Acta Derm Venereol (Stockh) 1998; 78: 214^219

Interdependence of Eye and Hair Colour, Skin Type and Skin

Pigmentation in a Caucasian Population
JÒRGEN LOCK-ANDERSEN1, HANS CHRISTIAN WULF1 and NIELS DITLEV KNUDSTORP2
1
Department of Dermatology, Bispebjerg Hospital, University of Copenhagen and 2Department of Orthopaedics, Rigshospitalet, National
University Hospital, Copenhagen, Denmark

Eye colour, hair colour and skin colour are important risk fac- respectively (6). In a multicentre study on BCC and SCC from
tors for malignant melanoma and non-melanoma skin can- southern Europe, involving 1,549 BCC cases, 228 SCC cases
cers. There are few studies in which the distribution of these and 1,795 controls, pale eye colour compared to dark eye colour
pigmentary factors in risk populations has been assessed. The had an OR of 1.4 for BCC and 1.8 for SCC, and red hair colour
purpose of this study was to investigate the prevalence of the compared to black hair colour held an OR of 2.4 for BCC and
major eye and hair colours and the distribution of skin types 18.0 for SCC (9). Univariate estimations of OR without adjusting
and skin pigmentation in a Caucasian population. In 892 Dan- for exposure may be to underestimate the OR in sun-sensitive
ish Caucasians, eye colour, hair colour and skin type were persons because these individuals probably expose themselves
assessed and facultative and constitutive skin pigmentation less to the sun than more sun-resistant persons.
were measured objectively using skin re£ectance spectro- In most brown-skinned and black-skinned populations there
scopy. Blue eye colour and blond hair colour and skin type II is little variation in hair and eye colour, but in Caucasians a
were the most frequent (60%, 67% and 33% of subjects, great diversity of hair and eye colours is seen in connection
respectively). All four major eye colours and four major hair with various degrees of skin pigmentation. Despite eye colour
colours (with the exception of red hair colour) were found and hair colour being well-de¢ned risk factors for melanoma
within skin types I^IV and we could not predict the skin type and non-melanoma skin cancers, and are readily observable
or the constitutive skin pigmentation. Skin type could not be characteristics, surprisingly few studies assess the distribution
taken to classify individuals reliably according to their facul- of these pigmentary traits in Caucasian risk populations and
tative or constitutive skin pigmentation. Key words: epidemiol- the association of these pigmentary traits to skin type and to
ogy; risk factor; skin re£ectance. skin pigmentation.
(Accepted December 8, 1997.) The purposes of the present study were (i) to establish the
distribution of hair colour and eye colour and skin type and
Acta Derm Venereol (Stockh) 1998; 78: 214^219.
skin pigmentation in a Danish Caucasian population, which
JÖrgen Lock-Andersen, Department of Dermatology D92, has a high risk for development of melanoma and non-mela-
Bispebjerg Hospital, Bispebjerg Bakke 23, DK-2400 noma skin cancers and (ii) to investigate whether eye colour
Copenhagen NV, Denmark. and hair colour are true predictors of skin type and skin pig-
mentation.

Persons of fair complexion, e.g. with blond or red hair and blue SUBJECTS AND METHODS
eyes and lightly pigmented skin, are clinically more prone to be The population sample consisted of patients and relatives attending the
sunburned and to develop sun-induced skin damage. In some outpatient clinic of the Department of Orthopaedics at the National
phototest studies the phenotypic characteristics eye colour and University Hospital in Copenhagen. This clinic was chosen because
hair colour have been found to be related to skin type and to persons of all ages and from a wide geographic area, including urban
minimal erythema dose (MED) (1^3). The incidence of skin and rural areas, attended the clinic and because we assumed that pig-
cancer, cutaneous malignant melanoma (CMM), basal cell mentary traits such as eye and hair colour and skin type would be unre-
lated to clinic attendance. In the study period, February to April 1995,
carcinoma (BCC) and squamous cell carcinoma (SCC), is
approximately 1,000 persons were approached randomly by one of two
increasing in white-skinned populations worldwide and skin
trained interviewers in the clinic's waiting area and invited to partici-
cancer is now becoming a major health problem. The fair pate in the study. We wanted the study to be a Danish population study
skinned Caucasians in Scandinavia are at considerable risk of and therefore excluded persons not holding Danish citizenship. Of
skin cancer and the CMM incidence rates are higher in Scandi- those approached, approximately 10% declined to participate for var-
navians than in populations from southern Europe (4). It has ious reasons. The study was approved by the local Ethics Committee
been calculated that, given the current incidence rates, 1 in and all participants gave informed consent before entering. Eight-hun-
every 100 Danish males and 1 in every 80 Danish females will dred-and-ninety-two persons were included (518 females and 374
develop a CMM during their lifetime (5). males), with a mean age of 37 years (SD~18 years; range 0^85 years).
In analytic studies of skin cancer it has been found that the Seven persons (0.8%) stated that they had previously had a skin cancer
or might have had a skin cancer. Since such information is doubtful,
phenotypic traits eye colour and hair colour and skin type and
and we could not verify this, they were not excluded.
skin colour are major risk factors for CMM (6, 7), BCC (8^10)
and SCC (8, 9, 11). In a meta-analysis of 10 case-control studies Study design
of CMM with more than 3,000 cases and almost 4,000 controls Study subjects were interviewed in a special study room in accordance
the odds ratio (OR) for blue eyes compared to brown eyes was with a standardized questionnaire on age, area of living, use of arti¢cial
calculated to be 1.6 and for red hair colour and blond hair colour tanning devices, hair colour and skin type. Eye colour was assessed and
compared to black or brown colours the ORs were 2.4 and 1.8, skin pigmentation was measured objectively at UV-exposed sites and in

Acta Derm Venereol (Stockh) 78 # 1998 Scandinavian University Press. ISSN 0001-5555
 Skin pigmentation 215

Table I. Distribution of eye colour, hair colour and skin type about average; skin type IV: rarely burn, tan more than average; skin
type V: brown-skinned persons, and skin type VI: black-skinned per-
Pigmentary trait Study sample Subgroup sons. In children, the skin type was indicated by their parents but we
did not estimate skin type in persons younger than 6 years because we
Males, % Females, % MzF, % MzF, % consider anamnestic skin typing to be too uncertain or impossible
before this age.
Eye colourA
Blue 64.4 56.0 59.6 61.6 Skin pigmentation
Green 14.2 20.1 17.6 15.1 To objectively measure the skin pigmentation we used a portable skin
Grey 6.3 6.0 6.1 5.5 re£ectance spectroscope (PBI UV-Optimize, Model 550/660, PBI
Brown 13.7 17.9 16.1 16.8 Medical, Ringsted, Denmark) which utilizes wavebands of visible light
NA 1.4 0.0 0.6 1.0 at 555 and 660 nm to independently measure the skin pigmentation
Hair colourB (the melanin content) and the skin redness (the haemoglobin content)
Red 3.2 3.9 3.6 3.5 and gives the results on continuous scales from 0 to 100% (14). Zero
Blond 68.0 66.8 67.3 66.6 percent pigmentation corresponds to skin with no pigmentation at
Brown 23.8 27.5 26.0 25.4 all, as in an extremely white person; 100% pigmentation corresponds
Black 4.6 1.8 3.0 4.2 to no light re£ected back, as in theoretical absolutely black skin. The
NA 0.4 0.0 0.1 0.3 resolution of the pigmentation scale and precision of the instrument
Skin typeC allows for at least 40 steps for measurements of skin pigmentation in
I 21.0 24.7 23.1 24.9 European Caucasians (15). The measuring probe is held lightly against
II 32.8 33.9 33.4 29.0 the skin and after a few seconds the readings, which are the average of
III 19.0 15.8 17.2 16.1 three independent measurements, are given on the display. Before each
IV 26.3 22.8 24.3 27.9 measurement session the apparatus was calibrated on a white reference
V 0.6 1.0 0.8 0.8 tile. Pigmentation measurements were performed at UV-exposed sites:
NA 0.3 1.8 1.2 1.3 on the forehead; on the upper chest in the infraclavicular region; on the
upper back in the suprascapular region; at the lateral aspect of the
A subgroup of 414 persons, with an age and gender distribution as in upper arm 7 cm above the lateral epicondyle; at the medial aspect of
the Danish population, was constructed from the total sample of 892 the upper arm 7 cm above the medial epicondyle, and at the UV-
persons. A Eye colour assessed in study persons older than 1 year of age shielded site of the medial and upper quadrant of the buttocks. The
(study sample: n~868; subgroup: n~398). B Hair colour at 25 years of study period of February to April was carefully selected because the
age (study sample: n~667; subgroup: n~287). C Skin type in persons e¡ect of seasonal variation on facultative skin pigmentation is least
older than 5 years of age; in children indicated by their parents (study during these months of the year (16).
sample: n~856; subgroup: n~390). M~males; F~females; NA~not
available.
Statistics
The distribution of age and gender in the study sample was compared
using the Chi-squared test to the expected distribution calculated from
UV-shielded buttock skin. At the end of the examination study persons the Danish population as recorded by January 1996 (17). Analysis of
were given general advice on the risk known to be associated with eye colour, hair colour and skin type by gender was performed using
excessive sun exposure and on sun protection. The interview and the the Chi-squared test, and skin pigmentation by gender was analysed
pigmentation measurements took approximately 20 min for each per- using the unpaired t-test. Skin pigmentation according to skin types
son. The standardized questionnaire was pretested prior to the study, was analysed using the Mann-Whitney test. We considered p values less
and during the study the two interviewers at regular intervals over- than 0.05 to be signi¢cant.
heard each other's interview to ensure that interviews were being per-
formed consistently during the study period. Before starting the study
and at regular intervals during the study the two interviewers per- RESULTS
formed simultaneous assessments of eye colour in the same study per-
sons to ensure that eye colour assessment was uniform. Study sample versus the Danish population

Eye and hair colour

As expected, the majority of study persons came from the
Copenhagen area, with 48% from the central part of Copenha-
Study persons were asked to indicate their natural (un-dyed) hair col-
our at 25 years of age in accordance with a 5-point scale: red, light gen (Copenhagen municipality; population January 1996:
blond, dark blond, brown or black. Eye colour (colour of the iris) was 476.751 persons (17)), but all of the 14 Danish counties were
assessed by the interviewers on a 4-point scale: blue, green, grey or represented in the study sample. The observed age and sex dis-
brown. We did not assess eye colour in infants younger than one year tribution of the study sample was compared with the expected
of age because we considered eye colour to be too unstable before this age and sex distribution calculated from the Danish population
age. distribution in 1995 and was found to be di¡erent (pv0.01) due
to too few persons older than 59 years (approximately 50% of
Skin type the expected number) and too many women in the age group
The skin phototype of the participants (anamnestic recall of burning 20^59 years (approximately 160% of the expected number).
tendency and tanning ability) was estimated using the Fitzpatrick
From the total sample of 892 persons a subgroup of 414 per-
classi¢cation (12) based on a person's memory of their reaction to
sons was selected (204 males and 210 females) with a similar
2 h of sun on a sunny day at noon at the beginning of the summer
period. This exposure in Denmark (situated at 56³ latitude N) corre-
age and sex distribution as in the Danish population by ran-
sponds to a UV dose of 9 SED (one SED~100 J/m2 /298 nm, CIE dom reduction of persons in overrepresented age-gender
erythema action spectrum (13)). The following six skin types were groups. In this subgroup the distribution of eye and hair colour
available; skin type I: always burn, never tan; skin type II: usually and skin type was also extracted (Table I). Compared to the
burn, tan less than average; skin type III: sometimes mild burn, tan total sample, only minor di¡erences were found for the distri-

Acta Derm Venereol (Stockh) 78

216 J. Lock-Andersen et al.

eye colour and more females having brown eye colour

(p~0.03). All four eye colours were found in skin types I^IV
(Fig. 1) and even brown eye colour was found in 13% of persons
with skin type I. There was no obvious trend for any eye colour
to be associated with a particular skin type. When eye colour
was compared to objective measurements of constitutive skin
pigmentation at the buttocks we also found all four eye colours
within the four groups of pigmentation (Fig. 2), but with a
slight tendency for brown eye colour to be more associated
with darker pigmented skin. Eye colour was therefore found
to be an unreliable predictor of anamnestic skin type and of
constitutive skin pigmentation.

Hair colour
Only 4% of the sample of Caucasians had red hair colour at the
age of 25 years and only 3% had black hair colour (Table I),
with blonds (light blond and dark blond) being the dominating
colours and with no statistical gender di¡erence (p~0.15). All
hair colours were found in the four skin types I^IV except red
hair, which was only found in skin types I and II (F|g. 1). There
was a tendency for brown and black hair colour to be asso-
ciated with the more sun-resistant skin types III and IV and
with a higher degree of constitutive skin pigmentation (Fig.
2), while blond and red hair were more associated with the
sun-sensitive skin types I and II and to lesser levels of constitu-
tive pigmentation. However, these were only weak trends and
overall hair colour was not found to be a predictor of skin type
or of the degree of constitutive skin pigmentation.
Fig. 1. Eye colour and hair colour according to anamnestic skin type.
Eye colour in 868 persons older than 1 year of age and hair colour at
Skin type and skin pigmentation
25 years of age in 667 persons.
Skin type II was the dominant skin type found in 33% of the
study persons (Table I), but a surprising 24% indicated skin
bution of eye and hair colour and skin types in the age and gen- type IV. There was no gender di¡erence in the distribution of
der population comparable subgroup (Table I). skin types (p~0.41). Facultative skin pigmentation at UV-
exposed sites was higher in females than in males except for
the forehead (Table II). Buttock skin pigmentation was also
statistically higher in females, but when solarium users were
Eye colour
excluded there was no signi¢cant di¡erence between constitu-
The dominating eye colour in the sample was blue with the fair tive skin pigmentation at the buttocks in either females or
eye colours (blue, green and grey) constituting the majority in males (13.3 vs. 12.8, p~0.18). More females than males were
both males and females; only 16% had brown eye colour (Table solarium users (33% of females and 14% of males). Considering
I). There was a gender di¡erence with more males having blue only persons that were not solarium users, females still had sig-

Table II. Skin pigmentation at various body sites

Body site Males Females MzF

mean SD mean SD mean SD

Upper arm, lateral 23.9 5.4 p~0.02 24.8 5.6 24.5 5.5
Chest 19.9 5.9 pv0.01 22.1 5.9 21.1 6.0
Back 19.8 5.5 pv0.01 21.2 5.6 20.6 5.6
Upper arm, medial 18.9 4.5 pv0.01 22.1 5.0 20.7 5.0
Forehead 18.1 5.4 p~0.39 17.8 5.0 17.9 5.2
ButtockA 13.3 5.5 pv0.01 14.8 5.6 14.2 5.6
ButtockB 12.8 4.8 p~0.18 13.3 4.8 13.1 4.8

Skin pigmentation measured by skin re£ectance spectroscopy at UV-exposed skin sites in all persons (facultative pigmentation) and at buttock skin
in all personsA (n~889) and in buttock skin of personsB not using arti¢cial tanning devices (n~666). Only in persons not using arti¢cial tanning
devices can the buttock pigmentation be considered to be the constitutive pigmentation. Statistical analyses of gender di¡erences by the unpaired
t-test.

Acta Derm Venereol (Stockh) 78

 Skin pigmentation 217

skin colour scales which at best have a resolution of seven steps

in Caucasians (15). Furthermore, we have previously estab-
lished that measurements of skin pigmentation by re£ectance
spectroscopy in not-previously UV-exposed buttock skin will
predict the UV sensitivity, as determined by a MED phototest,
with a high degree of precision (18).
Anamnestic skin type is per de¢nition dependent on the
recall precision of the study person's memory and cannot be
objectively estimated. Furthermore, skin typing involves the
combination of two di¡erent questions: burning susceptibility
and tanning ability. This is unfortunate from a scienti¢c point
of view with only four possible combinations of burning ten-
dency and tanning ability and may force Caucasians into
choosing a wrong skin type because the right combination of
burning tendency and tanning ability is not available (19).
Moreover, the questions used in skin typing may be interpreted
di¡erently in di¡erent populations and cultures. Despite this
obvious weakness, skin typing has found widespread use in
clinical and epidemiological studies because it is quick and
easy and can be performed without any equipment.
We found that the major eye colour in the Danish population
was blue, which compares well with the study of Pälsson et al.
(20). Females had darker eye colours than males, which has
also been found in other studies (21). Eye colour is not static,
however, and in a longitudinal study in children from 3 months
to 6 years of age Matheney & Dolan (21) have found an age
trend with a decrease in the lighter eye colours and a corre-
sponding increase in the darker colours with age. In parallel
with age-associated changes in eye colour it has been shown
that the hair of Caucasians tends to darken with age (22). We
Fig. 2. Eye colour and hair colour according to constitutive skin pig- chose to look at hair colour at the age of 25 years to avoid this
mentation. Eye colour in 619 persons (older than 1 year of age) and
age-related trend because our study sample had a wide age
hair colour at 25 years of age in 506 persons according to constitutive
range from 0 to 85 years.
skin pigmentation measured by skin re£ectance spectroscopy in per-
sons not using arti¢cial tanning devices. The four pigmentation quar- In a study in 809 indoor o¤ce workers by M. Berg per-
tiles divided the persons into four pigmentation groups (median formed in four Swedish cities (23), it was found that the most
buttock skin pigmentation~12.5% pigment). frequent skin type was type III, with 61% of the study persons;
the least represented was skin type I, with only 2% . The great-
est variation in skin types between the four cities in the Swedish
ni¢cantly higher pigmentation at the chest and at the inside of
study was in skin type IV, with 4% of the study persons in
the upper arm (pv0.01 for both regions), while pigmentation
Gothenburg in the south region of Sweden and 16% in Sunds-
at the forehead, the back and the outside of the upper arm did
wall in the northern region (23). In the German skin cancer
not di¡er statistically in females and males. F|g. 3 shows objec-
screening campaign ``Give Skin Cancer No Chance'' (24) it
tively measured skin pigmentation according to subjectively
was found that approximately 84% of the 24,432 screened per-
assessed skin type for the lateral arm and the back and but-
sons were of skin type II or III.
tocks. For the arm there were signi¢cant di¡erences in pigmen-
We found that eye and hair colours were evenly distributed
tation between adjacent skin types but for the back and the
within almost all skin types and Ho¡man et al. (24) found a
buttocks there were no signi¢cant di¡erences between pigment
comparable distribution of hair and eye colours within the four
distribution in adjacent skin types except for skin types II and
Caucasian skin types. Eye and hair colour cannot therefore be
III. The box-and-whisker plots demonstrate that there is con-
used as a substitute for skin typing in studies of sun sensitivity.
siderable overlapping in pigmentation range for all skin types
However, eye and hair colour seem to carry additional infor-
and that anamnestic skin cannot be used as an indicator to
mation about sun reactivity which is not contained in the Fitz-
clearly distinguish between di¡erent degrees of facultative
patrick classi¢cation. These two pigmentary traits have been
and constitutive skin pigmentation in individual subjects.
shown to be independent risk factors of skin cancer after
adjustment for skin type (6, 9) and should therefore be included
in analytic studies of skin cancer aetiology.
DISCUSSION
Our objective measurements of facultative and constitutive
In this study, we have used simple but clinically relevant colour skin pigmentation in all skin types showed a considerable dar-
scales to assess eye colour and hair colour rather than elabo- kening of the skin at UV-exposed sites compared to non-UV-
rate arti¢cial scales which are of doubtful clinical relevance. exposed buttock skin. Females had more pigmented skin at the
For assessing skin pigmentation we bene¢ted from having skin exposed sites (Table II) but not at UV-unexposed buttock skin,
re£ectance spectroscopy equipment giving objective pigmenta- suggesting that females had been more UV-exposed than
tion measurements on a continuous scale rather than arti¢cial males. The four major eye and hair colours were found within

Acta Derm Venereol (Stockh) 78

218 J. Lock-Andersen et al.

there seems to be an association between the two melanine

types in hair and skin (31, 32).
Fig. 3, which correlates skin types to skin pigmentation,
demonstrates clearly that F|tzpatrick skin types in Caucasians
cannot reliably be used to classify individual subjects according
to their constitutive skin pigmentation, since persons with a
broad range of constitutive pigmentation are found within
each skin type and there is wide overlapping between adjacent
skin types. This is in accordance with our previous results on
UV sensitivity in 248 persons of skin types I^V (15) and other
phototest studies (19, 33). We have even found that reproduci-
bility of anamnestic skin type is quite unsatisfactory because
one of every three persons could not classify themselves in the
same skin type 6 to 12 months after ¢rst being questioned (15).
These observations stress the need for a new skin type classi¢-
cation system.

ACKNOWLEDGEMENTS
We thank Mariane Bauer for invaluable help with data collection and
Age VÖlund for help with data analysis. The study was supported by a
grant from the Danish Cancer Society (Grant No. 94^037).

REFERENCES
1. Amblard P, Beani J, Gautron R, Reymond J, Doyon B. Statistical
study of individual variations in sunburn sensitivity in 303 volun-
teers without photodermatosis. Arch Dermatol Res 1982; 274:
195^206.
2. Andreassi L, Simoni S, F|orini P, F|miani M. Phenotypic charac-
ters related to skin type and minimal erythemal dose. Photoderma-
tol 1987; 4: 43^46.
3. Azizi E, Lusky A, Kushelevsky AP, Schewach-Millet M. Skin type,
hair color, and freckles are predictors of decreased minimal
erythema ultraviolet dose. J Am Acad Dermatol 1988; 19: 32^38.
4. Òsterlind A. Epidemiology of malignant melanoma in Europe.
Reviews in Oncologica 1992; 5: 903^908.
5. Òsterlind A. De¢ning individual risk factors for skin melanoma.
In: Altmeyer P, Ho¡mann K, StÏcker M, eds. Skin cancer and
Fig. 3. Objectively measured skin pigmentation in UV-exposed and UV radiation. Berlin: Springer Verlag, 1997: 562^570.
unexposed skin. Facultative skin pigmentation in persons not using 6. Bliss JM, Ford D, Swerdlow AJ, Armstrong BK, Cristofolini M,
arti¢cial tanning devices at the outside of the upper arm (668 persons) Elwood JM, et al. Risk of cutaneous melanoma associated with
and the upper back (667 persons) and the medial and upper quadrant pigmentation characteristics and freckling: systematic overview
of the buttocks (666 persons) according to their anamnestic skin type. of 10 case-control studies: the International Melanoma Analysis
Data are shown as box-and-whisker diagrams where the box extends Group (IMAGE). Int J Cancer 1995; 62: 367^376.
from the 25th percentile to the 75th percentile and with a horizontal 7. Evans RD, Kopf AW, Lew RA, Rigel DS, Bart RS, Friedman RJ,
line at the median value (50th percentile). Whiskers extend down to et al. Risk factors for development of malignant melanoma^I:
the minumum value and up to the maximum value. Note the extensive review of case-control studies. J Dermatol Surg Oncol 1988; 14:
overlapping in range of skin pigmentation between all skin types, not 393^408.
just for the extreme pigmentation values but also for the main core of 8. Suärez-Varela MM, Gonzälez AL, Caraco EF. Non-melanoma
values (25th percentile to 75th percentile) in all skin types. Statistical skin cancer: an evaluation of risk in terms of ultraviolet exposure.
analysis of skin pigmentation in skin type I vs. type II, type II vs. type Eur J Epidemiol 1992; 8: 838^844.
III, and type III vs. type IV by Mann-Whitney's test. 9. Zanetti R, Rosso S, Martinez C, Navarro C, Schraub S, Sancho-
Garnier S, et al. The multicentre south European study ``Helios''.
I: Skin characteristics and sunburns in basal cell and squamous cell
carcinomas of the skin. Br J Cancer 1996; 73: 1440^1446.
all four pigmentation groups (Fig. 2) except red hair colour, 10. Gallagher RP, Hill GB, Bajdik CD, F|ncham S, Coldman AJ,
which was not found in the most pigmented group. Hair colour McLean DI, et al. Sunlight exposure, pigmentary factors, and risk
of nonmelanocytic skin cancer. I. Basal cell carcinoma. Arch Der-
and eye colour and skin pigmentation are regulated by di¡erent
matol 1995; 131: 157^163.
genes by complex genetics (25^28) and may result in a wide
11. Gallagher RP, Hill GB, Bajdik CD, Coldman AJ, F|ncham S,
variety of phenotypes. There are two di¡erent types of melanin McLean DI, et al. Sunlight exposure, pigmentation factors, and
present in human hair [the brownish black eumelanin and the risk of nonmelanocytic skin cancer. II. Squamous cell carcinoma.
reddish yellow pheomelanin (29)] which di¡er not only chemi- Arch Dermatol 1995; 131: 164^169.
cally but also physically (30). Recent studies have now demon- 12. F|tzpatrick TB. The validity and practicality of sun-reactive skin
strated the presence of pheomelanin in human skin (31) and type I through VI (Editorial). Arch Dermatol 1988; 124: 869^871.

Acta Derm Venereol (Stockh) 78

 Skin pigmentation 219

13. Wulf HC, Lock-Andersen J. Standard erythema dose. Skin Res 24. Ho¡mann K, Gebler A, Trampisch HJ, Ho¡mann A, Lueg A, Alt-
Tech 1996; 4: 192. meyer P, et al. Do not give skin cancer a chance. In: Altmeyer P,
14. Wulf HC. Method and apparatus for determining an individual's Ho¡mann K, StÏcker M, eds. Skin cancer and UV radiation. Ber-
ability to stand exposure to UV. US Patent 1986; 14: 882, 598: 1^32. lin: Springer Verlag, 1997: 820^849.
15. Wulf HC, Lock-Andersen J. Measurements of constitutive skin 25. Kalla AK. Human skin pigmentation, its genetics and variation.
phototypes. In: Altmeyer P, Ho¡mann K, StÏcker M, eds. Skin Humangenetik 1974; 21: 289^300.
cancer and UV radiation. Berlin: Springer Verlag, 1997: 169^180. 26. Eiberg H, Mohr J. Major genes of eye color and hair color linked to
16. Lock-Andersen J, Wulf HC. Seasonal variation of skin pigmenta- LU and SE. Clin Genet 1987; 31: 186^191.
tion. Acta Derm Venereol (Stockh) 1997; 77: 219^221. 27. Eiberg H, Mohr J. Major locus for red hair color linked to MNS
17. Amternes folketal efter kÖn og alder. Tabel 48. In: Statistisk Ðrbog blood groups on chromosome 4. Clin Genet 1987; 32: 125^128.
1996. Danmarks Statistik. KÖbenhavn, 1996: 69. 28. Robins AH. Biology of the pigment cell. In: Robins AH, ed. Bio-
18. Lock-Andersen J, Gniadecka M, de F|ne Olivarius F, DahlstrÖm logical perspectives on human pigmentation. Cambridge: Cam-
K, Wulf HC. UV induced erythema evaluated 24h post-exposure bridge University Press, 1991: 1^24.
by skin re£ectance and laser Doppler £owmetry is identical in 29. Jimbow K, Ishida O, Ito S, Hori Y, Witkop CJ, King RA. Com-
healthy persons and patients with cutaneous malignant melanoma bined chemical and electron microscopic studies of pheomelano-
and basal cell cancer. J Photochem Photobiol (B) 1997; 41: 30^35. somes in human red hair. J Invest Dermatol 1983; 81: 506^511.
19. Rampen FHJ, Fleuren BAM, de Boo TM, Lemmens WAJG. 30. Chedekel MR. Photochemistry and photobiology of epidermal
Unreliability of self-reported burning tendency and tanning ability. melanins. Photochem Photobiol 1982; 35: 881^885.
Arch Dermatol 1988; 124: 885^888. 31. Thody AJ, Higgins EM, Wakamatsu K, Ito S, Burchill SA, Marks
20. Pälsson JOèP, Eriksson AW, Forsius H, Fellman J. Comparison of JM. Pheomelanin as well as eumelanin is present in human epider-
hair and iris pigmentation in Scandinavian populations. Arctic mis. J Invest Dermatol 1991; 97: 340^344.
Med Res 1994; 53: 52. 32. Valverde P, Healy E, Jackson I, Rees JL, Thody AJ. Variants of the
21. Matheny AP, Dolan AB. Changes in eye colour during early child- melanocyte-stimulating hormone receptor gene are associated
hood: sex and genetic di¡erences. Ann Hum Biol 1975; 2: 191^196. with red hair and fair skin in humans. Nat Genet 1995; 11:
22. Matheny AP, Dolan AB. Sex and genetic di¡erences in hair colour 328^330.
changes during early childhood. Am J Phys Anthropol 1975; 42: 33. Snellman E, Jansën CT, Leszynski K, Visuri R, Milan T, Jokela K.
53^56. Ultraviolet erythema sensitivity in anamnestic (I^IV) and photo-
23. Berg M. Epidemiological studies of the in£uence of sunlight on the tested (1^4) Caucasian skin phototypes: the need for a new classi-
skin. Photodermatology 1989; 6: 80^84. ¢cation system. Photochem Photobiol 1995; 62: 769^772.

Acta Derm Venereol (Stockh) 78