prevalence found in different populations is strongly associated with race and with population size rather than geographic latitude.
Changes in gene frequency can be caused by mutation or migration. There is good evidence from the surveys described in this review that differences in prevalence in small populations are caused by genetic drift. Gene mutation in humans is extremely slow and it might take 40 generations (1000 years) for a gene that improves survival fitness to reach equilibrium . The influence of migration is more rapid and occurs on a time scale that is proportional to the population size. A population can absorb new genes from a nearby or visiting population that maintains a different gene frequency within one or two generations. This depends on the difference in frequency between the two populations and the proportion of migrant genes that are absorbed in each generation. Variable sampling of the gene pool (random genetic drift) is accelerated by reduction in population size. The effective population size depends on the number of parents rather than the number of founder members or the number of people in the tribe. Not all founder members become parents. For example 45% of the residents of Tristan de Cunha in 1961 were descended from five founder members that settled there in 1816 . The frequency of any inherited abnormality therefore depends on the number of parents that have the genetic determinant and the number of children they have. This provides a plausible explanation for the differences in the male prevalence of red-green deficiency, between 1% and 10%, found in small isolated ethnic groups.
This review confirms that the prevalence of deficiency is remarkably constant in European Caucasian populations that conform to the Hardy–Weinberg Law but that slightly different results are obtained in local areas that have limited Hardy–Weinberg equilibrium. Emphasis on the most precise and accurate survey data confirms that the male prevalence of deficiency in European Caucasians is about 8% and that the prevalence in Asian populations is between 4% and 5%. There is no strong evidence that the prevalence of deficiency in African populations and African-American men is 4%. Small surveys in isolated populations that provide less precise data strongly suggest that the male prevalence rises after migrants with a higher frequency of deficiency enter the area and intermarry. These changes are due to random genetic drift. There is particular evidence for this in the mixed tribes of Central America, India, and in coastal areas of every continent settled by European traders [37,38,43,47]. The low prevalence previously reported in small groups of indigenous people is more likely to be due to the effective population size and the small number of founder members with the genetic determinant than to natural selection, cultural development or lifestyle.
1. F. Clements, “Racial differences in color blindness,” Am. J. Phys. Anthropol. 14, 417–432 (1930).
2. T. R. Garth, “The incidence of colour blindness among races,” Science 77, 333–334 (1933). [CrossRef]
3. S. Sato, “Statistical observations on congenital abnormalities in colour vision in Japan,” Acta Soc. Ophthalmol. Jpn. 38, 2227–2230 (1935).
4. E. Chan and W. S. Mao, “Colour blindness among Chinese,” Brit. J. Ophthalmol. 39, 744–745 (1950).
5. J. François, G. Verriest, V. Mortier, and R. Vanderdonck, “De la fréquence des dyschromatopsies congénitales chez l’homme,” Ann. Ottalmol. Clinica Oculistica 190, 5–16 (1957).
6. C. Yang, J.-C. Chiang, P.-H. Feng, and N.-H. Yang, “Colour blindness among Chinese,” Chin. Med. J. 76, 283–284 (1958).
7. R. Kherumian and R. W. Pickford, Heredité et Fréquence des Dyschromatopsies (Vigot Frères, 1959), pp. 63–67.
8. J. Koliopoulos, P. Iordanides, G. Palimeris, and E. Chimonidou, “Data concerning colour vision deficiencies among 29,985 young Greeks,” Mod. Probl. Ophthalmol. 17, 161–164 (1976).
9. T. A. Ahmad and S. L. Chai, “Colour Vision in the Singapore male,” Sing. Med. J. 21, 639–642 (1980).
10. P. E. Vernon and A. Straker, “Distribution of colour blind men in Great Britain,” Nature 152, 690 (1943). [CrossRef]
11. R. G. Frey, “Welche pseudoisochromatishen Tafeln sind fur die Praxis am besten geeignet?” Graefes Arch. Ophthalmol. 160, 301–320 (1958).
12. S. J. Belcher, K. W. Greenshields, and W. D. Wright, “A colour vision survey,” Brit. J. Ophthalmol. 42, 355–359 (1958).
13. L. L. Sloan and A. Habel, “Tests for color deficiency based on the pseudoisochromatic principle—a comparative study,” A.M.A. Arch. Ophthalmol. 55, 229–239 (1956).
14. J. Birch, “Efficiency of the Ishihara plates for identifying red–green colour deficiency,” Ophthalmol. Physiol. Opt. 17, 403–408 (1997).
15. G. Rand and M. C. Rittler, “An evaluation of the AO HRR pseudoisochromatic plates,” A.M.A. Arch. Ophthalmol. 56, 736–742 (1956).
16. J. Birch, “Identification of red-green colour deficiency: Sensitivity of the Ishihara and American Optical Company (Hardy, Rand and Rittler) pseudoisochromatic plates to identify slight anomalous trichromatism,” Ophthalmol. Physiol. Opt. 30, 667–671 (2010).
17. R. C. Baraas, “Poorer color discrimination by females when tested with pseudoisochromatic plates containing vanishing designs on neutral backgrounds,” Vis. Neurosci. 25, 501–505 (2008).
18. R. H. Post, “Population differences in red and green colour vision deficiency: A review and a query on selection relaxation,” Eug. Quart. 9, 131–146 (1962).
19. I. Iinuma and Y. Handa, “A consideration of the racial incidence of congenital dyschromats in males and females,” Mod. Probl. Ophthalmol. 17, 151–157 (1976).
20. L. G. Kilburn and Y. T. Beh, “The incidence of colour blindness among Chinese,” Science 79, 34 (1934). [CrossRef]
21. H. B. Kim, S. G. Lee, M. D. Choe, J. K. Lee, and B. H. Ahn, “The incidence of congenital colour deficiency among Koreans,” J. Kor. Med. Sci. 4, 117–120 (1989).
22. G. H. M. Waaler, “Ueber die Erblichkeitsverhaltnisse der verschiedenen Arten von angeborener Rotgrunblindheit,” Acta Ophthalmol. 5, 309–345 (1927). [CrossRef]
23. K. Feig and H.-H. Ropers, “On the incidence of unilateral and bilateral colour blindness in heterozygous females,” Hum. Genet. 41, 313–323 (1978). [CrossRef]
24. P. von Planta, “Die Haufigkeit der angeborenen Farbensinnstorungen bei Knaben un Madchen und ihre Feststellung durch die ublichen klinischen Proben,” Graefes Arch. Clin. Exp. Ophthalmol. 120, 253–281 (1928).
25. M. Weiland, “Utersuchungen uber Farbenschwache bei Konduktorinnen,” Graefes Arch. Clin. Exp. Ophthalmol. 130, 444–462 (1933). [CrossRef]
26. J. H. Nelson, “Anomalous trichromatism and it’s relation to normal trichromatism,” Proc. Phys. Soc. Lond. 50, 661–702 (1938). [CrossRef]
27. H. Forsius, A. W. Eriksson, and J. Fellman, “Colour blindness in Finland,” Acta Ophthalmol. 46, 542–552 (1967).
28. R. A. Crone, “Incidence of known and unknown colour vision defects,” Ophthalmol. Basel 155, 37–55 (1968). [CrossRef]
29. W. R. Miles, “100 cases of color blindness detected with the Ishihara plates,” J. Gen. Psychol. 2, 535–543 (1929). [CrossRef]
30. W. R. Miles, “Color vision in eleven thousand museum visitors,” Yale J. Biol. Med. 16, 59–76 (1943).
31. H. C. Thuline, “Color vision defects in American school children,” J. Am. Med. Assoc. 188, 514–518 (1964).
32. U.S. Department of Health, Education and Welfare, National Centre for Health Statistics Vital and Health Statistics-Series 11-No. 134. DHEW Publication No. (HRA) 74-1616, “Color vision deficiencies in youths 12-17 years of age” (1974).
33. M. Modarres, M. Mirasamadi, and G. A. Peyman, “Prevalence of congenital colour deficiencies in secondary school students in Tehran,” Intern. Ophthalmol. 20, 221–222 (1996).
34. M. Citirik, G. Acaroglu, C. Batman, and O. Zilelioglu, “Congenital colour blindness in young Turkish men,” Ophthalmol. Epidemiol. 12, 133–137 (2005).
35. H. Kalmus, Diagnosis and Genetics of Defective Colour Vision (Pergamon, 1965), pp. 63–65.
36. S. Naresh, “Study of colour blindness in Jat Sikhs,” Indian J. Physiol. Pharmacol. 39, 127–130 (1995).
37. K. S. Mehra, “Incidence of colour blindness among Indians,” Brit. J. Ophthalmol. 47, 486–487 (1963).
38. P. C. Dutta, “Variability and regional differences of colour blindness in India,” Hum. Genet. 2, 204–206 (1966).
39. K. R. Dronamraju and P. M. Khan, “Frequency of colour blindness among tribal and non-tribal peoples of Andhra Pradesh,” Ann. Hum. Genet. 27, 17–21 (1963).
40. I. Iinuma and Y. Handa, “Mass examination of colour sense using the anomaloscope,” Acta Soc. Ophthalmol. Jpn. 76, 549–557 (1972).
41. H. Ichikawa and A. Majima, “Genealogical studies on interesting families of defective colour vision discovered by mass examination in Japan and Formosa,” Mod. Probl. Ophthalmol. 13, 265–271 (1974).
42. G. O. Williams, F. E. Taylor, I. I. Odidika, and K. A. Musa, “Frequency of colour blindness among Nigerian school children in Lagos,” Am. J. Hum. Biol. 10, 283–288 (1998).
43. P. G. C. Odeigah and E. E. Okon, “Colour vision defects and gene flow in Nigerians,” East Afr. Med. J. 63, 666–671 (1986).
44. K. B. M. Crooks, “Further observations on colour blindness among Negroes,” Hum. Biol. 8, 451–458 (1936).
45. D. R. Sweeney, R. T. Zegers, and W. E. Collins, “Color blindness in male negro children,” J. Soc. Psychol. 62, 85–91 (1964).
46. R. Cruz-Coke, Colour Blindness—An Evolutionary Approach (Thomas, 1970), pp. 106–111.
47. I. Mann and C. Turner, “Colour vision in native races in Australasia,” Am. J. Ophthalmol. 41, 797–800 (1956).
48. M. Cruz, H. G. S. Cerdana, A. M. B. Cabrera, C. B. Garcia, E. T. Santos-Morabe, and M. L. R. Nanagas, “Prevalence of color-vision deficiency among male high-school students,” Phil. J. Ophthalmol. 35, 20–24 (2010).
49. G. P. Grassivaro, L. Romana, M. Mongogna, and F. Viviani, “Origin and distribution of Daltonism in Italy,” Am. J. Hum. Biol. 15, 566–572.
50. S. S. Deeb, “The molecular basis of variation in human color vision,” Clin. Genet. 67, 369–377 (2005).
51. J. Neitz and M. Neitz, “The genetics of normal and defective color vision,” Vis. Res. 51, 633–651 (2011).
52. A. L. Jorgenson, S. S. Deeb, and A. G. Motulsky, “Molecular genetics of chromosome-linked color vision among populations of African and Japanese ancestry: High frequency of a shortened red pigment gene among Afro-Americans,” Proc. Natl. Acad. Sci. USA 87, 6512–6516 (1990).
53. H. Ueyama, S. Hayashi, S. Tanabe, Y. Tanaka, T. Hayashi, S. S. Deeb, S. Yamada, and I. Ohkub, “ Number and arrangement of the red and green visual pigment genes in color-normal Japanese males,” Color Res. Appl. 26, S84–88 (2001).
54. C. McMahon, J. Carroll, S. Awua, J. Neitz, and M. Neitz, “The cone ratio in males of African descent with normal color vision,” J. Vision 8(2):5, 1–9 (2008).
55. H. Ueyama, Y. H. Li, G. L. Fu, P. Lertrit, L. O. Atchaneeyasakul, S. Oda, S. Tanabe, Y. Nishida, S. Yamada, and I. Ohkubo, “An A-71C substitution in a green gene at the second position in the red/green visual-pigment gene array is associated with deutan color deficiency,” Proc. Natl. Acad. Sci. USA 100, 3357–3362 (2003).
56. M. Rodriguez-Carmona, L. T. Sharpe, J. A. Harlow, and J. L. Barbur, “Sex related differences in chromatic sensitivity,” Vis. Neurosci. 25, 433–440 (2008).
57. G. Jordan and J. D. Mollon, “A study of women heterozygous for colour deficiencies,” Vision Res. 33, 1495–1508 (1993).
58. E. Konstantakopoulou, M. Rodriguez-Carmona, and J. L. Barbur, “Processing of colour signals in female carriers of colour vision deficiency,” J. Vision submitted.
59. J. M. Steward and B. L. Cole, “What do colour vision defectives say about everyday tasks?” Optom. Vis. Sci. 66, 288–295 (1989).
60. T. E. Reimchen, “Human color vision deficiencies and atmospheric twilight,” Soc. Biol. 34, 1–11 (1987).
61. J. H. Gillespie, Population Genetics (Johns Hopkins University, 2004), pp. 6–21.
62. M. W. Strickberger, Genetics, 3rd ed. (MacMillan, 1990), pp. 705–706.