Egyptian Journal of Forensic Sciences (2012) 2, 1828

Contents lists available at SciVerse ScienceDirect

Egyptian Journal of Forensic Sciences

journal homepage: www.ejfs.org

Morphometric sex determination from various sternal widths of Northwest Indian sternums collected from autopsy cadavers: A comparison of sexing methods
Jagmahender Singh
a b c

a,*

, R.K. Pathak b, Dalbir Singh

Department of Forensic Medicine and Toxicology, Govt. Medical College and Hospital, Chandigarh 160 030, India Institute of Forensic Science and Criminology, Panjab University, Chandigarh 160 014, India Department of Forensic Medicine, Postgraduate Institute of Medical Education and Research, Chandigarh 160 012, India

Received 19 October 2011; revised 8 December 2011; accepted 9 December 2011 Available online 27 February 2012

KEYWORDS Forensic anthropology; Sternum; Sternal widths; Sex determination; Sexing methods; Discriminant function analysis

Abstract The sternum is one of the least studied bones of human thoracic skeleton for sex determination both in forensic anthropology as well as in archeology. In the present study, the relationship between various sternal widths and known sex of the deceased was studied in a sample of 343 adult human sternums based on anthropometric measurements. A comparison was also made with different widely used methods of sex determination. The sternubrial width index was found to be the poorest indicator of sex, whereas all the other variables were equally suitable for sex determination. The criteria of demarking points (calculated from mean 3SD) failed in sex determination of the studied sternums. The identication points correctly sexed more percentage of bones than the demarking points which are considered more reliable covering more than 99.67% of cases. The limiting points proved to be a best method for almost all the variables, sexing about 7080% of the sternums examined for this purpose. The multivariate discriminant function analysis (DFA) of various sternal widths correctly sexed 84% sternums (Males = 82.1%, Females = 89%). The sex of a sternum of Northwest India can be estimated using equation F (D) = 13.166 + 0.087 BB1 + 0.088 BB-2 + 0.150 BB-4 + 0.040 MBX + 0174 WS-1 (group centroids i.e., 0.508 for male and 1.408 for female). From the logistic regression analysis, 86.6% sternums (93.3% males and 68.1% females) were assigned the correct sex category, thus giving slightly better results than

* Corresponding author. Address: # 905 B, Sector-43A, Chandigarh 160 022, India. Tel.: +91 172 2665253 58x1064, mobile: +91 9417048690; fax: +91 172 2608488. E-mail address: jagmindera@yahoo.com (J. Singh). 2090-536X 2012 Forensic Medicine Authority. Production and hosting by Elsevier B.V. All rights reserved. Peer review under responsibility of Forensic Medicine Authority. doi:10.1016/j.ejfs.2011.12.002

Production and hosting by Elsevier

Morphometric sex determination from various sternal widths of Northwest Indian sternums collected from autopsy cadavers: A comparison of sexing methods

19

DFA. Comparatively more males were correctly sexed than females using almost all the criteria except DFA. In conclusion, regression analysis was found the best statistical method of sex determination from various sternal widths studied in the present study.
2012 Forensic Medicine Authority. Production and hosting by Elsevier B.V. All rights reserved.

1. Introduction Sex determination of unidentied skeletal remains encountered in forensic or archeological situations, is one of the prime tasks of experts like forensic anthropologists and bioarcheologists involved in such examinations. In case of mass-disasters such as armed conicts, terrorist massacres, aeroplane crashes, war related crimes etc., when badly decomposed, mutilated or damaged human remains consisting of only a few bones or their fragments are recovered from the site, it becomes crucial to establish the biological identity, especially the sex of a missing individual. Though, a number of bones such as pelvis, skull femur etc., have contributed signicantly to this endeavor, these sex-specic or sex indicative bones are not always found at such scenes. In such situations, forensic anthropologists and bioarcheologists have to depend on less sexually dimorphic elements of human skeleton such as sternum.17 Nevertheless, such studies can provide useful clues and information to scientically prepare experts to face such situations in the future wherein bones like pelvis and skull may not be part of the skeletal remains recovered for examination. The anterior thoracic cage, with sternum in its median aspect, is frequently encountered in the forensic anthropological contexts.3,8 Though, the sex predictive value of human sternum (from its morphometric traits) has been studied extensively both in the Indian2,916 as well as in foreign contexts,38,1725 the manubrial breadths and inter-costal mesosternal breadths have been very rarely studied for this purpose. Furthermore, different workers have applied different criteria/methods/approaches for determination of sex from different sets of sternal measurements studied by them. In the present study, an effort has been made to estimate sex from the combination of all the possible widths of the freshly collected, cleaned but wet sternums taken from autopsy cases. Furthermore, comparative number and percentage of correctly sexed sternums from different approaches have been assessed in the present study. 2. Materials and methods After obtaining written consent (well-informed and explained) from the next kin or the legal administrators of the deceased, 343 sternums of both the sexes (252 males; 91 females) were collected from the adult cadavers (1794 years) brought for postmortem examination at Postgraduate Institute of Medical Education and Research (PGIMER), Chandigarh, India. The subjects belonging to the states of Punjab, Haryana, Himachal Pradesh, western part of Uttar Pradesh (UP) and the union territory of Chandigarh (states/UT geographically falling in Northwest of India) were included in the study (Table 1) to avoid any distortion in the sternal measurements previously reported by different workers for different zones of India. The entire sample area falls in the northwestern part of India whose inhabitants were basically agriculturists belonging to the same distinctive cluster and sharing similar food habits, gene

frequencies and other distinguishing traits,26 though with some morphological diversities. Each sternum was carefully dissected by giving incisions at the acromio-clavicular, sterno-olecranon and cost-sternal junctions, along a standard linear midline incision. Each sternum was prepared by placing it in boiling water, cleaned and washed for taking measurements as per the standard scientic procedures available in the literature.27 Any sternum with fractures, pathology, surgical repairs, grossly visible deformity or skeletal abnormality was excluded from the study sample. While boiling and cleaning, repeated careful inspection was done to avoid separation of the three segments of the sternum. Muscular attachments, if any, remaining even after boiling, were removed manually by scraping carefully with a blunt scalpel. The collected sternums were measured and replaced into the body before handing over the bodies to the relatives of the deceased. Each linear measurement was taken thrice on the anatomical position of the sternum using Mitutoyo digital vernier calipers to the nearest millimeter as shown in Fig. 1 and their average was recorded, the indices were calculated according to the landmarks and techniques prescribed by previous workers in some earlier research publications 12,914,16,4,2829: 2.1. Manubrial breadths The maximum manubrial breadth was measured by placing the two arms of vernier calipers at the points just below the rst costal notches of right and left sides of the manubrium, whereas, the minimum manubrial breadth was taken between the two points at superior ends of the second costal notches immediately above the sternal synchondrosis. These two points on the sternum are taken to represent as the greatest and the least manubrium breadth, respectively. 2.2. Sternebrial breadths These are the minimum distances measured at the waist of the rst and third sternebrae, respectively, in the midline instead of at the maximum width of the sternubrium as the latter may be obscured by the exaggerated costal processes which are frequently seen in aged specimens. The width of only rst and third sternebrae was measured in the present study because these two segments of mesosternum are found to be variable in shape1,20,30. 2.3. Ratio of rst and third sternebrae (Sternubrial width index) It is the ratio of breadth of 1st sternubrium at its waist to the breadth of 3rd sternubrium at its waist and multiplied by 100. SWI Width of first Sternubrium WS 1 100 Width of third Sternubrium WS 3

20
Table 1 State-wise distribution of study population. [Males (n = 252 pairs), Females (n = 91 pairs)].
Punjab M 1725 2635 3645 4655 5665 6675 >75 Total 19 24 21 12 10 5 2 93 137 (39.94%) F 19 8 2 8 4 2 0 44 Haryana M 28 30 13 6 2 3 3 85 112 (32.65%) F 9 10 0 4 3 1 0 27 Himachal Pradesh M 14 11 8 4 4 1 0 42 54 (15.74%) F 4 3 1 2 1 1 0 12 Chandigarh M 6 6 3 1 1 0 0 17 23 (6.70%) F 4 1 0 0 0 1 0 6 Western U.P. M 6 4 4 1 0 0 0 15 17 (4.96%) F 1 0 0 0 1 0 0 2

J. Singh et al.

Age interval (years)

Total

110 (32.07%) 98 (28.57%) 52 (15.16%) 38 (11.08%) 26 (7.58%) 14 (4.08%) 5 (1.46%) 343 (100%)

Figure 1 A typical human sternum showing landmarks for measuring various sternal widths i.e., MB-MAX = maximum manubrial breadth, MB-MIN = minimum manubrial breadth, WS-1 = width of 1st sternubrium, WS-3 = width of 3rd sternubrium, BB = intercostal breadth of mesosternum.

2.4. Inter-costal breadths of mesosternum These are the distances measured between the mid-points of second and third, third and fourth, fourth and fth and fth and sixth costal facets/notches of the mesosternum as per the technique adopted by Selthofer et al.4 These distances between the consecutive costal notches were measured with the help of spreading caliper with pointed ends The inter-costal breadth between sixth and seventh could not be recorded correctly in most of the cases (due to its inherent variability) and hence

was not considered for the observations and in the nal analysis. It may be pertinent to state that this measurement is different from that recorded by Torwalt and Hoppa8 on chestplate radiographs, wherein the width of mesosternum was recorded mid-way between second and third, third and fourth, fourth and fth and fth and sixth costal cartilage notches. Statistical descriptive, identication points, demarking points and limiting points were calculated for each variable as depicted in the Tables 25. The maximum value of a measurement for a variable in females is taken as male identication

Morphometric sex determination from various sternal widths of cadavers: A comparison of sexing methods point (IP) and the minimum value of a measurement for that variable in males constitutes the female identication point.31 All the bones having values less than minimum value of males were treated as female bones and the bones having values more than maximum value of females were treated as male bones. The demarking points (DP) were calculated as the mean value of a variable in the opposite sex plus (male) or minus (female) three times its standard deviation.32 The limiting points (LP) were calculated by dividing the sum of male and female demarking points by two which gives the most signicant classication of bones based on sex.11 The concept of limiting point has been used for the rst time in sex determination of sternum from various sternal widths. The number of sternums falling beyond these points particular/specic for a sex constitutes the correctly classied bones based on these points. Additionally, all the variables were subjected to stepwise discriminant function analysis (DFA) using SPSS software, version 16.0 33 to select the single most important variables which discriminate between the two sexes with higher percentage. A discriminant function equation was devised which may be useful for sexual dimorphism of unknown sternums found in forensic or archeological situations. Since logistic regression analysis gives better probability results than DFA with lesser statistical assumptions for predicting dichotomous entities like sex,8,3436 the entire data were also subjected to logistic regression analysis. A comparison was made between the number and percentage of correctly classied sternums using all the criteria used in the present study. 3. Results Table 2 shows that maximum manubrial breadth has comparatively more sample homogeneity than the minimum manubrial breadth as the latter has greater coefcient of variation. Using identication points, about 2737% sternums of both sexes were correctly classied for the maximum manubrial breadth, whereas only 27% sternums were correctly assigned to their sex from minimum manubrial breadth. The demarking points could correctly sex only about 6% male and about 1% female sternums for both the minimum as well as maximum manubrial widths. The limiting points calculated from both the identication points as well as demarking points were found to be the better estimator of sex for both these widths, correctly sexing about 7275% males and about 6267% female sternums of the present study. The difference in the mean values of two sexes was found to be statistically highly signicant for the maximum as well as minimum manubrial breadth (P < 0.01). Thus, limiting point was found to be a best criterion of sex discrimination from the maximum and minimum widths. The width of the third sternubrium at its waist was found to exhibit greater variability in its measurements than the rst sternubrium as observed in the studied sternums (Table 3). Though, identication points using two sternebrial widths differentiated sex of the sternums to a satisfactory level but demarking points failed to discriminate for the female sternums but correctly classifying sex of about 610% male sternums. So, a majority of sternal values were found lying in the overlapping zone. The limiting points were again found to be better estimator of sex for both the sternebrial breadths, though with slightly varying percentages. Studying sex

Northwest Indian sternums collected from autopsy

21

differences on the basis of Students t-test between means of two sexes, it was found to be statistically highly signicant. The values of sternebrial width index greatly deviated from its mean value and thus had a greater degree of variability (Table 4). It was found to be the poorest indicator of sex on the basis of all the criteria used in the present study. Both the identication points as well as demarking points almost completely failed to discriminate the sex of the collected sternums, though, the limiting points correctly classied a satisfactory number of the studied bones. Among the four mesosternum inter-costal breadths (Table 5), the third and fourth ones were found to exhibit greater variability (coefcient of variation = 1721) in their measurements than the rst and second inter-costal widths. Using identication and demarking points, comparatively greater percentage of male sternums were assigned to their sex than the female sternums (12) and hence most of sternal values were found in the overlapping zone of males and females. The limiting points for all these inter-costal breadths classied about 6080% sternums, comparatively more in male than in the female sternums (Table 6). 4. Discriminant function analysis Except the third inter-costal width of mesosternum (BB-3), sternebrial width index (SWI) and the minimum breadth of manubrium (MBN), all the other measurements were selected in the multivariate discriminant function analysis for sex determination of the studied variables of the sternum (Table 7). The sternebrial width index was totally eliminated by the multivariate discriminant analysis as a tool for sexing the studied sternums. The females showed a slightly greater tendency of being correctly classied. Table 8 shows that DFA correctly classied about 84% sternums (males = 82.1%, females = 89.0%) and on cross-validation, about 82.8% cases were correctly assigned to their correct sex. So, we can say that if an unknown sternum is encountered in any forensic or archeological situation, then its sex can be estimated with about 84% accuracy by placing the values of its measurements in the following DFA equation by matching the group centroids i.e., 0.508 (males) and 1.408 (females) and with a sectioning point of 1.96. F D 13:166 0:087 BB-1 0:088 BB-2 0:150 BB-4 0:040 MBX 0174 WS-1: The various widths of an unknown sternum can be put into this equation and if the resultant value is toward 0.508, it is a male sternum and vice-versa. The results of logistic regression analysis show that 235 males (93.3% and 62 females (68.1%) were correctly classied to the original sex category, giving an overall correct classication of 86.6% sternums (Table 9). 5. Discussions Proper identication of an individual is important not only for the family or friends of the deceased, but is also necessary for proper completion of a variety of documents and for the purpose of settling insurance claims, estate issues; criminal proceedings etc. Means of identifying an individual and the techniques and standards for identication are varied and depend upon the circumstances of each particular case and the nature and condition of the exhibit available.

22
Table 2
Statistics DESCRIPTIVE Mean Std. deviation Coe. of variation Variance Std. Error Range Identication and demarking points Number & percentage of sternums beyond identication points Number & percentage of sternums in overlapping zone Limiting point Number & percentage of sternums beyond limiting point t-test p-value

J. Singh et al.
Statistical descriptive of maximum and minimum manubrial breadths of sternum (mm) Males (n = 252), females (n = 91).
Maxi. manubrial breadth (MBX) M 54.91 5.44 9.91 29.57 0.34 44.9975.60 38.5971.23 P56.52 P63.64 93 (36.90%) 159 (63.10%) 237 (94.05%) 50.76 51.12 191 (75.79%) 184 (73.02%) 10.39 <0.0001 F 48.31 5.11 10.58 26.06 0.54 35.6656.52 32.9863.64 644.99 638.59 25 (27.47%) 66 (72.53%) 90 (98.90%) Mini. manubrial breadth (MBN) M 32.19 4.67 14.49 21.76 0.29 21.2346.03 18.1846.20 P36.76 P40.27 42 (16.67%) 210 (83.33%) 237 (94.05%) 29.00 29.33 189 (75.0%) 182 (72.22%) 7.34 <0.0001 F 28.48 3.93 14.34 13.80 0.41 13.1236.76 16.6940.27 621.23 618.18 2 (2.2) 89 (97.80%) 90 (98.90%)

Normal Mean 3SD IP DP IP IP DP IP DP IP DP

60 (65.93%) 61 (67.03%)

57 (62.64%) 59 (64.84%)

Table 3
Statistics

Statistical descriptive of widths 1st and 3rd sternubrae (mm); males (N = 252), females (N = 91).
Width of 1st sternubrium (WS-1) M 27.03 3.58 13.24 12.80 0.23 20.0538.81 16.2937.77 P29.24 P31.66 65(25.79%) 187(74.21%) 227(93.65%) 24.65 23.98 184(73.02%) 203(80.56%) 10.46 <0.0001 F 23.11 2.85 12.33 8.14 0.30 16.7629.24 14.5631.66 620.05 616.29 14(15.38%) 77(84.62%) 91 (100%) Width of 3rd sternubrium (WS-3) M 33.53 5.49 16.37 30.10 0.35 23.9654.01 17.0650.00 P38.65 P42.53 38(15.08%) 214(84.92%) 237(94.05%) 31.31 29.80 155(61.51%) 185(73.41%) 9.00 <0.0001 F 28.01 4.84 17.27 23.42 0.51 18.0138.65 13.4942.53 623.96 617.06 22(24.18%) 69(75.82%) 91 (100%)

Descriptive Mean Std. deviation Coe. of variation Variance Std. error Range Identication and demarking points Number & percentage of sternums beyond identication points Number & percentage of sternums in overlapping zone Limiting point Number & percentage of sternums beyond limiting point T-test P-value

Normal Mean 3sd IP DP IP IP DP IP DP IP DP

69(75.82%) 62(68.13%)

69(75.82%) 63(69.23%)

Identication of the deceased from the skeleton remains is one of the most crucial and the challenging problems faced by forensic experts as the common bones of interest are found either missing or are highly damaged from the forensic or bioarcheological contexts. Accordingly, they have to base their analysis on other less indicative bones with a wide range of maturations and subtle variations. Sex determination is one of the most important big fours of biological prole to be established from skeletal remains. The pelvic and cranial elements of human skeleton along with long bones are considered most sexually dimorphic.3738 Often an expert has to extract as much information as possible from such limited and heterogeneous materials available in the form of fragmentary skeletal

remains6,38 and even a single bone like sternum becomes important from a forensic anthropological view point when discriminant functions or regression equations are available for the particular population. The quantum and duration of growth spurt, growth patterns, strength of muscular attachments and exposure to differential biomechanical loadings in two sexes etc., might be the factors responsible for the bones that exhibit good sexual dimorphism, which, in turn, make them suitable for sex determination.3940 Both the quantitative and qualitative human skeletal features have been explored for estimating sex of an individual. However, both these methods have their own strengths and weaknesses; the former being more objective, reproducible or

Morphometric sex determination from various sternal widths of Northwest Indian sternums collected from autopsy cadavers: A comparison of sexing methods
Table 4
Sex Mean Std. Deviation Coe. of Variation Variance Std. Error Range Identication and Demarking Points Number & Percentage of sternums beyond identication points Number & Percentage of sternums in overlapping zone Limiting point Number & Percentage of sternums beyond limiting point t-test results P-Value

23

Descriptive statistics for ratio of width of rst and third sternubrae Males (N = 252), Females (N = 91).
Ratio of width of 1st & 3rd sternubrae (SWI) M 81.80 11.87 14.51 140.90 0.75 49.60120.28 46.19117.41 P115.94 P120.79 1(0.40%) 249(98.81%) 252 (100%) 82.77 83.49 115(45.63%) 108 (42.86%) 1.45(N.S.) <0.151 F 83.95 12.28 14.63 150.70 1.29 56.43115.94 47.11120.79 649.60 646.19 0(0.00%) 91(100.0%) 91 (100%)

Descriptive

Normal Mean 3SD IP DP IP IP DP IP DP IP DP

43(47.25%) 45 (49.45%)

duplicatable and applicable to fragmentary remains,4143 though requiring specialized techniques, expertise and instruments (walker). Whereas, the non-metric methods are simple, quick and easy to apply,4445 though they are more subjective in nature with higher inter-observer errors, thus not considered suitable for forensic applications.4647 With the introduction of discriminant function analysis in forensic osteological studies, the subjectivity levels and the expertise needed has been reduced substantially4849 and even 100% sexing accuracy can be obtained from it.50 Though discriminant function analysis is extremely useful in sex determination of fragmentary or incomplete skeletal remains,5152 such estimates are not universal and are highly population specic.5355 Furthermore, the logistic regression analysis gives much better results than the discriminant function analysis for sex estimation of skeletal remains.7 In Indian context, the recovery of unknown human skeletal remains is common, probably because of increasing naxalite or terrorist massacres, clandestine murders, natural calamities like tsunamis, earthquakes, plane/train/road accidents, maa conicts, etc. Sexual dimorphism in human sternum has remained a matter of discussion among forensic anthropologists since long times. Both the morphologic as well as morphometric sternal features have been utilized in this endeavor; however, the quantitative metric analysis of sternum is preferred to the morphological observations as the former analysis is prone to statistical analyses.3,6,22 Few workers have used the individual or the combined length/s of sternum (as a rule) for estimating sex, while others have devised either identication, demarking or limiting points; or have calculated discriminant functions and regression equations based on such measurements for this purpose.9,1112 Average maximum manubrial breadth in male as well as females of present study is signicantly more than those reported by Atal et al.,13 for a Delhi population (India), but is signicantly less than the Croatian and Spanish samples.4,23 Accuracy of sexing from maximum manubrial breadth for the present study varies between 1% and 6% (using demarking points) and between 67% and 73% (using limiting points).

The only study available for comparison of minimum manubrial breadth is of Croatians by Selthofer et al.4 It is seen that minimum breadth of manubrium in the present study is significantly less in both sexes as compared to the Croatian study. Accuracy of sexing from minimum manubrial breadth is almost similar to that observed for maximum manubrial breadth above. As shown in Table 10, the width of 1st sternubrium is more in males than in females of all the studies listed in the available literature, barring in the study on Maharashtrian subjects (India) by Dahiphale et al.,9. In the present study, this width is signicantly greater in both sexes than the Delhi sample (India) but is signicantly lesser in females than the studies of Jit et al.,2 and Dahiphale et al.,9 A similar study was conducted by Jit et al2 about two decades earlier, but signicant differences occur in some dimensions between the two studies as shown in Table 10, probably because of increased nutritional status, environmental changes etc. between the two different times. Using width of 1st sternubrium, the calculated limiting points gave accurate sex determination about 80% males and 68% females in the present study. From the demarking points, comparatively more percentage of sternums was correctly classied in the present study. The third sternubrium width on an average is more in males of present study than all other studies compared, though signicant difference is seen only with the Delhi sample (India), but, the same does not hold true in case of females. Using width of 3rd sternubrium, the calculated limiting points gave accurate sex determination about 73% males and 69% females in the present study. From the demarking points, comparatively lesser percentage of sternums was correctly classied in the present study. The mean ratio of two widths of the sternubrium in both sexes is observed to be less in the present study than that listed in the table; however, statistically signicant differences are found only for males. The limiting points for this ratio gives less percentage (males = 42%, females = 49%) of correctly sexed bones as compared to when the two widths of sternubrium are treated separately. No sternum could be sexed correctly using demarking points in the present study, though a

24

Table 5

Statistical descriptive for rst four inter-costal breadths of mesosternum (mm); Male (N = 252), Female (N = 91).
First inter-costal breadth (BB-1) Male 29.00 3.64 12.54 13.23 0.23 12.6337.78 18.0839.92 P35.11 P35.53 6(2.38) 245 (97.22) 245 (97.22) 23.87 26.81 230 (91.27) 200 (79.37) 8.13 <0.0001 Female 25.63 3.30 12.89 10.91 0.35 16.5735.11 15.7335.53 612.63 618.08 0(0.00) 91 (100.0) 89 (97.80) Second inter-costal breadth (BB-2) Male 24.03 3.07 12.78 9.44 0.19 14.1733.29 14.8233.24 P31.37 P30.98 1(0.44) 250 (99.21) 248 (98.41) 22.77 22.90 172 (68.25) 171 (67.86) 7.37 <0.0001 Female 21.11 3.29 15.57 10.81 0.34 14.8131.37 11.2430.98 614.17 614.82 0(0.00) 91 (100.0) 89 (97.80) Third inter-costal breadth (BB-3) Male 18.72 3.35 17.91 11.25 0.21 9.1830.84 8.6728.77 P28.12 P25.04 1(0.44) 251 (99.60) 248 (98.41) 18.65 16.86 126 (50.00) 180 (71.43) 6.88 <0.0001 Female 16.13 2.97 18.42 8.82 0.31 8.6128.12 7.2225.04 69.18 68.67 2(2.20) 89 (97.80) 89 (97.80) Fourth inter-costal breadth (BB-4) Male 14.20 3.02 21.29 9.14 0.19 7.8627.93 5.1423.26 P18.62 P18.30 22(8.73) 230 (91.27) 228 (90.48) 13.24 11.72 146 (87.94) 209 (82.94) 9.1 <0.0001 Female 11.40 2.30 20.19 5.30 0.24 6.2918.62 4.5018.30 67.86 65.14 4(4.40) 87 (95.60) 91 (100)

Parameters Sex Mean Std. deviation Coe. of variation Variance Std. error Range Identication and demarking points Number & percentage of sternums beyond identication points Number & percentage of sterna falling in the overlapping zone

Original Mean 3SD IP DP IP IP DP IP DP IP DP

Limiting point Number & percentage of sterna beyond limiting point

25 (27.47) 56 (61.54)

71 (78.02) 71 (78.02)

77 (84.62) 56 (61.54)

78 (85.71) 54 (59.34)

t-test results P-Value

J. Singh et al.

Morphometric sex determination from various sternal widths of Northwest Indian sternums collected from autopsy cadavers: A comparison of sexing methods
Table 6 Wilks lambda, classication and discriminant function coefcients.
Wilks lambda F P-value Fishers classication coecients Male BB-1 0.595 BB-2 0.581 BB-4 0.699 MB-MAX 0.777 WS-1 0.644 Constant Group centroids Sectioning points 4.815 9.302 3.411 4.221 3.349 0.0001 0.0001 0.0001 0.0001 0.0001 1.423 1.437 1.113 0.967 1.971 99.684 0.508 0.196 Female 1.257 1.269 0.824 0.89 1.638 75.315 1.408 0.087 0.088 0.151 0.04 0.174 13.166

25

Measurement

Canonical discriminant Function coecients

Table 7
Results Original

Classication results of discriminant function analysis.

Sex Male Female No. &% N % N % N % N % Male 207 82.1 81 89.0 205 81.3 79 86.8 Female 45 17.9 10 11.0 47 18.7 12 13.2 Total 252 91 252 91

Predicted group membership

Cross-validated

Male Female

Table 8
Observed

Classication results of logistic regression analysis.

Predicted group membership Sex Male Female 17 62 93.3 68.1 86.6 Percentage correct

Sex Overall Percentage

Male Female

235 29

Table 9
Author

Comparative statistics of maximum and minimum manubrial breadth, mm (mean SD) studied by different workers.
Ethnicity/region Croatian Italian Delhi (Indian)
6

Sex M F M F M F M F M F M F

N 55 35 20 20 50 50 123 83 197 143 252 91

Maximum manubrial breadth 68.20 8.0 60.60 8.10 68.90 7.10 57.10 5.90 40.72 2.68 35.80 4.07 61.4 6.0 54.1 5.2 52.08 4.76 44.6 3.93 54.91 5.44 48.31 5.11

t-test 11.65** 8.25** 27.41** 6.03** 27.41** 15.77** 11.81** 8.34** 5.16** 5.39** 0.00 0.00

Minimum manubrial breadth 36.8 5.5 31.3 3.3 32.19 4.67 28.48 3.93

t-test 5.74** 4.04**

Selthofer et al.4 Vella et al. Atal et al

23

13

Ramadan et al Macaluso
7

Raymond Dart/Pretoria bone collection South-African Chandigarh (Indian)

Present study

Signicant at 0.05 level. ** Signicant at 0.01 level. t-test values for comparison with present study.

26
3.32** 0.14 2.99** 1.41

J. Singh et al. negligible percentage of sternums were assigned their correct sex in the previous studies. The descriptive statistics presented above for the metric traits of the sternum clearly indicate the presence of sample homogeneity of the studied material (coefcient of variation ranging between 9.91 and 21.29). Out of the nine absolute measurements, maximum intra-sample variability is seen for fourth inter-costal length of mesosternum, followed by the third intercostal length and width of third sternubrium. Most of the measurements as studied by Macaluso7 and Ramadan et al6 are bigger again because of nutritional, genetic or environmental differences between these populations and present study Northwest Indian population. Studying sex differences in the metric traits on the basis of Students t-test, it is found that, except the ratio of width of 1st and 3rd sternubrae, all metric traits signicantly discriminate the male and female sternums. The demarking points of various metric traits placed a majority of the male and female sternums in the overlapping zone. The analysis of Tables 25 shows that, except the ratio of width of 1st and 3rd sternubrae (SWI), all the variables correctly discriminated the sex of the sternums ranging between 2and 10%. On the basis of demarking points, width of 1st sternubrium and the fourth inter-costal length of mesosternum happened to be the comparatively better sex determining variables which correctly sexed about 10% of male sternums only. On the other hand, the poorest estimators of sex (on the basis of demarking points) happened to be the ratio of width of 1st and 3rd sternubrae. Male sternums were correctly sexed with comparatively more percentage than the female sternums using the demarking points. It is clear from the limiting points presented for the widths of six metric traits studied, the ratio of width of 1st and 3rd sternubrae turned out to be weak estimators of sex (correctly sexing about 4245% of the sternums). All other measurements are found to be the best estimators of sex (correctly sexing about 6080% of the sternums). For most of the variables, the demarking points were found to discriminate male sternums with a greater reliability though a few female sternums were also correctly classied. It may also be mentioned here that for the majority of the metric traits considered, the limiting points help in classifying the male sternums with a greater accuracy than the female sternums. 6. Conclusions The morphometric study of sternal widths of northwest Indian subjects concluded that sternubrial width index completely failed in sexual dimorphism of the studied bones. While estimating sex based on the criteria of demarking points, it was revealed that except a small percentage of sternums, most of them were lying in the overlapping zone. Efforts to discriminate sex from the results evolved through usage of either identication points or demarking points were met with frustration, the limiting points calculated from both the identication points as well as demarking points and the discriminant function analysis showed promising results and therefore, found to be the best methods for determination of sex from studied metric measurements of the sample. The discriminant function analysis and limiting points are the best suitable methods of determination of sex from the various sternal widths considered in the present study. Further studies are needed to substantiate or reject the ndings of the present study.
Signicant at 0.05 level. Signicant at 0.01 level. t-test values for comparison with present study.
** *

t-test SWI Width of 3rd sternubrium Width of 1st sternubrium Sex N t-test t-test

Comparative statistics of the width of 1st and 3rd sternubrae, mm (mean SD) studied by different workers.

Raymond Dart/ Pretoria bone collection

Maharashtra (Indian)

Chandigarh (Indian)

Dahiphale et al.9

25

Fernandez et al.

Ramadan et al

Present study

Atal et al.14

Macaluso

Table 10

Jit et al.

Author

Chandigarh (Indian)

Ethnicity/region

Delhi (Indian)

South-African

Spanish

M F M F M F M F M F M F M F

312 88 96 47 43 40 50 50 123 83 197 143 252 91

27.45 3.72 24.32 3.94 21.17 3.89 24.44 4.13 26.67 3.78 23.54 3.17 25.88 2.01 21.93 1.64 28.7 3.8 25.2 3.3 34.9 5.9 30.7 5.1 27.03 3.58 23.11 2.85

1.35 2.34* 12.72** 1.96* 0.57 0.73 3.13** 3.10** 5.28** 3.07** 4.73** 5.70**

32.58 5.80 29.19 5.55 31.95 4.13 28.24 4.72 32.81 5.18 30.29 4.95 28.52 2.64 24.66 1.95 24.95 21.83 31.77 27.30 33.53 5.49 28.01 4.84

1.98 1.51 2.87 0.27 0.83 2.42 9.74* 5.76 3.04** 1.05 2.51* 4.04**

85.16 12.01 84.29 18.66 86.29 12.68 87.16 12.42 81.80 11.87 83.95 12.28

Morphometric sex determination from various sternal widths of Northwest Indian sternums collected from autopsy 27 cadavers: A comparison of sexing methods 15. Mahajan A, Batra APS, Khurana BS, Sharma SS. Sex determi7. Ethical standards The author declares that valid consent was obtained from the next kin of the deceased for the removal and examination of sternums and the whole study was conducted in compliance with the applicable current laws of India. 8. Conict of interest
18. 16. nation of human sterna in North Indians. J Punjab Acad Forensic Med Toxicol 2009;9(1):125. Mukhopadhyay PP. Determination of sex from the adult sternums by discriminant function analysis on autopsy sample of Indian Bengali population: a new approach. J Ind Acad Forensic Med 2010;32:3214. Hyrtl J, der Handbuch. Topographischan anatomie. Vienna: Wilhelm Braumdler; 1893. Dwight T. The sternum as an index of sex and age. J Anat Physiol (London) 1881;15:32730. Dwight T. The sternum as an index of sex, height and age. J Anat Physiol (London) 1890;24:52735. Paterson AM. The human sternum. London: William and Norgate; 1904, pp. 4267. Rother P, Hunger H, Liebert U, Seidermann D. Sex differences in the human sternum. Gegenbaur Morphol Jahrb 1975;121(1):2937. Stewart JH, McCormick WF. The sex predictive value of sternal length. Am J Forensic Med Path 1983;4:21720. G.D. Vella, C.P. Campobasso, M. Dragone, Jr F Introna, Sex determination by sternal measurements on a contemporary Southern Italian population. In: Bergamini PR, editor. Proceedings of IVth International Meeting on Forensic Medicine, AlpeAdria-Pannonia, Grado, Italy, 1214 May (1994); Gorgia: Tipographia, Usl/2 Gorziana. 127133. Queiroz A, Wada RB, Preza A, Spadacio C, Cruz BDL. Sex determination by weight verication of sternum bone Brasil. Forensic Med Costa Rica 2004;22:17. Fernandez ED, Saez AS, Moro JIM. Determinacion antropologica del sexo a traves del esternon: Anthropological determination of sex by studying the sternum. Revista de la Escuela de Medicina Legal 2007;116. Bowles GT. The people of Asia. London: Weidenfeld and Nicolson; 1977. Fenton FW, Birkby WH, Camelison J. A fast and safe nonbleaching method for forensic skeletal preparations. J Forensic Sci 2003;48(2):2746. Martin R, Saller K. In: Fischer G, editor. lehrbuch der Anthropologie, Stuttgart, 1959. Steele DG, Bramblett CA. The anatomy and biology of the human skeleton. College Station: Texas, A&M University Press; 1988. Marcowski J. Ueber die Varietaten der Ossication des Menschl. Brustbeins u. uber dren morphologische Bedeutung. Palnisches Arch. F. boil. U. Med. Wissensch., Bd. 1, 1902, cited by G.T. Ashley. The relationship between the pattern of ossication and the denitive shape of the mesosternum in man. J Anat 1956;91:87105. Singh S, Potturi BR. Greater sciatic notch in sex determination. J Anat 1978;125:61924. Asala SA, Mbajiorgu FE, Papandro BA. A comparative study of femoral head diameters and sex differentiation in Nigerians. Acta Anat 1998;162:2327. SPSS Science Inc. 2007. SPSS versions 16.0 for Windows. Chicago: SPSS. Albanese J. A metric method for sex determination using the hipbone and the femur. J Forensic Sci 2008;48:26373. Acharaya AB, Prabhu S, Muddapur MV. Odontometric sex assessment from logistic regression analysis. Int J Leg Med 2011;125:199204. Saunders S, Hoppa, R. Sex determination of human skeleton from long bone measurements (abstract) 42nd Annual CSFS Meeting, Toronto, Ontario; Sept 2630, 1995. Scheuer L. Application of osteology to forensic medicine. Clin Anat 2002;15:297312. Iscan MY. Osteometric analysis of sexual dimorphism in the sternal end of the rib. J Forensic Sci 1985;30:10909. Rogers TL. Determination of sex of human remains through cranial morphology. J Forensic Sci 2005;50:493500.

17.

The author declares that there is no conict of interest.

19. 20. 21. 22. 23.

Acknowledgements The authors would like to acknowledge the guidance, encouragements and support extended by Dr. KH Chavali, Associate Professor, Department of Forensic Medicine and Toxicology, Govt. Medical College and Hospital, Chandigarh, India, and Sh. Beldeu Singh, a PhD scholar in the Institute of Forensic Science and Criminology, Panjab University in the writing of this manuscript. We are also highly thankful to anonymous reviewers for their valuable comments. References
1. Ashley GT. Typing of the human sternum: the inuence of age and sex on its measurements. J Forensic Med 1956;3:2743. 2. Jit I, Jhingan V, Kulkarni M. Sexing the human sternum. Am J Phys Anthropol 1980;53:21724. 3. McCormick WF, Stewart JH, Langford LA. Sex determination from chest plate roentgenograms. Am J Phys Anthropol 1985;68:17395. 4. Selthofer R, Nicolic V, Mrcela T, Radic R, Leksan I, Rudez I, Selthofer K. Morphometric analysis of the sternum. Coll Antropol 2006;30:437. 5. Osunwoke EA, Gwunireama IU, Orish CN, Ordu KS, Ebowe I. A study of sexual dimorphism of the human sternum in the southern Nigerian population. J Appl Biosci 2010;26:16369. 6. Ramadan M, Turkmen N, Dolgan NA, Gokharman D, Menezes M, Kacar M, Kosar U. Sex determination from measurements of the sternum and fourth rib using multislice computed tomography of the chest. Forensic Sci Int 2010;197:120.e15. 7. Macaluso PJ. The efcacy of sternal measurements for sex determination in South African blacks. Forensic Sci Int 2010;202, 111e.111e.7. 8. Torwalt CRRM, Hoppa RD. A test of sex determination from chest plate radiographs. J Forensic Sci 1995;50:78590. 9. Dahiphale VP, Baheete BH, Kamkhedkar SG. Sexing the human sternum in Marathwada region. J Anat Soc Ind 2002;51: 1627. 10. Gautam RS, Shah GV, Jadav HR, Gohil BJ. The human sternum: as an index of age and sex. J Anat Soc Ind 2003;52:203. 11. Hunnargi SA, Menezes RG, Kanchan T, Lobo SW, Binu VS, Uysal S, Kumar HRS, Baral P, Herekar NG, Garg RK. Sexual dimorphism of the human sternum in a Maharashtrian population of India: A morphometric analysis. Leg Med 2008;10:610. 12. Hunnargi SA, Menezes RG, Kanchan T, Lobo SW, Uysal S, Herekar NG, Krishan K, Garg RK. Sternal index: is it a reliable indicator of sex in the Maharashtrian population of India? J Forensic Leg Med 2009;16:568. 13. Atal DK, Murari A, Rani Y, Naik SK. Sex differentiation from sternum: a postmortem metric study. Int J Med Toxicol Leg Med 2008;11:538. 14. Atal DK, Murari A, Naik SK. Sex differentiation from sternal widths. J Ind Acad Forensic Med 2009;30:198201.

24.

25.

26. 27.

28. 29. 30.

31. 32.

33. 34. 35.

36.

37. 38. 39.

28
40. Stewart TD. Essentials of Forensic Anthropology: Especially as developed in the United States. Springeld: CC Thomas; 1979. 41. Dibennardo R, Taylor JV. Multiple discriminant function analysis of sex and race in the postcranial skeleton. Am J Phys Anthropol 1983;61:30514. 42. Garpert R, Black S, Last J. Sex determination from occipital condyle: discriminant function analysis in a eighteenth and nineteenth century British sample. Am J Phys Anthropol 2009;138:38494. 43. Walker P. Sexing skulls using discriminant function analysis of visually assessed traits. Am J Phys Anthropol 2008;136: 3950. 44. Rhine S. Non metric skull racing. In: Gill GW, Rhine S, editors. Skeletal attribution of race. Albuquerque: University of New Mexico Press; 1990. p. 920. 45. Pertoldi C, Loeschcke V, Braun A, Madsen AB, Randi F. Craniometrical variability and developmental stability: Two useful tools for assessing the population viability of Eurasian otter (Lutra lutra) populations in Europe. Biol J Linnean Soc 2000;70:30923. 46. Sun YX, Zhao GC, Yang W. Age estimation on the female sternum by quantication theory I and stepwise regression analysis. Forensic Sci Int 1995;74:5762.

J. Singh et al.
47. White TD. Human osteology. New York: Academic Press; 2000. 48. Bennett KA. On the expression of sex dimorphism. Am J Phys Anthropol 1981;56:5961. 49. Bidmos MA, Dayal MR. Sex determination from the talus of South African whites by discriminant function analysis. Am J Forensic Med Pathol 2003;24:3228. 50. Armitage P. Statistical methods in medical research. Oxford: Blackwell; 1971, p. 332335. 51. Allen JC, Bruce MF, McLaughlin MF. Sex determination from the radius in humans. Hum Evol 1987;2:3738. 52. Holman DJ, Bennett KA. Determination of sex from arm bone measurements. Am J Phys Anthropol 1991;84:4216. 53. Krogman WM, Iscan MY. The Human Skeleton in Forensic Medicine. Springeld: CC Thomas; 1986, p. 551. 54. Barrio PA, Trancho GA, Sanchez JA. Metacarpal sexual determination in a Spanish population. J Forensic Sci 2006;51:9905. 55. Steyn M, Patriquin ML. Osteometric sex determination from the pelvis-Does population specicity matter? Forensic Sci Inter 2009;191:113.e15.