AMERICAN JOURNAL. OF PHYSICAL ANTHROPOLOGY 89:431440 (1992)
Evaluation of the Obstetric Significance of Some Pelvic Characters in an 18th Century British Sample of Known Parity Status MARGARET COX AND ANDREW SCOTT Department of Palaeontology, The Natural History Museum, London SW7 5BD, England (M.C.)and Department of Applied Statistics, University of Reading, Whiteknights, Reading, RG6 ZAN, England ( A S . )
KEY WORDS Parity status
Preauricular sulcus, Pubic pitting, Pubic tubercle,
ABSTRACT The excavations at Christ Church, Spitalfields (1984-86) produced a sample of 968 human skeletons which were interred between 1729 and 1859. Of these, 387 were recovered in association with coffin plates stating name, age at death, and date of death. There are 138 adult females in the named sample and the obstetric histories of 94 have been reconstructed from historical documentation. Such variables as birth spacing, number of children, and age at first and last births are known for the majority of this sample. Any individual about whose history there is any doubt has been excluded from the analysis. A middle-class group, they were largely of high nutritional status and, by the standards of the day, lived in sanitary and comfortable conditions. Both males and females have a mean age a t death of 56 years. The presence or absence, the typology, the severity, the width and the length of the preauricular sulcus, the presence or absence and the number of pits on the dorsal aspect of the pubic body, sulci along the anterior sacrum adjacent t o the auricular facet, and the extension of the pubic tubercle were evaluated in relation to the obstetric histories of these females. Statistical analysis has demonstrated a relationship between the presence of pubic tubercle extension and parity status and between the degree of extension and the number of children borne ( P < .02). Statistically there is no significant relationship ( P > .05) between either the preauricular sulcus or pubic pitting and parity status. Sacral scarring is significantly associated ( P < .05) with parity status, but as it was evident in only eight females it has little practical application. Sample sizes are small, and it must be considered that statistical evaluation of larger samples might detect associations between variables not demonstrated here. There is no significant relationship between any of the cortical variants under consideration and age at death. o 1992 Wiley-Liss, Inc. The distribution of cortical resorption and remodelling adjacent to the ventral preauricular margins of the ilium and sacrum and on the dorsal aspect of the pubic body is strongly sexually dimorphic. The deep and “scooped” preauricular sulcus is found almost exclusively in (Derry, Dunlap, 1981; Spring et al., 1989) while pU0 1992 WILEY-LISS, INC.
bic pits are occasionally seen in males (Suchey et al., 1979; Tague, 1988). Following the work of Angel (1969).Stewart (1970), -
~~
Received January 9,1990, accepted May 26, 1992 Address communlcatlon to Dr Margaret Cox, Grey Gables, Summerfield Road, Beacon H111, Bath, Avon, BA15QB, England
432
M. COX AND A. SCOTT
Houghton (1974, 19751, and Ullrich (1975) area of research and includes in its analysis on samples of unknown parity status, this such factors a s the number of births, age a t dimorphism has resulted in such features first and last birth, and birth spacing. becoming known collectively a s “scars of MATERIALS AND METHOD parturition” (i.e., Tague, 1988:252). Evaluation of the obstetric significance of Excavation of the vaults beneath Christ these changes has been hampered by the Church, Spitalfields, London between 1984 lack of reliably documented skeletal sam- and 1986 yielded 968 discrete human skeleples. The Hamann-Todd collection, for ex- tons. Interred between 1729 and 1859, 387 ample, where parity status was largely de- of the skeletons were excavated with associtermined by soft tissue analysis, has been ated coffin plates from secure contexts. subject to a number of studies (Andersen, These are referred to in this text a s “t,he 1988; Holt, 1978; Kelley, 1979). Of the stud- named sample.” The coffin plates are inies based on more reliably documented sam- scribed with name, age a t death, and date of ples with information known on the number death. Occasionally other information was --A ----:---L--L:lA--- S..-L--. - A - 1 i l O c i O \ _i1IcIuueu _-1.. A J bUCl1-l as Lllt:- Y U I I o f . . . ” or ‘‘Lord allu apaLlllg U l L l l l l U l C l l ) UL11C:y Gl. a1. ( l d I d ) found a weak correlation between full term Mayor of London.” The basic biographical pregnancies and pubic pitting in a n autopsy data have facilitated research into many assample of known parity status. Bergfelder pects of each individual’s life history and life and Herrmann (1980), using a dissecting style. room sample with known obstetric histories, The named sample interred beneath found that there was no unequivocal Christ Church were largely of French derelationship between pubic pitting and scent and mainly engaged in the silk induspregnancy and parturition. Bergfelder and try. Their presence in Spitalfields was the Herrmann (1980) also evaluated the rela- result of religious intolerance in France and tionship between the degree of extension of the fact that Spitalfields was the contempothe pubic tubercle in relation to obstetric rary silk weaving centre of London. They events. They found that there was no strong were middle class, well fed, and rarely emassociation with parity status but that a dis- ployed in manual labour. Among the named tinctly prominent tubercle was usually asso- sample there are 138 female skeletons, aged ciated with more than three births. 12 (the age of the youngest mother in the Dunlap (1981), analysing a “willed body” sample) and above. sample, all of postmenopausal age, conThe skeletal material varied in condition cluded that accurate prediction of obstetric from excellent t,o very poor. In some cases events was not possible but that Houghton’s data were available only from one 0s coxae “groove of pregnancy” (1974), a deep and and in a small number of cases no data at all rugged preauricular sulcus, was associated could be collected. The variation in sample with parturition. Conversely, using a sam- size evident in the results reflects the state ple studied radiographically with known ob- of preservation of the material. stetric histories, Spring et al. (1989) found The obstetric histories of 94 of the 138 no relationship between the deep preauricu- females were reconstructed from historical lar sulcus and pregnancy and parturition. records. An example of the method emThe results of assessments of the relation- ployed is illustrated in Appendix A. For full ship between so called “parturition scars” details of the methodology employed in this and parity status in different samples of reconstruction see Cox, 1989. Any obstetric known parity status are to date conflicting. histories which were slightly dubious were Interpopulation variability could be the ex- excluded from the sample. planation for this a s could the accuracy of The reconstruction of the obstetric histoeach sample’s obstetric data. ries utilised data derived from death certifiThe named sample from Christ Church, cates, family genealogies, and wills but rests Spitalfields (for further information see largely upon the baptism of infants. The Molleson and Cox, in press) provides an- 18th and early 19th century Huguenots who other opportunity to evaluate this important settled in Spitalfields were religious refuI__^
- I
1CLL
I . .
PARTURITION SCARS IN AN 18TH CENTURY SAMPLE
gees or their descendants and it seems unlikely that they would ignore baptism, a n important religious rite. Furthermore, the sobriety and moral fastidiousness of these people was such that Hogarth satirized them in “Noon” (1738) when he compared them with their less pious English contemporaries. As most stillbirths were not baptised, they cannot be detected in the historical record and they represent a n area of potential error in the data. Of the 94 females of known parity status, 73 (77.6%) had borne children; the remaining 21 (22.4%) appear to have been nulliparous. Of the nulliparous group, approximately half died aged below the normal age of marriage (twenty-six years) and may well have borne children had they lived to a greater age. Of the remainder, half were married but childless, the rest being nulliparous and unmarried. Family size ranged from 0 to 15 baptised infants; the mean was 2.7 infants (SD 2.82, SE mean 0.29). The age a t first birth ranged from 12 to 45 years; the mean was 27.1 years (SD 7.1, SE mean 0.9). The age at last birth ranged from 14 to 47 years, the mean being 35.8 years (SD 7.2, SE mean 0.9). Birth spacing ranged from 10 to 141 months; the mean was 29.4 months (SD 24.2, SE mean 3.3). Three females (4.3%) are known to have experienced a twin birth and five (7.1%) died within a few days of childbirth, probably from postpartum sepsis, Of the latter group, four had experienced previous confinements and all were aged 29 and below. The observations on the 138 adult female skeletons were made “blind to the obstetric data, the skeletal analysis being undertaken prior to the historical research. The presence or absence of a preauricular sulcus was observed, as was the type of sulcus. Four categories were used; the first and second were a s described by Houghton (1974). The first type, Houghton’s “groove of pregnancy”, appears to have been formed by the coalescence of a series of pits. It has a “scooped floor and a n uneven inferior margin (Fig. l a ) . The second, Houghton’s “groove of ligament”, has a smooth floor and a smooth inferior margin (Fig. lb). Fitting neither of Houghton’s categories, the third sulcus type is one observed by Dunlap (1981)
433
Fig. 1. a: A deep preauricular sulcus with a “scooped”appearance to the floor (Houghton’s“groove of pregnancy”). Inhumation number CAS 2438, a 56-year old of unknown parity status. b: A “groove of ligament” as defined by Houghton (1974).This preauricular sulcus is shallower than that illustrated in a,and has a smooth floor and smooth inferior margin. Inhumation number CAS 2438, a 56-year old of unknown parity status.
and by the first author when examining the females from the Poundbury Camp sample (Dorchester, England). This sulcus is comparatively wide, with an indistinct inferior margin and a “grainy” floor (Fig. 2a). The fourth sulcus was first described by Derry in 1909 (Fig. 2b). Usually seen in males, it is only occasionally found in females. Rather than being a true sulcus, that is a groove formed by cortical resorption, a n accentuated pyriformis tubercle with associated cortical deposition is creating the appearance of a sulcus. The severity of the preauricular sulcus was observed. Scored from 0 to 4, it is a subjective measure of size (not a series of precise mathematical increments) which considers depth a s well a s width and length.
434
M. COX AND A. SCOTT
Fig. 2. a: A preauricular sulcus type fitting neither of Houghton’s (1974) definitions. This sulcus is shallow, has a grainy floor, and an ill-defined inferior margin. Inhumation number CAS 2956, a 45-year old mother of three. h: The effect of a sulcus is created by the pyriformis tubercle and associated cortical deposition. Inhumation number CAS 2602, a 51-year old male.
The maximum width and maximum length of the sulcus were also measured. The presence or absence of sulci along the anterior margins of the sacrum adjacent to the auricular facet (Ullrich, 1975) was scored as present or absent. Pitting on the dorsal aspect of the pubic body was scored (Fig. 3a) both in terms of its presence or absence and in the number of discrete pits. The extension of the pubic tubercle (Fig. 3b) was scored from 0 to 3 in terms of severity (after Bergfelder and Herrmann, 1980). Examination of the cortical variants described above in association with measurements of pelvic capaciousness was also undertaken as a part of this study (Cox, 1989). The results of statistical analysis show that the presence of a preauricular sulcus (irrele-
Fig. 3. a: Pitting (cortical resorption) on the dorsal aspect of the pubic body. Inhumation number 2070, a 35-year old mother of five. b: An example of a greatly extended (grade 3) pubic tubercle (dorsal view). Inhumation number CAS 2566, a 74-year old mother of one.
vant of type) and the size of the sulcus are significantly. positively related t o each o f the following pelvic dimensions: sciatic notch width (Day, 19751, transverse diameter (Segeberth-Orban, 1980), bispinous diameter (Cox, 19891, inlet circumference (Cox, 1989), pubo-sacro-iliac diameter (Segeberth-Orban, 19801, and sacral width at the level of the caudal limit of the auricular facets. The presence of pubic pitting is significantly associated with bispinous diameter. (For full details see Cox, 1989). STATISTICAL ANALYSIS The type of test of significance used to evaluate the relationships between different variables was largely determined by the nature of the data in question. The distribution of continuous variables was examined using ranks to plot normal
PARTURITION SCARS IN AN 18TH CENTURY SAMPLE
435
scores and rootgrams. One way ANOVA was of those with extension were nulliparous. used to estimate analysis of variance where The data presented in Table 3b illustrates the data were normally distributed. Mann the distribution of the grades of tubercle exWhitney rank sum test and Kruskal Wallis tension in relation to parity status. A higher statistic were used where the data were not percentage of parous than nulliparous fenormally distributed as, for example, where males have grades one and two and no the analysis included number of births, age nulliparas have grade three extension. Table 3c illustrates that the degree of extenat death, or age at first or last birth. Chi square tests were used to examine the sion of the tubercle is significantly associdistribution of either two discrete variables ated with the number of births (P < .05), or one discrete and one continuous variable. those with advanced extension having a meThe expected cell frequencies in some tabu- dian of five births each. lations were occasionally very small (less There is a significant association between than one). In order to ensure valid results sacral scarring and parity status (Table 4), a the significance levels of all chi square tests higher proportion of nulliparous than parhave been evaluated using exact conditional ous females exhibiting this feature. However, a s scarring was present in only 8 of 63 tests (Kreiner, 1987). Pearson's r correlation coefficient was (12.7%) females this association has little used to determine the extent to which two practical value. Table 5 describes the results of analysis of continuous variables changed in relation to one another. the cortical variants under consideration As is traditional in such studies, the re- and age a t death. In no case was the associasults reported are those from the left 0s tion statistically significant ( P > .05). coxae. This alleviates problems caused by incomplete preservation and occasional DISCUSSION asymmetry in the distribution and type of It is and always has been a fact of life that preauricular sulcus and pubic pits (see also the birth of a child in certain circumstances Houghton, 1974; Tague, 1988). Figures l a is undesirable. This can lead to denial and and l b , illustrating different preauricular concealment and could affect the validity of sulcus types, are from the same individual. data retrieved from documentation. A further potential error lies in the fact that most stillbirths were not baptised. This results in RESULTS their absence from the historical record. Appropriate statistical examination of In order to take account of these two facparity status in relation to either the abtors, each cortical variant was examined sence or presence, severity, type or size of only in those females known to have borne the preauricular sulcus (Table 1)or the abchildren. In all cases, as can be seen by refersence or presence, or number of pubic pits ence to Tables 1, 2, 3, and 4,no feature oc(Table 2) indicates that none of these varicurs consistently among the parous females. ants is significantly associated with parity The presence of a n extended pubic tubercle status in this sample. In all cases P > .05. is the most consistent feature evident Preauricular sulcus types one to three are among the parous females in this sample. found in both parous and nulliparous females and there were only two cases of type four, both seen in parous females. Interestingly, all of the females with more than one pubic pit were parous but many parous females had no pitting or only one pit. Table 3a illustrates that the presence or absence of pubic tubercle extension is significantly associated with parity status ( P < .01). Of those without extension 66.66% were nulliparous, while only 12.5%
Scars of parturition and maternal mortality
Houghton (1975:658) states that following the delivery of a child ". . . processes of reorganisation and repair gradually return the ligaments to their normal state within a few weeks. Thereafter, a slow and variable replacement of bone appears to occur over years a t the sites of previous osteoclastic ac-
M. COX AND A. SCOTT
436
TABLE 1. Parity status and the preauricular su1cu.s l a . Presence or absence and parity status
Present
Nulliparous Parous
4 6
22.2 9.5
N
%'
14 57
77.8 90.5
x2, 2.09; D.F., 1;P , 0.223; N, 81. l b . Sulcus type and parity status
Type 1
Absent
2_ _ _Type ___ N _ _9% _
~~
~~
N
Q
4
22.2 9.6
N
5%
7 29
38 9 46 7
~
Nulliparous Parous y2, 2.9: D.F.. 4:
6 P. 0.57; N. 80.
lc. Sulcus severity and Daritv status
Nulliparous Parous
%
4
22.2 9.5
6
8 35
~ _ N_
22 2 17 7
4 11
_ Grade ~ 1_ N Q
Absent N
Type 3
_ _ _ _ _ _ _ _ ~
3 14
44.4 55.6
Q
N
%
16 7 22 G
0 2
32
Grade 2
_
Grade 3 & 4l
N
%
N
3 15
16.7 23.8
3 7
~-
Type 4 ______
3 _____
16.7 11.1
x2, 2.826; DF, 3; P , 0.46; N, 81. Id. Preauricular sulcus length and parity status
N
Mean
SD
SE Mean
13 57
23.54 27.05
7.74 7.43
2.1 0.98
l e . Preauricular sulcus width and nnrit,v status
N
Mean
SD
SE Mean
Nulliparous Parous
14 57
6.07 7.16
1.98 2.64
0.53 0.35
Nulliparous Parous T, -1.49; P , 0.16, DF, 17; N, 70.
T, -1.71; P , 0.099; DF, 25; N, 71. Grades 3 and 4 have been combined for the purposes of this test; grade 4 had nu values in the nulliparous cell and only one in the parous cell
TABLE 2. Pubic pitting and parity status
_Prewnl _ _ ~ -
2a Presence or absenw of pubic pitting and parity status
Absent
N
70
N
%
Nulliparous Parous
6 24
66.67 60.00
3 16
33.33 40.00
~
x2, 0.138; DF, 1;P , 1.00; N, 49. 2b. Number of pits and parity status
Absent _____ N
%
N
One
%$Nulliparous Parous
6 24
66.67 60.00
3 10
33.33 25.00
~~
~ Two _ N Q 0 3
_ Three N
Four %
N
?r
5.00
0 1
2.50
0
7.50
2
xa, 1.6; DF, 4; P , 0.88; N, 49.
tivity." If he is correct in this assumption, it seems that the sites of attachment of the inferior portion of the ventral sacroiliac ligament and the dorsal transverse ligament of the pubis should be unusual in females who die shortly after childbirth.
Five of the females among the Christ Church sample are known t o have died within days of parturition. None of these had any skeletal abnormality, such as a rachitic pelvis, a flat sacrum (Posner et al., 1955), nor any sign of pclvic disproportion
PARTURITION SCARS IN AN 18TH CENTURY SAMPLE
437
TABLE 3. Pubic tubercle extension and naritv status 3a Presence or absence of tubercle extension _and _ _parity __ _ _ _ -_ _ ~
____
N
- - _ -- - - -
Nulliparous Parous x‘, 9 682, DF, 1,P, 0 07, N, 33 3b Degree of tubercle extension and parity status
_ Absent ~ _ _~- __ _
_
76
__
--
6 3
66 7 12 5
2 _Grade - -- - .
_ _Absent ___-
_ _Grade _ _ -1_
-.
N
%
N
92
N
[k
N -__ __ _ ~ ~ _
%,
66 66 12 50
2 8
22 22 33 33
1 8
11.11 33.33
0 5
00.00 20.83
~
~~~~
Nulliparous 6 Parous 3 x’, 10 37, DF, 3, P, 0 02, N, 33 3c Degree of tubercle extension and number of births
(Kruskal Wallis) Grade
N
Grade 3 _ -
Median ~~
~_ _ _ _ _Avg - - _Rank ~
9 10 9 5 33
86 18 0 21 0 23 1 17 0
0 3 4 5
H, 10.49; ADJ H, 10.84; P, 0.013.
which would have prevented vaginal delivery. Four of these females had borne children previously. The skeletons of four of this group have survived well enough for their preauricular areas to be examined. All four had a preauricular sulcus, two were type 1 (Fig. l a ) and two were type 2 (Fig. lb). Three were slight and one moderately severe. The pubis survived in only one female; this had one pit on the dorsal aspect of each pubic bodj7 and slightly extended pubic tubercles. Xone had sacral scarring. This evidence suggests that the theory that “scars of parturition” represent cortical resorption which follows trauma incurred, by the mother, during pregnancy and parturition is unlikely to be correct. The macroscopic morphological changes evident on the pelvic bones of these four females were not in any way different from that seen on other females, some who died several decades after giving birth. Age at death and scars of parturition
The work of Suchey et al. (1979) and T a p e (1988) found that pubic pitting was significantly associated with the age a t death of the female. With one exception, the
TABLE 4. Sacral scarring and naritv status Absent
Present N %
~
N Nulliparous 8 Parous 47 x‘, 5.694; DF, 1;P, 0.04; N, 63.
% 66.67 92.16
4 4
33.33 7.84
age of the Christ Church females was known from their coffin plates. In many caws thi. could be checked against their baptism records and proved to be correct in all cases except one (where it was corrected for analytical purposes). The mean age a t death of this sample was 56.7 years (STD = 18.9); 75% were post-menopausal a t death (the oldest mothers were 47 years old at their last confinements). Appropriate statistical tests (Table 5) were used to examine the possibility of a n association between age at death and the presence of a pre-auricular sulcus, sacral scarring, pubic pitting, and a n extended pubic tubercle. Only for pubic tubercle extension ( P = 0.09) was the result even close to significance for this sample. Nevertheless, it was observed that in the young adult females whose iliac crests were unfused, there
M. COX AN1I A SCOTT
438
TABLE 5. A g e a t death and cortical remodelling of the aelvss
5a. Preauricular sulci _ and _ age _ _at ~death ___ ~~~
(Mann Whitney) W, 724.0; P = 0.56; N
=
112
5b.~Preauricular sulcus severity and age a t death _ _ _ _ _ _ _ _ ___ _ _ _ _ _ _ _ _ _ _ ~ ~ ~
~
~
(Kruskal Wallis) H, 3.346; A D J H, 3.348; P
=
0.50; N
=
115.
5c. Preauricular width and length and age a t death ~
(Correlation coefficient) Width 101 Length 98
0.001 0.117
:>.05 >.05
5d. Pubic pitting and age a t death (Mann Whitney) W, 1362.5; P = 0.65; N = 69. at _ death 5e. Number of pubic pits and age _ _~ (Kruskal Wallis) H, 1.560; A D J H, 1.561; P
=
0.67; N
=
69.
5f. Sacral scarring and age a ___ t death _____(Mann Whitney) W, 3255.0; P = 0.67; N
=
CONCLUSION
88.
5g. Pubic tubercle extension and _ age_a t death _ __ (Mann Whitney) W, 263.0; P = 0.09; N
=
~
47.
5h. Degree of tubercle extension and age at death_ _ _
~~~
~~
(Kruskal Wallis) H, 2.919; A D J H, 2.922; P
=
0.40; N
between the two sets of results lends credibility to the obstetric histories of the Christ Church sample and to the obstetric significance of an extended pubic tubercle. It must be considered though that 33.3% of the nulliparous females also had an extended tubercle. In view of this, the strong statistical association between the presence of an extended tubercle and parous females cannot be considered as conclusive proof of a causal relationship. Pregnancy may not be the only cause of tubercle extension, as the presence of an extended tubercle in 90.3%of the Christ Church males (n = 31) attests. It must be stressed though that the morphology of the extended tubercle is different in males and females, the male tubercle being far more robust and less fragile than the female tubercle.
=
47.
was never more than the slightest trace of either a preauricular sulcus or pubic pitting. The pubic tubercle and pregnancy and parturition
The association between the extended tubercle with parity status (Table 3) and the number of births could be explained physiologically. A portion of abdominus rectus attaches to the pubic tubercle. It is possible that the extended tubercle might represent bone remodelling in response to the stress imposed on the muscle insertion during the later stages of pregnancy, when this muscle is greatly extended. The association between the pubic tubercle and parity status and the number of births in the Christ Church sample concurs with that obtained by Bergfelder and Herrmann (1980). Their sample was from a dissecting room and obstetric histories were obtained from medical records and from relatives of the deceased. The concurrence
The results obtained by statistical analysis of the preauricular sulcus, pubic pitting, and sacral scarring in relation to the parity status of the Christ Church, Spitalfields sample (n = 94) indicates that none of these cortical variants is consistently associated with parity status. In this sample, the small number of females with more than one pubic pit had all borne children, but the lack of pits, or only one pit, is not associated with nulliparity. Unfortunately, the number of nulliparous females in this sample is small. It should be considered that statistical analysis of larger samples might detect associations between variables not apparent here. A significant association is, however, demonstrated between the presence of an extended pubic tubercle and parity status. Eighty-seven percent of the parous females had an extended pubic tubercle and the greater the degree of extension, the greater the number of births. However, 33.3%of the nulliparous females also had an extended tubercle and it must be considered that pregnancy may not be the only causal factor. ACKNOWLEDGMENTS
I thank Theya Molleson of The Natural History Museum, London and Don Brothwell, Simon Hilson, and Clive Orton of the Institute of Archaeology, University College
PARTURITION SCARS IN AN 18TH CENTURY SAMPLE
London for their help and support with this project. Phillip Crabb of The Natural History Museum took the excellent photographs. The Christ Church Spitalfields project was funded by English Heritage and The Nuffield Foundation.
439
Suchey JM, Wisely DV, Green RF, and Noguchi TT (1979)Analysis of dorsal pitting in the 0s pubis in an extensive sample of modern American females. Am. J. Phys. Anthropol. 51t517-540. Tague RG (1988) Bone Resorption of the Pubis and Preauricular Area in Humans and Nonhuman Mammals. Am. J. Phys. Anthropol. 76:251-267. Ullrich H (1975) Estimation of fertility by means of pregnancy and childbirth alterations a t the pubis, the ilium. and the sacrum. OSSA 2.23-39.
LITERATURE CITED Andersen BC (1988) Pelvic scarring analysis: Parturition or excess motion. Am. J. Phys. Anthropol. 75t181. Angel J L (1969)The bases of palaeodemography. Am. J. Phys. Anthropol. 30:427437. Bergfelder T, and Herrmann B (1980)Estimating fertility on the basis of birth traumatic changes in the pubic bone. J. Hum. Evol. 93511-613. Cox MJ (1989) PhD Thesis. Institute of Archaeology, University College London. An Evaluation of the Significance of “Scars of Parturition” in the Christ Church Spitalfields Sample. Day MH (1975) Sexual differentiation in the innominate bone studied by multivariate analysis. Ann. Hum. Biol. 2t143-151. Derry DE (1909)Note on the innominate bone as a factor in the determination of sex. J. Anat. Physiol. 43:266276. Dunlap SS (1981) PhD Thesis, Michigan State University. A study of preauricular sulci in a cadaver population. Holt CA (1978)A re-examination of the parturition scars on the female pelvis. Am. J. Phys. Anthropol. 4 9 9 94. Houghton P (1974) The relationship of the preauricular groove of the ilium to pregnancy. Am. J. Phys. Anthropol. 4It381-390. Houghton P(1975) The bony imprint of pregnancy. Bull. W A c a d . Med. 51t655-661. Kelley MA (1979)Parturition and pelvic changes. Am. J. Phys. Anthropol. 51t541-546. Kreiner S 1,1987)Anaiysis of multidiniensional contingency tables by exact conditional tests: Techniques and strategies. Scand. J . Statist. 14r97-112. Molleson TI, and Cox MJ (In Press) Spitalfields: The Middling Sort. London: Council for British Archaeology. Posner A, Bloch NR, and Posner NS (1955) The flat sacrum: Its importance in obstetrics. Am. J. Obstet. Gynecol. 70: 1021-1025. Segeberth-Orban R (1980) An evaluation of the sexual dimorphism of the human innominate bone. J. Hum. Evol. 9:GOl-607. Spring DB, Lovejoy CO, Bender GN, and Duerr M (1989) The Radiographic Preauricular Groove: Its Non-Relationship to Past Parity. Am. J. Phys. Anthropol. 79t247-252. Stewart TD (1970) Identification ofthe scars of parturition in the skeletal remains of females. In TD Stewart (ed.):Personal identification in mass disasters. Washington D.C.: National Museum ofNatural History, pp. 127-135.
APPENDIX A
The processes involved in reconstructing the females’ obstetric histories were complex and involved working from, and building upon, the initial data base, the coffin plates. These were inscribed with the individual’s name, age at death, and date of death. Starting from this primary data, information was collected and collated piece by piece until, where possible, a n understanding of each individual’s obstetric history had been achieved. The first procedure was to establish if the females were married, when, and to whom. That a female was titled “Mrs.” on her coffin plate did not necessarily signify that she was married. The terms “Mrs.” and “Miss” are both contracted from the term “Mistress” and it is not clear historically when the use of the terms to distinguish married from unmarried women crystallised. In order to illustrate the complexity of the inquiry one individual case is described below. Mrs. Grace Wells (CAS 2666) died aged 62 on June 8th, 1811. Reference to the appropriate Christ Church Spitalfields Burial Register (Greater London Reference Library) was made to verify the details and to collect additional information contained therein. Grace was buried on June 16th, 1811 and her “abode at death” was Dorset Street, Spitalfields. Grace’s year of birth would have been 1749, therefore her fecund period would have extended from approximately 17601800. Any marriage she entered into after this date would be irrelevant to this inquiry a s there would have been no issue. Reference to the International Genealogical Index for London (Guildhall Library) within the period of concern revealed only one marriage in which a male with the surname
M. COX AND A. SCOTT
440
T A B L E 6. Baptism data relating to the children o f Grace and Christopher Wells Child’s name
Date of birth
Date of baotism
Father’s occuo~ ition
19/06/1768 13/08/1769 06/09/1772 09/04/1775 20/07/1777 06/09/1778 3010511781 19/05/1782 07/05/1784 16/09/1785 16/07/1788 11/07/1790
Painter Painter Painter Painter Painter Painter Painter Painter Painter Painter Painter Painter
~~
Christopher Wi11ia m Christopher George Ann Robert Grace Ann Charlotte Elizabeth John Mary Richard
~
25/07/1769 14/08/1772 05/03/1775 24/06/1777 09/08/1778 02/05/1781 26/04/1782 07/04/1784 28/08/1785 23/06/1788 01/07/1790
Wells married a female with the Christian name Grace. Christopher Wells of the parish of Christ Church, Spitalfields married Grace Duxbury on May 19th, 1767 a t St. Leonard’s Church, Shoreditch. (Grace would have been aged eighteen years old a t the time of her marriage.) The place of marriage suggests that Shoreditch had been the bride’s place of residence. That the groom’s parish of residence was Christ Church indicated that this could be the correct marriage. The International Geneaological Index also lists the baptisms at Christ Church Spitalfields of several children born to this couple. The next step was t o refer to the original
Parents’ address ~~~
Paler Noster Row Paler Noster Row Paler Noster Row Brown’s Lane Brown’s Lane Brown’s Lane Brown’s Lane Brawn’s Lane Brawn’s Lane Brawn’s Lane Brown’s Lane Darset Street
Baptism Register in order to see if it contained any information which would confirm that the mother was CAS 2666. The data derived from the registers relating to the children of Grace and Christopher is listed in Table 6. The address given at the baptism of Richard, Grace’s last baptised child, was the same as that on the registration of Grace’s death. Reference to the Land Tax Returns (Guildhall library) for the appropriate years indicated that there was no other householder with the same surname living in Dorset Street. With this number of positive indicators it was considered that the obstetric data did relate to the female in question.
Evaluation of the Obstetric Significance of Some Pelvic Characters in an 18th Century British Sample of Known Parity Status MARGARET COX AND ANDREW SCOTT Department of Palaeontology, The Natural History Museum, London SW7 5BD, England (M.C.)and Department of Applied Statistics, University of Reading, Whiteknights, Reading, RG6 ZAN, England ( A S . )
KEY WORDS Parity status
Preauricular sulcus, Pubic pitting, Pubic tubercle,
ABSTRACT The excavations at Christ Church, Spitalfields (1984-86) produced a sample of 968 human skeletons which were interred between 1729 and 1859. Of these, 387 were recovered in association with coffin plates stating name, age at death, and date of death. There are 138 adult females in the named sample and the obstetric histories of 94 have been reconstructed from historical documentation. Such variables as birth spacing, number of children, and age at first and last births are known for the majority of this sample. Any individual about whose history there is any doubt has been excluded from the analysis. A middle-class group, they were largely of high nutritional status and, by the standards of the day, lived in sanitary and comfortable conditions. Both males and females have a mean age a t death of 56 years. The presence or absence, the typology, the severity, the width and the length of the preauricular sulcus, the presence or absence and the number of pits on the dorsal aspect of the pubic body, sulci along the anterior sacrum adjacent t o the auricular facet, and the extension of the pubic tubercle were evaluated in relation to the obstetric histories of these females. Statistical analysis has demonstrated a relationship between the presence of pubic tubercle extension and parity status and between the degree of extension and the number of children borne ( P < .02). Statistically there is no significant relationship ( P > .05) between either the preauricular sulcus or pubic pitting and parity status. Sacral scarring is significantly associated ( P < .05) with parity status, but as it was evident in only eight females it has little practical application. Sample sizes are small, and it must be considered that statistical evaluation of larger samples might detect associations between variables not demonstrated here. There is no significant relationship between any of the cortical variants under consideration and age at death. o 1992 Wiley-Liss, Inc. The distribution of cortical resorption and remodelling adjacent to the ventral preauricular margins of the ilium and sacrum and on the dorsal aspect of the pubic body is strongly sexually dimorphic. The deep and “scooped” preauricular sulcus is found almost exclusively in (Derry, Dunlap, 1981; Spring et al., 1989) while pU0 1992 WILEY-LISS, INC.
bic pits are occasionally seen in males (Suchey et al., 1979; Tague, 1988). Following the work of Angel (1969).Stewart (1970), -
~~
Received January 9,1990, accepted May 26, 1992 Address communlcatlon to Dr Margaret Cox, Grey Gables, Summerfield Road, Beacon H111, Bath, Avon, BA15QB, England
432
M. COX AND A. SCOTT
Houghton (1974, 19751, and Ullrich (1975) area of research and includes in its analysis on samples of unknown parity status, this such factors a s the number of births, age a t dimorphism has resulted in such features first and last birth, and birth spacing. becoming known collectively a s “scars of MATERIALS AND METHOD parturition” (i.e., Tague, 1988:252). Evaluation of the obstetric significance of Excavation of the vaults beneath Christ these changes has been hampered by the Church, Spitalfields, London between 1984 lack of reliably documented skeletal sam- and 1986 yielded 968 discrete human skeleples. The Hamann-Todd collection, for ex- tons. Interred between 1729 and 1859, 387 ample, where parity status was largely de- of the skeletons were excavated with associtermined by soft tissue analysis, has been ated coffin plates from secure contexts. subject to a number of studies (Andersen, These are referred to in this text a s “t,he 1988; Holt, 1978; Kelley, 1979). Of the stud- named sample.” The coffin plates are inies based on more reliably documented sam- scribed with name, age a t death, and date of ples with information known on the number death. Occasionally other information was --A ----:---L--L:lA--- S..-L--. - A - 1 i l O c i O \ _i1IcIuueu _-1.. A J bUCl1-l as Lllt:- Y U I I o f . . . ” or ‘‘Lord allu apaLlllg U l L l l l l U l C l l ) UL11C:y Gl. a1. ( l d I d ) found a weak correlation between full term Mayor of London.” The basic biographical pregnancies and pubic pitting in a n autopsy data have facilitated research into many assample of known parity status. Bergfelder pects of each individual’s life history and life and Herrmann (1980), using a dissecting style. room sample with known obstetric histories, The named sample interred beneath found that there was no unequivocal Christ Church were largely of French derelationship between pubic pitting and scent and mainly engaged in the silk induspregnancy and parturition. Bergfelder and try. Their presence in Spitalfields was the Herrmann (1980) also evaluated the rela- result of religious intolerance in France and tionship between the degree of extension of the fact that Spitalfields was the contempothe pubic tubercle in relation to obstetric rary silk weaving centre of London. They events. They found that there was no strong were middle class, well fed, and rarely emassociation with parity status but that a dis- ployed in manual labour. Among the named tinctly prominent tubercle was usually asso- sample there are 138 female skeletons, aged ciated with more than three births. 12 (the age of the youngest mother in the Dunlap (1981), analysing a “willed body” sample) and above. sample, all of postmenopausal age, conThe skeletal material varied in condition cluded that accurate prediction of obstetric from excellent t,o very poor. In some cases events was not possible but that Houghton’s data were available only from one 0s coxae “groove of pregnancy” (1974), a deep and and in a small number of cases no data at all rugged preauricular sulcus, was associated could be collected. The variation in sample with parturition. Conversely, using a sam- size evident in the results reflects the state ple studied radiographically with known ob- of preservation of the material. stetric histories, Spring et al. (1989) found The obstetric histories of 94 of the 138 no relationship between the deep preauricu- females were reconstructed from historical lar sulcus and pregnancy and parturition. records. An example of the method emThe results of assessments of the relation- ployed is illustrated in Appendix A. For full ship between so called “parturition scars” details of the methodology employed in this and parity status in different samples of reconstruction see Cox, 1989. Any obstetric known parity status are to date conflicting. histories which were slightly dubious were Interpopulation variability could be the ex- excluded from the sample. planation for this a s could the accuracy of The reconstruction of the obstetric histoeach sample’s obstetric data. ries utilised data derived from death certifiThe named sample from Christ Church, cates, family genealogies, and wills but rests Spitalfields (for further information see largely upon the baptism of infants. The Molleson and Cox, in press) provides an- 18th and early 19th century Huguenots who other opportunity to evaluate this important settled in Spitalfields were religious refuI__^
- I
1CLL
I . .
PARTURITION SCARS IN AN 18TH CENTURY SAMPLE
gees or their descendants and it seems unlikely that they would ignore baptism, a n important religious rite. Furthermore, the sobriety and moral fastidiousness of these people was such that Hogarth satirized them in “Noon” (1738) when he compared them with their less pious English contemporaries. As most stillbirths were not baptised, they cannot be detected in the historical record and they represent a n area of potential error in the data. Of the 94 females of known parity status, 73 (77.6%) had borne children; the remaining 21 (22.4%) appear to have been nulliparous. Of the nulliparous group, approximately half died aged below the normal age of marriage (twenty-six years) and may well have borne children had they lived to a greater age. Of the remainder, half were married but childless, the rest being nulliparous and unmarried. Family size ranged from 0 to 15 baptised infants; the mean was 2.7 infants (SD 2.82, SE mean 0.29). The age a t first birth ranged from 12 to 45 years; the mean was 27.1 years (SD 7.1, SE mean 0.9). The age at last birth ranged from 14 to 47 years, the mean being 35.8 years (SD 7.2, SE mean 0.9). Birth spacing ranged from 10 to 141 months; the mean was 29.4 months (SD 24.2, SE mean 3.3). Three females (4.3%) are known to have experienced a twin birth and five (7.1%) died within a few days of childbirth, probably from postpartum sepsis, Of the latter group, four had experienced previous confinements and all were aged 29 and below. The observations on the 138 adult female skeletons were made “blind to the obstetric data, the skeletal analysis being undertaken prior to the historical research. The presence or absence of a preauricular sulcus was observed, as was the type of sulcus. Four categories were used; the first and second were a s described by Houghton (1974). The first type, Houghton’s “groove of pregnancy”, appears to have been formed by the coalescence of a series of pits. It has a “scooped floor and a n uneven inferior margin (Fig. l a ) . The second, Houghton’s “groove of ligament”, has a smooth floor and a smooth inferior margin (Fig. lb). Fitting neither of Houghton’s categories, the third sulcus type is one observed by Dunlap (1981)
433
Fig. 1. a: A deep preauricular sulcus with a “scooped”appearance to the floor (Houghton’s“groove of pregnancy”). Inhumation number CAS 2438, a 56-year old of unknown parity status. b: A “groove of ligament” as defined by Houghton (1974).This preauricular sulcus is shallower than that illustrated in a,and has a smooth floor and smooth inferior margin. Inhumation number CAS 2438, a 56-year old of unknown parity status.
and by the first author when examining the females from the Poundbury Camp sample (Dorchester, England). This sulcus is comparatively wide, with an indistinct inferior margin and a “grainy” floor (Fig. 2a). The fourth sulcus was first described by Derry in 1909 (Fig. 2b). Usually seen in males, it is only occasionally found in females. Rather than being a true sulcus, that is a groove formed by cortical resorption, a n accentuated pyriformis tubercle with associated cortical deposition is creating the appearance of a sulcus. The severity of the preauricular sulcus was observed. Scored from 0 to 4, it is a subjective measure of size (not a series of precise mathematical increments) which considers depth a s well a s width and length.
434
M. COX AND A. SCOTT
Fig. 2. a: A preauricular sulcus type fitting neither of Houghton’s (1974) definitions. This sulcus is shallow, has a grainy floor, and an ill-defined inferior margin. Inhumation number CAS 2956, a 45-year old mother of three. h: The effect of a sulcus is created by the pyriformis tubercle and associated cortical deposition. Inhumation number CAS 2602, a 51-year old male.
The maximum width and maximum length of the sulcus were also measured. The presence or absence of sulci along the anterior margins of the sacrum adjacent to the auricular facet (Ullrich, 1975) was scored as present or absent. Pitting on the dorsal aspect of the pubic body was scored (Fig. 3a) both in terms of its presence or absence and in the number of discrete pits. The extension of the pubic tubercle (Fig. 3b) was scored from 0 to 3 in terms of severity (after Bergfelder and Herrmann, 1980). Examination of the cortical variants described above in association with measurements of pelvic capaciousness was also undertaken as a part of this study (Cox, 1989). The results of statistical analysis show that the presence of a preauricular sulcus (irrele-
Fig. 3. a: Pitting (cortical resorption) on the dorsal aspect of the pubic body. Inhumation number 2070, a 35-year old mother of five. b: An example of a greatly extended (grade 3) pubic tubercle (dorsal view). Inhumation number CAS 2566, a 74-year old mother of one.
vant of type) and the size of the sulcus are significantly. positively related t o each o f the following pelvic dimensions: sciatic notch width (Day, 19751, transverse diameter (Segeberth-Orban, 1980), bispinous diameter (Cox, 19891, inlet circumference (Cox, 1989), pubo-sacro-iliac diameter (Segeberth-Orban, 19801, and sacral width at the level of the caudal limit of the auricular facets. The presence of pubic pitting is significantly associated with bispinous diameter. (For full details see Cox, 1989). STATISTICAL ANALYSIS The type of test of significance used to evaluate the relationships between different variables was largely determined by the nature of the data in question. The distribution of continuous variables was examined using ranks to plot normal
PARTURITION SCARS IN AN 18TH CENTURY SAMPLE
435
scores and rootgrams. One way ANOVA was of those with extension were nulliparous. used to estimate analysis of variance where The data presented in Table 3b illustrates the data were normally distributed. Mann the distribution of the grades of tubercle exWhitney rank sum test and Kruskal Wallis tension in relation to parity status. A higher statistic were used where the data were not percentage of parous than nulliparous fenormally distributed as, for example, where males have grades one and two and no the analysis included number of births, age nulliparas have grade three extension. Table 3c illustrates that the degree of extenat death, or age at first or last birth. Chi square tests were used to examine the sion of the tubercle is significantly associdistribution of either two discrete variables ated with the number of births (P < .05), or one discrete and one continuous variable. those with advanced extension having a meThe expected cell frequencies in some tabu- dian of five births each. lations were occasionally very small (less There is a significant association between than one). In order to ensure valid results sacral scarring and parity status (Table 4), a the significance levels of all chi square tests higher proportion of nulliparous than parhave been evaluated using exact conditional ous females exhibiting this feature. However, a s scarring was present in only 8 of 63 tests (Kreiner, 1987). Pearson's r correlation coefficient was (12.7%) females this association has little used to determine the extent to which two practical value. Table 5 describes the results of analysis of continuous variables changed in relation to one another. the cortical variants under consideration As is traditional in such studies, the re- and age a t death. In no case was the associasults reported are those from the left 0s tion statistically significant ( P > .05). coxae. This alleviates problems caused by incomplete preservation and occasional DISCUSSION asymmetry in the distribution and type of It is and always has been a fact of life that preauricular sulcus and pubic pits (see also the birth of a child in certain circumstances Houghton, 1974; Tague, 1988). Figures l a is undesirable. This can lead to denial and and l b , illustrating different preauricular concealment and could affect the validity of sulcus types, are from the same individual. data retrieved from documentation. A further potential error lies in the fact that most stillbirths were not baptised. This results in RESULTS their absence from the historical record. Appropriate statistical examination of In order to take account of these two facparity status in relation to either the abtors, each cortical variant was examined sence or presence, severity, type or size of only in those females known to have borne the preauricular sulcus (Table 1)or the abchildren. In all cases, as can be seen by refersence or presence, or number of pubic pits ence to Tables 1, 2, 3, and 4,no feature oc(Table 2) indicates that none of these varicurs consistently among the parous females. ants is significantly associated with parity The presence of a n extended pubic tubercle status in this sample. In all cases P > .05. is the most consistent feature evident Preauricular sulcus types one to three are among the parous females in this sample. found in both parous and nulliparous females and there were only two cases of type four, both seen in parous females. Interestingly, all of the females with more than one pubic pit were parous but many parous females had no pitting or only one pit. Table 3a illustrates that the presence or absence of pubic tubercle extension is significantly associated with parity status ( P < .01). Of those without extension 66.66% were nulliparous, while only 12.5%
Scars of parturition and maternal mortality
Houghton (1975:658) states that following the delivery of a child ". . . processes of reorganisation and repair gradually return the ligaments to their normal state within a few weeks. Thereafter, a slow and variable replacement of bone appears to occur over years a t the sites of previous osteoclastic ac-
M. COX AND A. SCOTT
436
TABLE 1. Parity status and the preauricular su1cu.s l a . Presence or absence and parity status
Present
Nulliparous Parous
4 6
22.2 9.5
N
%'
14 57
77.8 90.5
x2, 2.09; D.F., 1;P , 0.223; N, 81. l b . Sulcus type and parity status
Type 1
Absent
2_ _ _Type ___ N _ _9% _
~~
~~
N
Q
4
22.2 9.6
N
5%
7 29
38 9 46 7
~
Nulliparous Parous y2, 2.9: D.F.. 4:
6 P. 0.57; N. 80.
lc. Sulcus severity and Daritv status
Nulliparous Parous
%
4
22.2 9.5
6
8 35
~ _ N_
22 2 17 7
4 11
_ Grade ~ 1_ N Q
Absent N
Type 3
_ _ _ _ _ _ _ _ ~
3 14
44.4 55.6
Q
N
%
16 7 22 G
0 2
32
Grade 2
_
Grade 3 & 4l
N
%
N
3 15
16.7 23.8
3 7
~-
Type 4 ______
3 _____
16.7 11.1
x2, 2.826; DF, 3; P , 0.46; N, 81. Id. Preauricular sulcus length and parity status
N
Mean
SD
SE Mean
13 57
23.54 27.05
7.74 7.43
2.1 0.98
l e . Preauricular sulcus width and nnrit,v status
N
Mean
SD
SE Mean
Nulliparous Parous
14 57
6.07 7.16
1.98 2.64
0.53 0.35
Nulliparous Parous T, -1.49; P , 0.16, DF, 17; N, 70.
T, -1.71; P , 0.099; DF, 25; N, 71. Grades 3 and 4 have been combined for the purposes of this test; grade 4 had nu values in the nulliparous cell and only one in the parous cell
TABLE 2. Pubic pitting and parity status
_Prewnl _ _ ~ -
2a Presence or absenw of pubic pitting and parity status
Absent
N
70
N
%
Nulliparous Parous
6 24
66.67 60.00
3 16
33.33 40.00
~
x2, 0.138; DF, 1;P , 1.00; N, 49. 2b. Number of pits and parity status
Absent _____ N
%
N
One
%$Nulliparous Parous
6 24
66.67 60.00
3 10
33.33 25.00
~~
~ Two _ N Q 0 3
_ Three N
Four %
N
?r
5.00
0 1
2.50
0
7.50
2
xa, 1.6; DF, 4; P , 0.88; N, 49.
tivity." If he is correct in this assumption, it seems that the sites of attachment of the inferior portion of the ventral sacroiliac ligament and the dorsal transverse ligament of the pubis should be unusual in females who die shortly after childbirth.
Five of the females among the Christ Church sample are known t o have died within days of parturition. None of these had any skeletal abnormality, such as a rachitic pelvis, a flat sacrum (Posner et al., 1955), nor any sign of pclvic disproportion
PARTURITION SCARS IN AN 18TH CENTURY SAMPLE
437
TABLE 3. Pubic tubercle extension and naritv status 3a Presence or absence of tubercle extension _and _ _parity __ _ _ _ -_ _ ~
____
N
- - _ -- - - -
Nulliparous Parous x‘, 9 682, DF, 1,P, 0 07, N, 33 3b Degree of tubercle extension and parity status
_ Absent ~ _ _~- __ _
_
76
__
--
6 3
66 7 12 5
2 _Grade - -- - .
_ _Absent ___-
_ _Grade _ _ -1_
-.
N
%
N
92
N
[k
N -__ __ _ ~ ~ _
%,
66 66 12 50
2 8
22 22 33 33
1 8
11.11 33.33
0 5
00.00 20.83
~
~~~~
Nulliparous 6 Parous 3 x’, 10 37, DF, 3, P, 0 02, N, 33 3c Degree of tubercle extension and number of births
(Kruskal Wallis) Grade
N
Grade 3 _ -
Median ~~
~_ _ _ _ _Avg - - _Rank ~
9 10 9 5 33
86 18 0 21 0 23 1 17 0
0 3 4 5
H, 10.49; ADJ H, 10.84; P, 0.013.
which would have prevented vaginal delivery. Four of these females had borne children previously. The skeletons of four of this group have survived well enough for their preauricular areas to be examined. All four had a preauricular sulcus, two were type 1 (Fig. l a ) and two were type 2 (Fig. lb). Three were slight and one moderately severe. The pubis survived in only one female; this had one pit on the dorsal aspect of each pubic bodj7 and slightly extended pubic tubercles. Xone had sacral scarring. This evidence suggests that the theory that “scars of parturition” represent cortical resorption which follows trauma incurred, by the mother, during pregnancy and parturition is unlikely to be correct. The macroscopic morphological changes evident on the pelvic bones of these four females were not in any way different from that seen on other females, some who died several decades after giving birth. Age at death and scars of parturition
The work of Suchey et al. (1979) and T a p e (1988) found that pubic pitting was significantly associated with the age a t death of the female. With one exception, the
TABLE 4. Sacral scarring and naritv status Absent
Present N %
~
N Nulliparous 8 Parous 47 x‘, 5.694; DF, 1;P, 0.04; N, 63.
% 66.67 92.16
4 4
33.33 7.84
age of the Christ Church females was known from their coffin plates. In many caws thi. could be checked against their baptism records and proved to be correct in all cases except one (where it was corrected for analytical purposes). The mean age a t death of this sample was 56.7 years (STD = 18.9); 75% were post-menopausal a t death (the oldest mothers were 47 years old at their last confinements). Appropriate statistical tests (Table 5) were used to examine the possibility of a n association between age at death and the presence of a pre-auricular sulcus, sacral scarring, pubic pitting, and a n extended pubic tubercle. Only for pubic tubercle extension ( P = 0.09) was the result even close to significance for this sample. Nevertheless, it was observed that in the young adult females whose iliac crests were unfused, there
M. COX AN1I A SCOTT
438
TABLE 5. A g e a t death and cortical remodelling of the aelvss
5a. Preauricular sulci _ and _ age _ _at ~death ___ ~~~
(Mann Whitney) W, 724.0; P = 0.56; N
=
112
5b.~Preauricular sulcus severity and age a t death _ _ _ _ _ _ _ _ ___ _ _ _ _ _ _ _ _ _ _ ~ ~ ~
~
~
(Kruskal Wallis) H, 3.346; A D J H, 3.348; P
=
0.50; N
=
115.
5c. Preauricular width and length and age a t death ~
(Correlation coefficient) Width 101 Length 98
0.001 0.117
:>.05 >.05
5d. Pubic pitting and age a t death (Mann Whitney) W, 1362.5; P = 0.65; N = 69. at _ death 5e. Number of pubic pits and age _ _~ (Kruskal Wallis) H, 1.560; A D J H, 1.561; P
=
0.67; N
=
69.
5f. Sacral scarring and age a ___ t death _____(Mann Whitney) W, 3255.0; P = 0.67; N
=
CONCLUSION
88.
5g. Pubic tubercle extension and _ age_a t death _ __ (Mann Whitney) W, 263.0; P = 0.09; N
=
~
47.
5h. Degree of tubercle extension and age at death_ _ _
~~~
~~
(Kruskal Wallis) H, 2.919; A D J H, 2.922; P
=
0.40; N
between the two sets of results lends credibility to the obstetric histories of the Christ Church sample and to the obstetric significance of an extended pubic tubercle. It must be considered though that 33.3% of the nulliparous females also had an extended tubercle. In view of this, the strong statistical association between the presence of an extended tubercle and parous females cannot be considered as conclusive proof of a causal relationship. Pregnancy may not be the only cause of tubercle extension, as the presence of an extended tubercle in 90.3%of the Christ Church males (n = 31) attests. It must be stressed though that the morphology of the extended tubercle is different in males and females, the male tubercle being far more robust and less fragile than the female tubercle.
=
47.
was never more than the slightest trace of either a preauricular sulcus or pubic pitting. The pubic tubercle and pregnancy and parturition
The association between the extended tubercle with parity status (Table 3) and the number of births could be explained physiologically. A portion of abdominus rectus attaches to the pubic tubercle. It is possible that the extended tubercle might represent bone remodelling in response to the stress imposed on the muscle insertion during the later stages of pregnancy, when this muscle is greatly extended. The association between the pubic tubercle and parity status and the number of births in the Christ Church sample concurs with that obtained by Bergfelder and Herrmann (1980). Their sample was from a dissecting room and obstetric histories were obtained from medical records and from relatives of the deceased. The concurrence
The results obtained by statistical analysis of the preauricular sulcus, pubic pitting, and sacral scarring in relation to the parity status of the Christ Church, Spitalfields sample (n = 94) indicates that none of these cortical variants is consistently associated with parity status. In this sample, the small number of females with more than one pubic pit had all borne children, but the lack of pits, or only one pit, is not associated with nulliparity. Unfortunately, the number of nulliparous females in this sample is small. It should be considered that statistical analysis of larger samples might detect associations between variables not apparent here. A significant association is, however, demonstrated between the presence of an extended pubic tubercle and parity status. Eighty-seven percent of the parous females had an extended pubic tubercle and the greater the degree of extension, the greater the number of births. However, 33.3%of the nulliparous females also had an extended tubercle and it must be considered that pregnancy may not be the only causal factor. ACKNOWLEDGMENTS
I thank Theya Molleson of The Natural History Museum, London and Don Brothwell, Simon Hilson, and Clive Orton of the Institute of Archaeology, University College
PARTURITION SCARS IN AN 18TH CENTURY SAMPLE
London for their help and support with this project. Phillip Crabb of The Natural History Museum took the excellent photographs. The Christ Church Spitalfields project was funded by English Heritage and The Nuffield Foundation.
439
Suchey JM, Wisely DV, Green RF, and Noguchi TT (1979)Analysis of dorsal pitting in the 0s pubis in an extensive sample of modern American females. Am. J. Phys. Anthropol. 51t517-540. Tague RG (1988) Bone Resorption of the Pubis and Preauricular Area in Humans and Nonhuman Mammals. Am. J. Phys. Anthropol. 76:251-267. Ullrich H (1975) Estimation of fertility by means of pregnancy and childbirth alterations a t the pubis, the ilium. and the sacrum. OSSA 2.23-39.
LITERATURE CITED Andersen BC (1988) Pelvic scarring analysis: Parturition or excess motion. Am. J. Phys. Anthropol. 75t181. Angel J L (1969)The bases of palaeodemography. Am. J. Phys. Anthropol. 30:427437. Bergfelder T, and Herrmann B (1980)Estimating fertility on the basis of birth traumatic changes in the pubic bone. J. Hum. Evol. 93511-613. Cox MJ (1989) PhD Thesis. Institute of Archaeology, University College London. An Evaluation of the Significance of “Scars of Parturition” in the Christ Church Spitalfields Sample. Day MH (1975) Sexual differentiation in the innominate bone studied by multivariate analysis. Ann. Hum. Biol. 2t143-151. Derry DE (1909)Note on the innominate bone as a factor in the determination of sex. J. Anat. Physiol. 43:266276. Dunlap SS (1981) PhD Thesis, Michigan State University. A study of preauricular sulci in a cadaver population. Holt CA (1978)A re-examination of the parturition scars on the female pelvis. Am. J. Phys. Anthropol. 4 9 9 94. Houghton P (1974) The relationship of the preauricular groove of the ilium to pregnancy. Am. J. Phys. Anthropol. 4It381-390. Houghton P(1975) The bony imprint of pregnancy. Bull. W A c a d . Med. 51t655-661. Kelley MA (1979)Parturition and pelvic changes. Am. J. Phys. Anthropol. 51t541-546. Kreiner S 1,1987)Anaiysis of multidiniensional contingency tables by exact conditional tests: Techniques and strategies. Scand. J . Statist. 14r97-112. Molleson TI, and Cox MJ (In Press) Spitalfields: The Middling Sort. London: Council for British Archaeology. Posner A, Bloch NR, and Posner NS (1955) The flat sacrum: Its importance in obstetrics. Am. J. Obstet. Gynecol. 70: 1021-1025. Segeberth-Orban R (1980) An evaluation of the sexual dimorphism of the human innominate bone. J. Hum. Evol. 9:GOl-607. Spring DB, Lovejoy CO, Bender GN, and Duerr M (1989) The Radiographic Preauricular Groove: Its Non-Relationship to Past Parity. Am. J. Phys. Anthropol. 79t247-252. Stewart TD (1970) Identification ofthe scars of parturition in the skeletal remains of females. In TD Stewart (ed.):Personal identification in mass disasters. Washington D.C.: National Museum ofNatural History, pp. 127-135.
APPENDIX A
The processes involved in reconstructing the females’ obstetric histories were complex and involved working from, and building upon, the initial data base, the coffin plates. These were inscribed with the individual’s name, age at death, and date of death. Starting from this primary data, information was collected and collated piece by piece until, where possible, a n understanding of each individual’s obstetric history had been achieved. The first procedure was to establish if the females were married, when, and to whom. That a female was titled “Mrs.” on her coffin plate did not necessarily signify that she was married. The terms “Mrs.” and “Miss” are both contracted from the term “Mistress” and it is not clear historically when the use of the terms to distinguish married from unmarried women crystallised. In order to illustrate the complexity of the inquiry one individual case is described below. Mrs. Grace Wells (CAS 2666) died aged 62 on June 8th, 1811. Reference to the appropriate Christ Church Spitalfields Burial Register (Greater London Reference Library) was made to verify the details and to collect additional information contained therein. Grace was buried on June 16th, 1811 and her “abode at death” was Dorset Street, Spitalfields. Grace’s year of birth would have been 1749, therefore her fecund period would have extended from approximately 17601800. Any marriage she entered into after this date would be irrelevant to this inquiry a s there would have been no issue. Reference to the International Genealogical Index for London (Guildhall Library) within the period of concern revealed only one marriage in which a male with the surname
M. COX AND A. SCOTT
440
T A B L E 6. Baptism data relating to the children o f Grace and Christopher Wells Child’s name
Date of birth
Date of baotism
Father’s occuo~ ition
19/06/1768 13/08/1769 06/09/1772 09/04/1775 20/07/1777 06/09/1778 3010511781 19/05/1782 07/05/1784 16/09/1785 16/07/1788 11/07/1790
Painter Painter Painter Painter Painter Painter Painter Painter Painter Painter Painter Painter
~~
Christopher Wi11ia m Christopher George Ann Robert Grace Ann Charlotte Elizabeth John Mary Richard
~
25/07/1769 14/08/1772 05/03/1775 24/06/1777 09/08/1778 02/05/1781 26/04/1782 07/04/1784 28/08/1785 23/06/1788 01/07/1790
Wells married a female with the Christian name Grace. Christopher Wells of the parish of Christ Church, Spitalfields married Grace Duxbury on May 19th, 1767 a t St. Leonard’s Church, Shoreditch. (Grace would have been aged eighteen years old a t the time of her marriage.) The place of marriage suggests that Shoreditch had been the bride’s place of residence. That the groom’s parish of residence was Christ Church indicated that this could be the correct marriage. The International Geneaological Index also lists the baptisms at Christ Church Spitalfields of several children born to this couple. The next step was t o refer to the original
Parents’ address ~~~
Paler Noster Row Paler Noster Row Paler Noster Row Brown’s Lane Brown’s Lane Brown’s Lane Brown’s Lane Brawn’s Lane Brawn’s Lane Brawn’s Lane Brown’s Lane Darset Street
Baptism Register in order to see if it contained any information which would confirm that the mother was CAS 2666. The data derived from the registers relating to the children of Grace and Christopher is listed in Table 6. The address given at the baptism of Richard, Grace’s last baptised child, was the same as that on the registration of Grace’s death. Reference to the Land Tax Returns (Guildhall library) for the appropriate years indicated that there was no other householder with the same surname living in Dorset Street. With this number of positive indicators it was considered that the obstetric data did relate to the female in question.
