AMERICAN JOURNAL OF PHYSICAL ANTHROPOLOGY 117:157–168 (2002)

A Method for Visual Determination of Sex,

Using the Human Hip Bone
Jaroslav Bruzek*

U.M.R. 5809 du C.N.R.S., Laboratoire d’Anthropologie des Populations du Passe´ Universite´ Bordeaux I, 33405
Talence, France

KEY WORDS human pelvis; sex determination; morphological traits; method

ABSTRACT A new visual method for the determina tion of sex DOI 10.1002/ajpa.10012
using the human hip bone (os coxae) is pro posed, based on a identify sex in only 3%. The advantage of this new method is a
revision of several previous approaches which scored isolated reduction in observer subjectivity, since the evalua tion procedure
characters of this bone. The efficacy of the methodology is tested cannot involve any anticipation of the re sult. In addition, this
on a sample of 402 adults of known sex and age of French and method of sex determination in creases the probability of a correct
Portuguese origins. With the simultaneous use of five characters diagnosis with isolated fragments of the hip bone, provided that a
of the hip bone, it is possible to provide a correct sexual diagnosis combination of elements of one character is found to be typically
in 95% of all cases, with an error of 2% and an inability to male or female. Am J Phys Anthropol 117:157–168, 2002.
© 2002 Wiley-Liss, Inc.

Correct sex identification of the human skeleton is

important in bioarcheological and forensic practice. Current The method proposed by Iscan and Derrick (1984)
opinion regards the hip bone (os coxae) as providing the provides an accuracy level of 90% (Iscan and Dun lap,
highest accuracy levels for sex deter mination. However, 1983), but it cannot be regarded as equivalent to the results
“simple” observations of the hip bone without any scoring found with methods using the entire hip bone. The accuracy
of related traits should not be normally considered proper, of the method of Ferembach et al. (1980) has not been
despite the fact that the results may be surprisingly studied. Bruzek and Ferem bach (1992) found 93% correct
accurate. Only those procedures which define a precise and sex assignment, using a set of eight variables of the hip
unambiguous methodology leading to an accurate diagnosis bone. However, this method necessitates highly trained
of sex should be employed. observers with experience in morphological variability,
Three techniques for the visual evaluation of traits of the because it is based on an ordinal scale of evaluation.
hip bone are: 1) the method of Phenice (1969), which uses When (Novotny´, 1981, 1988) studied the discrim
three traits on the pubis, 2) the method of Iscan and Derrick ination power of 14 sexually dimorphic traits of the hip
(1984) using the poste rior pelvis, and 3) the method of bone frequently used in sex determination, he replaced a
Ferembach et al. (1980) of sexing the entire pelvis through descriptive or ordinal evaluation of char acters (e.g., small,
an evalu ation of eleven traits. However, even if it is wide, shallow) with binary scor ing (yes or no) along an
generally accepted that these methods provide satisfactory intermediate category mor phology. He then selected the
ac curacy, only a few studies have tested their reliabil ity in characters with the highest diagnostic value, reducing the
known-sex samples and, surprisingly, the re sults are often analysis to three complex variables: the preauricular
ignored. surface, the sciatic notch, and the inferior aspect of the hip
The results from the method of Phenice (1969) have been bone. These characters reduce misclassification (Bruzek,
inconsistent, as accuracy levels range from 59% 1991). In addition, Rogers and Saunders (1994) tested
(MacLaughlin and Bruce, 1990) to 96% (Schone, quoted in accuracy and reliability in a set of 17 individual traits on a
Sutherland and Suchey, 1991). Such inconsistency is due to small sample of identified pel ves. Accuracy varied from
the fact that sexual dimorphism of the whole hip bone 6.1–94.1% for each indi vidual pelvic trait, and the greatest
should be consid ered, and observations should not be accuracy was obtained by combining the scoring of three
restricted to the pubis (Novotny´, 1986; Bruzek, 1992). traits.
Following Lovell (1989), MacLaughlin and Bruce (1990),
and Bruzek (1991), the reliability of the method of Phenice
*Correspondence to: Jaroslav Bruzek, Laboratoire d’Anthropologie des
(1969) is probably ca. 80%. Furthermore, pubic Populations du Passe´, UMR 5809 du CNRS, Universite´ Bordeaux I, Avenue
preservation rarely exceeds 30% in archaeo logical samples des Faculte´s, 33 405 Talence, France.
(Waldron, 1987). E-mail [email protected]

Received 2 September 1997; accepted 21 August 2001.

© 2002 WILEY-LISS, INC.
158 J. BRUZEK
The level of intraobserver error for all pelvic traits in In these techniques and evaluations of them, the
combination was 11.3%. frequently cited drawbacks of visual techniques for
determining the sex of an individual are: 1) the high degree sexual dimorphism in the sacroiliac complex is visible
of observer subjectivity, 2) a lack of consis tency in the during the fetal period (Boucher, 1955; Fazekas and Kosa,
evaluation of traits, and 3) a strong dependence on the 1978; Holcomb and Konigsberg, 1995). The fourth and
results of previous experiences of the observer. fifth charac ters (the ischiopubic complex) reflect the
Nonetheless, there is an advantage in visual techniques, adaptation of the female pelvic canal to the requirements of
which emanate from their rapid ity of use as well as their reproduction. In addition, this region reflects differ ences in
ability to be used when damage does not allow complete set the anatomy of the external genitalia in both sexes.
if measure ments. Manifestations of sexual dimorphism in the ischiopubic
To address these problems, this study proposes a method complex do not begin until puberty (Coleman, 1969).
for the visual allocation of sex from the hip bone,
emphasizing four aspects insufficiently consid ered in General procedure
previous studies. First, there is a reduction of observer
subjectivity during the evaluation of se lected traits by Table 1 presents a formula for sex diagnosis and the
using only three possibile scores (present, indeterminate, principles of evaluation. The first column lists the five
absent), contrary to ordinal scoring in which it is always characters used. Sexing can be divided into three steps: 1)
difficult to make a decision between two neighboring scoring of the traits; 2) evaluation of the form of the
categories. Sec ond, the method eliminates confusion character; and 3) sexual diagnosis based on the five
between traits (e.g., the preauricular and paraglenoid characters (last column).
grooves, which represent separate elements; Kurihara et al., Each character is defined either by one condition
1996). Third, when necessary for complex charac ters, it (simple: third and fifth characters) or three condi tions
uses a rigorous evaluation of three relatively independent (complex: first, second, and fourth characters). The
characters reflecting the sex of the in dividual. When at observer determines whether each character is male or
least two of these elements lead in the same direction, it is female or indeterminate. For the simple characters, this
possible to decide whether a determination is possible. A score corresponds directly to the sexual form of character,
reliable male or female diagnosis is considered possible designated by symbols “F” for female form, “0” for
when at least two variables are concordant. This approach intermediate form, or “M” for male form. The sexual form
is the oppo site to one attempting to score a primary sexual of complex characters results from the combination of the
characteristic, thereby anticipating the presence or absence three separate subcharacters. When at least two answers
of its secondary manifestations. And fourth, this method indicate either “male” or “female,” these are used as a basis
can be applied to damaged or incom plete hip bones. to determine sexual morphology. A majority of “F” or “M”
sexual forms indicates female and male sex, respectively. If
MATERIALS equal amounts of the sexual forms are present, sex is
considered as indeterminate. We should underscore that all
This study is based on an examination of 402 specimens
the figures given in this text are provided for strictly
of adult individuals of known sex from two European
illustrative purpose and do not correspond at all to practical
collections (Paris, France: 98 males and 64 females;
diagnostic tools.
Coimbra, Porugal: 106 males and 134 females). The French
collection dates to the first half of the twentieth century
(Carre, 1978; Olivier, 1972); the Portuguese collection Preauricular surface
includes people born between 1820–1920, for which a On the preauricular surface, three morphological
number of sociological, medical, and anthropological data structures may occur independently of each other, all of
are also available (Bocquet et al., 1978). which are sexually dimorphic (Fig. 1). The first structure
METHODS was originally described by Lo¨hr (1894) as the
“paraglenoid groove” and is characterized by a depression
Five characters of the hip bone were used. They are: 1) or groove along the antero-inferior edge of the auricular
aspects of the preauricular surface (Novotny´, 1981), 2) surface. The arc of the groove is open, or makes up less
aspects of the greater sciatic notch (No votny´, 1981), 3) than half the circumference of a circle (Hoshi, 1961). The
the form of the composite arch (Genove´s, 1959), 4) the depression is a result of ligaments exerting tension on the
morphology of the inferior pelvis (margo inferior ossis sacroiliac joint capsule (Weisl, 1954). This groove is
coxae; Novotny´, 1981), and 5) ischiopubic proportions. related to gen eral skeletal robusticity and thus is more
These characters were chosen to reflect the func tional frequent in males (Lazorthes and Lhe´s, 1939).
morphology of two very distinct areas of the The second structure is the preauricular groove (Zaajer,
pelvis: the sacroiliac complex and the ischiopubic complex. 1866), a closed depression in the bone in the form of one or
The first three characters are sex-specific adaptations of the several pits (Hoshi, 1961), each with a closed
sacroiliac complex to bipedal lo comotion. Expression of circumference (arc is greater than half of a

ih

ht
fo

s
r

la

S
o

it

e
F

)B

C(

dr
o

hc

ir
e

)B

f
h

c
f

il

id
n

e
F0M
m

dn

u
f

i
F0M

hc

if

de
s

do

ht

EL

AT
n

it

il

fo

ir
o

S
e

id
n

hC
)

i
p

il

ll

is
s

f o

le
v

d:

id

iF

ir

P
m

de

il

it

fr

s
f

il

it

lt

il
s

ht

f
e

il

fr

ir

p
e

ic

lc

iw

ip

f o

id

S
m

de

it
t

ip

g
e

ic

o
h

iw

is
s

d,

m
e

lc
r

fo

kc

le
v

d:

id

ih

T
m

de

il

is
o

fo
e

lc

lc
r

br

fr

ir

p
o

la
u

ht

l)

A(

dr
o

hc

ir
e

f o

it
r

id
n

iF

it

G
m

de

dr
o

hc

it

fo

ht

n
C(

dr
o

hc

ir
e

ht

hs

)C
A(

dr
o

hc

ir
e

m
it

fo

it
r

ht

it

f o

f:

id
n

S
m

de

dr
o

hc

c
h

it

fo

ht

it

y
r

m
p

dr
o

hc

ir
e

fo

)r

il
t

f o

id

ih

T
m

de

dr

hc

ir
e

p
n

il

id

dr
o

hc

ir
e

s
o

fo

il

it

r
ht

it

t
e

lb

it

id

is
o

C
m

de

it

fo

il
t

te

b
hc

a
e

lg

iS

f o

il
t

n
e

fr

ir

a
n

la

id
n

iF

ir
e

I
m

de

fo

it

ir
e

hc

s
i

le

p
tr

de

fo

id

is
s

ir
e

i
o

lt
t

il

fo

di
r

il
la

hp

ht

fo

kc

ba

id
n

S
m

de

di
r

ill

hp

ht

fo

e
r

p
e

di
r

ill

hp

ht

fo

lc

m
tc

li

ib

hc
s

i:

id
n

ih

T
m

de

tc

r
tc

m
m

hc
s

ht

ib

it

l
e

id

ib

hc
s

I
m

de

hc
s

ib

it
r

p
s

ib

ht

hc
s

ht

160 J. BRUZEK
second step evaluates the outline border of the pit or
the groove; if an absence of negative relief is ob
served in the first step, then the score is “m” in the
second step as well. In case of negative relief, if the
outline border of this depression is closed, one scores
“f.” Forms of the depression with open borders are
classified as “m.” The third step evaluates the pres
ence and relief of the piriform tubercle. Only an f-f-f
combination determines a true preauricular groove
 and a reliable female sex diagnosis. Other combina
tions yield the theoretically highest risks of misclas
sifications.

Greater sciatic notch

The sexual characteristics of the sciatic notch are
exceedingly difficult to use. Not only is the observer
influenced by the size of the pelvis (which undoubt
edly affects the morphological shape of the sciatic
notch), but by the development of marginal struc
tures (e.g., ischial spine, piriform tubercle) which
exhibit high levels of sexual dimorphism. Direct
evaluation of notch shape is thus subjective. Fur
thermore, a number of authors (Lazorthes and Lhe´s,
Fig. 1. Preauricular surface. as, auricular surface; sn, sciatic notch; t, 1939; Glanville, 1967; Singh and Potturi, 1978; No votny´,
piriform tuubercle; pg, preauricular groove; f, female condition; m, male 1981) demonstrated that the degree of sexual dimorphism is
condition. F: Specific female shape (f-f-f), showing the deep depression,
with closed circumference (true preauricular groove), associated with lack most distinct along the posterior chord of the sciatic notch.
of tubercle. M: Typical male form (m-m-m), showing the relief as smooth In order to achieve a more objective evaluation of the
or very slight, with open border, associated with presence of tubercle. Two notch shape, Novotny´ (1981) proposed the use of a
ex amples of other female forms: f-f-i, deep depression, pits or groove with “shadow image” of the contour. This is a modifica tion of
closed border, with intermediate aspects of positive relief; i-f-f, little
depression with closed border, associated with lack of tubercle. Two the technique employed by Hanna and Wash burn (1953),
examples of other male forms: m-m-i, little groove opened laterally, with whereby a contour of the sciatic notch is obtained by the
intermediate aspect of positive relief (paraglenoid groove); m-f-m, smooth exposure of photosensitive paper onto which the hip bone is
relief with closed circumfer ence and very prominent tubercle. placed, with the lat erodorsal iliac surface on the paper.
After exposure, a contour is drawn by marking points
circle) along the anteroinferior edge of the auricular corresponding to the scheme outlined in Figure 2. The line
surface. According to Houghton (1974), this forma tion of the notch width (A–B) is defined from the base of the
results from obstetrical trauma during the course of ischial spine (point B, frequently broken off but its base
delivery. This trait, in its most typical form, would allow preserved) to the top of the piriform tubercle (point A). In
for differentiation between two groups, nulliparous women case of its absence, the line should be drawn to the posterior
plus males vs. females who have given birth (Ullrich, inferior iliac spine (point A ). A perpendicular line (the
1975). It is possible, however, that for parous females, the depth) is then drawn from the deepest point of the notch to
flexibility of the pelvis and the relationship between their the notch width, separating it (at point C) into two chords,
pelvic dimensions and the size of the foetus at parturition posterior and anterior.
do not always produce this obstetrical trauma dur ing The three aspects can then be evaluated. The first
delivery; therefore, some parous females do not present a evaluates visually the proportions between the two
preauricular sulcus. segments (A–C and C–B) of the A–B chord. The second
The third structure in this region is the highly variable, trait evaluates the symmetry (female ele ment “f”) or
so-called piriform tubercle (tuberculum musculi piriformis). asymmetry (male element “m”) of the notch base according
The presence of this feature is likely associated with high to the depth line. The third aspect follows the course of the
levels of muscular activ ity, and it occurs more frequently contour above the posterior chord, using a line parallel with
in males than in females (Genove´s, 1959). the depth line extended through point A of the sciatic notch
For sexual assessment, evaluation of all three structures width.
is necessary. Three steps are required to determine the
sexual morphology of the preauricu lar surface (Novotny´, Composite arch
1981). The first is to determine the presence or absence of The trait proposed by Genove´s (1959) evaluates the
negative relief (pits). The course of the contour of the anterior arm of the
VISUAL SEXING OF THE HUMAN PELVIS 161
 Fig. 3. Composite arch. Outline of anterior sciatic notch chord
(2), relative to outline of anterior segment of auricular surface (1).
as, auricular surface; sn, sciatic notch. M: Absence of composite
arch (single curve, 1 2). F: Presence of composite arch (double
curve, 1 2).

Inferior pelvis (inferior margin of the os coxae)

Sexual dimorphism of the ischiopubic ramus, the caudal
Fig. 2. Greater sciatic notch. as, auricular surface; sn, sciatic notch; A, region of the hip bone that results from the junction
top of piriform tubercle, in the case of its absence; A , top of posterior between ischium and pubis, is concentrated on its lateral
inferior iliac spine; AB and A B, sciatic notch breadth; CD, sciatic notch
depth; AC and A C, posterior chord of sciatic notch breadth; CB, anterior
margin (Fig. 4). In the pelvis, both the left and right
chord of sciatic notch breadth; AP (A P), perpendicular at point A (A ) to ischiopubic rami form strongly sexu ally differentiated
the line formed by the sciatic notch breadth. F: Extreme female form (f-f- subpubic angles: a closed angle (angulus pelvis) in males,
f), showing posterior chord segment AC (or A C) longer (or close to and an obtuse angle (ar cus pelvis) in females. Sexual
equality) than the anterior chord CB, symmetry of notch contour associated differences of the is chiopubic ramus are related to the
with no crossing of the line A–P (A –P) with the contour of the posterior
chord. M: Extreme male form (m-m-m), showing the posterior chord shape of the pelvic outlet and to the anatomy of the
segment AC (A C) shorter than the anterior chord CB, asymmetry of the urogenital tract. The angle of the pelvis, however, cannot be
outline chords associated with the crossing the line A-P (A P) with the evaluated directly on the hip bone.
contour of the posterior chord. Two examples of other female forms: f-i-f, Following Novotny´ (1981), this region can be sexed by
posterior chord segment AC equal the anterior chord CB, associated with
intermediate pat tern of the outline chords relative to the deep of the notch,
a simultaneous evaluation of three condi tions. The first
and outline of the posterior chord does not cross the perpendicular line AP. concerns the external eversion of the inferior border. The
f-f-i, posterior chord segment AC equal to anterior chord CB, presence of second consists of a morpholog ical analysis of the middle
relative symmetry of outline chords to the deep of notch and the line AP part of the ischiopubic ramus, which is frequently broad in
tangent to the outline of the posterior chord. Two examples of other male males and has a laterally directed swelling known as the
forms: m-m-i, posterior chord segment AC shorter than anterior chord CB,
associated with asymmetry of outline chords and line AP tangent to the crista phallica (phallic crest). In certain female specimens, a
outline of the posterior chord; m-m-f, posterior chord segment AC shorter narrowing occurs in the middle part of the ramus, which is
than anterior chord CB, associated with asymmetry of outline chords; quite similar to males who display the presence of the
finally, outline of posterior chord doe not cross perpendic ular line AP. phallic ridge (Poulhes, 1948; Genove´s, 1959). When the
phallic ridge is not sepa rate, then the course of the inferior
pelvic aspect (inferior margin of the os coxae) may seem to
be identical in both sexes (Francis, 1952; Smout and
Jacoby, 1948). The third condition refines the two
auricular surface and the contour of the anterior chord of preceding ones evaluating the gracility (f) or robus ticity
the sciatic notch (Fig. 3). (m) of the inferior pelvic aspect. It is recog nized that the
In females, both contours are sections of two dis tinct scoring of this character is subjective. It depends on the
circles with different radii: the composite arch is present. experience of the investigator and on the nature of the
Among males, both contours are a part of one circle: the population studied. For these
composite arch is absent.
162 J. BRUZEK
 the ischiopubic index (Schultz, 1930; Washburn,
1948).
RESULTS
The reference samples were scored blind, without
knowledge of actual sex. The level of accuracy cor
responds to the percentage of correct sex identifica
tion. The level of error corresponds to the percentage
of individuals whose sex was incorrectly identified
without taking into account specimens recorded as
“indeterminate.”

Preauricular surface
The results of the evaluation of the preauricular
surface are shown in Table 2. The female morphol
ogy of the preauricular surface occurred in 76% of
French and 70% of Portuguese females. In three
male specimens, the feminine morphology of this
character was observed, in conjunction with the
presence of the piriform tubercle, resulting in the
combination (f-f-m). Similar results were obtained
by Novotny´ (1981) in females of a Czech-German
sample (71% of females displayed the female mor
phology of the character). These results also corre
spond to the frequency of 71% for a “GP-type” named
by Houghton (1974), and 70% for the “cavity type” of
pits by Hoshi (1961). These results cannot be com
Fig. 4. Inferior pelvis. pus, pubic symphysis; itu, ischial tu berosity; cp, pared with those of other authors, who considered any
phallic ridge. Horizontal lines limit the middle part of the ischiopubic depression as the preauricular groove and did not
ramus. Broken line represents major axis of this structure. F: Typical
female form (f-f-f), showing external eversion of ischiopubic ramus and differentiate it from the paraglenoid groove. In the two
absence of the phallic ridge, associated with gracility of the bones. M: groups studied here, the f-f-f combination (true preauricular
Typical male form (m-m-m), showing direct course of medial part of groove) was observed in 23% of females from Paris and
ischiopubic ramus and presence of phallic ridge, associated with 28% of females from Coim bra. Novotny´ (1981) reported
robustness of the bones. Two examples of other female forms: f-f-i,
external eversion of ischiopubic ramus and absence of phallic ridge, asso the presence of this combination in 31% of women. The
ciated with intermediate aspect of ischiopubic ramus; f-i-f, exter nal combination at tributed to male morphology is shared with
eversion of ischiopubic ramus and doubt about absence of crista phallica, nullip arous women. This combination alone, therefore,
associated with gracility of the bones. Two exam ples of other male forms: cannot result in an accurate diagnosis of sex when
m-i-m, direct course of medial part of ischiopubic ramus and doubt about fragments of hip bone are examined.
presence of phallic ridge, associated with robustness of bones; m-m-i,
direct course of me dial part of ischiopubic ramus and presence of phallic Recently, Schemmer et al. (1995) refuted the con
ridge, associated with intermediate aspect of ramus morphology. clusions of Spring et al. (1989) and Cox and Scott (1992)
about the lack of a relationship between the preauricular
groove and parity. These authors doc umented a deep
groove in the preauricular region due to past pregnancy.
reasons, in many cases, the intermediate category is These authors nonetheless conceded that genetic and
selected. The sexual form of the character is given by a behavioral factors con tribute to variability in the
combination of the results of the three steps, ac cording to expression of the deep groove.
Table 1.
Greater sciatic notch
Ischiopubic proportions
The results of the shape of the greater sciatic notch are
The sexual differences of the ischiopubic propor tions shown in Table 2. In the sample from Paris, the female
reflect the adaptation of the female pelvis to parturition, shape of the notch was observed in 67% of the women, and
and result in the remodeling of the fe male bone pelvis in the Coimbra sample, it was found in 70% of the women.
during puberty. These changes can be observed in the The male condition of the character occurred in 67% of
proportions of the length of the pubis and ischium (Fig. 5). French men and 80% of Portuguese men. Previous studies
In women, the pubic length/ischium length ratio is using this methodology reported 58% of women and 85%
higher than in men. Practically, this ratio is as sessed by a of men with the correctly corresponding sexual form of the
visual inspection only and corresponds to notch (Novotny´, 1981). Misclassifications oc
VISUAL SEXING OF THE HUMAN PELVIS 163
 Fig. 5. Ischiopubic proportions. Proportions of length of pubis (pu) and ischium (is). M: Male morphology (pu is). F: Female morphology (pu is).

curred in 8% of the men and in 21% of the women of the sample.

French group, but in only 6% of men and 16% of women of It would appear difficult to attribute the observed
the Portuguese collection. Novotny´ (1981) reported only differences in the composite arch solely to pelvic sexual
1% of males with the feminine form of the notch, and 10% dimorphism. Interpopulational differences suggest that it is
of females with the masculine form of the notch. When the unlikely that all pelvic differences would be attributed to
two samples are grouped, the typical female combination differences between sexes. The concordance of the
(f-f-f) was present in only one male, while the typical male determined and actual sex of the specimens studied in this
combination (m-m-m) was observed in less than 5% of series approaches 80%. This success rate falls within those
women from Paris and less than 7% of women from obtained by Genove´s (1959) (83%) and Novotny´ (1981)
Coimbra. Novotny´ (1981) did not find this com bination in (61%). The occurrence of erroneously attributed sex (12%)
women. The precision of sex determina tion using this set appears high in comparison to that of Genove´s (1959)
of traits to describe the shape of the sciatic notch directly (6%) and Novotny´ (1981) (8%). The percent age of
corresponds to an estima tion of the discrimination power specimens with indeterminate sex (7%) closely
of this character, which varies by around 70%, as found by approximates that of Genove´s (1959) (11%), while
Washburn (1948), Genove´s (1959), Novotny´ (1981), and Novotny´ (1981) noted a much greater percent age of
Bruzek and Ferembach (1992). indeterminate cases (31%).
This trait confers greater accuracy and precision when
Composite arch sexing a female pelvis, as indicated by the higher
The concordance of the observations with known sex proportion of correctly sexed females. But it is important to
(Table 2) were found in 67% of males and 92% of females note that the use of the composite arch in sexing individual
of the French series. In the Portuguese sam ple, observation specimens has its greatest potential in association with
of the male pattern was in agree ment with the known sex other features of the pelvis, rather than as an independent
of 78% of males, while female morphology was observed in indicator.
87% of the fe males. Genove´s (1959) demonstrated Inferior pelvis
concordance with known sex in 80% of males and 87% of
females in an English population. In contrast to this, No Table 3 shows the frequency of male, female, and
votny´ (1981) observed a single circle in 67% of males indeterminate morphology of the inferior border of the
1
and a double circle in only 56% of females in the Czech ischiopubic ramus. Male morphology was found
e

fo

s
fo

fo

T
s

it

ic
S

fr

ir

P
e

lp

ir

P
e

lp

io

lp

ir

lp

io

lp

ir

P
a

la

la

la

la

la

la

la

M
s

la

la

M
n

%n

%n

%n

%n

%n

%n

%n

%n

%n

8.

23

44

6.

45

0.

93

52

1.

11

4.

82

83

5.

32

51
3.

11

5.
21

8
5.

02

82

9.

01

2.

71

11

1.

33

5.

12

6.

11

5.

46

9.

01

7.

75

0.

22

3.

34

2.

32

6.

11

4.

82

83

8.

23

4.

53

32
5.
46

7.

69

6.

11

0.

22

95

4.

12

12

4.

96

39

4.

67

1.

76

34

2.

88

9.

96

39

8.

23

2.

67

94

0.

00

7.

01

9.

01

0.

34

9.

01

2.

32

1.

11
6.

11

7.

01

2.

32

3.

34

9.

01

2.

32

1.

33

2.

57

9.

01

6.

11

5.

12

8.

12

6.

11

1.

33

5.

46

8.

23

3.

64

1.

33

1.

11

8.

23

8.

12
0.

44

9.

01

6.

11

0.

22

1.

11

6.

68

6.

01

5.

64

3.

41

41

6.

11

3.

34

9.

01
1

2.

11

11

0.

51

91

2.

31

41

3.

11

5.

42

42

6.

51

12

2.

89

1.

91

21

4.

88

1.

33

7.

35

1.

71

81

4.

21

3.

02

02

7.

69

8.

45

1.

33

6.

89
3.

02

02
7.

01

8.

12

1.

11

7.

01

7.

45

0.

44

5.

46

7.

22

42

6.

43

3.

02

02

7.

01

4.

21

31

1.

21

21
7.

69

8.

83

14

6.

43
6.

52

52

7.

69

5.

85

26

7.

43

0.

94

84
4

3.

76

66

6.

51

12

4.

08

58

6.

12

41

3.

76

66

8.

41

02

98

49

7.

43

6.

99

09
.

la

de

la

la

la
m

de

la
la

S
f o

f
s

ti

t
f-
f-
f
i-
f-
f
f

i-
f
f-
f

i
m

f-
f
f-

f
f-
f-

m
F
i-

f
m

i-
f
m

i
i

i-
f
i-
f

i
f

i
i

i
m

i
i-

i
i

i-

m
f-

i
f

i-

m
i-

f-

m
0
m

f
m

f-

m
f-

m
m

i
m

i-

m
i-

m
m-

m-

m
M
m

VISUAL SEXING OF THE HUMAN PELVIS 165

TABLE 3. Frequencies of scoring of elements and sexual form distribution in two traits of ischiopubic complex of hip bone 1 Inferior pelvic
aspects Ischiopubic proportion aspects

Sexual form of traits Males Females Males Females Males Females n % n % n % n % n%n%n %
Paris sample Coimbra sample Paris sample Coimbra sample Males Females n %

f-f-f 30 47.6 2 1.9 68 51.1

f-f-i 10 15.9 1 1.0 9 6.8
f-i-f 6 4.5
i-f-f 9 14.3 2 1.9 28 20.9
f-f-m
f-m-f
m-f-f 2 2.1 7 6.6 2 1.5
F 2 2.1 49 77.8 12 11.4 113 84.8 0 0.0 53 87.0 8 7.8 102 77.9 f-m-i 3 2.2
f-i-m 2 3.2 6 5.9
i-f-m
f-i-i 2 2.1 7 11.1 7 6.6 7 5.3
i-f-i 1 1.0 1 0.8
i-i-f 1 1.5
i-i-i 1 1.1 1 1.0 1 0.8
i-i-m 1 1.1 2 3.2 1 0.8
i-m-i 4 4.3 1 1.0
m-i-i 16 17.3 13 12.7
i-m-f 4 3.9
m-i-f 6 5.9
m-f-i 2 2.1
0 26 28.0 12 19.0 39 38.0 13 9.9 27 29.0 8 13.0 22 21.3 23 17.2 f-m-m
m-f-m
m-m-f
i-m-m 4 4.3 8 7.8 1 0.8
m-i-m 18 19.4 2 3.2 20 19.5 5 3.7
m-m-i 10 10.7 8 7.8
m-m-m 33 35.5 16 15.5 1 0.8
M 65 69.9 2 3.2 52 50.6 7 5.3 66 71.0 0 0.0 73 70.4 4 4.6 1 f, female; i, intermediate; b, male. Sexual forms: F, female; 0, intermediate; M, male.

to occur in 70% of males from Paris and in 51% of males Employing a visual evaluation of the ischiopubic
from Coimbra. The female form was observed in 78% of proportion aspects (Table 3), it is possible to cor rectly
French females and 85% of Portuguese females. Incorrect estimate sex in 71% of French and Portuguese males.
allocations were more frequent in Portuguese males (11%) Among females, the pubis is relatively longer in 87% of
than in French males (2%). The frequency of correctly cases in the French series, but only in 78% of cases in the
diagnosed male and fe male morphology in the Czech- Coimbra sample. A high frequency of the intermediate
German series of No votny´ (1981) falls between values for morphology is observed (one third of all specimens), but
the French and Portuguese series. Furthermore, the male the margin of error (the wrong sexual attribution) is very
form was found in 3% of French and 5% of Portuguese small. In the Portuguese sample, misclassification was
women. The typical female form of this character, denoted encountered in only 8% of males and in 5% of females. In
by the combination (f-f-f), accurately determined 36– 51% the Paris sample, no false classifications occurred. The high
of the female individuals examined and was found in only reliability of this trait is attributed to the introduc tion of the
two male specimens from the Portu guese series. Similarly, category “intermediate morphology,” which reduces
the typical male combina tion (m-m-m) accurately subjectivity. This character is suit able as a preliminary
determined 15–36% of the men and was observed only in indicator of the sex and, to gether with the other characters,
one female hip bone. In the sample of Novotny´ (1981), contributes to an accurate evaluation of sexual dimorphism
forms typical for one sex did not occur in individuals of the of the hip bone overall.
other sex. This character, therefore, makes it possible to allo
cate sex with an error which does not exceed 12%. The Sacroiliac pelvic complex
presence of a typical sexual form for this trait provides a
high probability of correctly estimating the sex of bone The combination of the first three characters
fragments. But the subjectivity of the character evaluation, (preauricular surface aspects, sciatic notch aspects, and
suggested by as many as 38% of cases of the intermediate morphology of the composite arch) correctly de termines
form in certain groups, is high. sex in 91% of individuals (Table 4A). The
Ischiopubic proportions
166 J. BRUZEK
TABLE 4. Results of sexual diagnoses1
Sacroiliac complex sex assessment
Correct Undete. Errors
Series A n %n%n%
Paris, males n 98 89 91.8 4 3.1 5 5.1 Paris, females n 64 58 90.6 3 4.7 3 4.7 Paris, males and females n 162 147 90.7 7 4.3 8 5.0 Coimbra, males n
106 101 95.5 3 2.8 2 1.7 Coimbra, females n 134 117 87.3 8 6.0 9 6.7 Coimbra, males and females n 240 218 91.0 11 4.5 11 4.5 Paris and Coimbra,
males n 204 190 93.0 7 3.5 7 3.5 Paris and Coimbra, females n 198 175 88.4 11 5.6 12 6.0 Paris and Coimbra, males and females n 402 365 90.8 18
4.5 19 4.7
Ischiopubic pelvic complex sex assessment
Correct Undete. Errors
Series B n %n%n%
Paris, males n 93 84 90.8 9 9.2 0 0.0 Paris, females n 61 59 96.7 0 0.0 2 3.3 Paris, males and females n 154 143 92.8 9 5.8 2 1.4 Coimbra, males n
104 73 70.2 20 19.2 11 10.6 Coimbra, females n 129 118 91.4 6 4.7 5 3.9 Coimbra, males and females n 233 191 82.0 26 11.6 16 6.4 Paris and
Coimbra, males n 197 157 79.7 29 14.7 11 5.6 Paris and Coimbra, females n 190 177 93.2 6 3.2 7 3.6 Paris and Coimbra, males and females n 387
334 86.3 35 9.0 18 4.7
Entire hip bone sex assessment
Correct Undete. Errors
Series C n %n%n%
Paris, males n 93 90 96.8 2 2.1 1 1.1 Paris, females n 63 60 95.2 1 1.6 2 3.2 Paris, males and females n 156 150 96.2 3 1.9 3 1.9 Coimbra, males n
103 97 94.2 5 4.8 1 1.0 Coimbra, females n 133 123 92.6 5 3.7 5 3.7 Coimbra, males and females n 236 220 93.3 10 4.2 6 2.5 Paris and Coimbra,
males n 196 187 95.4 7 3.6 2 1.0 Paris and Coimbra, females n 196 183 93.3 6 3.1 7 3.6 Paris and Coimbra, males and females n 392 370 94.4 13
3.3 9 2.3
1
A, sacroiliac pelvic complex (combination of three characters). B, ischiopubic pelvis complex (combination of two characters). C, pelvic bone as a
whole (combination of the five characters of two pelvic complexes). Correct, concordance with sex of specimens. Undete., undeterminable, equality of
forms “F” and “M.” Errors, discordance of sexual diagnosis regarding known sex of specimens.

lowest success was encountered in Portuguese fe males encountered in

(87%), in spite of the fact that the success in determining French men. The largest error occurred with Portu guese
sex using all three characters is higher than using each men (11%). In the pooled sample, the total error ws lower
variable separately. On the other hand, the highest than 5%. The highest frequency of indeterminate cases was
percentage of correctly deter mined sex was achieved in in men (9%, Paris; 19%, Coimbra). The overall occurrence
Portuguese males (95%). The error of sexual diagnosis of errors for women did not exceed 5%.
oscillates between 2–7% of cases in both groups. It is
higher in females than in men, but in the whole collection Sexual diagnosis of entire hip bone
(pooled groups) achieves a value of less than 5%. The per
centage of intermediate cases varies between 3–6% in With the combination of all five traits (Table 4C), the
individual groups, and when the two groups are pooled it results become very homogeneous. When the two series
does not exceed 5%. were considered separately, a correct diagno sis was
achieved in 93–98% of all cases. The largest margin of
Ischiopubic pelvic complex error was seen in groups of females, yet the difference
The combination of the last two traits (morpholo gy of between females and males was only 2%. For both
the inferior pelvic aspects, and the ischiopubic proportion samples, the frequency of correctly diagnosed cases was
aspects) permits proper assessment of sex in 86% of all 95%. The error of determining the sex did not exceed 4% in
specimens studied (Table 4B). How ever, the success in any subgroup, and for the sample as a whole was only 2%.
individual groups varied from 70% (males, Coimbra) to The percentage of indeterminate cases did not exceed a
97% (females, Paris). No errors in sexual diagnosis were level of 5% for the separate samples, and was only 3%
when the two series were pooled.
VISUAL SEXING OF THE HUMAN PELVIS 167

Sexual assessment using fragmentary guarantees a 92% chance of the correct sex allocation,
hip bones while for the typical male morphology (m-m-m) only 85%
of the cases with this combination of scores were
The contribution of each trait towards a correct sex accurately diagnosed.
diagnosis differs considerably (Tables 2 and 3). With The presence of the typical female morphology of the
respect to simple traits such as the composite arch or inferior pelvic aspects (f-f-f) correctly determined 98% of
ischiopubic proportion, the level of success varies between all specimens with this combination of scores; the typical
60–80%. Error for these traits be comes very important, and male morphology (m-m-m) also correctly determined 98%.
reduces the accuracy of the diagnosis. The other subscores have reduced reliability for the
With the more complex traits (preauricular sur face, diagnosis of fragments.
greater sciatic notch, and inferior pelvic), the subscores
(i.e., the combination of the three condi tions) which are DISCUSSION
included in the complex trait may be of significance and
may lead to a reliable sexual diagnosis. This is the case With respect to the accuracy of sex determination in men
with the morphology of the preauricular surface, whereby and women, Meindl et al. (1985) considered that the female
the diagnosis of a true preauricular groove (f-f-f) may lead skeleton is seldomly incorrectly di agnosed. This is
to a correct diagnosis of the female sex. Some of the other apparently due to less variability in female pelvic size,
pos sible combinations which describe the variety of since women are under strong selective pressures related to
preauricular groove, such as f-f-i, f-i-f, or i-f-f, have only adaptations in locomo tion as well as reproduction.
been found in females; for this character, then, the above Reduced variability in pelvic size of women has not,
combinations may be absolute indicators of female however, been statis
specimens. The subscore (f-f-m) preauric ular groove with tically confirmed (Tague, 1989). However, Pachner (1937)
presence of a piriform tubercle is rare in males. However, demonstrated that the greater pelvis is more variable in men
three males manifest this triplet, which can be explained by than in women. Inversely, the lesser pelvis is more variable
a false praeau ricular groove diagnosis in association with a in women. Therefore, this variability has nothing to do with
very deep paraglenoidal groove with a strongly developed different levels of accuracy in determining the sex of males
piriform tubercle. Theoretically, it is possible that a strong and fe males. When one sex is more frequently identified
injury comparable to obstetric trauma could cause the than the other, it should be considered a manifesta tion of
rupture of the anterior sacroiliac ligament in a few males or an insufficiency in the method used, rather than a
nulliparous females. Furthermore, the absence of any consequence of various biological character istics of the
grooves or the presence of the paraglenoid groove (m-m-m material under study.
or m-m-i and f-m-m) does not assist in the diagnosis, for its It is difficult to admit that sexual traits of the skeleton
existence is common in both sexes. could be more clearly expressed in one sex than the other.
With respect to the form of the sciatic notch, it is only The total degree of sexual dimor phism of the hip bone is a
those subscores which resemble the unique morphology function of the interaction of the partial dimorphism of the
which actually contribute to a correct sex diagnosis. The two main regions of the pelvis. Thus, according to the
observation of the typical female morphology (f-f-f) concept of func tional integration, lower levels of sexual
 dimorphism in one morpho-functional pelvic complex (i.e., above yields an accuracy rate close to 98% using the whole
open ness vs. closure of the greater sciatic notch) can be hip bone, while the misclassification rate is 2%. Three
functionally compensated by higher levels of dimor phism percent of the sample was classi fied as indeterminate. The
in the other morpho-functional pelvic com plex (i.e., advantage of the method proposed is the potential of sexing
ischiopubic proportions). The intersegment size with high accu racy, even with fragmentary hip bones,
relationships are sex- and population-specific. This can be when estab lishing the specific male (m-m-m) or specific
seen in the reduction of the frequency of misdiagnosed female (f-f-f) combination for complex characters. The bi
specimens between subgroups, from isolated characters nary scoring method of Novotny´ (1981) for sex eval uation
through isolated complex or functional divisions (sacroiliac compares favorably with other methods which use criteria
pelvic complex and ischiopubic pelvic complex) to the hip such as “smaller than” or “larger than.” Binary scoring
bone as a whole. The inconsistency of the success of the rather than subjective assign ment allows for the systematic
method of Phenice (1969) demonstrated by Lovell (1989), accomodation of more of the inherent variation seen in
MacLaughlin and Bruce (1990), and Bruzek (1991) can be pelvic mor phology.
considered in the same manner using different population
samples, since the dimorphism of the pubis does not reflect ACKNOWLEDGMENTS
the total dimorphism of the hip bone.
I thank M. Laranjeira Rodrigues de Areia for ac cess to
the material in the Museu Antropolo´gico da Universidade
CONCLUSIONS de Coimbra (Portugal). I am grateful to F. Demoulin for
These results demonstrate that the method out lined granting me access to the collec
168 J. BRUZEK
tion of the Laboratoire d’Anthropologie de l’Universite´ Hoshi H. 1961. On the preauricular groove in the Japanese pelvis.
Paris 7 (France). Discussions with D. Ferembach and V. Okajima Folia Anat Jpn 37:259–269.
Houghton P. 1974. The relationship of the pre-auricular groove of the
Novotny´ were very helpful in the preparation of this study. ilium to pregnancy. Am J Phys Anthropol 41:81–390. Kurihara Ya,
I also thank B. Schumann, D. Walrath, P. Murail, J. Braga, Kurihara Yo, Ohashi, K, Kitazgawa A, Miyasaka M, Okamoto E, Ishikawa
and A. Schmitt for help with the preparation of the English- T. 1996. Radiologic evidence of sex dif ferences: is the patient a woman or
language version of the manuscript. I thank A.-M. Tillier, a man? Am J Radiol 167: 1037–1040.
Iscan MY, Derrick K. 1984. Determination of sex from the sacro iliac
B. Arensburg, and P. Murail for comments, and J. Cathalaa joint: a visual assessment technique. Florida Sci 47:94–98.
and M. Seurin for technical assistance. Three anonymous Iscan MY, Dunlap SS. 1983. Sexing the human sacroiliac joint. Am J Phys
reviewers provided helpful comments, and E. Trinkaus Anthropol 60:208–209.
edited the manuscript. Lazorthes G, Lhe´s A. 1939. La grande e´chancrure sciatique: e´tude de la
morphologie et de ses caracte`res sexuels. Arch Anat Histol Embryol
LITERATURE CITED (Strasb) 27:143–170.
Lo¨hr P. 1894. Ueber den Sulcus praeauricularis des Darmbeins und
Bocquet J-P, Neto MAM, Da Rocha Tavares MA, De Morais Xavier MH. a¨hnliche Furchen anderer Knochen. Anat Anz 9:521– 536.
1978. Estimation de l’aˆge au de´ce`s des squelettes d’adultes par re Lovell NC. 1989. Test of Phenice’s method for determining sex from the
´gressions multiples. Contrib Estud Antropol Port 10:7–163. os pubis. Am J Phys Anthropol 79:117–120. MacLaughlin SM, Bruce MF.
Boucher BJ. 1955. Sex differences in the foetal sciatic notch. J Forensic 1990. The accuracy of sex identifi cation in European skeletal remains
Med 2:51–54. using the Phenice charac ters. J Forensic Sci 35:1384–1392.
Bruzek J. 1991. Fiabilite´ des proce´de´s de de´termination du sexe a` Meindl RS, Lovejoy CO, Mensford RS, Don Carlos L. 1985. Accu racy
partir de l’os coxal. Implications a` l’e´tude du dimorphisme sexuel de and direction of error in the sexing of the skeleton: impli cation for
l’homme fossile. Ph.D. thesis, Institut de Pale´ontolo gie Humaine, paleodemography. Am J Phys Anthropol 68:79–85.
Muse´um National d’Histoire Naturelle, Paris. Novotny´ V. 1981. Pohlavn´ı rozd´ıly a identifikace pohlava´ p´ınevn´ı
Bruzek J. 1992. Fiabilite´ des fonctions discriminantes dans la de
kosti [Sex differences and identification of sex in pelvic bone]. Ph.D.
´termination sexuelle de l’os coxal. Critiques et propositions. Bull Mem
thesis, Purkyneˇ University, Brno.
Soc Anthropol Paris 4:67–104.
Novotny´ V. 1986. Sex determination of the pelvic bone: a system
Bruzek J, Ferembach D. 1992. Fiabilite´ de la me´thode visuelle de de
´termination du sexe a` partir du bassin du «Groupe de travail approach. Anthropologie (Brno) 24:197–206.
d’Anthropologues europe´ens». Application sur l’os coxal. Arch Novotny´ V. 1988. Parturition scars on the human female pelvis and other
Antropol Etnol 72:145–161. modifications of bones: problems with identification and interpretation.
Carre Y. 1978. A propos de la biome´trie d’une collection de man dibules In: Proc of Twelfth Intern Cong Anthrop Ethnol Sci, Zagreb.
de Franc¸ais adultes. The`se de Doctorat en Sciences Odontologiques, Olivier G. 1972. Profils sagittaux et mensuration de craˆnes de Franc¸ais.
Universite´ Paris 7, Paris. Documentation anthropologique, Laboratoire d’Anthro pologie
Coleman WH. 1969. Sex differences in the growth of the human bony biologique, Universite´ Paris 7, Paris.
pelvis. Am J Phys Anthropol 31:125–151. Pachner P. 1937. Pohlavn´ı rozd´ıly na lidske´ pa´nvi [Sex differ ences of
Cox M, Scott A. 1992. Evaluation of the obstetric significance of some ˇ
pelvic characters in an 18th century British sample of known parity human pelvis]. Prague: C eska´ Akademie Veˇd a Umeˇn´ı. Phenice TW.
status. Am J Phys Anthropol 89:431–449. 1969. A newly developed visual methods of sexing the os pubis. Am J
Fazekas GI, Kosa F. 1978. Forensic foetal osteology. Budapest: Phys Anthropol 30:297–301.
Akademiai Kiado. Poulhes MJ. 1948. La branche ischio-pubienne, ses caracte`res sexuels.
Ferembach D, Schwidetzky I, Stloukal M. 1980. Recommenda tions for Bull Mem Soc Anthropol Paris 9:191–201. Rogers T, Saunders S. 1994.
age and sex diagnoses of skeletons. J Hum Evol 9:517– 549. Accuracy of sex determination using morphological traits of the human
Francis CC. 1952. The human pelvis. London: Mosby, Kimpton. Genove´s pelvis. J Forensic Sci 39: 1047–1056.
S. 1959. L’estimation des diffe´rences sexuelles dans l’os coxal; diffe Schemmer D, White PG, Friedman L. 1995. Radiology of the paraglenoid
´rences me´triques et diffe´rences morphologiques. Bull Mem Soc sulcus. Skeletal Radiol 24:205–209.
Anthropol Paris 10:3–95. Schultz AH. 1930. The skeleton of trunk and limbs of higher primates.
Glanville EV. 1967. Sexual dimorphism in the pelvic bones of the Telem, Hum Biol 2:303–348.
a medieval negro population from the Mali Republic. Konikl Nederl Singh S, Potturi BR. 1978. Greather sciatic notch in sex determi nation. J
Akad Wetesch Amsterdam 70:368–377. Anat 125:619–624.
Hanna RE, Washburn SL. 1953. The determination of the sex of skeletons Smout CFV, Jacoby F. 1948. Gynaecological and obstetrical anat omy of
as illustrated by a study of the Eskimo pelvis. Hum Biol 25:21–27. the female pelvis. London: Edward Arnold & Co. Spring DB, Lovejoy
Holcomb SMC, Konigsberg LW. 1995. Statistical study of sexual CO, Bender GN, Duerr M. 1989. The radio graphic preauricular groove:
dimorphism in the human fetal sciatic notch. Am J Phys An thropol
its non-relationship to past parity. Am J Phys Anthropol 79:247–252.
97:113–125.
Sutherland LD, Suchey JM. 1991. Use of the ventral arch in pubic sex
determination. J Forensic Sci 36:501–511.
Tague RG. 1989. Variation in pelvic size between males and females. Am
J Phys Anthropol 80:59–71.
Ullrich H. 1975. Estimation of fertility by means of pregnancy and child
birth alterations at the pubis, ilium, and the sacrum. Ossa 2:23–29.
Waldron T. 1987. The relative survival of the human skeleton: implication
for palaeopathology. In: Boddindton A, Garland AN, Janaway RC,
editors. Death, decay and reconstruction. Manchester: Manchester
University Press. p 55–64.
Washburn SL. 1948. Sex differences in the pubic bone. Am J Phys
Anthropol 6:199–207.
Weisl H. 1954. The ligaments of the sacro-iliac joint examined with
particuliar reference to their function. Acta Anat (Basel) 20:201–213.
Zaajer T. 1866. Untersuchungen u¨ber die Form des Beckens javanischen
Frauen. Naturrk Verhandel Holland Maatsch Ven tensch Haarlem 24:1–
42.